Treatment of Schizophrenia

Family Assessment and Intervention

Edited byM. J. Goldstein, I. Hand, andK. Hahlweg

With contributions by
C. M. Anderson, B. Barbaro, R. Berkowitz, G. Buchkramer,
A. Cechnicki, L. J. Cozolino, B. Dulz, R. Eberlein-Vries,
I. R. H. Falloon, M. J. Goldstein, K. Hahlweg, I. Hand,
J. M. Hooley, J. H. Jenkins, M. Karno, L. Kuipers, J. Leff,
L. Lewandowski, R. P. Liberman, D. J. Miklowitz,
H. J. Moller, K. Nuechterlein, D. Ploog, F. Santana,
W. Schmid-Bode, H. Schulze-Monking, A. de la Selva,
A. M. Strachan, F. Strian, D. Sturgeon, A. Szymusik,
C. E. Vaughn, H. U. Wittchen, C. Wittgen, L. C. Wynne,
J. Zadecki, K. Zawadzka, D. v. Zerssen

With 28 Figures and 18 Tables

Springer-Verlag
Berlin Heidelberg New York Tokyo
Prof. Michael J. Goldstein, Ph.D.
University of California, L.A.
Department of Psychology
1283 Franz Hall
Los Angeles, CA 90024
USA

Prof. Iver Hand, Dr. med.

University of Hamburg
Department of Psychiatry
Martinistrasse 52
D-2000 Hamburg
FRG

Dr. phil. habil. Kurt Hahlweg

Max Planck Institute of Psychiatry
Kraepelinstrasse 2
D-8000 Munchen 40
FRG

ISBN-13: 978-3-540-16628-3 e-ISBN-13: 978-3-642-95496-2

DOl: 10.1007/978-3-642-95496-2

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically those of translation, reprinting, reuse of illustrations, broadcasting, reproduc-
tion by photocopying machine or similar means, and storage in data banks.
Under § 54 of the German Copyright Law, where copies are made for other than private use, a fee is
payable to Verwertungsgesellschaft Wort, Munich.
© by Springer-Verlag Berlin Heidelberg 1986

Product Liability: The publisher can give no guarantee for information about drug dosage and
application thereof contained in this book. In every individual case the respective user must check
its accuracy by consulting other pharmaceutical literature.
The use of registered names, trademarks, etc. in this publication does not imply, even in the absence
of a specific statement, that such names are exempt from the relevant protective laws and regulations
and therefore free for general use.

2119/3321-543210
Contents

Contributors VII
h~. ~
Introduction:
Treatment of Schizophrenia: Historical Aspects (D. Ploog and F. Strian) 1

A. Prediction ofthe Course of Schizophrenia

1. Patient Attributes
1. Outcome and Prediction of Outcome in Schizophrenia: Results from the Litera-
ture and from Two Personal Studies (H. J. Moller, W. Schmid-Bode, H. U.
Wittchen, and D. v. Zerssen) 11

II. Family Attributes

2. An Introduction to EE Measurement and Research (J. M. Hooley) 25
3. Expressed Emotion in Cross-Cultural Context: Familial Respones to Schizo-
phrenic Illness Among Mexican Americans (1. H. Jenkins, M. Karno, A. de la
Selva, and F. Santana) 35
4. Do Relatives Express Expressed Emotion? (A. M. Strachan, M. J. Goldstein,
and D. J. Miklowitz) 51
5. Short-Term Relapse in Young Schizophrenics: Can It Be Predicted and Affect-
ed by Family (CFI) , Patient, and Treatment Variables? An Experimental Study
(B. Dulz and I. Hand) . 59
Comment on Chapter 5 (C. Vaughn) 76
6. Emotional Atmosphere in Families of Schizophrenic Outpatients: Relevance of
a Practice-Oriented Assessment Instrument (G. Buchkramer, H. Schulze-Mon-
king, L. Lewandowski, and C. Wittgen) 79
7. The Marriages and Interaction Patterns of Depressed Patients and Their
Spouses: Comparison of High and Low EE Dyads (J. M. Hooley and K. Hahl-
weg) 85
8. Patterns of Emotional Response in the Families of Schizophrenic Patients (C. E.
Vaughn) 97

B. Modification of the Course of Schizophrenia by Family Interventions

9. Working with Families of Acute Psychotics: Problems for Research and Recon-
sideration (L. C. Wynne) . 109
10. Family Education as a Component of Extended Family-Oriented Treatment
Programs for Schizophrenia (L. J. Cozolino and M. J. Goldstein) . 117
11. Pilot Study of the Impact of a Family Education Program on Relatives of Re-
cent-Onset Schizophrenic Patients (L. J. Cozolino and K. Nuechterlein) 129
12. Psychoeducational Family Therapy (c. M. Anderson) 145
13. Controlled Trial of Social Intervention in the Families of Schizophrenic Patients
(J. Leff, L. Kuipers, R. Berkowitz, R. Eberlein-Vries, and D. Sturgeon) 153
14. Behavioral Family Therapy for Schizophrenia: Clinical, Social, Family, and
Economic Benefits (I. R. H. Falloon) . 171
VI Contents
15. The Impact of Family Intervention Programs on Family Communication and the
Short-Term Course of Schizophrenia (M. J. Goldstein and A. M. Strachan). 185
16. The Psychosocial Program of Treatment of Schizophrenic Patients in the
Crakow Psychiatric Clinic (B. Barbaro, A. Cechnicki, A. Szymusik, K. Zawad-
zka, and J. Zadecki) 193

Epilogue
17. Coping and Competence as Protective Factors in the Vulnerability-Stress Model
of Schizophrenia (R. P. Liberman) 201

Subject Index . 217

Contributors

Carol M. Anderson, Ph.D., University of Department of Psychiatry, Martinistrasse

Pittsburgh, Western Psychiatric Institute, 52, D-2000 Hamburg, FRG
3811 O'Hara Street Pittsburgh, PA Jill M. Hooley, Ph.D., Harvard University,
15213, USA Department of Psychology and Social
Bogdan Barbaro, M.D., University of Cra- Relations, William James Hall, 33 Kirk-
cow Medical Academy, Zwierzyniecka land Street, Cambridge, Massachusetts
15/8,31-108 Cracow, PL 02138, USA
Ruth Berkowitz, Ph.D., MRC Social Janis H. Jenkins, Ph.D., University of Cali-
Psychiatry Unit, Institute of Psychiatry, fornia, Los Angeles, Department of Psy-
De Crespigny Park, GB-London SE 5, chiatry, 760 Westwood Plaza, Los Ange-
GB les, CA 90024, USA
Gerhard Buchkramer, Dr. med., Psychia- Marvin Karno, M.D., University of Cali-
trische und Nervenklinik der Westfali- fornia, L.A., Department of Psychiatry,
schen Wilhelms-Universitat, Albert- 760 Westwood Plaza, Los Angeles, CA
Schweitzer-StraBe, D-4400 Munster, 90024, USA
FRG Liz Kuipers, B.Sc., M.Sc., MRC Social
Andrzej Cechnicki, M.D., University of Psychiatry Unit, Institute of Psychiatry,
Cracow Medical Academy, Zwierzy- De Crespigny Park, GB-London SE 5,
niecka 15/8, 31-108 Cracow, PL GB
Louis J. Cozolino, Ph.D., Department of Julian Leff, M.D., MRC Social Psychiatry
Psychology, University of California, Unit, Institute of Psychiatry, De Cres-
L.A., 1283 Franz Hall, Los Angeles, CA pigny Park, GB-London SE 5, GB
90024, USA Ludwig Lewandowski, Dr. phil., Psychia-
Birger Dulz, Dr. med., Allgemeines Kran- trische und Nervenklinik der Westfali-
kenhaus Ochsenzoll, Langenhorner schen Wilhelms-Universitat, Albert-
Chaussee 560, D-2000 Hamburg 62, FRG Schweitzer-StraBe, D-4400 Munster,
Rosemarie Eberlein-Vries, Dr. med., Psy- FRG
chiatrische Universitatsklinik, NuB- Robert P. Liberman, M.D., University of
baumstraBe, D-8000 Munchen 2, FRG California, L.A., Brentwood VA Medi-
Ian R. H. Falloon, M.D., University of cal Center, 11301 Wilshire Blvd., Los
Southern California, School of Medicine Angeles, CA 90073, USA
Los Angeles; Correspondence: Bucking- David J. Miklowitz, Ph.D., University of
ham Hospital, High Street, Buckingham, California, L.A., Department of
MK181NV,GB Psychology, 1283 Franz Hall, Los Ange-
Michael J. Goldstein, Ph.D., University of les, CA 90024, USA
California, L.A., Department of Psycho- Hans-Jurgen Moller, Dr. med. Psychiatri-
logy, 1283 Franz Hall, Los Angeles, CA sche Klinik und Poliklinik Rechts der
90024, USA Isar, Technische UniversiHit Munchen,
Kurt Hahlweg, Dr. phil. Max Planck Insti- MohlstraBe 26, D-8000 Munchen 80,
tute of Psychiatry, Kraepelinstrasse 2, D- FRG
8000 Munchen 40, FRG Keith Nuechterlein, Ph.D., University of
Iver Hand, M.D., University of Hamburg, California, L.A., Department ofPsychia-
VIII Contributors
try, 760 Westwood Plaza, Los Angeles,
CA 90024, USA
Detlev Ploog, Dr. med., Max Planck Insti-
tute of Psychiatry, Kraepelinstrasse 2,
D-8000 Munchen 40, FRG
Felipe Santana, Ph.D., University of Cali-
fornia, L.A., Department of Psychiatry,
760 Westwood Plaza, Los Angeles, CA
90024, USA
Willi Schmid-Bode, Dr. med., Max Planck
Institute of Psychiatry, Kreapelinstr. 10,
D-8000 Munchen 40, FRG
Heinrich Schulze-Monking, Dr. med., Psy-
chiatrische und Nervenklinik der Westfa-
lischen Wilhelms-Universitat, Albert-
Schweitzer-StraBe, D-4400 Munster,
FRG
Aurora de la Selva, M.Ed., University of
California, L.A., Department ofPsychia-
try, 760 Westwood Plaza, Los Angeles,
CA 90024, USA
Angus M. Strachan, Ph.D., University of
California, L.A., Department of Psycho-
logy, 1283 Franz Hall, Los Angeles, CA
90024, USA
Friedrich Strian, Dr. med., Max Planck
Institute of Psychiatry, Kraepelinstr. 10,
D-8000 Munchen 40, FRG
David Sturgeon, M.D., Department of
Mental Health, School of Medicine, Uni-
versity College, London WC 1, GB
Adam Szymusik, M.D., University of Cra-
cow Medical Academy, Zwierzyniecka
15/8,31-108 Cracow, PL
Christine E. Vaughn, Ph.D., Salfort Com-
munity Health Project, Department of
Clinical Psychology, Prestwich Hospital,
GB-Manchester; Correspondence: 20
Beverly Rd., GB-Cholchester, Essex C03
3NG,GB
Hans-Ulrich Wittchen, Dr. phil., Max
Planck Institute of Psychiatry, Kraepelin-
str. 2, D-8000 Munchen 40, FRG
Connie Wittgen, Dipl.-Psych., Psychiatri-
sche und Nervenklinik der Westfalischen
Wilhelms-Universitat, Albert-Schweit-
zer-StraBe, D-4400 Munster, FRG
Lyman C. Wynne, M.D., Ph.D., Univer-
sity of Rochester, Medical Center,
Department of Psychiatry, Wing R, 601
Elmwood Avenue, Rochester, NY
14642, USA
Jerzy Zadecki, M.D., University of Cracow
Medical Academy, Zwierzyniecka 15/8,
31-108 Cracow, PL
Krystyna Zawadzka, M.D., University of
Cracow Medical Academy, Zwierzy-
niecka 15/8, 31-108 Cracow, PL
Detlef von Zerssen, Dr. med., Max Planck
Institute of Psychiatry, Kraepelinstr. 10,
D-8000 Munchen 40, FRG
Preface
A group of people are seated together in a
room. Recently, they have shared two
important experiences - they have lived
with a close relative, usually an offspring or
spouse, who has been through an episode
of a major mental disorder, most often of a
schizophrenic type, and they have also been
participants in an experimental program
designed to assist them and their disturbed
relative in modifying family patterns to
foster rehabilitation. The experimental
programs involve maintenance pharmaco-
therapy as well as a combination of family
education and therapy. Since a number of
the relatives had experienced the problems
of dealing with a mentally ill relative before
such a program existed, they were asked to
compare their current experiences with
earlier ones. An interviewer asket about
previous experiences with mental health
professionals when the relatives asked for
information about their offspring or
spouse's mental disorder. One mother an-
swered:
Everything seemed to be kept such a dark secret
from the parents, from the people that are
involved - whether that's the thing to do, I dont't
know, but I think that the parents need a lot of
support in whatever the program, the psychiatrist
didn't give us that much. He was dealing with the
patient but not with the family.
A father from another family echoed these
feelings of isolation, "We had the same
problems, they would deal with the patient
but not with the family at all. Everyone that
our son went to, 1 mean, we went but that
was it. We were there, but we never saw
anybody." The interviewer asked, "Did
you ask to see somebody?" The father
replied, "Yeah, you could see them, but
they never had time to talk to you."
Another mother from still another family
stated, "I first got absolutely nowhere, they
would tell me, 'You won't understand if 1
tried to explain it to you, you wouldn't un-
derstand,' and this went on and on in a big
circle, nowhere, - no information at all!"
The feelings of these relatives are not
unique. Until recently, many relatives of
schizophrenic patients experienced some
form of rebuff by mental health profes-
sionals while their relative was treated as an
inpatient and little or no involvement in the
aftercare process when the relative re-
turned home. These attitudes of mental
health personnel reflected both the prevail-
ing wisdom of the time that the family, be-
cause of their pernicious impact, needed to
be isolated from the patient and his treat-
ment, as well as a sense of helplessness
concerning what information and advice
could be provided that would be useful and
helpful.
The tendency to isolate and dissociate the
relatives of severely disordered mental
patients fit rather well with the treatment
pattern existing in the United States and
elsewhere from the end of World War II
until the early 1960s. Patients, when symp-
tomatic, were removed from their families
for an extended inpatient stay. If and when
released, they rarely returned to parental or
spousal homes. Also, when patients were
released during this period, they frequently
had achieved reasonable levels of remission
from their psychosis so that when they did
return home relatives did not have to deal
with active psychotic symptoms.
The patterns of care which have evolved
since the mid-1960s have forced both men-
tal health professionals and families of men-
tal patients to rethink their respective roles.
Especially in the United States, the pattern
of community care for schizophrenic pa-
tients has evolved into a two-phase ap-
proach in which a relatively brief inpatient
stay is followed by extended aftercare in the
community. Because adequate transitional
X M. J. Goldstein et al.
living environments are rare, and even
when they do exist patients resist staying in
them, patients frequently return to a family
environment in varying states of partial re-
mission. Examination of these family envi-
ronments and the problems of reentry of a
disturbed and frequently disturbing relative
has revealed that families often experience
considerable difficulty in fulfilling the
newly assigned role of primary caretaker.
Even when families appear to cope well,
there is a notable impact on the mental
health status of relatives. In one study by
Hawks (1975), 50% of the relatives stated
that their own mental health had been ad-
versely affected in the process of coping
with the patient's mental disorder.
Further research has indicated that atti-
tudes expressed by the relatives toward the
returning mental patient are highly predic-
tive of the likelihood that the patient will
remain in the community 9 months after
discharge from the hospital. The work of
Brown and his associates (Brown, 1959;
Brown, Birley, and Wing, 1972) followed
up by Vaughn and Leff (1976) has gone be-
yond informal clinical observation to pro-
vide empirical evidence that those delete-
rious aspects of the family environment
associated with high risk for relapse could
be specified and measured with standard-
ized procedures. Their work on expressed
emotion (EE) revealed that relatives' atti-
tudes of criticism, hostility, and/or emo-
tional overinvolvement possess powerful
prognostic information about the likeli-
hood of relapse. These attitudes, assumed
to reflect ongoing family transactions, sug-
gest a particular sensitivity of the schizo-
phrenic patient to discriminable attributes
ofthe family environment.
While research on EE attitudes has been
very provocative, there are still a number of
unanswered questions about the construct.
First, are EE attitudes associated with the
course of other psychiatric disorders such as
depression? Second, are the results gener-
alizable to national or subcultural groups
other than the original English samples in
which the measure was developed? Third,
are attitudes elicited during the interview
used to evaluate EE actually expressed in
family transaction? Finally, does thera-
peutic manipUlation of these EE attitudes,
associated with a high probability of re-
lapse, actually alter the short-term course
of the disorder? The first half of the book is
oriented to reports which address these
significant questions.
Although the focus of this book is on the
prognostic value of family attitudes, we
have not ignored the significance of individ-
ual patient attributes as predictors of the
course of the disorder. Therefore a chapter
discussing this issue is presented.
The awareness that family members are
increasingly responsible for the aftercare of
schizophrenic relatives and that specific
attributes of the family environment may be
particularly pernicious has led to renewed
interest in family intervention programs
designed to support the family system, pre-
vent relapse, and foster the social recovery
of the patient. Many of these family inter-
vention programs have utilized measures of
EE to identify the relapse-prone family
units in special need for treatment. Previous
attempts at using traditional family therapy
with families of schizophrenics were not
successful or widely adopted. Further, these
programs were, in the past, viewed as alter-
natives to maintenance pharmacological
treatment. The newer models described in
this volume are not based on such an either/
or conceptualization, and all, in fact, ac-
knowledge the importance of antipsychotic
drugs as maintenance agents during the
posthospialization phase of treatment.
These newer, experimental programs,
which are frequently very pragmatic in
orientation, attempt to provide an addition-
al psychosocial dimension to aftercare be-
yond the protection against relapse provid-
ed by pharmacological agents.
In recent years, several models of family
treatment have been developed, which at-
tempt to include an educational component
as 'part of a larger program of intervention
for schizophrenic patients and their famil-
ies. These models have been classified as
psychoeducational family therapies and
have been designed to help patients out of
hospitals and to return them to a productive
life in the community. Working in tandem
with medication programs, they teach pa-
tients and families about the disorder and
its management. Treating the family as a
resource rather than as a stress, the pro-

grams focus on concrete problem - solving
and specific helping behaviors for coping
with stress. These models recognize the
likelihood of a biological basis of the illness
and the negative impact of stress on the
patient's ability to function. The underlying
assumption of all psychoeducational family
programs developed to date is that giving
families information about the nature of the
disorder, along with specific suggestions for
coping with it effectively, can decrease the
intensity and conflict inherent in family life
and thus reduce the likelihood of (a) relapse
in the index patient and (b) the emergence
of mental disorders in the previously non-
affected relatives.
These programs appear to have certain
common components; (a) the establishment
of a collaborative relationship between
therapist( s) and family members, (b) the
provision of information and support, and
(c) the creation of highly structured pre-
dictable environments in the treatment set-
ting and in the home.
While these commonalities exist, there
are also substantial differences in the pro-
grams developed to date in terms of the
type and detail of information provided
about schizophrenia; the format of these
family education programs (patients pres-
ent or not), the format of the extended fam-
ily therapy phase (relatives only vs family
and patient), the context of the extended
family program (home vs clinic based), and
the duration of these programs ranging
from 6 weeks following discharge to
programs lasting over 1 year.
The variations in key dimensions of the
programs tested in controlled trials to date
warranted closer scrutiny to evaluate what
direction future research on preventive in-
tervention programs for families with a
mentally ill relative should take.
The second half of the book deals with
some of the more innovative family treat-
ment programs that have been tested in
controlled clinical trials in the United States
and Great Britain. In addition, we present
some data concerning the impact of one
such program on family interaction, using
procedures described earlier in this volume
for measuring the correspondence between
EE attitudes and directly observed family
behavior.
Preface XI
One chapter of this section describes a
unique intervention program for schizo-
phrenia developed by a clinical research
team from Poland. Since reports from
Eastern Europe of innovative treatment
programs for schizophrenia rarely appear in
conjunction with those from Western Eu-
rope and the United States, the editors felt
that the inclusion of such a report would
broaden the readers' appreciation of the
diversity of family-oriented programs for
schizophrenia which are emerging world-
wide.
These topics were extensively discussed
and evaluated in two consecutive research
planning workshops, which were held on
May 5-7, 1982 in San Diego, California
(USA) and on June 21-24, 1983 at SchloG
Ringberg, Bavaria (FRG). The first work-
shop was sponsored by the Department of
Psychology, UCLA, and the Office of Pre-
vention, Center for the Study of Schizo-
phrenia and Psychopharmacological and
Somatic Research Branch, all of the Na-
tional Institute of Mental Health (NIMH).
The second workshop was sponsored by the
Max Planck Institute of Psychiatry, Munich
(FRG).
Most of the chapters in this book repre-
sent papers that were presented at the
workshops. However, many of them have
been altered or extended to reflect issues
raised during the formal and informal
discussions that followed each presenta-
tion. The epilogue in particular was written
to provide a comprehensive discussion of
these various issues.
It is difficult to characterize, on a printed
page, the emotional climate of both work-
shops. However, it was tremendously hear-
tening to us (as the organizers) to partici-
pate in the sessions with professionals who
were not only extremely knowledgeable
about their subject matter, but who were so
enthusiastic in sharing their ideas and ex-
periences with each other with an absolute
minimum of defensiveness and posturing.
There was a general feeling among the par-
ticipants that both workshops were a pro-
foundly enriching experience.
It is our hope that some of that flavor will
carryover to the reader and will stimulate
both clinical practice and research in the
difficult area of schizophrenia.
XII M. J. Goldstein et al.
References
Brown, G. W. (1959). Experiences of discharged
chronic schizophrenic mental hospital patients
in various types of living group. Millbank Me-
morial Fund Quarterly, 37, 105-131.
Brown, G. W., Birley, J. L. T., and Wing, J. K.
(1972). Influence of family life on the course of
schizophrenic disorders: A replication. British
Journal of Psychiatry, 121,241-258.
Hawks, D. (1975). Community care. An analysis
of assumptions. British Journal of Psychiatry,
127,276-285.
Vaughn, C. E., and Leff, J. P. (1976). The in-
fluence of family and social factors on the
course of psychiatric illness. British Journal of
Psychiatry, 129, 125-137.
M. J. Goldstein, I. Hand, and K. Hahlweg
Introduction: Treatment of Schizophrenia:
Historical Aspects
D. Ploog and F. Strian
The history of the treatment of schizophre-
nia reflects the continuing problems of es-
tablishing the etiology, even though today
there is agreement that there has to be a
genetic disposition before schizophrenia
will develop. As with all etiologically un-
clarified diseases, there are numerous etio-
logical hypotheses and therapeutic treat-
ment methods, with a corresponding dan-
ger of treatment methods being based on a
priori assumptions or uncritical prag-
matism. Even the assessment of the effec-
tiveness of treatment, however, is closely
linked with valid diagnostic and assessment
criteria, which have only recently become
the object of intensive study (Carpenter,
Heinrichs, and Hanlon, 1981).
In the history of the treatment of schizo-
phrenia, the uncertainties with respect to
diagnosis, etiology, and therapy are re-
flected not only in the relatively large varie-
ty of therapy practiced but also in the
marked influence of contemporary trends
and cultural aspects, with swings between
biologically and psychologically biased ap-
proaches. It can clearly be seen how fatal
any extreme position with regard to diagno-
sis and therapy (e.g., a heavily biased genet-
ic or psychoanalytic approach) can be for
both patient and psychiatrist. In between
these extreme positions, however, there is a
wide range of somatic and psychological
therapies that individually and in combina-
tion have without a doubt considerably
ameliorated the situation of the schizo-
phrenic patient.
When schizophrenic psychoses first be-
gan to be medically understood there were
virtually no somatic treatment methods
available. Opium in high doses, also bella-
donna, quinine, and even digitalis were
tried (Griesinger, 1861; Kahlbaum, 1874).
At the same time, however, it was pointed
out that in the case of an acute state of agita-
tion, hypnotics and narcotics "as a rule are
of little use" (Kraepelin, 1899). Since schiz-
ophrenic disorders were considered to be
"the result of compound influences acting
over a long period of time and gradually
having an increasingly profound effect," it
was not expected that rapid recovery would
result from administering drugs (Kahl-
baum, 1874). On the other hand, in the pre-
vious century psychiatry already had at its
disposal a wide variety of nursing methods
and psychological, social, and ergothera-
peutic approaches. An important part was
played by what was known as "moral treat-
ment," which evolved as early as the first
half of the nineteenth century in England,
France, and the United States and which
essentially corresponded to present-day
sociotherapy. Even before the turn of the
century, "no restraint" treatment had not
only shown itself to be of therapeutic value,
but the abandonment of rigorous restric-
tions and physical restraint had also led to a
decrease in states of dramatic agitation.
Committal was practiced to nurse the pa-
tient in an acute phase, but also to shield
him from provocative environmental influ-
ences. In some cases, psychodynamic and
behavioral therapy approaches seem to
have been anticipated. Kahlbaum, for ex-
ample, wrote in 1874 that the goal of every
therapy should be "to strengthen and rein-
force the ego and to save it from suppres-
sion and disintegration." To that end, the
individual's premorbid imagination and
emotions were to be stimulated, without
however provoking the psychotic charac-
teristics. On the patient's return to his home
environment, the avoidance of emotional
conflicts was considered to be of vital im-
2 D. Ploog and F. Strian
portance. An atmosphere of understanding
and patience would also help to prevent re-
lapses.
It was not, however, these psychosocial
aspects, but the fact that schizophrenia in
the final analysis was considered to be an as
yet unclarified brain disease that was deci-
sive for the further development of research
on psychoses. Kraepelin, in particular, called
for an intensive search for underlying neu-
ronal mechanisms. Looking back at the
somatic forms of treatment, it becomes
clear that, with certain reservations, these
represented the most important advances in
the treatment of schizophrenic disorders.
This is substantiated by a number of factors
such as, for example, the improvement of
the patient's condition, the shortening of
the hospitalization period, the decrease in
the number of readmissions, and better
vocational and social reintegration. Thus,
for example, the introduction of insulin and
convulsive treatment in the 1930s (Cerletti
and Bini, 1938; Sakel, 1935) and the intro-
duction of neuroleptic drugs in the 1950s
(Delay and Deniker, 1952, 1956) drasti-
cally, almost exponentially, reduced the
period schizophrenic patients spent in hos-
pital (Davis, Tanicak, Chang, and Kler-
man, 1982; Meyer, 1984; Meyer, Simon,
and Stille, 1964). In approximately the
same way, it was possible to reduce the se-
verity of the illness (Saarma, 1983). Since
the introduction of neuroleptic therapy in
particular, the once common practice of
physically restraining agitated patients as
well as compulsory committals have be-
come the exception rather than the rule
(Brill and Patton, 1959). In addition,
neuroleptic treatment has entirely replaced
insulin treatment, and electroconvulsive
therapy is practiced in only a limited num-
ber of cases. The neuroleptic spectrum was
broadened, and depot preparations in par-
ticular have led to marked progress in long-
term treatment. It is now possible to identi-
fy or predict side effects with greater cer-
tainty, and in the case of long-term treat-
ment, it has become easier to balance out
desirable and undesirable neuroleptic ef-
fects.
It has recently been shown that in addi-
tion to their immediate antipsychotic effect,
neuroleptic drugs also have a long-term
prophylactic effect. In a review of 35 dou-
ble-blind studies already published, Davis
et al. (1982) established that long-term
neuroleptic therapy resulted in significantly
fewer relapses. Of the total of 3606 schizo-
phrenic patients investigated in the studies,
20% of those treated with neuroleptic drugs
but 53% of those taking placebos had re-
lapses over a 4- to 6-month time period. Pa-
tients seemed to gradually relapse over
time, that is, the number of relapses in each
month of the treatment appeared to be con-
stant. Neuroleptic drugs were shown to re-
duce the expected relapse rate by a factor of
2.5. When noncompliance patients were ex-
cluded, this factor rose to approx. 5.0. One
year after hospital discharge, the average
relapse rate was 41 % for patients receiving
active medication with a corresponding rate
of 68% for patients receiving placebo treat-
ment (Hogarty, 1984).
Maintenance treatment has been shown
to produce significant results only when suf-
ficiently high doses are administered. In a
group of schizophrenic patients treated
with low doses of Fluphenazin decanoate
(1.25-5.0 mg biweekly), the relapse rate
was significantly higher than in a group
treated with a standard dosage (12.5-50.0
mg biweekly) (Kane, Rifkin, Quitkin, and
Klein, 1979a, 1979b). However, relapses
tended to be less severe when compared to
relapses in the standard dosage group
(Kane et aI., 1983).
Some words of caution are necessary with
regard to long-term neuropleptic treat-
ment. First, individual prognostic criteria
are still lacking. Consequently, treatment is
not prescribed on the basis of individual
parameters but of overall criteria such as
the chronicity of the disease (Gaebel and
Pietzcker, 1983). Second, drug-induced
tardive dyskinesia presents a serious prob-
lem. It is estimated that after 4 years of
neuroleptic treatment about 12% of the pa-
tients develope these largely irreversible
side effects (Kane, Woerner, Weinhold,
Wegner, and Kinon, 1982). To avoid tar-
dive dyskinesia, new treatment approaches
were tested, mainly low dose treatment
(Kane et al., 1983) or intermittent treat-
ment in which the neuroleptics are with-
drawn and reintroduced whenever the pa-
tient complains about prodromal signs

(Carpenter, Stephens, Rey, Hanlon, and
Heinrichs, 1982; Herz, Szymanski, and
Simon, 1982). Further research will show
how effective these approaches will be.
Although in recent years new antipsycho-
tic drugs such as molindone, loxapine, and
pimozide have been developed, expecta-
tions in the field of psychopharmacology as
to the development of selective or basically
new substances have not yet been fulfilled.
Except in a few specific cases, propranolol
seems to be mainly effective through poten-
tiation of neuroleptic drugs (Peet, 1981).
The effectiveness of lithium in schizophre-
nia, excluding the schizoaffective forms,
does not yet seem to be sufficiently estab-
lished (Delva and Letemendia, 1982), and
hemodialysis in schizophrenia has been
shown to be a mistake (Skrabanek, 1982).
No actual new alternatives to neuroleptic
treatment are yet in sight.
Parallel to somatotherapy, psycho- and
sociotherapeutic measures, though with
varying temporal and regional importance,
always played a part in the treatment of
schizophrenia. An early example of the ef-
fectiveness of social support comes from
American psychiatry around the mid-nine-
teenth century (approx. 1830-1860) where
"moral treatment" in small, open institu-
tions where there was no discrimination
against the psychiatrically ill was also of
considerable benefit to sufferers from schi-
zophrenia (Savino and Mills, 1967). These
early experiences have been corroborated
anew by very recent, well-controlled trans-
cultural studies in which patients remain in-
tegrated in extended families and the mani-
festations of psychosis are tolerated more
readily by society. Here, the prognoses for
schizophrenic patients turned out to be
more favorable in Nigeria, India, on the is-
land of Mauritius, and in Sri Lanka by com-
parison with Denmark and the USSR
(Murphy and Raman, 1971; Waxler, 1979;
World Health Organization, 1979). The re-
sults from research on twins and adopted
children are also of relevance in this con-
text: on the one hand, they clearly show the
importance of genetic factors in the genesis
of schizophrenia, and on the other, do not
exclude the influence of environmental fac-
tors (Kety, 1983; Kringlen, 1981; Rosenthal
and Ketty, 1968).
Introduction 3
Although even before the turn of the cen-
tury and before Freud there was a great deal
of interest in neurotic and hysterical disor-
ders as well as in appropriate therapeutic
measures (such as hypnosis), this did not
extend to psychotic disorders. It was quite a
while before even Freud began to study the
problem of the psychodynamics of psycho-
ses, and then he applied mechanisms that
he had established in connection with neu-
roses. [In the case of Schreber, incidentally,
according to today's diagnostic criteria, the
patient was not suffering from paranoid
schizophrenia, but from endogenous de-
pression, probably of the bipolar type
(Koehler, 1981)].
The absence of a completely independent
theoretical basis for psychoanalytic psycho-
sis therapy is probably also part of the ex-
planation for the generally discouraging re-
sults. It can in general be seen that among
the psychotherapeutic methods of treating
schizophrenic psychoses, psychoanalytic
methods have become much less popular
and methods aiming at the patient's social
integration and rehabilitation have gained
in importance. Here the treatment and
management of so-called negative symp-
toms is once again in the foreground.
Group therapy should not begin before the
acute psychotic symptoms have abated and
at the same time should be oriented toward
concrete everyday problems (Donlon,
Rada, and Knight, 1973; O'Brien, et al.,
1972). Even group therapy is more success-
ful on the basis of neuroleptic drug treat-
ment (Claghorn, Johnstone, Cook, and
Itschner, 1974). In general, there will have
to be a more precise definition of the tlrera-
peutically effective variables in psycho-
therapy, situational therapy, and psycho-
social rehabilitation. In their survey, Keith
and Matthews (1982) hence came to the
conclusion that all psychosocial therapeutic
methods are generally assessed more opti-
mistically than the level of knowledge about
them merits.
On the other hand, a relatively clear pic-
ture is obstained by the methodologically
excellent therapy studies in which the ef-
fectiveness of neuroleptic and psychothera-
peutic forms of treatment are compared.
These comparative studies, however, are
mainly concerned with the effects treat-
4 D. Ploog and F. Strian
ment has on acute psychotic symptoms and
only partially with long-term effects. In
these investigations the methodological
problems are not insignificant. A follow-up
study, for instance, on a group of patients
first described by Rosen (1947, 1953)
showed that a large number of borderline
cases had been included, and that in addi-
tion, in the long-term outcome of 75% of
the patients two to five readmissions had
occured (Horwitz, Polatin, Kolb, and
Hoch, 1958). Results which significantly
demonstrated the greater effectiveness of a
combination of psychotherapeutic and neu-
roleptic drug treatment than of psycho-
therapy unsupported by drug treatment
were obtained in a number of comparative
studies (e.g., Grinspoon, Ewalt, and
Shader, 1972; Hogarty, Goldberg, and
Schooler, 1974a, 1974c; Hogarty, Gold-
berg, Schooler, and Ullrich, 1974b; Lind-
berg, 1981; O'Brienet al., 1972). It is signif-
icant, for example, that in the study by
Grinspoon et al. treatment with phenothia-
zine proved to be no obstacle to psycho-
therapy, but was rather found to result in
better therapeutic cooperation and at the
same time fewer withdrawal tendencies on
the part of the patient. Greenblatt, Solo-
mon, Evans and Brooks (1965), in a com-
parison of neuroleptic and social therapies,
had already demonstrated that patients
who had no drug treatment in the acute
phase of the illness were subsequently less
able to profit from long-term psychothera-
py than patients who had been treated with
drugs.
Particularly sobering results were ob-
tained in the effectiveness studies of May,
Tuma, and Dixon (1981), where compari-
sons were made between neuroleptic drugs,
psychotherapy, and psychotherapy with
basic neuroleptic drug treatment and a con-
trol group with neither psychotherapy nor
neuroleptic drug treatment. Particularly
when long-term comparisons were made,
substantially better results were obtained
for both the groups receiving drugs than for
the groups that had not received drugs. Al-
though the comparison is open to criticism
on account of the limited experience of
some of the therapists, as well as the short
duration of psychotherapy, once again the
long-term effects resulting from initial neu-
roleptic treatment were noticeably better
than those resulting from psychotherapeu-
tic treatment.
The question as to which of the psycho-
therapeutic methods used for schizophrenia
gives the best results remains open due to a
lack of methodologically reliable studies as
well as of criteria for indicating when to
apply a particular method and establishing
how effective it is. With regard to the effec-
tiveness oflong-term treatment, moreover,
factors such as the length of time spent in
the hospital (Glick and Hargreaves, 1979;
Herz and Melville, 1980; Platt, Hirsch, and
Knight, 1981), treatment on an inpatient
basis only or a mixture of inpatient and out-
patient treatment (Vanicelli, Washburn,
Scheef, and Longabaugh, 1978), and partic-
ularly the type of clinical aftercare (Linn,
Caffey, Klett, and Hogarty, 1977; Stein and
Test, 1980) appear to be important.
Although a large number of valid com-
parative studies demonstrated the necessity
of neuroleptic drug treatment, these results
are by no means an argument against psy-
chotherapeutic measures. It is far more a
question of when which psychotherapeutic
method should be used and with what aim
in view. The results of the comparative
studies above all point toward therapeutic
measures that can improve psychosocial in-
tegration and acceptance in the period fol-
lowing hospitalization. In this respect, the
situation today is in many ways analogous
to the historical developments outlined
above.
Also on a pragmatic, rehabilitative level,
social therapy and aftercare today are es-
sential components of the treatment of schi-
zophrenia. This is particularly shown by the
experiment in deinstitutionalizing schizo-
phrenic patients since this was only achiev-
ed without disastrous consequences for the
patient when he could be reintegrated into
family or social structures or received ap-
propriate aftercare (Braun et al., 1981;
Johnstone, Owens, Gold, Crow, and Mac-
millan, 1981). The importance of including
the family in aftercare is clearly demon-
strated by the family assessment and thera-
py studies which are included in this vol-
ume. It is therefore not necessary to review
these results. However, it must be pointed
out that it is not very likely that communica-

tive structures are of specific significance in
connection with schizophrenia since social
support also improves prognosis in other
psychiatric and even in most organic dis-
eases (Strian, 1983, 1984). Whether the
particular sensitivity and anxiety on the part
of the schizophrenic patient in response to a
critical, overprotective, and adverse envi-
ronment ("expressed emotion") is of spe-
References
Braun, P., Kochansky, G., Shapiro, R., Green-
berg, S., Gudeman, J. E., Johnson, S., and
Shore, M. F. (1981). Overview: Deinstitution-
alization of psychiatric patients, a critical re-
view of outcome studies. American Journal of
Psychiatry, 138,736-749.
Brill, H., and Patton, R. E. (1959). Analysis of
population reduction in N ew York state mental
hospitals during the first four years of large
scale therapy with psychotropic drugs. Ameri-
can Journal of Psychiatry, 116,495-508.
Carpenter, W. T., Heinrichs, D. W., and Han-
lon, T. E. (1981). Methodologic standards for
treatment outcome research in schizophrenia.
American Journal of Psychiatry , 138,465 -471.
Carpenter, W. T., Stephens, J. H., Rey, A. C.,
Hanlon, T. E., and Heinrichs, D. W. (1982).
Early intervention vs. continuous pharmaco-
therapy of schizophrenia. Psychopharmacolo-
gy Bulletin, 18,21-23.
Cerletti, U., and Bini, L. (1938). L'electroshock
[Electric shock therapy]. Archives of General
Neurology and Psychiatry, 19,266.
Claghorn, J. L., Johnstone, E. E., Cook, T. H.,
and Itschner, L. (1974). Group therapy and
maintenance treatment of schizophrenics. Ar-
chives General Psychiatry, 31, 361-365.
Davis, J. M., Janicak, P., Chang, S., and Kler-
man, K. (1982). Recent advances in the phar-
macologic treatment of the schizophrenic disor-
ders. In: L. Grinspoon (Ed.), Psychiatry 1982
(pp. 239-252). Washington, DC: American
Psychiatric Press.
Delay, J., and Deniker, P. (1952). Le traitement
des psychoses par une methode neurolytique
derivee de l'hiberno-therapie. [Treatment of
psychoses by a neurolytic method derived from
artificial hibernation therapy]. In: Comptes
Rendus du Congres des Medicins Alcemistes et
Neurologistes de Langue Fran~aise, Luxem-
burg.
Delay, J., and Deniker, P. (1956). COlloque in-
ternational sur la chlorpromazine et les medic-
aments neuroleptiques en therapeutique psy-
chiatrique. [Special issue]. Encephalo, 1-982.
Introduction 5
cial importance is something that still has to
be proved. In any case, such concrete pat-
terns of interaction present an opportunity
of bringing about behavior changes without
the individuals involved having to be bur-
dened with unproductive questions as to
circumstances and persons in their environ-
ment that might be responsible for the ill-
ness.
Delva, N. J., and Letemendia, F. J. (1982).
Lithium treatment in schizophrenia and schizo-
affective disorders. British Journal of Psychia-
try, 141, 387-400.
Donlon, P. T., Rada, R. T., and Knight, S. W.
(1973). A therapeutic aftercare setting for
"refractory" chronic schizophrenic patients.
American Journal of Psychiatry, 130, 682 - 684.
Gaebel, W., and Pietzcker, A. (1983). Indikation
zur neuroleptischen Langzeitmedikation: Stan-
dardverfahren oder individualprognostisch ge-
leitete Intervention? Eine prospektive Unter-
suchung zum klinischen Handlungsmodell
[Indications for long-term neuroleptic medica-
tion: standard procedure or intervention se-
lected on the basis of individual prognosis? A
prospective study of the clinical behavior
model]. Nervenarzt, 54, 467-476.
Glick, J. D., and Hargreaves, W. A. (1979). Psy-
chiatric hospital treatment for the 1980s.
Lexington: Heath.
Greenblatt, M., Solomon, M. H., Evans, A. S.,
and Brooks, G. W. (1965). Drug and social
therapy in chronic schizophrenia. Springfield,
Ill.: Thomas.
Griesinger, W. (1861). Pathologie und Therapie
der psychischen Krankheiten [Pathology and
therapy of mental illnesses]. Stuttgart: Krabbe.
Grinspoon, L., Ewalt, J. R., and Shader, R. I.
(1972). Schizophrenia: Pharmacotherapy and
psychotherapy. Baltimore: Williams and Wil-
kins.
Herz, M. I., and Melville, C. (1980). Relapse in
schizophrenia. American Journal of Psychia-
try, 137, 801-805.
Herz, M.I., Szymanski, H. V., and Simon, J. C.
(1982). Intermittent medication for stable schi-
zophrenic outpatients: An alternative to main-
tenance medication. American Journal of Psy-
chiatry, 139, 918-922.
Hogarty, G. E. (1984). Depot neuroleptics: The
relevance of psychosocial factors. A United
States perspective. Journal of Clinical Psychia-
try, 45, 36-42.
Hogarty, G. E., Goldberg, S. C., and Schooler,
6 D. Ploog and F. Strian
N. R. (1974a). Drug and sociotherapy in the
aftercare of schizophrenic patients. Psycho-
pharmacology Bulletin, 10,47.
Hogarty, G. E., Goldberg, S. C., and Schooler,
N. R., Ulrich, R. F. (1974b). P.rug and socio-
therapy in the aftercare of schizophrenic pa-
tients, II: Two-year relapse rates. Archives of
General Psychiatry, 31, 603-608.
Hogarty, G. E., Goldberg, S. C., and Schooler,
N. R. (1974c). Drug and sociotherapy in the
aftercare of schizophrenic patients, III: Adjust-
ment of nonrelapsed patients. Archives of
General Psychiatry, 31, 609-618.
Horwitz, W. A., Polatin, P., Kolb, L. C., and
Hoch, P. H. (1958). A study of cases of schizo-
phrenia treated by "direct analysis". American
Journal of Psychiatry, 114, 780-783.
Johnstone, E. E., Owens, D. G., Gold, A.,
Crow, T. J., and Macmillan, J. F. (1981). In-
stitutionalization and the defects of schizophre-
nia. British Journal of Psychiatry, 139,
195-203.
Kahlbaum, K. (1874). Die Katatonie oder das
Spannungsirresein. [Catatonia or tension de-
mentia). Berlin: Hirschwald. pp 100-104,
468-537.
Kane, J., Rifkin, A., Ouitkin, F., and Klein, D.
F. (1979a). A pilot study of "low dose" flu-
phenazine decanoate in outpatient schizophre-
nics. Psychophannacology Bulletin 15, 78.
Kane, J., Rifkin, A., Ouitkin, F., and Klein, D.
F. (1979b). Low dose fluphenazine decanoate
in maintenance treatment of schizophrenia.
Journal of Psychiatric Research, 1,341-348.
Kane, J., Woerner, M., Weinhold, P., Wegner,
B., and Kinon, B. (1982). A prospective study
of tardive dyskinesia development: Prelimi-
nary results. Journal of Clinical Psychophanna-
cology, 2, 345-349.
Kane, J., Rifkin, A., Woerner, M., Reardon,
G., Sarantakos, S., Schiebel, D., and Ramos-
Lorenzi, J. (1983). Low dose neuroleptic treat-
ment of outpatient schizophrenics, I: Prelimi-
nary results for relapse rates. Archives of Gen-
eral Psychiatry, 40, 893-896.
Keith, S. J., and Matthews, S. M. (1982). Group,
family, and milieu therapies and psychosocial
rehabilitation in the treatment of the schizo-
phrenic disorders. In: L. Grinspoon (Ed.), Psy-
chiatry 1982 (pp. 166-178). Washington, DC:
American Psychiatric Press.
Kety, S. S. (1983). Mental illness in biological
and adoptive relatives of schizophrenic adop-
tees: Findings relevant to genetic and environ-
mental factors in etiology. American Journal of
Psychiatry, 140,720-727.
Koehler, K. G. (1981). The Schreber case and
affective illness: A research diagnostic reassess-
ment. Psychological Medicine, 11,689-696.
Kraepelin, E. (1899). Psychiatrie. Ein Lehrbuch
fUr Studierende und Arzte [Psychiatry. A text
book for students and doctors]. Leipzig: Barth,
pp213-215.
Kringlen, E. (1981). Zum heutigen Stand der
Schizophrenieforschung. [The present status of
research into schizophrenia]. Nervenarzt, 52,
68-73.
Lindberg, D. (1981). A controlled study of 5
year's treatment with psychotherapy in combi-
nation with depot neuroleptics in schizophre-
nia. Acta Psychiatrica Scandinavica, 289:
(Suppl.),56-66.
Linn, M. W., Caffey, E. M., Klett, C. J., and
Hogarty, G. (1977). Hospital vs. community
(foster) care for psychiatric patients. Archives
of General Psychiatry, 34, 78-83.
May, P. R. A., Tuma, A. H., and Dixon, W. J.
(1981). Schizophrenia. A follow-up study of
the results of treatment. Archives of General
Psychiatry, 38, 776-784.
Meyer, J. E. (1984). Die Therapie der Schizo-
phrenie in Klinik und Praxis [Treatment of
schizophrenia in clinic and private practice set-
tings). Nervenarzt, 55, 221-229.
Meyer, J. E., Simon, G., and Stille, D. (1964).
Die Therapie der Schizophrenie und der endo-
genen Depression zwischen 1930 und 1960
[Treatment of schizophrenia and endogenous
depression between 1930 and 1960]. Archiv fiir
Psychiatrie und Nervenkrankheiten, 206,
165-179.
Murphy, H. B. M., and Raman, A. C. (1971).
The chronicity of schizophrenia in indigenous
tropical peoples. British Journal of Psychiatry,
118, 489-497.
O'Brien, C., Hamm, K., Ray, B., Pierce, J.,
Luborsky, L., and Mintz, J. (1972). Group vs.
individual psychotherapy with schizophrenics.
Archives of General Psychiatry, 27, 474-478.
Peet, M. (1981). Is propranolol antischizophre-
nic? Neuropharmacology, 20,1303-1307.
Platt, S., Hirsch, S. R., andKnight,D. C. (1981).
Effects of brief hospitalization on psychiatric
patients behavior and social functioning. Acta
Psychiatrica Scandinavica, 63,117-128.
Rosen, J. N. (1947). The treatment of schizo-
phrenic psychoses by direct analytic therapy.
Psychiatry, 21,117-119.
Rosen, J. N. (1953). Direct analysis. Selected
papers. New York: Grune and Stratton.
Rosenthal, D., and Kety, S. S. (1968). The trans-
mission of schizophrenia. New York: Perga-
mon.
Saanna, J. (1983). Katamnestische Untersu-
chungen von Behandlungsergebnissen bei
Schizophrenen [Study of results of treatment in
schizophrenic patients with reference to follow-
up histories]. Psychiatrie, Neurologie und
Medizinische Psychologie, Leipzig, 35,
587-597.

Sakel, M. (1935). Neue Behandlungsmethode
der Schizophrenie. [A new treatment method
for schizophrenia]. Wien: Perthes.
Savino, M., and Mills, A. (1967). The rise and
fall of moral treatment in California. Journal of
the History of the Behavioral Sciences, 3,
359-369.
Skrabanek, P. (1982). Haemodialysis in schizo-
phrenia: Deja vu or idee fixe. Lancet, 1,
1404-1405.
Stein, L., and Test, M. (1980). Alternative to
mental hospital treatment. Archives of Gener-
al Psychiatry, 37, 392-411.
Strian, F. (1983). Angst. Grundlagen und Klinik
[Fear. Bases and clinical practice]. Berlin Hei-
delberg New York: Springer.
Introduction 7
Strian, F. (1984). Medizinische Aspekte von
Sozialangst und fehlender mitmenschlicher
Hilfe. Medizinische Klinik 79, 278-280.
Vanicelli, M., Washburn, S., Scheef, B. J., and
Longabaugh, R. (1978). A comparison of usual
and experimental patients in a psychiatric cen-
ter. Journal of Consulting and Clinical Psycho-
logy, 46, 87-93.
Waxler, N. E. (1979). Is outcome forschizophre-
nia better in the nonindustrialized societies?
The case of Sri Lanka. Journal of Nervous and
Mental Disease, 167, 144-158.
World Health Organization. (1979). Schizophre-
nia: An international follow-up-study. Chiche-
ster: Wiley.
A. Prediction of the Course of Schizophrenia
I. Patient Attributes
1. Outcome and Prediction of Outcome in
Schizophrenia: Results from the Literature and
from Two Personal Studies
H. J. Moller, W. Schmid-Bode, H. U. Wittchen, and D. v. Zerssen
Short Review of the Literature
A large number of studies investigating the
outcome of schizophrenia have been car-
ried out. These studies have provided many
hypotheses about the course of the illness,
especially the characteristics of and the pre-
diction of the outcome. However, the find-
ings of the earlier studies have been ques-
tioned because of severe methodological
deficiencies. In most of these studies, ret-
rospective evaluation techniques or insuffi-
ciently standardized instruments were em-
ployed and many have introduced signifi-
cant bias. Some more recent studies, how-
ever, have avoided such methodological
problems by studying patients prospective-
ly. Unfortunately, these excellent studies
have evaluated patients only over short pe-
riods of time, generally from 1 to 2 years
(e.g., Schooler, Goldberg, Boothe, and
Cole, 1967; Astrachan, Brauer, Harrow,
and Schwartz, 1974; Hogarty, Goldberg,
Schooler, and Ullrich, 1974; Strauss and
Carpenter, 1974; Wittenborn, McDonald,
and Maurer, 1977; World Health Organiza-
tion, 1979). Only the 5-yearfollow-up study
of patients in the International Pilot Study
of Schizophrenia (IPSS) covers a longer pe-
riod of time. Thus far, data from this 5-year
follow-up study have only been published
by the Washington Research Center of the
IPSS (Hawk, Carpenter, and Strauss, 1975;
Strauss and Carpenter, 1977).
Although there are many discrepancies
between the results of different follow-up
studies - these discrepancies are caused,
among other things, by differences in sam-
pling, diagnostic process, assessment of
variables, length of follow-up period -
some findings have been replicated in many
of them. Very important is the result that,
in general, schizophrenia has a poor out-
come in less than 50% of the patients. This
was not only demonstrated by short-term
follow-up studies (e.g., Strauss and Car-
penter, 1974; Astrachan et al. 1974; World
Health Organization 1979) but also in long-
term follow-up studies covering 5 - 20 years
and more (e.g., Bleuler, 1972; Strauss and
Carpenter, 1977; Huber, Gross and Schutt-
ler, 1979; Achte, 1980; Watt, Karz, and
Schepherd, 1983). Comparing the results of
older and newer studies, there is some evi-
dence from follow-up studies that the better
outcome of schizophrenia in the recent de-
cades is due to modern biological treatment
methods, but also some factors, e.g., insti-
tutional changes, must be taken into consid-
eration (Hogarty, 1977).
Based on the results of studies on long-
term treatment with neuroleptics (Davis,
Schaffer, Killian, Kuard, and Chan, 1980),
it can be concluded that this therapeutic ap-
proach is effective in preventing relapses of
schizophrenic illness in a high percentage of
schizophrenic patients. Hitherto, it has
been rather uncertain whether the global
outcome concerning psychopathology and
social adjustment after many years is better
in those patients who took neuroleptic
drugs than in those who refused to do so.
This is especially true for treatment periods
longer than 3 years because there is no con-
trolled study on this topic. Another pro-
blem of neuroleptic long-term treatment is
that a~out 20% - 30% of the patients seem
to be nonresponders.
The knowledge on the effect of psycho-
therapeutical strategies in schizophrenia is
inconclusive up to now. The reported re-
sults of controlled studies concerning psy-
12 H. J. Moller et al.
choanalytical individual therapy, for exam-
ple, seem to indicate that a significant effect
on outcome variables was not measurable
(Cancro, Fox, and Shapiro, 1978; Feinsil-
ver and Gunderson, 1972; Grinspoon,
Ewalt, and Shader, 1968; May, Tuma, and
Dixon, 1977; Mosher and Keith, 1979).
The recent results of family intervention
strategies, based on the concept of high ex-
pressed emotion (Vaughn and Leff 1976),
seem to be more hopeful (see Anderson,
Falloon, and Leff et aI., this volume). High
expressed emotion of close relatives was
also described as an important predictor of
schizophrenic relapses (Chap. 2).
There is a long tradition concerning the
differentiation of schizophrenia with good
and poor outcome. In this context, the diag-
nostic concepts of schizophreniform psy-
chosis, reactive psychosis, atypical psy-
chosis, and schizoaffective psychosis - all
describing psychoses with a comparatively
good outcome - have to be mentioned
(Langfeld, 1957; Eitinger, Laane, and
Langfeld, 1958; Wellner and Stromgren
1958; Vaillant, 1962; Stephens and Astrup,
1965; Achte, 1967; Stephens, 1970;
Crougan, WeIner, and Robins, 1974; Hawk
et aI., 1975; Tsuang, Woolson, and Fle-
ming, 1979). Besides psychopathological
and anamnestic characteristics defining
these subtypes, numerous other psycho-
pathological, sociodemographic, and
anamnestic data have been described as
relevant concerning prognosis (Renton,
Affleck, Carstairs, and Forrest, 1965;
Achte, 1967; Strauss and Carpenter, 1974;
Tsuang and Winokur, 1974; Gunderson,
Carpenter, and Strauss, 1975; Hawk et aI.,
1975; Pokorny, Thornby, Kaplan, and Bau,
1976; Goldberg, Schooler, Hogarty, and
Roper, 1977; Hargreaves, Glick, Drues,
Showsrack, and Geigenbaum, 1977;
Strauss and Carpenter, 1977; Wittenborn
et aI., 1977; Bland, Parker, and Om, 1978;
Huber et al., 1979; World Health Organiza-
tion, 1979). Only some of these predictors
could be replicated by different research
groups, e.g., lower age at onset, insidious
onset, no precipitating factors, affective
blunting, disturbances of premorbid per-
sonality, or adjustment and length of epi-
sode prior to assessment are known as pre-
dictors of poor outcome. Besides this sub-
stantial agreement on certain in prognostic
criteria, some findings remain contradicto-
ry, such as the prognostic implications of
certain psychopathological symptoms.
In general, the percentage of outcome
variance explained by a single one of these
predictors was rather low. Therefore, in re-
cent studies, combinations of predictors
were analyzed by multivariate methods to
increase the percentage of explained out-
come variance (Strauss and Carpenter,
1977; WHO, 1979). However, such findings
of multivariate analyses must be considered
critically because the multivariate method
optimizes the result concerning the specific
sample. Consequently, the practical rele-
vance has to be proven by cross validation.
Replication studies concerning the results
of such multivariate analyses, however,
were usually not performed.
Another approach to improve the predic-
tion of outcome is the simple combination
of relevant predictors, either according to
the literature or identified by one's own ana-
lyses. Such combination scores should also
prove their predictive power in cross-vali-
dation studies. However, this attempt if
performed at all, was successful only in a
few studies (Kokes, Strauss, and Klorman,
1977; Strauss and Carpenter, 1974). The
total score of such a prognostic scale should
guarantee a better prediction than any sin-
gle predictive item, a presumption that does
not seem to be fulfilled by some of these
scales (Moller et aI., 1984 a, b, c). Several
prognostic scales for schizophrenic patients
have been described in the literature
(Kokes et aI., 1977). Some of them consist
only of data on premorbid adjustment -
such as the Phillips scale (Phillips, 1966) -
some others focus more on psychopatholog-
ical items - such as the Vaillant scale (Vail-
lant, 1962) or the Stephens scale (Stephens,
1970). The broadest approach was realized
by the Strauss-Carpenter prognostic scale,
a multidimensionaL approach combining
aspects of psychopathology, social adjust-
ment, and history of illness.
 Outcome and Prediction of Outcome in Schizophrenia 13

Results of Own Follow-up Studies on Schizophrenia

Outcome and Predictors, First Sample

Our first study was a 5- to 6-year follow-up
on about 100 patients who had been treated
as inpatients in the Max Planck Institute of
Psychiatry (Moller et al. 1981 a, b, 1982 a,
b). The study was performed using standard-
ized assessment procedures, among them
the Inpatient Multidimensional Psychiatric
Scale (IMPS) (Lorr, 1974) and the clinical
self-rating scales (von Zerssen, 1976) to de-
scribe the psychopathological state and the
Global Assessment Scale (GAS) (Spitzer,
Endicott, and Fleiss, 1976) for describing
the global outcome at follow-up.
At discharge 5 years earlier, the clinical
diagnosis was based on the leD classifica-
tion. Of all the patients, 77% met the leD
description for schizophrenia (leD 295; pa-
tients with schizo affective psychoses: leD
295.7 were excluded from this study) and
23% the leD descriptions for other types
of paranoid psychoses (leD 297, 298.2,
298.3,298.9). At follow-up, seven patients
had died, five of them by suicide. Sufficient
follow-up data for the statistical analyses
could be obtained from 81 patients, for
three of them only by relatives.
This sample (n = 81) included patients
with their first lifetime episode (55%) or re-
turn of the disorder at index admission.
Sixty percent of the patients had never pre-
viously been admitted to a psychiatric hospi-
tal before entering the Max Planck Institute
(index admission). The sample was almost
equally divided between men (49%) and
women (51%). Sixty-six patients (82%)
were 40 years old or younger, and only one
patient was over age 60. Only 22% were
married. Treatment consisted of neurolep-
tics (mostly haloperidol, in individualized
but not high dosages), supportive psycho-
therapy, and sociotherapy. Hospitalization
was not longer than 3 months for 92% ofthe
patients with only 8% staying for more than
4 months. Usually 1-2 weeks before dis-
charge, the oral neuroleptic medication was
changed to an intramuscular depot neuro-
leptic. The sample seems comparable to the
schizophrenic population of university clin-
ics, but not to that of county hospitals where
chronic patients are overrepresented. It
also appears comparable to the IPSS sam-
ple, which excluded chronic patients (with
psychotic symptoms of more than 2 years
duration in the 5 years before index admis-
sion or hospitalized more than 3 years).
At follow-up, 44% of the patients showed
serious psychopathological symptoms and/
or disturbances of social adjustment, indi-
cated by a score below 50 on the GAS; 31 %
were below a GAS score of 40, which means

Table 1. Level of Functioning (GAS Score) at Follow-up

Information from
Patients Relatives
Level of functioning No. % No. %
100-91 No symptoms 8 10 5 10
90-81 Transient symptoms 8 10 0 0
80-71 Minimal symptoms 4 5 5 10
70-61 Some mild symptoms 14 18 11 23
60-51 Moderatesymptoms 8 10 1 3
50-41 Any serious symptomatology 10 13 6 13
40-31 Majorimpairment 15 19 11 24
30- 21 Unable to function 9 12 8 17
20-11 Needs some supervision 0 0 0 0
10-1 Needs constant supervision 0 0 0 0
Not classifiable 2 3 0 0
Total 78 100 47 100
14 H. J. Moller et al.
according to the GAS definition that hospi-
talization would be expected if the patient
was not already hospitalized. During the
follow-up period, 34% of the patients were
unable to work for more than 1 year; 58%
had to be readmitted to a psychiatric hospi-
tal, 10% for more than 1 year. At follow-
up, 16% were living in a nursing home and
6% were psychiatric inpatients (Table 1).
Correlation analyses between the GAS
score and other outcome criteria demon-
strated that the GAS score best reflects the
general outcome (Table 2). Therefore, it
0\
V)
was chosen as the main criterion for the
o analysis of predictors. Among approxima-
I tely 40 potential predictor variables tested
in a product moment correlation analysis, a
O","N\O ,....;
number of variables proved to be statistical-
000"'""'"
""';000
V)
o
ly significant (P < 0.05) predictors of global
outcome (Table 3). The best predictors
(r > 0.35) were: duration of occupational
O<'l","NN
disintegration (inability to work, unemploy-
o \0\0 \0 V)
""';0000 ment) during the 5 years preceding index
admission, impairment of working ability in
the year before index admission, personali-
00\0\<'lV)""';
ty change before index admission, poor psy-
or-\Or-V)V)
""';00000 chopathological state on discharge, and the
self-rating factor of paranoid tendencies at
discharge. High scores on these variables
000","0\000\<'l
were correlated with a poor global out-
o \0 V) r- 00 <'l V)
""';000000 come. Besides these variables, some others
correlated significantly with the outcome
criterion, but mostly the percentage of ex-
ONr-r-Oo\\O""'; \0
plained variance was below 10%.
Or-<'lNV)V)N","
""';0000000
V)
o Most of the predictors found in our study
have already been mentioned in the litera-
ture dealing with the prognosis of schizo-
phrenia. Among those variables that
proved to be of greatest prognostic signifi-
cance in our own study, occupational dis-
integration (employment status) before in-
dex admission was a strong predictor in stu-
dies by Strauss and Carpenter (1974,1977).
The prognostic value of the psychopatho-
logical state at discharge was emphasized by
Renton et al. (1965), Affleck, Burns, and
Forrest (1976), and Wittenborn et al.
(1977). A personality change, especially in
the sense of a lack of emotion, is another
predictor that has been described by many
authors (Vaillant, 1964; Lindelius, 1970;
Stephens, 1970; WHO, 1979).
Analyses of correlations between the
predictor variables and other outcome
 Outcome and Prediction of Outcome in Schizophrenia 15
Table 3. Predictors of Global Outcome (GAS) (n = 74-78; for Self-Rating Factors, n = 45-50)
Correlation
Predictors of global outcome (GAS) coefficient
(-) Higher socioeconomic status of the parents .25
(-) Premorbid working dysfunction .27
(+) More advanced age at first manifestation .28
(+) More advanced age at first hospitalization .23
(+) Precipitating factors before first manifestation .28
(-) Duration of psychiatric hospitalization (5 years before index admission) .27
(-) Duration of occupational disintegration (5 years before index admission) .38
(+) Lasting heterosexual relationship .28
(- ) Impairment of working ability (1 year before index admission) .39
(-) Personality change (1 year before index admission) .42
(-) Diagnosis of schizophrenia .25
(- ) Poor psychopathological state at discharge .35
(-) IMPS superfactor of organic syndrome * at discharge .30
(- ) IMPS superfactor of depressive-apathetic syndrome" at discharge .24
(+) Ratio of amelioration of the IMPS superfactor of psychotic excitement .28
(-) Self-rating factor of paranoid tendencies at discharge .39
(+ ) Ratio of amelioration of the self-rating factor of paranoid tendencies .40
( + ) = good prognosis; ( - ) = poor prognosis
, IMPS factor "retardation and apathy" + IMPS factor "disorientation";
** IMPS factor "anxious depression" + IMPS factor "retardation and apathy" + IMPS factor "im-
paired functioning"

criteria revealed that the prognostic signifi-
cance of one variable for the GAS score did
not necessarily apply to other outcome cri-
teria (Table 4). Likewise, a variable that is
a good predictor for one outcome criterion
may not be predictive for the outcome on
the GAS. The inclusion of variables that are
predictive for other outcome criteria, but
not for the GAS, expands the number of
predictors.
Outcome and Predictors, Second Sample
Using partially the same assessment
procedures, a second follow-up study
(Schmid-Bode and Moller, in preparation)
on patients of the same diagnostic group
was performed to replicate the findings con-
cerning outcome and predictors, especially
the findings on prognostic scales (see p. 17)
on the one hand and to compare the out-
come results with affective psychoses and
schizoaffective psychoses on the other (see
p.19). This sample included 61 patients who
had been treated in the Max Planck Insti-
tute of Psychiatry. Most of the patients (55)
were diagnosed as schizophrenics (without
schizoaffective psychoses) and only six as
suffering from similar types of paranoid
psychoses (leD 297, 298.3). In this inves-
tigation, the follow-up period was 5-8
years.
At follow-up, two patients had com-
mitted suicide. Sufficient information for
the statistical analyses could be obtained
from 46 patients: 32% of them had never
been admitted to a psychiatric hospital
prior to their treatment in the Max Planck
Institute (index admission); 59% of the pa-
tients were male, 41 % female; 83% were 40
years old or younger. Only 24% were mar-
ried.
The outcome results are rather similar to
the first sample; concerning the longitudi-
nal criteria, they were somewhat worse,
perhaps due to the longer follow-up period.
Of the patients, 33% showed serious psy-
chopathological symptoms and/or disturb-
ances of social adjustment (GAS score be-
low 50); 26% were rated on the GAS below
40, indicating that they needed hospital
treatment; 49% of the patients were unable
to work for more than 1 year during the fol-
low-up period; 81 % had to be readmitted to
a psychiatric hospital; and: 30% were treat-
ed for more than 1 year as psychiatric in-
patients during 5 years after index treat-
ment. At follow-up, 26% were living in a
16 H. J. MOller et al.
Table 4. Product-Moment Correlations of the Most Important Predictors (Correlations with the G AS-
Score r > 0.35). Sample I (n = 74-78); Sample II = Replication Sample (n = 43-46)

Sample I Impair- Global Person- Impair- Para- Depres- Dura- Dura-

ment of psycho- ality mentof noid- sive- tion of tionof
level of patho- change work hallucin- apathet- occupa- psychi-
function-logical ability atory ic tional atric
ing state syn- syn- disinte- hospi-
(GAS) drome drome gration taliza-
(IMPS) (IMPS) tion
Duration of occupational
disintegration (5 years
beforelA) 0.38** 0.32** 0049*** 0048*** 0.28* 0.32* 0047*** 0.52***
Impairment of work
ability(lyearbeforeIA) 0.39** 0.34** 0046*** 0.30* 0.23* 0.23* 0.17 0.24
Personality change
(1 year before IA) 0042*** 0042*** 0.59*** 0040** 0.27* 0.19 0.29* 0.15
Psychopathological
state at discharge 0.35** 0047*** 0.19 0.22 0.26* 0.04 0.11 0.15

Sample II Impair- Global Person- Impair- Para- Depres- Dura- Dura-

ment of psycho- ality mentof noid- sive- tion of tionof
level of patho- change work hallucin- apathet- occupa- psychi-
function- logical ability atory syn- tional atric
ing state syn- drome desinte- hospi-
(GAS) drome gration taliza-
(IMPS) (IMPS) tion
Manic or depressive
component (first
manifestation) -0.38* -0.23 -0.26 -0.08 -0.26 -0.25 -0.06 -0.13
Lasting hetero-
sexual relation-
ship (IA) -0043* -0040* -0042* -0.32* 0.34 * -0.29 -0.22 -0040*
Duration of occu-
pational disinte-
gration (5 years
beforelA) 0.40* 0.38 0.52** 0048** 0.20 0.44** 0.33* 0.56***
Psychopatholog-
ical state at dis-
charge 0.36* 0046** 0049** 0.50** 0.12 0043* 0.44** 0.27
* = P < 0.05; ** = P < 0.01; *** = P < 0.001; IA = index admission

nursing home or were psychiatric inpa-
tients.
As to the predictors of outcome, some
findings could be replicated. Of special in-
terest is that the duration of professional
disintegration during the 5 years before in-
dex admission and the psychopathological
state at discharge again were found among
the best predictors (Table 3). Also, some
other characteristics again could prove their
predictive value, such as a lasting hetero-
sexual relationship. Some others were not
replicated as predictors concerning the
GAS score, e.g., age at onset of the illness
and age at first psychiatric hospitalization,
self-rating of paranoid tendencies at dis-
charge, personality change before index ad-
mission, and depressive-apathetic syn-
drome at discharge. Thus, the known fact
that results of predictor analyses depend in-
tensively on the specific characteristics of
the sample becomes obvious.
 Outcome and Prediction of Outcome in Schizophrenia
Optimizing the Predictive Possibilities by
Prognostic Scales
To increase the outcome variance explain-
ed by single predictors and to make the pre-
diction more stable concerning different
samples and different outcome criteria,
prognostic scales were applied (Moller,
Schad, and Zerssen, 1984a, 1984b). Four
scales communicated in the literature were
tested concerning their predictive validity:
1. The Gittelman-Klein scale of prem-
orbid adjustment (Gittelman-Klein and
Klein, 1969)
2. The Phillips scale of premorbid adjust-
ment (Phillips, 1966) in the short version
of Harris (Harris, 1975)
3. The Goldstein scale of pre morbid adjust-
ment (Rodnick and Goldstein, 1974)
4. The Strauss-Carpenter prognostic scale
(Strauss and Carpenter, 1974)
All these scales were used in the version
published in the appendix of the review by
17
Kokes et al. (1977), apart from the fact that
we used an inverse scoring for the Strauss-
Carpenter scale. The data for these scales
had to be gathered from the records so that
in this respect the study is complicated by
the typical problems of retrospective ap-
proaches. Fortunately, the records of the
Psychiatric Department of the Max Planck
Institute of Psychiatry (MPIP) are rather
well structured and informative so that in
most cases the necessary data were availa-
ble. In this context, it has to be mentioned
that some of these scales, such as the Phil-
lips scale, were developed with the special
aim of making it possible to obtain an evalu-
ation on the basis of data from case records.
To avoid a potential bias, the evaluation of
the records was made by a psychiatrist who
was not informed about the outcome data
of the patients.
In addition, four new prognostic scores
were developed and tested concerning their
predictive validity. Based on the results of

Table 5. Items of the Self-Constructed Prognostic Scores

Score 1 Score 2 Score 3 Score 4

Premorbid working dysfunction x
No precipitating factors before first manifestation x x x
Younger age at first hospitalization x
No signs of manic depressive disorder at first
manifestation x
Duration of occupational disintegration
(5 years before index admission) x x x
Duration of psychiatric hospitalization
(5 years before index admission) x
No lasting partnership at index admission x
Impairment of working performance
(during the year after index admission x x x
Residual syndrome (personality change)
prior to index admission x
Poor psychopathological state at discharge x x x
IMPS superfactors
"Organic syndrome" at discharge" x
"Depressive, phobic-compulsive
syndrome"*" at discharge (- ) x
IMPS factors
"Disorientation" at admission ( - ) x
"Obsessional-phobic" at admission ( - ) x
"Retardation and apathy" at discharge x
"Paranoid projection" at discharge x
"Motor disturbances" at discharge x
" IMPS factor "disorientation" + IMPS factor "retardation and apathy";
*" IMPS factor "anxious depression" + IMPS factor "obsessional phobic"
(-) = the value (10% of the theoretical score = 1,20% = 2, etc.) must be subtracted from the total
score
18 H. J. Moller et a1.
univariate and multivariate analyses of pre-
dictors concerning the first sample of pa-
tients (Moller et aI., 1981b, 1982a, 1982b),
these prognostic scores (Moller, Schmid-
Bode, and Zerssen, 1984c) were created by
combinations of five to eight unweighted
predictors (Table 5). The first one contain-
ed seven variables, the predictive signifi-
cance of which was proven by univariate
analyses. They all correlated relatively
highly (r > 0.35) with the GAS score and/or
they correlated significantly with several
other outcome criteria. The second score
consisted of five variables, which had emer-
ged from stepwise mUltiple regression anal-
ysis as the best predictor combination for
the GAS score. Score 4 was also based on
the results of stepwise multiple regression
analyses. It contained all those predictor
variables which had repeatedly appeared
among the combination of the five best pre-
dictors concerning the different outcome
criteria. The third score was based on the
stepwise multiple regression analysis of the
12 IMPS syndrome scores at admission and
discharge. It was constructed by the combi-
nation of the five best predictors concerning
the GAS score: total score of "disorienta-
tion" and "obsessional phobia" at admis-
sion, total score of "retardation and apa-
thy," "paranoid projection," and "motor
disturbances" at discharge. This was the
only scale that consisted of psychopatho-
logical data alone, whereas the other scores
contained psychopathological data as well
as sociodemographic data, data on social
adjustment, and illness-related informa-
tion.
Considering the main outcome criterion,
the GAS score, the Strauss-Carpenter scale
obtained the best results (r = 0.43) in the
first sample. The predictive value of the
Gittelman-Klein scale and of the Phillips
scale were somewhat lower (about
r = 0.30); the Goldstein scale failed to prove
a significant correlation with the global lev-
el of functioning (Table 6).

Table 6. Product-Moment Correlations of Prognostic Scales and Outcome Criteria

Variables Impair- Global Person- Impair- Para- Depres- Dura- Dura-

mentof psycho- ality mentof noid- sive- tionof tionof
level of patho- change work hallucin- apathet- occupa- psychi-
function- logical ability atory ic- tional atric
ing state syn- syn- desinte- hospi-
(GAS) drome drome gration taliza-
(IMPS) (IMPS) tion
Gittelmann- 0.28* 0.20 0.50*** 0.30* 0.21 0.36** 0.19 0.34*
Klein scale 0.28 0.28 0.35* 0.13 0.15 0.37* 0.26 0.34*
Goldstein -0.19 -0.07 -0.38** -0.31 -0.14 -0.38* -0.22 -0.31 *
scale -0.24 -0.33 -0.39* -0.12 0.19 -0.35 -0.29 -0.31
Phillips 0.29* 0.22 0.43*** 0.36** 0.19 0.38*** 0.28* 0.46***
scale 0.46** 0.48** 0.48** 0.22 0.12 0.48*** 0.34* 0.36*
Strauss-Car- 0.43*** 0.42*** 0.54*** 0.37** 0.38** 0.39*** 0.32** 0.49***
penter scale 0.41** 0.44** 0.52*** 0.40** 0.21 0.36* 0.40** 0.46**
Score 1 0.53*** 0.51*** 0.55*** 0.45** 0.36* 0.28 0.32* 0.36*
0.44* 0.50* 0.60* 0.49* 0.08 0.45* 0.47* 0.49*
Score 2 0.56*** 0.54*** 0.64*** 0.54*** 0.37** 0.29* 0.38** 0.52***
0.43* 0.49* 0.56* 0.43* 0.07 0.43* 0.39* 0.42*
Score 3 0.32* 0.27* 0.35** 0.39*** 0.33** 0.27* 0.24* 0.27*
0.21 0.30* 0.26 0.22 -0.08 0.14 0.23 0.36*
Score 4 0.62*** 0.55*** 0.72*** 0.76*** 0.39** 0.38* 0.53*** 0.51***
0.43* 0.57* 0.59* 0.54* 0.09 0.46* 0.49* 0.47*
Correlations above: sample I (n = 55-57); correlations below: sample II (n = 35-46).
* P<0.05 ** P < 0.01 *** P < 0.001
 Outcome and Prediction of Outcome in Schizophrenia
The Strauss-Carpenter scale had also the
highest predictive power from the view-
point of the magnitude and the number of
significant correlations with the other out-
come criteria. The scale reached prognostic
significance for all eight outcome criteria
within a range of r = 0.32 to r = 0.54. Also,
the highest correlation with an outcome cri-
terion was achieved by the Strauss-Carpen-
ter scale: the correlation with personality
change (r = 0.54). In comparison to this
scale, the Gittelman-Klein scale proved
only significant correlations with five, the
Phillips scale only with six and the Gold-
stein scale only with three outcome criteria.
The newly developed scores showed bet-
ter results in the first sample (Table 6).
There was a significant and rather close cor-
relation with nearly all outcome criteria.
Concerning the global level of functioning,
score 4 obtained the best results (r = 0.62),
closely followed by scores 1 and 2. Score 3,
which consisted only of psychopathological
data, was less significant. Considering the
other outcome criteria, scores 1, 2, and 4
also seemed to be rather equivalent in their
potential predictive power. They showed
rather close correlations to nearly all out-
come criteria. Only score 3 has a lower, but
mostly significant relationship to the other
outcome criteria.
It could be expected that the newly devel-
oped scales would show better results than
the older prognostic scales in the original
sample of patients, in which they have been
developed. To test their predictive value
more realistically, the data of the cross-vali-
dation study on the second sample have to
be considered (Table 6). Scores 1, 2, and 4
reached highly significant substantial cor-
relations to almost all outcome criteria.
Score 3 largely failed to prove predictive
significance. There were only significant
correlations between score 3 and the out-
come criteria "global psychopathological
state" and "duration of psychiatric hospi-
talization". Scores 1, 2, and 4 were of a bet-
ter predictive value considering the magni-
tude and frequency of significant correla-
tions with the outcome criteria than the best
single predictors. The superiority of score 4
which had been found in the original sample
was not replicable.
19
In comparison with the newly developed
scores, the prognostic scales from the litera-
ture obtained slightly less favorable results
(Table 6) also in the replication sample. In
general, however, they reached the same
predictive level as in the former sample.
Again, the Strauss-Carpenter scale proved
to have the highest prognostic value, fol-
lowed closely by the Phillips scale.
It can be summarized that the four prog-
nostic scales described in the literature as
well as three of the newly constructed prog-
nostic scores gave better results in general
than any single predictive variable. The cor-
relations with different outcome criteria
were mostly comparatively close, e.g., the
Strauss-Carpenter scale was able to explain
16% of outcome variance. It seems that a
multidimensional approach, taking into ac-
count many different aspects, ranging from
social functioning to illness-related varia-
bles, reaches a better prognostic value than
approaches that regard less complex con-
cepts, such as premorbid adjustment orpsy-
chopathological syndromes. Such scales
may have a comparatively high predictive
value if certain dimensions and aspects are
regarded. However, even by adding more
and more items to a prognostic scale, a cer-
tain maximum can usually not be exceeded.
Perhaps a higher number of items can guar-
antee a better stability of the predictive
power in different samples.
Prognosis of Schizophrenia Compared to
Other Endogenous Psychoses and to the
Average Population
The outcome of schizophrenic or similar
psychoses (second sample) was compared
to the outcome of affective or schizoaffec-
tive psychoses and to controls from the
average population.
Twenty-two former inpatients with schi-
zoaffective psychoses (ICD 295.7) (Zaudig
and Vogl, in prep.) and 35 former inpa-
tients with affective psychoses (ICD 296)
(Cording-T6mmel, in preparation) were
also followed-up 5-8 years after their hospi-
tal treatment at the MPIP. The mean age
was higher for patients with affective psy-
choses (44 years at the time of follow-up)
than for the patients with schizophrenia (37
years) and schizoaffective psychoses (38
20 H. J. Moller et al.
50%
40
30
20
10
10 20 30 40 50 60
o Schizophrenia
li'I Affective psychosis
• Schizoaffective psychosis
years). There were more female patients
among affective psychoses (57%) and schizo-
affective psychoses (68%) (schizophre-
nia: 41 % females). In addition, the schizo-
phrenic sample was compared concerning
social adjustment to 46 nonpatients
matched for sex, age, and marital status
from a random sample of 499 persons of the
general population who were assessed by
the same standardized method. To measure
social adjustment, a modified version of the
social interview schedule (SIS) (Clare and
Carins, 1978; Faltermaier, 1982) was ap-
plied.
At follow-up, 77% of the patients with
affective psychoses were married compared
to 17% of schizophrenic and 27% of schizo-
affective patients. The global level of func-
tioning as measured by the GAS also show-
ed marked differences between the diagnos-
tic groups. The schizophrenic patients
were, on an average rated in the lower half
of the 1oo-point GAS score range, whereas
patients with affective of schizo affective
psychoses were more frequently repres-
ented in the upper range of scores (Fig-
ure 1).
Impairment of working performance in-
cluding study and housekeeping during the
last year of the follow-up period can be seen
from Figure 2. No impairment was shown
by 95% of schizo affective patients, 71% of
Fig. 1. Global level of func-
tioning (GAS) 5-8 years after
discharge.
70 80 90 100 GAS
patients with affective psychoses, and 33%
of schizophrenics. In the worst group
("completely unable to work or help with
the housekeeping"), only schizophrenic pa-
tients (24%) were found. The duration of
occupational disintegration during the fol-
low-up period exhibited a similar tendency:
the schizophrenics had the longest duration
of inability to work.
These results might be partially biased by
the discrepancies of sociodemographic vari-
ables mentioned above. In this respect, the
comparison of SIS data between patients
and matched controls attracts more confi-
dence. Furthermore, the structured SIS al-
lows a more detailed view on social adjust-
ment (Figure 3). The objective conditions
of occupation (for the 50% of schizophrenic
patients who were still working or studying)
and the opportunities for social contacts
and leisure activities were not markedly dif-
ferent for schizophrenic patients and
matched controls. Regarding their social
management, however, significantly more
schizophrenic patients had marked or severe
problems in occupation, social contacts,
and leisure activities. Subjective dissatisfac-
tion was also more frequently expressed by
schizophrenic patients, but here the dif-
ferences were not as marked. This discrep-
ancy between the ratings on management
on the one side and the ratings on satisfac-
 Outcome and Prediction of Outcome in Schizophrenia 21

1::%1
80

70

60

50

40

30

20
Fig. 2. Impairment of working 10
performance 5-8 years after dis-
charge.
o ~~~ ______ ~~ ____ ~~ ____ ~~ ______ L L_ __ _ _

No impairment Moderate impairment Inability to work

Slight impairment Serious impairment
o Schizophrenia
G Affective psychosis
• Schizoaffective psychosis

Social contacts and

%
Occupation leisure activities
60
50
4fJ
Difficult
objective 30
conditions 20
10

Social Leisure
Occupation contacts activities
%
60
50
Problems .0
in social 30
management
20
10

Social Leisure
Occupation contacts activities
%
60
50

Fig. 3. Percentage of subjects with 4fJ

Dissatisfaction 30
marked or severe difficulties in
objective conditions, social man- 20
agement, and subjective satisfac- 10
tion according to the SIS.

• Patients (n = 43, occupation: n = 23)

o Controls (n = 46, occupation: n = 40)
22 H. J. Moller et al.
tion on the other side might be explained
partially by affective blunting and schizoid
personality traits.
In contrast to the schizophrenic patients,
the patients with schizo affective psychoses
or affective psychoses were not significantly
different from the controls concerning the
SIS dimensions mentioned above.
It can be concluded that patients with
schizophrenic psychoses have a poorer out-
come than patients with affective or schizo-
References
Achte, K. A. (1967). On prognosis and rehabili-
tation in schizophrenic and paranoid psycho-
ses. Acta Psychiatrica Scandinavica, Supple-
ment 196.
Achte, K. A. (1980). Verlaufund Prognose schi-
zophrener Psychosen in Helsinki. In: G. W.
Schimmelpfenning (Ed.), Psychiatrische Ver-
laufsforschung: Methoden und Ergebnisse (pp.
144-158). Bern: Huber.
Affleck, J. W., Bums, J., and Forrest, A. D.
(1976). Long-term follow-up of schizophrenic
patients in Edingburgh. Acta Psychiatrica
Scandinavica, 53, 227-237.
Angst, J., Felder, W., and Lohmeyer, B. (1980).
Course of Schizoaffective psychoses: Results
of a follow-up study. Schizophrenia Bulletin, 4,
579-585.
Astrachan, B. M., Brauer, L., Harrow, M., and
Schwartz, C. (1974). Symptomatic outcome in
schizophrenia. Archives of General Psychiatry,
31,37-42.
Bland, R. C., Parker, J. H., and Om, H. (1978).
Prognostic predictors and outcome. Archives
of General Psychiatry, 35, 72-77.
Bleuler, M. (1972). Die schizophrenen Geistes-
storungen im Lichte langjiihriger Kranken- und
Familiengeschichten. Stuttgart: Thieme.
Cancro, R., Fox, N., and Shapiro, L. E. (1978).
Behandlungstechniken bei Schizophrenie.
Munchen: Reinhardt.
Clare, A. W., and Cairns, V. E. (1978). Design,
development and use of a standardized inter-
view to assess social maladjustment and dys-
function in community studies. Psychological
Medicine, 8, 589-604.
Croughan, J. L., Weiner, A., and Robins, E.
(1974). The group of schizoaffective and relat-
ed psychoses - critique, record, follow-up and
family studies. II. Record studies. Archives of
General Psychiatry, 37: 632-637.
Davis, J. M., Schaffer, C. B., Killian, G. A.,
Kuard, C., and Chan, C. (1980). Important
issues in the drug treatment of schizophrenia.
affective psychoses. The outcome of schizo-
affective psychoses seems to be similar to
that of affective psychoses, which supports
the hypothesis that this diagnostic group
should be classified separately or as a sub-
group of the affective psychoses rather than
as a subgroup of schizophrenic psychoses, a
hypothesis based on the results of other fol-
low-up studies as well as on genetic findings
(Tsuang et al., 1979; Angst, Felder, and
Lohmeyer, 1980).
Schizophrenia Bulletin, 6, 70-87.
Eitinger, L., Laane, C. L., and Langfeldt, G.
(1958). The prognostic value ofthe clinical pic-
ture and the therapeutic value of physical treat-
ment in schizophrenia and the schizophreni-
form states. Acta Psychiatrica Scandinavica,
33,35-58.
Faltermeier, T. (1982). Zusammenhiinge zwi-
schen sozialen Faktoren und psychiatrischer
Morbiditat: Die Bedeutung von Skalen fUr
soziale Integration in epidemiologischer For-
schung und erste Ergebnisse mit einer deut-
schen Fassung der "Social Interview Sched-
ule". Paper presented at the 2nd Congress for
Clinical Psychology and Psychotherapy, Ber-
lin.
Feinsilver, D. B., and Gunderson, J. C. (1972).
Psychotherapy for schizophrenics - is it indi-
cated? A review of the relevant literature. Schi-
zophrenia Bulletin, 6, 11-23.
Gittelman-Klein, R., and Klein, D. F. (1969).
Premorbid asocial adjustment and prognosis in
schizophrenia. Journal of Psychiatric Re-
search, 7, 35-53.
Goldberg, S. C., Schooler, N. R., Hogarty, G.
E., and Roper, M. (1977). Prediction ofrelapse
in schizophrenic out-patients treated by drug
and sociotherapy. Archives of General Psy-
chiatry, 34, 171-184.
Grinspoon, L., Ewalt, J. R., and Shader, R. J.
(1968). Psychotherapy and Pharmacotherapy
in chronic schizophrenia. American Journal of
Psychiatry, 124, 1645-1651.
Gunderson, J. G., Carpenter, W. T., and
Strauss, J. S. (1975). Borderline and schizo-
phrenic patients: A comparative study. Ameri-
can Journal of Psychiatry, 132, 1257-1264.
Hargreaves, W. A., Glick, I. D., Drues, J.,
Showstack, J. A., and Geigenbaum, E. (1977).
Short vs. long hospitalisation: VI. Two-year
follow-up results for schizophrenics. Archives
of General Psychiatry, 34, 305-311.
Harris, J. G., Jr. (1975). An abbreviated form of
 Outcome and Prediction of Outcome in Schizophrenia
the Phillips rating scale of premorbid adjust-
ment in schizophrenia. Journal of Abnormal
Psychology, 84,129-137.
Hawk, A. B., Carpenter, W. T., and Strauss, J.
S. (1975). Diagnostic criteria and five year out-
come in schizophrenia: A report from the In-
ternational Pilot Study of Schizophrenia. Ar-
chives of General Psychiatry, 32, 343-347.
Hogarty, G. E. (1977). Treatment and the course
of schizophrenia. Schizophrenia Bulletin, 3,
587-599.
Hogarty, G. E., Goldberg, S. c., Schooler, N.
R., and Ullrich, R. F. (1974). Drug and socio-
therapy in the aftercare of schizophrenic pa-
tients. II. Two-year relapse rates. Archives of
General Psychiatry, 31, 603-607.
Huber, G., Gross, G., and Schiittler, R. (1979).
Schizophrenie. Eine Ver!aufs- und psychiatri-
sche Langzeitstudie. Berlin, Heidelberg, New
York: Springer.
Kokes, R. F., Strauss, J. S., and Klorman, R.
(1977). Premorbid adjustment in schizophre-
nia. Measuring premorbid adjustment: The
instruments and their development. Schizo-
phrenia Bulletin, 3,186-213.
Langfeldt, G. T. (1957). The prognosis in schizo-
phrenia. Acta Psychiatrica Scandinavica,
(Suppl. 110).
Lorr, M. (1974). Assessing psychotic behavior by
the IMPS. In: P. Pichot, P., and R. Olivier-
Martin (Eds.), Psychological measurements in
psychopharmacology. Modern problems in
Pharmacopsychiatry: Vol. 7 (pp. 50-63).
Basel: Karger.
Lindelius, R. (1970). A study of schizophrenica.
A clinical, prognostic and family investigation.
Acta Psychiatrica Scandinavica, (Suppl. 216).
May, P. R. A., Tuma, A. H., and Dixon, W. J.
(1977). For better or for worse? Outcome vari-
ance with psychotherapy and other treatments
for schizophrenia. The Journal of Nervous and
Mental Diseases, 165,231-239.
Moller, H. J., Zerssen, D. v., Werner-Eilert, K.,
and Wiischner-Stockheim, M. (1981a). Psy-
chopathometrische Verlaufsuntersuchungen
an Patienten mit Schizophrenien und verwand-
ten Psychosen. Archiv fUr Psychiatrie und Ner-
venkrankheiten, 230, 274-292.
Moller, H. J., Zerssen, D. v., Wiischner-Stock-
heim, M., and Werner-Eilert, K. (1981b). Die
prognostische Bedeutung psychopathometri-
scher Aufnahme- und Entlassungsbefunddaten
schizophrener Patienten. Archiv fUr Psychiat-
rie und Nervenkrankheiten, 231, 13-34.
Moller, H. J., Werner-Eilert, K., Wiischner-
Stockheim, M., and Zerssen, D. v. (1982a).
Relevante Merkmale fUr die 5-Jahres-Pro-
gnose von Patienten mit schizophrenen und
verwandten Psychosen. Archiv fUr Psychiatrie
und Nervenkrankheiten, 231, 305-322.
23
Moller, H. J., Zerssen, D. v., Werner-Eilert, K.,
and Wiischner-Stockheim, M. (1982b). Out-
come in schizophrenic and similar paranoid
psychoses. Schizophrenia Bulletin, 8, 99-108.
Moller, H. J., Schar!, W., and Zerssen, D. v.
(1984a). StOrungen der pramorbiden sozialen
Adaptation als Priidiktor fUr die Fiinfjahres-
prognose schizophrener Psychosen. Nerven-
arzt, 55, 358-364.
Moller, H. J., Schar!, W., and Zerssen, D. v.
(1984b). Strauss-Carpenter-Skala: Uberprii-
fungihresprognostischenWertesfUrdas5-Jah-
res-"Outcome" schizophrener Patienten. Eu-
ropean Archives of Psychiatric and Neuro-
logical Sciences, 234, 112-117.
Moller, H. J., Schmid-Bode, W., and Zerssen,
D. v. (1984c). Entwicklung neuer Prognose-
Indizes fUr Patienten mit schizophrenen
Erkrankungen und Uberpriifung ihrer pradik-
tiven Validitat. Fortschritte der Neurologie -
Psychiatrie, 52, 421-427.
Mosher, L. R., Keith, S. J. (1979). Research on
the psychosocial treatment of schizophrenics:
A summary report. American Journal of Psy-
chiatry, 136,623-631.
Phillips, L. (1966). Social competence, the proc-
ess-reactive distinction, and the nature of men-
tal disorder. In: P. H. Hoch, and J. Zubin
(Eds.), Psychopathology of Schizophrenia.
(pp. 471-484). New York: Grune and Strat-
ton.
Pokorny, A. D., Thornby, J., Kaplan, H. W.,
and Ball, D. (1976). Prediction of chronicity in
psychiatric patients. Archives of General Psy-
chiatry, 33, 932-937.
Renton, C. A., Affleck, J. W., Carstairs, G. M.,
and Forrest, A. D. (1965). A follow-up of schi-
zophrenic patients in Edinburgh. Acta Psy-
chiatrica Scandinavica, 39, 548-581.
Rodnick, E. H., and Goldstein, M. J. (1974).
Premorbid adjustment and the recovery of
mothering function in acute schizophrenic
women. Journal of Abnormal Psychology, 83,
623-628.
Schooler, N. R., Goldberg, S. c., Boothe, H.
and Cole, J. O. (1967). One year after dis-
charge: Community adjustment of schizophre-
nic patients. American Journal of Psychiatry,
123,986-995.
Spitzer, J., Endicott, R. L., and Fieiss, L. (1976).
The global assessment scale. A procedure for
measuring overall severity of psychiatric dis-
turbances. Archives of General Psychiatry, 33,
776-771.
Stephens, J. H. (1970). Long-term course and
prognosis in schizophrenia. Seminars in Psy-
chiatry, 2, 464-485.
Stephens, J. H., and Astrup, C. (1965). Treat-
ment outcome in "process" and "non-process"
schizophrenic treated by "A" and "B" types of
24 H. J. Moller et al.
therapists. Journal of Nervous and Mental Dis-
eases, 140,440-465.
Strauss, J. S., and Carpenter, W. T. (1974). The
prediction of outcome in schizophrenia. II:
Relationship between predictor and outcome
variable. Archives of General Psychiatry, 31,
37-42.
Strauss, J. S., and Carpenter, W. T. (1977). The
prediction of outcome in schizophrenia. III. 5-
year outcome and its predictors. Archives of
General Psychiatry, 34, 159-163.
Tsuang, M. T., and Winokur, G. (1974). Criteria
for subtyping schizophrenia. Clinical differen-
tiation of hebephrenic and paranoid schizo-
phrenia. Archives of General Psychiatry, 31,
43-47.
Tsuang, M. T., Woolson, R. F., and Fleming, J.
A. (1979). Long-term outcome of major psy-
choses. I. Schizophrenia and affective disor-
ders compared with psychiatrical symptom-
free surgical conditions. Archives of General
Psychiatry, 36, 1295 -1301.
Vaillant, G. (1962). The prediction of recovery in
schizophrenia. Journal of Nervous and Mental
Disease, 135, 534-543.
Vaillant, G. (1964). Prospective prediction of
schizophrenic remission. Archives of General
Psychiatry, 11,509-518.
Vaughn, C. E., and Leff, J. P. (1976). The influ-
ence of family and social factors on the course
of psychiatric illness: A comparison of schizo-
phrenic and depressed neurotic patients. Brit-
ish Journal of Psychiatry, 129, 125 -137.
Watt, D. c., Katz, K., and Shepherd, M. (1983).
The natural history of schizophrenia: A 5-year
prospective follow-up of a representative
sample of schizophrenics by means of a stand-
ardized clinical and social assessment. Psych-
ological Medicine, 13,663-670.
Wellner, J., and Stromgren, E. (1958). Clinical
and genetic studies on benign schizophreni-
form psychoses based on follow-up. Acta Psy-
chiatrica Scandinavica, 377 - 399.
Wittenborn, J. R., McDonald, D. c., and Mau-
rer, H. S. (1977). Persisting symptoms in schi-
zophrenia predicted by background factors.
Archives of General Psychiatry, 34,
1059-1061.
World Health Organization. (1979). Schizophre-
nia. An International Follow-up Study.
Chichester: Wiley.
Zerssen, D. v., and Koeller, D. M. (1976). Kli-
nische Selbstbeurteilungs-Skalen (KSb-S) aus
dem Munchner Psychiatrischen Informations-
System (PSYCHIS Munchen). Manuale.
Weinheim: Beltz.
II. Family Attributes
2. An Introduction to EE Measurement
and Research
J. M. Hooley
Introduction
Expressed emotion (EE) is currently one of
the most important measures of family
functioning available to clinicians and re-
searchers working with schizophrenic pa-
tients. Based as it is on the number of criti-
cal comments and level of emotional overin-
volvement spontaneously expressed by a
relative over the course of the Camberwell
Family Interview (CFI) , the EE rating is
considered to reflect the attitude of the re-
lative toward the psychiatrically ill family
member.
Research conducted in recent years has
An Historical Perspective
The initial stimulus for the development of
the EE construct came originally from work
begun in England. Working with more than
200 discharged male psychiatric patients,
most of whom had a diagnosis of schizophre-
nia, Brown and his collegues (Brown, 1959;
Brown, Carstairs, and Topping, 1958) ob-
served that the type of living group to which
patients were discharged seemed to be as-
sociated with how well or how poorly they
fared upon return to the community. Of
particular interest was the finding that pa-
tients who returned to live in lodgings or
with siblings were more psychiatrically
healthy over the course of the subsequent
year than patients who were discharged into
the parental or matrimonial home. Inter-
estingly, it was also noted that the amount
of time patients spent in contact with their
relatives was another important factor relat-
ed to outcome; patients who had only limit-
ed contact with their relatives were less like-
ly to suffer a relapse than patients who had
provided good evidence that the construct
of EE is both valid and can be measured
reliably. Moreover, not only do familial
levels of EE predict relapse in schizophre-
nic patients, but EE also appears to have
some predictive validity for other psy-
chiatric populations, specifically the depres-
sed. This introductory chapter will discuss
the development of the EE construct and its
measurement and focus on some of the re-
search which has made an important contri-
bution to our current understanding of this
valuable concept.
high levels of contact with their families.
That spending a lot of time with close re-
latives might not necessarily be beneficial to
schizophrenics was a rather unusual find-
ing, and in subsequent research Brown
(Brown, Monck, Carstairs, and Wing,
1962) attempted to document more preci-
sely those aspects of living situations
having most impact on patient outcome. In
a later study, which again used male schizo-
phrenics, Brown employed a prospective
design and, in addition, interviewed the key
female relative of each patient on three dif-
ferent occasions. Levels of "emotion ex-
pressed," "hostility," and "dominance"
were assessed for patients and relatives, al-
though later analyses revealed only the rat-
ings made for relatives to have any predic-
tive utility. Briefly, the results showed that
patients who returned to homes high in hos-
tility or EE (high emotional involvement
homes) were more likely to suffer a psychia-
tric relapse over the course of a 1-year fol-
26 J. M. Hooley
low-up than patients who returned to live in
low emotional involvement homes.
This confirmation of the original finding
was encouraging and, over subsequent
years, Brown and his co-workers began to
focus attention on improving the interview
Measuring EE
Family levels of EE are assessed by means
of the CFI, a nonschedule, standardized in-
terview typically taking between 1 and 2 h
to carry out. The interview is conducted
with the patient's closest relative and is au-
diotaped for later coding. Over the course
of the interview, the relative (who is most
frequently a parent or a spouse) is asked
questions about the emotional climate in
the home in the 3 months prior to the pa-
tient's admission to the hospital. Informa-
tion is obtained about the events which had
taken place during this time and about the
relative's feelings toward the patient and his
or her condition. It is on the basis of the
emotions expressed while talking about the
patient that ratings of EE in the relative are
made.
A total of five different ratings are made.
These include criticism, hostility, emotion-
al overinvolvement, warmth, and positive
remarks. While only the first three of these
have featured prominently in the literature
thus far, brief descriptive details of all five
scales will be given below.
Criticism:
Criticism is the sum total of critical remarks
the relative makes about the patient during
the course of the interview. Remarks are
judged as critical on the basis of (a) their
content or (b) voice-tone changes. For a re-
mark to be coded as critical on grounds of
content, there has to be clear evidence of
dislike, disapproval, or resentment on the
part of the speaker, e.g., "It's annoying
coming in at night and finding breakfest
things still on the table." A respondent can
"opt out" of a remark which is critical only
in content if he or she implies that imper-
fections in himlherself are in some way re-
sponsible for the reaction shown, e.g., "It
really annoys me when she does that but
assessments made and establishing reliable
techniques of EE measurement (Brown
and Rutter, 1966; Rutter and Brown,
1966). The result was the development of
the Camberwell Family Interview, describ-
ed below.
that's probably because I'm a rather intoler-
ant person." In these cases there is no
dicrect blame of the patient and the remark
is consequently not considered critical.
Remarks which are rated as critical be-
cause they involve changes in the speed, in-
flection, or pitch of voice of the speaker
cannot be retracted in this way, however.
While it is not easy to convey the essence of
critical remarks which involve tone change
using printed examples, the following il-
lustration, taken from Kuipers (1979), is
helpful.
The words 'every morning he goes to buy the
newspaper' have innocuous content but can be
said in a variety of ways to give different emo-
tional meanings. If they were intended to be criti-
cal, this would be conveyed by writing stress
marks - 'every morning he goes to buy the news-
paper.'
Emphasis on the first word thus conveys cri-
ticism - in this example, criticism of the
regularity with which the patient goes out to
buy the newspaper.
Hostility:
Unlike criticism, which, as can be seen from
the above illustration is situation specific,
hostility involves a greater generalization of
negative feeling and usually includes re-
marks critical of the patient himself rather
than of his particular actions or behaviors.
Examples of hostility include generalized
criticism (e.g., "He is stupid. Everything he
does is stupid. ") or of direct rejection ("It's
just unbearable sometimes to be with her").
While criticism is measured by simplyrecor-
ding the frequency of critical remarks the
relative makes about the patient, hostility is
rated using a three-point scale. Although
initially incorporated into early definitions
of EE (Brown, Birley, and Wing, 1972),
hostility ratings are now no longer used in
this way - principally because of the strong
 Introduction to EE Measurement and Research
association between hostility and criticism
and the lack of predictive power of the hos-
tility ratings when used alone.
Emotional Overinvolvement:
Emotional overinvolvement reflects a dra-
matic or exaggerated emotional response to
the patient's illness (e.g., "I said what hos-
pital? Just imagine how I felt! I thought I'd
have to go there myself from the shock.") or
behavior which is excessively overprotec-
tive ("I quit my job and went into debt so I
can be home in case he ever needs me.").
Like hostility, emotional overinvolvement
is rated on a point scale, although typically
this involves five or six points rather than
three.
The two final assessments, warmth and
positive remarks, have been given relatively
little attention in the EE literature, probab-
ly in large part because empirically they
have been found to add little to the predic-
tive power of EE ratings made using criti-
cism and overinvolvement. However, they
are mentioned here for completeness and
also because, at some future time, these
positive aspects of the patient-relative rela-
EE and Schizophrenia: Recent Research
Serveral studies have now provided evi-
dence that levels of relatives' expressed
emotion, measured while patients are still
in the hospital, are reliably associated with
patients' psychiatric relapse in a 9-month
period following discharge (Brown et aI.,
1972; Vaughn and Leff, 1976; Vaughn, Sny-
der, Freeman, Jones, Fallo, and Liberman,
1982). Typically, a cutoff of six or seven cri-
tical comments and/or marked evidence of
emotional overinvolvement is used to as-
sign relatives to the high or low EE groups.
While relapse rates for patients living in
high EE families are generally around 55%,
those for patients who live with low EE re-
latives approximate 15%.
In all EE research, relapse is determined
using the present state examination (PSE)
(Wing, Cooper, and Sartorius, 1974) - a
semistructured standardized interview
widely used for psychiatric assessment. Pa-
tients are considered to have relapsed if
27
tionship may be important in helping us un-
derstand why some patients do not relapse,
even when living within high EE families.
Like criticism, warmth is rated by focus-
ing on tonal change in the speaker. In this
case, however, the voice tone has to be
clearly positive. Positive remarks, on the
other hand, are defined primarily by con-
tent and reflect unambiguous praise or ap-
preciation for some characteristic or behav-
ior of the patient (e.g., "He's a wonderful
husband."). Although the rating for posi-
tive remarks is based on the number of re-
marks occurring during the interview,
warmth is rated using a five - or six-point
scale.
Even though the principal component of
the EE rating (criticism) is rated primarily
by focusing on what can often be quite sub-
tle tone changes in the voice of the relative,
with adequate training it is possible for
highly reliable ratings to be made. EE train-
ing typically takes about 3 weeks, and be-
fore a trainee is permitted to work inde-
pendently, interrater agreement on criti-
cism and emotional overinvolvement must
be at least 0.80 (Pearson product-moment
correlation) .
they show either a change in state (i.e.,
from non schizophrenic to schizophrenic)
during follow-up or exhibit a marked
exacerbation of symptoms. One important
point to note is that since relapse does not
necessarily involve readmission to a hospi-
tal, relatives' attitudes toward hospitaliza-
tion or their ability to cope with a pychiatri-
cally ill family member without medical as-
sistance are not factors which can confound
the association between relapse and levels
ofEE.
Since EE has been demonstrated to be
reliably associated with the risk of relapse in
English schizophrenic samples, it is particu-
larly encouraging to note that the associa-
tion between levels of EE and relapse re-
tains its predictive validity cross-culturally.
Work recently carried out on an American
population by Vaughn and her associates
(see Vaughn et aI., 1982; also Vaughn, Sny-
der, Jones, Freeman, and Falloon, 1984)
28 J. M. Hooley
used a sample of 54 schizophrenics from the
Los Angeles area. As with the British sam-
ples, family levels of EE were found to be
significantly associated with patients' 9-
month relapse rates, the percentages of re-
lapsing patients in the high and low EE
groups being remarkably similar to those
previously found for the English samples
(56% vs 17%).
Much work is now being conducted using
even more diverse cultural groups. As EE
Factors Interacting with EE Levels
At the present time, we know that a sub-
stantial number of patients fail to relapse,
even when exposed to high EE families. In-
creased knowledge about protective factors
which attentuate the impact of such rela-
tives is of evident benefit for preventive
psychiatry and might, in addition, lead to
the development of more sophisticated and
precisely targeted techniques of interven-
tion.
At the present time, there are data to sug-
gest that three factors do interact with rela-
tives' levels of EE and can serve to modify
the influence of high EE families. The first
of these is the amount of time patient and
relative spend in face-to-face contact. As
Total group
LowEE
(13%)
Subgroups
2 3
not
on drugs on drugs on drugs
(12%) (15%) (15%)
Fig. 1. Relapses rates (%) for patients who returned to high and low EE homes at 9-month follow-up
as a function of medication compliance and degree of relative-patient contact. Data combined from
Brown et aI. (1972) and Vaughn and Leff (1976) studies.
research extends, however, it is likely that
some instances of nonreplications will be
reported in the near future. The findings of
studies reporting a failure to replicate will
do much to benefit EE research generally
since they may provide some insights into
the precise mechanism by which EE opera-
tes within the relapse process. More de-
scriptive information about the types of pa-
tients who do or do not respond to high EE
influences will be of clear value.
Brown's earlier work had suggested
(Brown et al., 1962), in cases where schizo-
phrenic patients spend only limited
amounts oftime (less than 35 h/week) with
high EE relatives, some protection from
their negative effects seems to be afforted
(see Brown et aI., 1972; Vaughn and Leff,
1976).
Medication also seems to be beneficial
for patients exposed to a high EE relative.
When Vaughn and Leff (1976) analyzed
data from their own study together with
that collected by Brown et aI. (1972), they
found that while 78% of patients living in
high EE families and not taking phenothia-
zines relapsed during follow-up, only 25%
Low EE = 71 patients
High EE = 57 patients
High EE
(51%)
~
< 35 hrs > 35 hrs
A 4 5
A 6
not not
on drugs on drugs on drugs
(42%) (53%) (92%)
 Introduction to EE Measurement and Research
of patients also living in high EE homes but
regularly taking drugs did so.
Moreover, the protective effects of de-
creased contact time and regular mainte-
nance medication appear to be additive. As
Figure 1 shows, while patients in the high
EE group who take drugs regularly and
spend less than 35 h1week in direct contact
with their families have a relapse rate com-
parable to patients from low EE homes
(15%), the risk of relapse increases mar-
kedly in the absence of these two protective
factors. Thus, the highest rate of relapse
(92%) is found in patients who have a high
degree of contact with their high EE fami-
lies and who in addition are not taking neu-
roleptic drugs.
The third factor which appears to interact
with levels of EE concerns independent life
Psychophysiological Investigations
In an attempt to learn more about how high
EE effects might operate to influence re-
lapse rates in schizophrenic patients, re-
search has, in recent years, turned to psy-
chophysiological studies. These typically
assess peripheral autonomic responses to
high and low EE relatives, and recently
some interesting if conflictual data have
emerged.
In a study carried out in 1979 in a group of
schizophrenic patients in remission, Tarrier,
Vaughn, Lader, and Leff were able to de-
monstrate that while both high and low EE
patient groups showed highly aroused phy-
siological response patterns during a IS-min
period spent talking to a psychiatrist, short-
ly after their relatives entered the room, the
rates of spontaneous fluctuations in skin
conductance in the patients with low EE re-
latives quickly habituated and approached
normal levels. Patients with high EE rela-
tives on the other hand showed no such
habituation upon entry of the relative. In-
stead, their arousal levels remained high
throughout the full 30 min of recording.
While Sturgeon, Kuipers, Berkowitz, Tur-
pin, and Leff (1981) later claimed to have
replicated Tarrier et al.'s earlier finding, a
recently published reanalysis of these initial
data, together with data subsequently col-
lected on an additional ten patients (total
29
events. Research published in 1980 by Leff
and Vaughn investigated the factors asso-
ciated with a schizophrenic episode in 37
patients. While episodes of schizophrenia
in patients coming from high EE families
did not seem to be triggered by an excess of
independent life events, such events do
seem to be implicated in the relapses of pa-
tients coming from low EE homes. These
data thus suggest that a schizophrenic epi-
sode can be associated either with contin-
ued exposure to a high EE relative or by an
elevated number of independent life events
in the months preceding the onset of the
episode. Thus, both life events and critical
attitudes in relatives seem to constitute un-
acceptable forms of stress for such psychia-
trically vulnerable individuals.
n = 30), no longer appears to do so. Stur-
geon, Turpin, Kuipers, Berkowitz, and
Left's (1984) data now show no differential
effect of EE on the decline in electrodermal
activity after the relative joins the patient
and interviewer. Also, in contrast to Tarrier
et al., Sturgeon and his collegues report a
mean rate of skin conductance in the high
EE patient group which is almost double
that found in the low EE patients.
Since Sturgeon was using an acutely ill
patient group, all of whom had high levels
of face-to-face contact with their relatives,
and since all his recordings were made while
the patient was in the hospital rather than at
home, there are a number of differences
between the two studies which might ac-
count for the discrepant results. While it is
difficult to understand the implications of
the studies in the area of psychophysiology
to date (many of the results seem to depend
on the physiological index chosen and the
method of statistical analysis employed), it
does seem that patients do differ in their
electrodermal activity according to the EE
levels of their relatives. Whether these dif-
ferences appear after the relative enters the
room or whether they are more long-stand-
ing is not yet clear however.
Data from Valone, Goldstein, and Nor-
ton (1984) even suggest that mere antic-
30 J. M. Hooley
ipation of an interaction with a high EE rel-
ative can be arousing. Disturbed adoles-
cents who were waiting to interact with a
high EE parent showed higher levels of psy-
chophysiological reactivity than adoles-
cents who were anticipating interactions
with a low EE parents. Moreover, within
direct interactions, high EE parents and
adolescents become significantly more phy-
siologically aroused than parents and off-
spring in the low EE group, although prior
to the interaction no differences in reactivi-
ty level were evident in the parents.
Valone et al.'s data provide good evi-
dence that encounters with high EE parents
are more emotionally arousing than compa-
rable interactions involving low EE family
members. Given this, an hypothesis pro-
posed by Sturgeon and his co-workers is of
some interest. One explanation these re-
searchers offer for their finding that SCR
frequencies differ between the high and low
EE patient groups concerns a factor which
often precipitates relapse or onset in the
two groups. As has been mentioned earlier,
within the low EE group, a schizophrenic
episode is most frequently preceded by the
acute stress of an independent life event
(Leff and Vaughn, 1980). In the high EE
patient group, however, life events seem to
play a less important role in relapse, the
Characteristics of mgh and Low EE Relatives
High EE relatives are not easy to spot - at
least by the untrained observer (see Hoo-
ley, 1984). Despite being able to discrimi-
nate maritally dissatisfied individuals from
those who reported satisfactory marriages
by ratings made of couples' videotaped in-
teractions, ratings of positive and negative
behavior made by an observer blind to the
EE levels of the individuals concerned did
not reveal any significant differences across
EE groups. High EE spouses were thus not
rated as any less positive or any more nega-
tives in their interactions with patients than
were low EE relatives, nor were any signifi-
cant differences found in the ratings made
by patients interacting with such spouses.
Whatever it is that high EE relatives are
doing that is so potentially detrimental to
patients's progress after hospital discharge
more chronic stress associated with high
levels of contact with a high EE relative
seeming most crucial. Sturgeon et al. hy-
pothesize that the rise in SCR frequency
which takes place in low EE patients during
a psychiatric episode is a more transient
phenomenon than the rise which occurs in
high EE patients. They suggest this is be-
cause the former have been exposed to
more acute stress prior to the episode while
the latter group have endured life stress
which is more chronic. If electrodermal
activity is thus in some way associated with
vulnerability to relapse, this slow recovery
rate, argued to be characteristic of patients
in the high EE group, would render such
patients at increased risk of subsequent re-
lapse until remission levels can be estab-
lished.
While the above is an attractive formula-
tion and provides an alluring link between
several areas of EE research, it must be
stressed that at the present time the issue of
the relation between electrodermal activity
and vulnerability to schizophrenia cannot
be resolved. Clearly, however, psychophy-
siological studies have a valuable role to
play in EE research and are likely to pro-
vide some important insights into the
dynamics of the concept in future years.
is thus not clearly apparent to an untrained
observer. EE is therefore not measuring
anything obvious to the layman. It is by con-
trast a rather esoteric measure, the subtle
nature of which is reflected in the long train-
ing period necessary to rate reliably.
Nevertheless, insofar as it is tapping im-
portant aspects of family functioning, EE
remains a valuable construct. Given this,
however, the issue of what exactly the
measure is reflecting is central to our under-
standing of the construct. While the psycho-
physiological data described above and the
behavioral research to be described in
Chap. 8 are of evident value in illuminating
the nature of the construct, a content analy-
sis of the critical remarks made by relatives
has also been enlightening (Vaughn, 1977;
Vaughn and Leff, 1981).
 Introduction to EE Measurement and Research 31
On the basis of a detailed investigation of She's tending to walk the way she sees all the de-
interview material collected from relatives pressive patients walk here. I think she's very im-
of patients participating in their 1976 study, pressionable.
Vaughn and Leff (1981) identified four re- The final difference between high and low
sponse styles which discriminated high from EE individuals described by Vaughn and
low EE individuals. The first of these con- Leff concerns tolerance. Probably to a large
cerned intrusiveness. While low EE rela- degree due to their conviction that the pa-
tives tend to respect the patient's desire for tient is suffering from a legitimate illness,
social distance, high EE relatives seem to low EE relatives appear much more accept-
find this difficult. They frequently invade ing of the low levels of functioning typically
the patient's privacy and undermine feel- found in schizophrenic patients. High EE
ings of autonomy by making repeated eff- relatives, in contrast, often exert considera-
arts to "get through to" the patient. Often ble pressure on the patient to behave in
this takes the form of unsolicited help or ad- ways they consider more normal.
vice. The example below is typical.
He got up one Sunday morning and he sat in the She wouldn't go out voluntarily. We dragged her
chair and I said, "What's the matter? Don't you out sometimes and it was virtually dragging her
feel well?" "No," he said. So 1 said, "Can 1 get out.
you a drink?" "No." "Well it's not very warm.
Don't you think 1 ought to get something to put With reference to this last point, it is warth
on you?" "No." So I left him. 1went back again to noting that Miklowitz, Goldstein, and Fal-
see that he was alright. He was still in the same
position. This was nearly an hour later. So 1 loon (1983) report that patients with high
thought, it's not very warm, he must be getting EE relatives generally have quite good lev-
cold. So 1went out and got a blanket without ask- els of premorbid adjustment. Conse-
ing him and put it on him. "I don't want it," he quently, it is interesting to speculate wheth-
said. And then 1 did break down because 1 er the contrast between this premorbid lev-
thought well what the devil can 1do for him? And el and the level of functioning exhibited
that's when it upset me - when 1 can't do any- during the illness phase might in some way
thing to help him like that. contribute to the critical responses of the
High EE relatives also tend to respond relatives. The idea that high EE relatives
more dramatically to the patient's illness may in some instances be trying to exert
and lack the more controlled coping style control over the patient's behavior in an ef-
found in low EE individuals. Moreover, fort to restore premorbid levels of function-
they are less likely to believe that the pa- ing is developed further elsewhere (Hoo-
tient is genuinely ill. ley, 1985).

Interventions Involving Relatives

EE researchers have always been fully
aware of the correlational nature of their
data and the consequent problems in deter-
mining the direction of the causal arrow
which arise from this. Both Brown et al.
(1972) and Vaughn and Leff (1976) used a
variety of statistical techniques to eliminate
the possibility that mediating factors such as
the level of behavioral disturbance mani-
fested might both lead to high EE attitudes
in relatives and also increase relapse risk.
Miklowitz et al. (1983) have also recently
addressed this issue. In a study designed to
examine the relation between relatives' EE
and key patient attributes, symptomatolog-
ical data were collected on 42 schizophre-
nics during hospitalization and also 2-4
weeks after discharge. In keeping with pre-
vious findings (Brown et aI., 1972; Vaughn
and Leff, 1976), patients living in high EE
families did not differ significantly from
those living in low EE households with re-
gard to their premorbid adjustment or the
symptoms and syndromes they exhibited
during hospitalization. Moreover, postdis-
charge measures also failed to discriminate
between patients with low or high EE rela-
tives. Thus, the general pattern of results
suggests that patients with high or low EE
families are essentially indistinguishable on
32 J. M. Hooley
measures of the severity of the clinical con-
dition. The available evidence therefore
points to the fact that high EE may cause re-
lapse, although the data can hardly be de-
scribed as unequivocal.
To demonstrate the direction of the caus-
al arrow convincingly, the hypothesized in-
dependent variable (EE) must be external-
ly manipulated, and the consequent effect
on the dependent variable (patient's re-
lapse rates) observed. Only when it can be
demonstrated that a modification of EE
levels leads to a resulting change in relapse
rates can any issues of causality be fully re-
solved.
Clearly, ethical considerations render
studies designed to increase EE levels total-
ly unacceptable. However, one recent inter-
vention trial, designed to do exactly the op-
posite (see Leff, Kuipers, Berkowitz, Eber-
lein-Vries, and Sturgeon, 1982), has had an
important influence on the directionality
question. Since this study is described in
Chap. 17 of this volume and Sect. B of the
book deals specifically with family inter-
ventions, little further remains to be said
here. It is evident, however, that social in-
terventions of this nature have a very im-
portant role to play in relapse prevention
and will continue to form an active area of
EE research in coming years.

EE and Depression
So far in this chapter, only research carried between EE and relapse might not be a phe-
out in schizophrenic patients or individuals nomenon uniquely associated with schizo-
considered to be at risk for schizophrenia phrenic samples.
spectrum disorders has been considered. Since the cutoff of two critical comments
While this is only appropriate, given that utilized by Vaughn and Leff was deter-
schizophrenia has been central to almost all mined post hoc, a replication of the study
the research involving EE, some mention was clearly needed before more generality
should perhaps at this point be given to two for the EE construct could be claimed. The
studies which have examined the relation first full replication and extension of
between EE and relapse rates in depressed Vaughn and Leff's 1976 study has recently
populations. been completed by the author (see Hooley,
Vaughn and Leffs study of 1976 was the Orley, and Teasdale, in press) in a sample of
first to suggest that depressed patients, like 39 patients admitted to psychiatric hospitals
schizophrenics, might be vulnerable to the suffering from an episode of major depres-
influences of high EE relatives. Using a sive disorder. All patients involved in the
group of 30 depressed hospital inpatients, study were married, and consequently all
whose symptom pattern involved no delu- key relatives interviewed were spouses.
sions or hallucinations, Vaughn and Leff While no demographic or illness-related
were able to show that while the thresholds variables, such as number of symptoms,
of six or seven critical remarks which had severity of depression, or duration of the
been so useful in schizophrenic populations key depressive episode, were significantly
did not reveal any significant associations associated with 9-month relapse rates, a
with relapse when used on data from the de- significant relation between spouses' EE
pressed sample, decreasing the cutoff to levels and probability of relapse was evi-
two critical remarks did produce significant dent. Although 59% (14120) of patients liv-
effects. Specifically, when this lowered ing with spouses rated as high EE (two or
threshold was used to assign relatives as more critical comments made during CFI)
high or low EE, 67% (141:21) of patients liv- relapsed during follow-up, none (0/5) of the
ing with high EE relatives relapsed over the patients living with low EE spouses did so.
9-month follow-up period. In contrast, only These results thus provide confirmation of
22 % (2/7) of patients living with low EE re- Vaughn and Leff's original findings and
latives did so. These results were thus inter- suggest that the EE construct not only has
esting since they were the first to suggest predictive validity with respect to schizo-
that the previously well-established relation phrenic patients, but is also a potentially
 Introduction to EE Measurement and Research
valuable predictor of relapse in depressed
populations. While no intervention studies
involving EE have yet been conducted with
Concluding Remarks
Despite the strength of the association be-
tween EE and relapse, we still know rela-
tively little about the mechanism by which
EE might operate to bring about a return of
symptoms in psychiatric patients. Although
we now know more about the correlates of
high and low expressed emotion attitudes,
we still need to know much more about how
EE can operate to affect patients when they
return home. High EE relatives seem in
some way to constitute a form of increased
social stress for vulnerable individuals. Al-
ternately, low EE relatives can be construc-
ted as individuals capable of lowering stress
levels in vulnerable patients. Regardless of
which perspective is taken, the psychiatric
patient is seen as an individual who, for
whatever reason, seems excessively sensi-
tive to the effects of social stress.
While it is important to know what fac-
tors contribute to this sensitivity in patients,
it is also important for us to learn more
about the development of EE in relatives.
What makes a relative high in EE? Is it a
trait? Are certain individuals born high in
EE or does it arise as a consequence of ex-
posure to a psychiatrically ill patient? If so,
what kinds of patients and in response to
what kinds of symptoms? While practically
References
Brown, G. W. (1959). Experiences of discharged
chronic schizophrenic mental hospital patients
in various types of living group. Millbank Me-
morial Fund Quarterly, 37, 105-13l.
Brown, G. W., Birley, J. L. T., and Wing, J. K.
(1972). Influence of family life on the course of
schizophrenic disorders: A replication. British
Journal of Psychiatry, 121,241-258.
Brown, G. W., Carstairs, G. M., and Topping,
G. C. (1958). The post hospital adjustment of
chronic mental patients. The Lancet, II,
685-689.
Brown, G. W., Monck, E. M., Carstairs, G. M.,
and Wing, J. K. (1962). The influence offamily
life on the course of schizophrenic illness. Brit-
33
depressed patients, this is clearly an impor-
tant next step.
related issues focus attention quite rightly
on techniques of intervention - techniques
to reduce high EE levels - we should also
consider that high EE in relatives may not
invariably be a bad thing. Perhaps, as a con-
sequence of the attitude they adopt and
the behavior they show, high EE relatives
promote higher levels of patient function-
ing than low EE relatives. True, patients
may relapse more often, but they may also
achieve higher levels of intermorbid adjust-
ment which might, in turn, increase their
overall psychosocial stress levels and conse-
quently their relapse risk. Thus, while inter-
vention studies such as that of Leff et al.
have demonstrated that high EE relatives
can learn much from exposure to the coping
styles of low EE relatives, some of the po-
tentially beneficial aspects of the high EE
attitude might also be valuably conveyed to
low EE relatives. Many challenging ave-
nues of research still await researchers in
the field, and may important and fundamen-
tal questions still remain to be answered.
The next few years will be an important
time for EE research. Our understanding of
the dynamics of this valuable construct is
only just beginning.
ish Journal of Preventative Social Medicine,
16,55-68.
Brown, G. W., and Rutter, M. L. (1966). The
measurement of family activities and relation-
ships. Human Relations, 19,241-263.
Hooley, J. M. (1984). Criticism and depression.
Unpublished D. Phil. thesis, University of
Oxford.
Hooley, J. M. (1985). Expressed emotion: A
review of the critical literature. Clinical Psy-
chology Review, 5, 119 - 139.
Hooley, J. M., Orley, J., and Teasdale, J. D. (in
press). Levels of expressed emotion and
relapse in depressed patients. British Journal of
Psychiatry .
34 J. M. Hooley
Kuipers, L. (1979). Expressed emotion: A
review. British Journal of Social and Clinical
Psychology, 18,237-243.
Leff, J., Kuipers, L., Berkowitz, R., Eberlein-
Vries, R., and Sturgeon, D. (1982). A control-
led trial of social intervention in the families of
schizophrenic patients. British Journal of Psy-
chiatry, 141, 121-134.
Leff, J., and Vaughn, C. (1980). The interaction
of life events and relatives' expressed emotion
in schizophrenia and depressive neurosis. Brit-
ish Journal of Psychiatry, 136, 146-153.
Miklowitz, D. J., Goldstein, M. J., and Falloon,
I. R. H. (1983). Premorbid and symptomatic
characteristics of schizophrenics from families
with high and low levels of expressed emotion.
Journal of Abnormal Psychology, 92 (3),
359-367.
Rutter, M., and Brown, G. W. (1966). The re-
liability and validity of measures of family life
and relationships in families containing a psy-
chiatric patient. Social Psychiatry, 1,38-53.
Sturgeon, D., Kuipers, L., Berkowitz, R., Tur-
pin, G., and Leff, J. (1981). Psychophysiologi-
cal responses of schizophrenic patients to high
and low expressed emotion relatives. British
Journal of Psychiatry, 138,40-45.
Sturgeon, D., Turpin, G., Kuipers, L., Berko-
witz, R., and Leff, J. (1984). Psychophysiologi-
cal responses of schizophrenic patients to high
and low expressed emotion relatives: A follow-
up study. British Journal of Psychiatry, 145,
62-69.
Tarrier, N., Vaughn, c., Lader, M. H., and Leff,
J. P. (1979). Bodily reactions to people and
events in schizophrenia. Archives of General
Psychiatry, 36, 311-315.
Valone, K., Goldstein, M. J., and Norton, J. P.
(1984). Parental expressed emotion and psy-
chophysiological reactivity in an adolescent
sample at risk for schizophrenia spectrum dis-
orders. Journal of Abnormal Psychology, 93,
448-457.
Vaughn, C. E. (1977). Patterns of interaction in
families of schizophrenics. In H. Katschnig
(Ed.). Schizophrenia: The other side. Vienna:
Urban and Schwarzenberg.
Vaughn, C. E., and Leff, J. P. (1976). The in-
fluence of family and social factors on the
course of psychiatric illness. British Journal of
Psychiatry, 129, 125 - 137.
Vaughn, C. E., andLeff, J. P. (1981). Patterns of
emotional response in relatives of schizophre-
nic patients. Schizophrenia Bulletin, 7 (1),
43-44.
Vaughn, C. E., Snyder,K., Freeman, W.,Jones,
S., Falloon, I., and Liberman, R. (1982). Fami-
ly factors in schizophrenic relapse: A replica-
tion. Schizophrenia Bulletin, 8 (2), 425-426.
Vaughn, C. E., Snyder, K. S., Jones, S., Free-
man, W. B., and Falloon, I. R. H. (1984).
Family factory in schizophrenic relapse. Archi-
ves of General Psychiatry, 41,1169-1177.
Wing, J. K., Cooper, J. E., and Sartorius, N.
(1974). The description of psychiatric symp-
toms: An introduction manual for the PSE and
catego system. London: Cambridge University
Press.
3. Expressed Emotion in Cross-Cultural Context:
Familial Responses to Schizophrenic Illness
Among Mexican Americans*
J. H. Jenkins, M. Karno, A. de la Selva, and F. Santana
Culture and Schizophrenic Outcome
Over the past several decades, a variety of
cross-cultural studies have been under-
taken to explore the relationship of culture
to schizophrenia (cf. Draguns, 1980; Ken-
nedy, 1974; Sauna, 1980). Several investi-
gators (e.g., Murphy, 1982, p. 78) have con-
cluded that schizophrenia is "widely distrib-
uted around the world" and "takes a rather
similar form in all the diverse societies in
which it is found." In light of these data, it
may have seemed reasonable to assume
that schizophrenia is not significantly in-
fluenced by sociocultural forces. However,
in the wake of recent findings by the World
Health Organization, such an assumption
now appears unfounded.
There is strong evidence that the out-
come of schizophrenia varies across cul-
tures. Furthermore, this variation does not
seem to occur without regularity or pattern.
The International Pilot Study of Schizo-
phrenia (IPSS) 2-year follow-up of 1202 pa-
tients in nine countries revealed that pa-
tients from the three developing nations of
Nigeria, India, and Colombia had a more
favorable course than did patients from the
* The authors wish to acknowledge the research
projects and sources of funding upon which this
chapter is based: "The Course of Schizophrenia
among Mexican-Americans," Marvin Kamo,
M.D., Principal Investigator, National Institute
of Mental Health, MH-33502 and 39011; "Family
Factors," Christine Vaughn, Ph.D., Principal In-
vestigator, recently reported upon by Vaughn
et al. (1984); "Developmental Processes in Schi-
zophrenic Disorders," Keith Nuechterlein,
Ph.D., Principal Investigator, MH-30911.
Anglo-American data from the two latter pro-
jects were generously made available to Jenkins
(1984), for which we are grateful.
six more industrialized countries of Czecho-
slovakia, Denmark, Taiwan, the United
Kingdom, the United States, and the USSR
(World Health Organization, 1979).
The systematic methods of data collec-
tion of the IPSS lend a high degree of cred-
ibility to their findings. However, even
prior to publication of the IPSS outcome
data, there had existed a body of evidence
to suggest that the course of schizophrenia
is, indeed, cross-culturally variable. Several
reports of a favorable, nondisabling course
for the disorder have come from various
parts of the non-Western world, including
Africa (Jilek and Jilak-Aall, 1970; Kenne-
dy, 1974; Lambo, 1955), China (Rin and
Lin, 1962), Sri Lanka (Waxler, 1977), and
Mauritius (Murphy and Raman, 1971).
Review of the IPSS findings concerning
differential prognosis led Sartorius, Jablon-
sky, and Shapiro (1978) to conclude that
social and cultural factors may playa role in
schizophrenic outcome. Although the IPSS
investigators did not collect data specifical-
ly designed to analyze the sociocultural con-
text, they nonetheless concluded that the
family may have an important influence on
the patient. In particular, they have sug-
gested that "differences in the intensity of
family bonds, in the type of family structure
... may make it more or less difficult for a
schizophrenic patient to return to the com-
munity and remind in remission" (World
Health Organization, 1979, p. 371).
Other investigators have made similar in-
terpretations related to benign schizophren-
ic outcome that support the need for fami-
ly and community studies. For example,
Rin and Lin (1962) found that a relative
lack of stigma was attached to mental illness
36 J. H. Jenkins et al.
among a Taiwanese population. The au-
thors speculated that the community might
contain therapeutic resources such as "the
intimate and close emotional ties between
members of a family and also a clan or vil-
lage, and the abundance of opportunities
for group and community participation in
respect to daily life, farming, and festivals"
(Rin and Lin, 1962, p. 145). Similarly, Wax-
ler has proposed that the apparently benign
outcome for schizophrenic patients in Sri
Lanka rested in part on close familial in-
volvement and expectations of responsibili-
ty for the afflicted family member.
As Kleinman (1980) has recently pointed
out, the family sector of health care has yet
to be fully appreciated in the daily manage-
ment of illness episodes worldwide. For
persons suffering from schizophrenia, who
appear to be particularly sensitive to their
immediate social environment (Chapman
and Chapman, 1973; Leff, 1976), the home
environment may prove influential to the
course of illness. H. B. M. Murphy (1978)
has posited that a biologically based infor-
mation processing deficit in those vulnera-
ble to schizophrenia may express itself as
overt illness under conditions of stress
(within or without the family setting),
which overburden the individual's deficient
information-processing capacity while at
the same time preventing him or her from
escaping the demands of the situation. He
theorizes that different cultures impose dif-
ferent degrees of difficulty on those who are
vulnerable to schizophrenia by means of
varying combinations of expectations and
supports. Investigation of these sociocultur-
al processes has often been discussed gener-
ally at a community or societal level of
analysis, but can be more specifically ad-
dressed within family settings.

Family Studies of Expressed Emotion

For well over 2 decades, researchers from from relapse if they regularly took antipsy-
London, England have conducted studies chotic medication or were not in frequent
related to the sensitivity of the schizophren- contact with their high EE relative.
ic individual to the emotional atmosphere Vaughn, Snyder, Jones, Freeman, and Fal-
of the family environment (Brown, Birley, loon (1984) reported remarkably similar
and Wing, 1972; Brown and Rutter, 1966; findings for Anglo-American families living
Vaughn and Leff, 1976a). Their work has in Southern California, although some dif-
centered on the concept of expressed emo- ferences were observed with respect to the
tion (EE). The development of the concept presence of protective factors. In the Unit-
has been reviewed elsewhere (Kuipers, ed Kingdom, the high EE relapse rate was
1979; Leff, 1976; Hooley, this volume). partially mitigated if at least one of two pro-
A study by Vaughn and Leff (1976a) con- tective influences (medication compliance
firmed the earlier findings of Brown, Bir- or reduced contact with a high EE relative)
ley, and Wing (1972) that high EE was the was present. Schizophrenics within the Cali-
single most potent predictor of clinical re- fornia sample were similarly protected only
lapse, although patients who lived with high if both factors were present.
EE relatives were substantially protected

Family Factors and Schizophrenia Among Mexican-Americans

The influence of family factors on schizo-
phrenia among Mexican-Americans is un-
known but the subject of much specula-
tion. Until recently, Mexican-Americans
have been significantly underrepresented
as psychiatric patients in proportion to their
numbers in the general population through-
out the Southwestern United States in both
inpatient and outpatient settings (Karno,
1966; Karno and Edgerton, 1969; Keefe,
Padilla, and Carlos, 1978; Lopez, 1981; Sue,
1977; Weaver, 1973). Based on similar data
from the state of Texas and from his obser-
vation of Hispanic culture, Jaco (1959) hy-
pothesized that the Mexican-American
family provided a closely meshed psycho-
 Expressed Emotion in Cross-Cultural Context
social support network which tended to
protect its members against the develop-
ment of psychotic disorders. Madsen's eth-
nographic studies among Mexican-Ameri-
can families in the Rio Grande Valley in
South Texas led him to support Jaco's hypo-
thesis. As summarized by Madsen, "The
most important role of the individual is his
familial role and the family is the most val-
ued institution in Mexican-American socie-
ty. The individual owes his primary loyalties
to the family, which is also the source of
most affective relations" (Madsen, 1964,
p. 17).
Regarding the frequently cited orienta-
tion of "familism" among Mexican-Ameri-
cans, an Hispanic scholar has noted that
"For the Chicano, the family is likely to be
the single most important social unit in life.
It is usually at the core of his thinking and
behavior and is the center from which his
view of the rest of the world extends. Even
with respect to identification, the Chicano
self is likely to take second place after the
family" (Murillo, 1976, p. 19). Fabrega
(1969) has commented that "Extended
family-of-origin ties, family allegiances and
closeness to relatives are felt to be sources
of psychological support to the individual
and he is expected to value these relations-
ships. Great respect is shown to older rela-
tives and separate interests of family mem-
bers are subordinated fo family considera-
tions."
Doubt concerning the Jaco-Madsen hy-
pothesis of a Mexican-American familial
factor providing protection against the de-
velopment of major mental disorders led
Edgerton and Karno (1971) to study percep-
tions of and response to mental illness in the
East Los Angeles Mexican-American com-
munity. Based on 668 in-home interviews,
they found: (a) that primary language usage
differentiated between unacculturated
Mexican-Americans on the one hand and
acculturated (English-speaking) Mexican-
Americans on the other, the latter being
very like Anglo-Americans in perceptions
of mental illness, and (b) that a provocative
cultural difference was expressed in re-
sponse to a hypothetical case of an acutely
schizophrenic young woman. Spanish-spea-
king respondents expressed the belief that
such a person would best recover by remain-
37
ing with the family, whereas English-speak-
ing Mexican-American respondents agreed
with Anglo-American respondents that re-
covery would most likely occur by having
the afflicted individual removed to an out-
side treatment setting (Edgerton and
Karno, 1971). The Edgerton-Karno
group's findings supported the belief that
Spanish-speaking Mexican-American fami-
lies would, in general, attempt to cope with
the problem of psychotic illness within the
solidarity of the family. Beyond the nuclear
family's felt obligation of support, the ex-
tended family may provide additional sup-
port for a psychotic family member among
Mexican-American families. Keefe, Padil-
la, and Carlos report that in Southern Cali-
fornia "Mexican-Americans are much more
likely than Anglos to have large numbers of
their relatives living in the community ...
Anglos are more likely to seek help from
friends, neighbors, coworkers and groups.
Mexican-American's main resource, on the
other hand, is their extended kin network.
... " (Keefe, Padilla and Carlos, 1978b).
Based on the preceding lines of evidence,
an investigation of EE among Mexican-
American families of schizophrenic pa-
tients seemed particularly appropriate. A
research project entitled "The Course of
Schizophrenia Among Mexican-Ameri-
cans" (COSAMA) was designed to carry
out a transcultural extension of the British
findings on EE at a greater "cultural dis-
tance" than was afforted by the Anglo-
American replication of Vaughn et al.
(1984). The study was intended to evaluate
levels of EE among Mexican-American
families living in Southern California in
which a schizophrenic patient would be re-
turning home after discharge from the hos-
pital and to determine the influence of EE
on schizophrenic outcome. The study,
which has completed both a pilot and for-
mal data collection phase over the past 4
years, has gathered materials for 70 patients
and their families. A discussion of the de-
sign and methodology of the study, togeth-
er with some results entailing data on the
first 30 patients and 49 key family members,
form the core ofthis initial report. Data will
also be presented for a matched cohort of 30
Anglo-American patients and 47 key rela-
tives who have been studied under the aus-
38 J. H. Jenkins et al.
pices of the UCLA-MHCRC. 1 Cross-cul-
tural comparisons of these Mexican-Ameri-
can and Anglo-American subsamples have
been more fully analyzed elsewhere (Jen-
kins, 1984) and will be referred to here in
abbreviated form for purposes of drawing
cultural contrasts between these two ethnic
groups.

Methodology
with a history which suggested that the pri-
Patients
mary disorder was affective, organic, or
Patients recruited to the COSAMA project drug- or alcohol-related were excluded
have met the following criteria: from the study, as were those who did not
1. Diagnosed schizophrenic by the Present return to live in the household in which they
State Examination (PSE) and DSM-III were living prior to the index admission. At
criteria the time of the PSE, the Brief Psychiatric
2. Of bilateral Mexican descent Rating Scale (Overall and Gorham, 1962)
3. Between 18 and 50 years of age was also completed. Once the patient had
4. Living with parent, spouse, or other been diagnosed as schizophrenic, home
close relative for at least 1 of the 3 visits were scheduled with the patient's fam-
months prior to hospital admission ily.
The patients were recruited exclusively
from public mental health services in the
Assessment of Relatives
counties of Los Angeles and Ventura. With
the exception of ethnicity, the inclusion cri- Within several days after the patient's ini-
teria employed in the Anglo-American stu- tial diagnosis, adult key relatives - those
dies ofEE were similar to those of the Mexi- with whom the patient had an ongoing rela-
can-American project. tionship within the residential household -
were administered the Camberwell Family
Assessment of Patients Interview (CFI), which has been described
in detail by Vaughn and Leff (1976b).
Screening
The CFI is a partially structured inquiry
All patients were initially interviewed by a into the experience of the patient's illness
fieldworker fluent in Spanish, who estab- and its influence upon the life of the family
lished rapport, determined whether the di- during the 3 months prior to index hospitali-
agnosis was likely to be one of schizophre- zation. The CFI typically yields lenghtly
nia, and established that the patient met all and detailed narrations of family events and
other inclusion criteria, and who then ob- the emotional atmosphere of the house-
tained informed consent. hold. The abbreviated version of the CFI
takes about 1.5- 2 hs to complete and is au-
Diagnosis diotape-recorded for later scoring. In par-
ticular, the interviewer probes for and ob-
All patients were interviewed as soon as serves the expression of emotions and atti-
possible after admission - in accord with tudes expressed by family members toward
their capacity to consent and participate - the patient. The initial CFI interviewers of
by a bilingual clinical psychologist who had the COSAMA project (de la Selva and Jen-
been trained to research reliability in the kins) were trained in a 2-week workshop led
PSE. The ninth edition of the PSE was used by Vaughn and Snyder~ followed by several
in its official IPSS Spanish version. Patients months of training in the rating of Anglo-
1 The UCLA Mental Health Clinical Research
Center for the Study of Schizophrenia, MH- 2In the winter of 1980, Christine Vaughn, Ph.D.,
30911, R. P. Liberman, M.D., Principal Investi- and Karen Snyder, M. A., conducted a training
gator. All three research projects reported upon workshop in the administration of the Camber-
here for the Mexican-American and Anglo-Ame- well Family Interview and ratings of expressed
rican data were supported by the UCLA- emotion at Camarillo State Hospital, Camarillo,
MHCRC. California.
 Expressed Emotion in Cross-Cultural Context
American and British "master" audiotapes
of the various CFI scales. Both achieved
high interrater reliability scores with the
original ratings (above Pearson r of 0.90).
Periodic reliability checks were completed
within the project to ensure the ongoing re-
liability of the ratings.
Translation and Development
of the Spanish eFI
Two independent Spanish translations of
the CFI were made, one by a professional
native Spanish-speaking translator and lan-
guage teacher, the other by de la Selva. The
two translations were integrated into a
single first draft translation, which was then
back-translated into English by Santana
(who was "blind" to the CPI in English).
The integrated first translation was re-
viewed line-by-line in comparison with the
back-translation and original English ver-
sion to resolve discrepancies and to produce
a semifinal translation that was utilized in
22 pilot interviews with Spanish-speaking
relatives of schizophrenic Mexican-Ameri-
cans in Ventura and Los Angeles Counties.
These tape-recorded interview experiences
were than reviewed to produce a final trans-
lation of the CPI, which was revised and
adapted for local, colloquial Mexican-
American usage.
All CFI interviews were scored for vari-
ous scales described by Vaughn and Leff
(1976b), but the major focus was on the
Sample Characteristics
Selected sociodemographic and clinical
characteristics of the Mexican-American
patients, along with those of the matched
Anglo-American comparison group, are
presented in Table 1. The samples are simi-
lar in many respects. Nearly all of the fami-
lies were of lower socioeconomic status ac-
cording to Hollingshead's (1957) index of
social position.
The majority (79%) of the Mexican-
American key relatives were first-genera-
tion Mexicanos, i.e., born in Mexico. A
similar percentage (73%) of the Mexican-
American patients were first-generation
immigrants. According to the acculturation
39
three component ratings of EE, viz., the
number of critical comments, hostility, and
emotional overinvolvement.
Additional Procedures
While the focus of this report is on data col-
lected within the CFI, data from several ad-
ditional procedures were also employed.
These materials are important for the inter-
pretation of EE within the Mexican-Ameri-
can family context. They include interview
schedules which inquire into following
areas: social-psychiatric histories of the pa-
tient; sociodemographic data related to the
composition of the household; social net-
works of the patient; level of acculturation
of key relatives and patients; and family
members' ethnopsychiatric understandings
of the nature of schizophrenia and patterns
of coping with the problem. Also, ethno-
graphic observations of the household,
neighborhood, and community were ob-
tained.
Since the primary goal of the project was
to assess the relationship of schizophrenic
outcome to family EE profiles, the clinical
status of the patient was regularly monitor-
ed. The follow-up phase of clinical status
has yet to be completed and therefore will
not be reported upon here. The purpose of
this initial report is the exploration of the
nature and meaning of EE among Mexican-
American families coping with schizophre-
nic illness.
scale developed by Cuellar, Harris, and J as-
so (1980), both the key relatives and pa-
tients tended to be relatively unacculturat-
ed. Most (65%) of the key relatives (often
parents) were predominantly or only Span-
ish-speaking, while only 43.4% of patients
were similarly monolingual. The vast ma-
jority (90%) of the sample were Catholic.
The mean size of household was signifi-
cantly different among the Mexican-Ameri-
can families compared to that of the Anglo-
American sample, 6.7 and 3.6 persons, re-
spectively (P < 0.001). Further, Mexican-
American households of the present sam-
ple are characterized by large networks of
40 J. H. Jenkins et al.
Table 1. Selected Sociodemographic and Clinical Features of the Mexican-American (MA) and
Anglo-American (AA) Patient Samplesa

Variable MA AA
n % n %
Patients 30 100.00 30 100.00
Sex
Females 11 36.7 11 36.7
Males 19 63.3 19 63.3
Mean age 27.4 25.2
Marital status
Never married 22 73.3 26 86.7
Married 4 13.3 0 0.0
Separated, divorced, widowed 4 13.3 4 13.3
Length of illness
(mean number years) 5.5 4.3
Number of hospitalizations
(mean) 3.1 3.4
Family type
Parental 22 73.3 28 93.3
Sibling 3 10.0 1 3.3
Marital 3 10.0 0 0.0
Other 2 6.7 1 3.3
Socioeconomic status
(Hollingshead two-factor index)
III 1 3.3 2 6.7
IV 14 46.7 13 43.3
V 15 50.0 15 50.0

a Due to rounding error, percentages do not always equal 100%.

Note. From Schizophrenia and the Family: Expressed Emotion Among Mexican-Americans and
Anglo-Americans by J. Jenkins, 1984, University of California, Los Angeles.

locally residing kin who frequently visit the mic support for the families.
household, providing emotional and econo-

EE in Cross-Cultural Context
The importance of the EE construct has
been demonstrated for British and Anglo-
American psychiatric patients and their
families; however, its relevance and mean-
ing in other cultural contexts among non-
English speakers must be established if the
EE concept is to gain a wider cross-cultural
validity. Variation in levels and types ofEE
needs to be explored. Moreover, since EE
is an index which taps a host of features re-
lated to familial response to schizophrenia,
factors underlying various patterns of EE
that are culturally distinctive as well as
more universal in nature need to be identi-
fied. Discovery of features associated with
EE profiles will contribute to the develop-
ment of theoretical models informative of
ways in which sociocultural processes may
serve to influence the course of schizophre-
nic illness.
 Expressed Emotion in Cross-Cultural Context
Cross-Cultural Variability ofEE
Vaughn et al. (1984) reported a signifi-
cant difference between the EE profiles of
British and Anglo-American families of
schizophrenic patients. Slightly over one-
half of the British households were low in
EE (Brown, Birley, and Wing, 1972;
Vaughn and Leff, 1976a) as against only
one-third of Anglo-American families.
While this report is based on comprehen-
sive analyses completed for the first 30 fam-
ilies included in the Mexican-American
study, the EE scores for the entire sample
(n = 70) will be briefly presented here.
These data, together with those collected
for the British and Anglo-American sam-
ples, establish the significant variation in
patterns of EE across sociocultural settings
and are presented in Table 2. A household
is defined as being high in EE if one or more
key relative was rated as high in EE.
The above percentage of high vs low EE
were made using the original cutoff points
(Six or more criticisms and/or a score of
4-5 on the EO! scale) as designated by the
British researchers. These have been em-
ployed here for purposes of cross-cultural
comparison. However, it should be noted
that upon completion of data collection for
all clinical assessments, a different baseline
could later prove important for the predic-
tion of schizophrenic relapse among Mexi-
can-American families.
41
The percentages of EE presented in
Table 2 reveal that EE was rated highest
among Anglo-American families and low-
est among Mexican-Americans. While this
may seem suggestive of cultural differences
in EE, it is necessary to take into considera-
tion other features that could account for
this variation. One such major concern is
that of socioeconomic status. Leff (1977,
p. 321) has highlighted "the importance of
matching for social class in studies of emo-
tional expression."
To investigate styles of EE that may be
related to features that are cultural in na-
ture, Jenkins (1984) compared the EE pro-
files of lower socioeconomic status families.
Thirty Mexican-American and 30 Anglo-
American families (Table 1) living in South-
ern California were studied. The Mexican-
American families were the first 30 recruit-
ed to the COSAMA project described
above, and the Anglo-American families
were selected from the project reported on
by Vaughn et al. (1984) and an ongoing
UCLA-MHCRC project.3
The results of overall household EE for
these matched subsamples are presented in
Table 3. A striking difference between
these two ethnic groups was found: Anglo-
American schizophrenic patients were con-
3 See Footnote at bottom of p. 38.

Table 2. EE Profiles of Anglo-American, British, and Mexican-American Households

LowEE HighEE
n % n %
Anglo-American (n = 69)a 23 33.3 46 66.7

British (n = 138)b 72 52.2 66 47.8

Mexican-American (n = 70) 41 58.6 29 41.4

a Note. From "Family Factors in Schizophrenic Relapse: A California Replication of the British Re-
search on Expressed Emotion" by C. E. Vaughn, K. S. Snyder, S. Jones, W. B. Freeman, and I. R.
H. Falloon, 1984, Archives of General Psychiatry, 41, pp. 1169-1177.
b Note. From "Influence of Family Life on the Course of Schizophrenic Disorders: A Replication" by
G. W. Brown, L. T. Birley, andJ. K. Wing, 1972, British Journal of Psychiatry, 121, pp. 241-258 and
Vaughn, C., and Leff, J. (1976). The influence of family and social factors on the course of psychiatric
illness: A comparison of schizophrenic and depressed neurotic patients. British Journal of Psychiatry,
129,125-137.
Overall x 2 = 9.90, d.f 2, P < 0.02; Anglo-American-British, Yates corrected x 2 = 5.84, d.f 1, P <
0.02; Mexican-American-British, Yates correctedx2 = 0.530, NS; Anglo-American-Mexican-Ameri-
can, Yates correctedx2 = 7.92, d.f 1, P < 0.01.
42 J. H. Jenkins et al.
Table 3. Comparison of Overall EE Profiles for Mexican-American and Anglo-American Households
of Lower Socioeconomic Status
LowEE
HigbEE
Totals
Note. From "Schizophrenia and the Family: Expressed Emotion Among Mexican-Americans and
Anglo-Americans" by J. Jenkins, 1984, University of California, Los Angeles.
siderably more likely to reside in high EE
home environments than were their Mexi-
can-American counterparts.
It should be noted that the EE percent-
ages reported upon here for the lower socio-
economic status Anglo-American compari-
son group are typical for the entire subsam-
pIe of such cases within the study completed
by Vaughn et al. (1984).
Thus, it appears that differences in EE
profiles are related not only to ethnicity but
also to the socioeconomic status of key rela-
tives. The higher percentage ofEE (83.3%)
found here, compared with that of the en-
tire Anglo-American sample (67%) of 69
families, is due to the selection of cases of
lower socioeconomic status. Analysis of
variance and covariance revealed that ob-
served differences in the EE profiles were
not significantly related to patient charac-
teristics (sex, age, marital status, edul::a-
tion, employment status, number of hospi-
talizations, length of illness, socially disrup-
tive symptomatology, or level of premorbid
adjustment). Furthermore, the type ofrela-
Sociocultural Nature of EE
A first step toward an understanding of the
variation in EE among different ethnic
groups is appreciation of the fact that EE is
inherently sociocultural and qualitative in
nature. This is not to suggest that EE is not
amenable to quantitative analysis, but rat-
her to underscore the point that EE indexes
a wide array of behaviors, emotions, and at-
titudes which are part of an individual's cul-
tural repertoire, developed through proces-
ses of socialization. As noted by Geertz
(1959, p. 225):
Mexican-American Anglo-American
(n = 30) (n = 30)
17 56.3 5 16.7
13 43.3 25 83.3
30 100.0 30 100.0
tionship of key relatives (mother, father,
sister, and so forth), type of household (pa-
rental, marital, sibling), or variation among
the lower levels of socioeconomic status
[Hollingshead (1957) levels IV or V] also
proved unrelated to EE profiles. Rather,
the highly significant difference is account-
ed for by the ethnic identity of the family
(Mexican-American or Anglo-American).
This is compelling evidence for major dif-
ferences in cultural styles of emotional re-
sponse to schizophrenic illness within the
family. We turn now to a discussion of some
qualitative aspects of these styles and an ex-
ploration of some of the features that seem
to characterize patterns of EE among Mexi-
can-Americans. Materials analyzed for the
comparably matched Anglo-American
sample will be employed to highlight cultur-
al contrasts. In light of the finding that eth-
nicity and social class figure prominently in
styles of emotional expression, it is evident
that a greater appreciation of sociocultural
factors is needed.
Every cultural system includes patterned ideas
regarding certain interpersonal relationships and
certain affective states, which represent a selec-
tion from the entire potential range of interper-
sonal and emotional experiences. The child, gro-
wing up within the culture and gradually inter-
nalizing these premises, undergoes a process of
socially guided emotional specialization. He
learns, in a sense, a special vocabulary of emo-
tion.
Explantations of cross-cultural differences
in the expression of emotion must include
 Expressed Emotion in Cross-Cultural Context
an understanding of the "vocabulary of
emotion" characteristically employed by
individuals.
Low EE Profiles Among
Mexican-Americans
The majority (69%) of Mexican-American
key relatives were rated low in EE. This fig-
ure varies remarkably from that for the
Anglo-American key relatives of the mat-
ched comparison group, as shown in
Table 4.
These percentages provide a more indi-
vidualized view of EE than does the overall
household index used for predicting re-
lapse. Clearly, most of the Mexican-Ameri-
can family members tended not to be highly
critical, overinvolved, or hostile toward
their ill relatives. While this may indicate
the absence of certain kinds of verbally ex-
pressed emotions or communications with-
in the family, it tells us relatively little about
what other sorts of patterns might be ac-
tively shaping the household environment.
Moreover, the EE percentages do not say
anything about whether low EE among
Mexican-Americans is similar in kind to low
EE profiles among other cultural groups.
We will now turn our attention to a discus-
sion of some of the contours and major
components of these low EE Mexican-
American relatives.
One of the most striking findings was that
feelings of sadness as conveyed through the
content and tone of the relatives' verbal be-
havior would often pervade the interviews.
This contrasted sharply with emotional re-
sponses characterized by anger, indigna-
tion, and frustration that were frequently
displayed by high EE relatives. While the
Table 4. Comparison of EE Profiles for Mexican-American and Anglo-American Key Relatives
LowEE
HighEE
Totals
Note. From "Schizophrenia and the Family: Expressed Emotion Among Mexican-Americans and
Anglo-Americans" by J. Jenkins, 1984, University of California, Los Angeles.
43
Anglo-American relatives reported feeling
sad (particularly those relatives rated low in
EE), this did not occur as frequently as
among Mexican-American family mem-
bers. Feelings of sadness (tristeza), sorrow
(pena), and pity (ltistima) were commonly
voiced. This was especially true of Mexican-
American women. Verbalizations of sad-
ness were often accompanied by evidence
of warmth and sympathy for the patient and
his or her condition.
Such expressions of sadness or sorrow
over the illness of a close family member
may reflect genuine feelings on the part of
the relative. On the other hand, they should
also be regarded as culturally appropriate
responses to schizophrenic illness within
the family. As Nichter (1981) has recently
pointed out, every culture provides its
members with a variety of ways to express
emotions, including distress. Faced with
what they may view as a serious illness or
condition - whether temporary or long-
term - Mexican-American relatives ex-
press feelings that acknowledge an inher-
ently distressing situation, and the Spanish
language provides a rich lexicon of emotion-
al terms to express feelings of unhappiness
or distress. While some low EE relatives ex-
pressed emotions related to their ill relative
that included anger or resentment, expres-
sions of sadness clearly predominated over
those of indignation or antipathy. How-
ever, several relatives reported that they at-
tempted to keep their feelings of sadness to
themselves, fearing that the patient might
be adversely affected by their feelings of
being distraught.
Low EE profiles of Mexican-Americans
are also linked to ethnopsychiatric views of
the nature of the problem that has beset
Mexican-American Anglo-American
(n = 49) (n=47)
34 69.4 18 38.3
15 30.6 29 61.7
49 100.0 47 100.0
44 J. H. Jenkins et al.
their ill relative. These ethnopsychiatric or
"folk" models for interpreting schizophren-
ic behavior are complex and merit in-
depth treatment that would be beyond the
scope of this chapter. However, we will at-
tempt to briefly summarize some of the
more salient aspects of these models and
their relationship to low EE. Of principal
significance is that Mexican-Americans do
tend to view the problem as one of illness; it
would appear that psychosis is commonly
considered to be an illness-related problem
cross-culturally (Murphy, 1982). Among
Mexican-Americans, the problem is often
considered to be associated with one's "ner-
ves" or mental condition. In whatever man-
ner the illness may be conceptualized, it is
typically believed to be legitimate. Fre-
quently, the relatives would refer to the "ill-
ness" independent of any prompting from
the interviewer and in advance of any que-
ries by the interviewer into the relative's
view of the nature of the problem. Most in-
formants were firm in their conviction that
the patient suffered from an illness prior to
the time that he or she was first hospitalized
for the present troubles.
Belief in the legitimacy of schizophrenia
as an illness is also linked to views that sym-
ptom-related behaviors generally lie out-
side the patient's control. Some of the fol-
lowing comments4 made during the course
of the interviews illustrate this point:
If he doesn't want to do anything it's because he's
sick; he always worked hard in school and jobs
before. He must be sick.
I knew right away that this wasn't Luis. I know
what he was doing wasn't his fault.
She would like to get well ... if she could.
These comments are similar to those made
by low EE Anglo-American relatives:
If he could really control himself, I think that he
would.
She's always looking for help. She tries the best to
help herself to get healthy. She tries her best-
looking for doctors to help her ... I don't think
she could do any more. She's a fighter. She wants
to be well.
Ethnopsychiatric views of the nature of the
problem also influenced the relatives' re-
4 The qualitative materials reproduced here are
verbatim quotations from CFI materials and orig-
inally appeared in Jenkins (1984).
sponses to symptom behaviors. A content
analysis of critical comments by Jenkins
(1984) revealed that Mexican-American
key relatives were far less likely to criticize
symptom behaviors than were their Anglo-
American counterparts. These data provide
support for the proposition current in the
literature that Hispanics tend to be relative-
ly tolerant of psychotic symptomatology
(Fabrega, Swartz, and Wallace, 1968; Rog-
Ier and Hollingshead, 1965; World Health
Organization, 1979). The Mexican-Ameri-
can relatives demonstrated a great deal of
tolerance, patience, and respect toward the
schizophrenic family member and would
seldom challenge or take exception to psy-
chotic behaviors. Many relatives reported
that confrontations with a sick family mem-
ber were unwise and did little to change
matters.
Low EE Mexican-American families
would sometimes interpret symptom behav-
iors in a sympathetic or favorable light. For
example, one young woman spent a great
deal of time sitting alone and laughing to
herself for no apparent reason. Her moth-
er's view of this behavior was not a critical
one; rather, she reported feeling glad that
her daughter was able to find some momen-
tary relief from the very difficult circum-
stances of her life by simply being able to
laugh at her troubles. In another instance, a
mother viewed her son's habit of sleeping
until early afternoon as potentially benefi-
cial and clearly understandable, given that
he was typically up quite late into the night,
unable to sleep because he was continually
bothered by voices. She believed that sleep
was crucial to his recovery and that he
should sleep whenever he was so inclined or
able. With respect to this same behavior -
sleeping late - another key relative's re-
sponse to his wife's routine of sleeping until
noon every day was to say, "Let her sleep as
late as she likes! She is my queen!" These
examples should not be taken as evidence
that Mexican-Americans do not recognize
these sorts of behavior as indicative of men-
tal illness, for as discussed above, they do.
Rather, these vignettes should be viewed as
exemplary of relatives' efforts to interpret
such behaviors in a noncritical manner.
Finally, low EE profiles among Mexican-
Americans are linked to strong notions of
 Expressed Emotion in Cross-Cultural Context
the importance of family bonds, which are
enduring and unchanging. The strength of
family ties demonstrated in the face of long-
term schizophrenic illness is often remarka-
ble, although the type of relationship (par-
ent, spouse, and so on) is often influential.
Mothers tend to be the most devoted, fol-
lowed by fathers, sisters and brothers, and
spouses. This may be common across a vari-
ety of cultural contexts. Much of the Mexi-
can-American relatives' compassion stem-
med from a deeply felt commitment to a
family member, particularly one who is seri-
ously ill. Many relatives reported that they
felt a great deal more affection for the pa-
tient as a result of the illness. Several made
passionate statements about how they
would never, indeed could never, leave
their ill relatives to fend for themselves "in
the streets."
Another feature that we believe is related
to the predominance of low EE profiles
among Mexican-Americans is the size and
structure of the families. As was discussed
earlier in connection with the sample charac-
teristics, Mexican-American nuclear house-
holds tended to be larger than those of the
Anglo-Americans. In nearly all ofthe Mexi-
can-American households, locally residing
kin visit the home frequently, providing
emotional and instrumental support and as-
sistance. This includes help in alleviating
the stress generated by living with schizo-
phrenic illness. In some families, caretaking
of the ill relative was shared among various
relatives, thereby allocating the burden of
responsibility among several individuals.
Large family size may also inhibit the devel-
opment of patterns of overinvolvement or
critical attitudes and behaviors. For exam-
ple, occasionally, key relatives would ex-
press regret over not having more time to
devote to their ill relative, since responsibili-
ties to other family members tended to de-
tract from their efforts in this regard.
To summarize, low EE among Mexican-
Americans is patterned along four dimen-
sions: (1) sadness as a predominant emo-
tional response to the problem, (2) ethno-
psychiatric models of schizophrenia that
view the problem as a serious, legitimate ill-
ness outside of the patient's locus of con-
trol, (3) styles of coping with troublesome
behaviors in a manner which avoids argu-
45
ments or confrontations, and (4) large kin-
based households and networks in which a
sense of the importance of family bonds
induces relatives to assume responsibility
for the patient's care and recovery.
High EE ProfIles Among
Mexican-Americans
The above positive features of low EE
households are not universally found
among Mexican-American families. In con-
trast to the low EE Mexican-American rela-
tives for whom sadness seemed to represent
a major emotional orientation toward the
illness, high EE respondents were likely to
express feelings of anger and indignation.
In fact, one of the most common ways a
high EE Mexican-American relative would
express criticism would be to comment on a
behavior by saying mi di6 mucho coraje (it
made me very angry). These relatives were
often distressed over behaviors they viewed
as unacceptable. For example, the father of
one female patient made seven critical com-
ments during the course of the interview,
and within four of these criticisms he made
reference to the anger that certain behav-
iors would provoke in him. These criticisms
centered on the patient's tendency to argue,
insult and swear at family members, and
throw objects about the house.
Moreover, some high EE relatives, un-
like those in low EE families, more often
doubted whether their family member was
truly ill. These relatives would point out
that the patient's symptoms would some-
times wax and wane. More frequent, how-
ever, were doubts that their relative could
possibly be mentally ill. This may indicate
that high EE relatives had a greater tend-
ency to deny the problem of mental illness
in their family or to experience greater feel-
ings of shame and stigma associated with it.
Thus, one Mexican-American father be-
lieved that his son might not be truly ill since
he sometimes seemed ill and other times
not; in any case, he claimed that it had gone
on long enough.
Among the high EE Anglo-American
relatives, often there was doubt or disbelief
in the legitimacy of the illness. These rela-
tives would express the view that even if
46 J. H. Jenkins et al.
their relative was in fact ill, he or she could
surely exercise powers of personal initiative
and control in determining the course of the
illness, should they be so inclined.
A lot oftimes it wouldn't have happened if she ...
saw it coming ... to me I think she could have
done something about it and you know with me I
can talk with someone and try to help them, but
if they don't help themselves and don't give a
damn ... that's where it bugs me ... She wouldn't
do anything - like she wanted to feel sorry for
herself ... She was just giving up ... I don't like to
see that. I don't like to see someone give up when
there's still something to fight for ... I say, "Why
aren't you trying, don't give up," because to me
people who do that I can't get along with ... I
can't get along with someone who sees the prob-
lem and does nothing, just lets it happen ... and
I start to get a dislike.
I don't know if she has a button she can push or
what ... that girl can fool you and fool anybody
... She can go like that and be as normal as you or
I (see Footnote 4).
There were few differences between high
and low EE Mexican-American relatives
with respect to criticism of symptom behav-
iors. Although generally speaking, high EE
relatives were more likely to be critical of
the patient, there was not a disproportion-
ate amount of attention focused on symp-
tom behaviors per se. This differed greatly
from comments made by high EE Anglo-
Americans. They expressed discomfort
concerning psychotic behaviors, as the fol-
lowing remarks about conversation illus-
trate:
... bugged the hell out of me, the thing that really
drove me up a wall was her starting a conversa-
tion and not finishing it. I mean it really bugged
me.
She'd talk, talk, talk, talk, talk, about God or
whatever - religion - you can't correct her and
I would not correct her ... but that constant
drone is enough to drive me up a wall.
I can't even remember some of the horrible
things he'd come out and talk about ... and actu-
ally he was making me nervous. The things that
he'd talk about (see Footnote 4).
Cultural· differences between Mexican-
Americans and Anglo-Americans were also
apparent in criticisms directed toward in-
activity or unemployment on the part of
schizophrenic relatives. While critical rela-
tives of both ethnic groups were likely to
complain about these behaviors, they dif-
fered in the nature of criticisms they typical-
ly expressed. For example, Mexican-Amer-
icans stated flatly that it was important that
their relatives work or be active for their
own sakes or in order that they be able to
take care of some of their own personal
needs.
He has to dress, to have money, work, so he can
go out, have a girlfriend, because we can't give it
all to him.
Look, my son, it's not good for someone to sleep
till 2:00 in the afternoon! We have to go to work,
do something! (see Footnote 4).
Anglo-American relatives, on the other
hand, displayed critical attitudes toward in-
activity or unemployment, frequently in-
ferring negative assessments of their rela-
tive's character or personality:
. .. he just didn't have no ambition to do anything
... 20 years old, right in his prime ... but now he
should be involved but he's not. He would just sit
there in the chair and watch TV all the time.
No ambition - doesn't care for nothing ... I
should say doesn't care for anything, class up my
English. The way I said it was to emphasize that
worthlessness of his. Useless.
She's too darn lazy. That's her trouble.
Do this today and you want to do this tomorrow,
or you never will make it. You've gotta make up
your mind and go after it (see Footnote 4).
Styles of coping and adaptation to schizo-
phrenic illness among high EE Mexican-
Americans were typically different from the
strategies employed by low EE relatives.
High EE relatives would attempt to change
undesirable behaviors through scolding,
admonitions, explanations, and avoidance.
In contrast to low EE coping styles, these
responses were reactive and combative.
Furthermore, some relatives tended to be
intrusive in ways that were uncharacteristic
of low EE relatives. For example, some rel-
atives would search through the patient's
personal things or listen in on telephone
conversations to gamer information con-
cerning their activities.
These kinds of responses were common
to high EE Anglo-Americans relatives, as
well. These high EE relatives, however,
were more likely to report frequent "nagg-
ing" of the patient, as the following exam-
ples demonstrate:
I am a nagger, well, normally [about] him doing
things wrong. I am a nagger. I've tried to break
myself of it, but I do nag him.
 Expressed Emotion in Cross-Cultural Context
Sometimes I nag. I'm so exasperated ... mostly
about money [she throws away] and the fact that
she always sits around ... (see Footnote 4).
Despite the critical attitude displayed by
high EE relatives, this did not alter their
sense of familial commitment toward their
ill relative. In this regard, they did not
markedly differ from low EE relatives.
There was only one instance in which a key
relative demonstrated a rejecting attitude.
This case was a martial household in which
the wife, very weary of her husband's con-
tinuously psychotic condition throughout
their 9 years of marriage, sought a divorce.
Her priest, however, counseled against such
a move and told her to remain with her hus-
Conclusion
Both quantitative and qualitative analyses
of EE among Mexican-American families
revealed a predominance of low EE styles
of response to a schizophrenic family mem-
ber. These low EE profiles were found to be
patterned along several dimensions, many
of which were notably different from those
characteristically employed by high EE rela-
tives. In general, low EE family members
provided their schizophrenic relatives with
a great deal of tolerance, sympathy, and
support. However, not all Mexican-Ameri-
can individuals responded in this fashion.
High EE relatives displayed anger, impa-
tience, and frustration related to the pa-
tient's troublesome behaviors. The strength
of family bonds and affection was typically
in evidence for both high and low EE rela-
tives alike, despite the difficulties generated
References
Brown, G. W., Birley, L. T., and Wing, J. K.
(1972). Influence of family life on the course of
schizophrenic disorders: A replication. British
Journal of Psychiatry, 121,241-258.
Brown, G. W., and Rutter, M. (1966). Themeas-
urement of family activities and relationships:
A methodological study. Human Relations, 19,
241-263.
Chapman, L. J., and Chapman, J. P. (1973). Dis-
ordered thought in schizophrenia. New York:
Appleton-Century-Crofts.
Cuellar, 1., Harris, L. c., and Jasso, R. (1980).
47
band since divorce was not possible within
the Catholic church. This she accepted, al-
though in one instance she reported that her
husband's incessant complaints concerning
the living conditions of the household drove
her to exclaim, "Look, if you're not happy
here, there is the door! I've had it! I'm 'up
to here!" The fact that only one such in-
stance occurred lends strong support to the
prevailing views concerning Mexican-Ame-
rican familism and to the view that Mexi-
can-American relatives display high levels
of acceptance toward schizophrenic rela-
tives in the home, even when high EE char-
acterizes the mode of response toward
them.
by schizophrenic illness.
Comparative data from a matched sam-
ple of Anglo-American families demon-
strated significant variations in EE across
cultures. While Mexican-Americans tended
to be typically low in EE, the converse was
true of Anglo-Americans, for whom high
EE profiles dominated. Although some
similarities across high and low EE styles for
these two ethnic groups were noted, there
were several aspects of EE styles (both high
and low) that were culturally distinctive to
Mexican-Americans. While these patterns
of response may not be unique to Mexican-
Americans when viewed in broader cross-
cultural perspective, they do suggest that
the sociocultural context may be influential
in shaping an individual's response to a schi-
zophrenic family member.
An acculturation scale for Mexican-American
normal and clinical populations. Hispanic Jour-
nal of Behavioral Sciences, 2 (3), 199-217.
Draguns, J. G. (1980). Psychological disorders of
clinical severity. In H. C. Triandis and J. G.
Draguns (Eds.), Handbook of cross-cultural
psychology: Psychopathology. Vol. 6, (pp.
99-174). Boston: Allyn and Bacon.
Edgerton, R. B., and Karno, M. (1971). Mexi-
can-American bilingualism and the perception
of mental illness. Archives of General Psychia-
try, 24, 286-290.
48 J. H. Jenkins et al.
Fabrega, H. (1969). Mexican-Americans of
Texas: Some social psychiatric features. In E.
G. Brody (Ed.), Behavior in new environ-
ments: Adaptation of migrant populations
(p. 256). Beverly Hills: Sage.
Fabrega, H., Swartz, J. D., and Wallace, C. A.
(1968). Ethnic differences in psychopathology
- II. Specific differences with emphasis on a
Mexican-American group. Journal of Psychia-
tric Research, 6, 221-235.
Geertz, H. (1959). The vocabulary of emotion: A
study of Javanese socialization processes. Psy-
chiatry, 22 (3), 225-237.
Hollingshead, A. B. (1957). Two-factor index of
social position. Unpublished manuscript.
Jaco, E. G. (1959). Mental health ofthe Spanish-
American in Texas. In M. K. Opler (Ed.), Cul-
ture and Mental Health. New York: Mac-
millan.
Jenkins, J. (1984). Schizophrenia and the family:
Expressed emotion among Mexican-Ameri-
cans and Anglo-Americans. Unpublished doc-
toral dissertation, University of California, Los
Angeles.
Jilek, W. G., and Jilek-Aall, L. (1970). Transient
psychosis in Africans. Psychiatric Clinic, 3,
337-364.
Karno, M. (1966). The enigma of ethnicity in a
psychiatric clinic. Archives of General Psychia-
try, 14, 516-520.
Karno, M., and Edgerton, R. B. (1969). Percep-
tion of mental illness in a Mexican-American
community. Archives of General Psychiatry,
20,233-238.
Keefe, S. E., Padilla, A. M., and Carlos, M. L.
(1978a). Emotional support systems in two cul-
tures: A comparison of Mexican-Americans
and Anglo-Americans (Occasional Papers No.
7). Los Angeles: University of California,
Spanish-Speaking Mental Health Resource
Center.
Keefe, S. E., Padilla, A. M., and Carlos, M. L.
(1978b). The Mexican-American extended
family as an emotional support system. In J. M.
Casas and S. E. Keefe (Eds.), Family and men-
tal health in the Mexican-American community
(Monograph No.7, pp. 65-65). Los Angeles:
University of California, Spanish-Speaking
Mental Health Research Center.
Kennedy, J. G. (1974). Cultural psychiatry. InJ.
J. Honigmann (Ed.), Handbook of social and
cultural anthropology. New York: Rand
McNally College Publishing.
Kleinman, A. (1980). Patients and healers in the
context of culture: An exploration of the bor-
derland between anthropology, medicine and
psychiatry. Berkeley: University of California
Press.
Kuipers, L. (1979). Expressed emotion: A re-
view. British Journal of Social and Clinical Psy-
chology, 15, 157-165.
Lambo, T. M. (1955). Theroleofculturalfactors
in paranoid psychosis among the Y oruba tribe.
Journal of Medical Science, 101,239-266.
Leff, J. (1976). Schizophrenia and sensitivity to
the family environment. Schizophrenia Bulle-
tin, 2 (4),566-574.
Leff, J. (1977). The cross-cultural study of emo-
tions. Culture, Medicine and Psychiatry, 1,
317-350.
Lopez, S. (1981). Mexican-American usage of
mental health facilities: Underutilization re-
considered. In A. Baron (Ed.), Explorations in
Chicano psycology (pp. 139-164). New York:
Praeger.
Madsen, W. (1964). Mexican-Americans of
South Texas. New York: Holt, Rinehart and
Winston.
Murillo, N. (1976). Mexican-American family.
In C. A. Hernandez, M. J. Haug, and N. N.
Wagner (Eds.), Chicanos: Social and psycho-
logical perspectives (2nd Ed., p. 19). St. Louis:
Mosby.
Murphy, H. B. M. (1978). Cultural influences in
incidence, course and treatment response. In
L. C. Wynne, R. L. Cromwell, and S. Matthys-
se (Eds.), The nature of schizophrenia (pp.
586-594). New York: Wiley.
Murphy, H. B. M., and Raman, A. C. (1971).
The chronicity of schizohrenia in indigenous
tropical peoples: Results of a twelve-year fol-
low-up in Mauritius. British Journal of Psychia-
try, 118,489-497.
Murphy, J. M. (1982). Cultural shaping and men-
tal disorders. In W. R. Gove (Ed.), Deviance
and mental illness (pp. 49-82). Beverly Hills:
Sage.
Nichter, M. (1981). Negotiation of the illness ex-
perience. Ayurvedic therapy and the psycho-
social dimension of illness. Culture, Medicine
and Psychiatry, 5, 5-24.
Overall, J. A., and Gorham, D. H. (1962). The
brief psychiatric rating scale. Psychological Re-
ports, 10,799-812.
Rin, H., and Lin, T. (1962). Mental illness
among Formosan Aborigines as compared with
Chinese in Taiwan. Journal of Mental Science.
108, 134-146.
RogIer, L. H., and Hollingshead, A. B. (1965).
Trapped: Families and schizophrenia. New
York: Wiley.
Sauna, V. (1980). Familial and sociocultural an-
tecedents of psychopathology. In H. C. Trian-
dis and J. G. Draguns (Eds.), Handbook of
cross-cultural psychology: Psychopathology
Vol. 6 (pp. 175-236). Boston: Allyn and
Bacon.
Sartorius, N., Jablonsky, A., and Shapiro, R.
(1978). Cross-cultural differences in the short-
term prognosis of schizophrenic psychosis.
 Expressed Emotion in Cross-Cultural Context
Schizophrenia Bulletin, 4,102-113.
Seymour, S. (1983). Household structure and
status and expressions of affect in India. Ethos,
2 (4), 263-277.
Sue, S. (1977). Community mental health serv-
ices to minority groups: Some optimism, some
pessimism. American Psychologist, 32,
616-624.
Vaughn, c., and Leff, J. (1976a). The influence
of family and social factors on the course ofpsy-
chiatrie illness: A comparison of schizophrenic
and depressed neurotic patients. British Jour-
nal of Psychiatry, 129, 125-137.
Vaughn, c., and Left, 1. (1976b). The measure-
ment of expressed emotion in the families of
psychiatric patients. British Journal of Social
and Clinical Psychology, 15, 157-165.
49
Vaughn, C. E., Snyder, K. S., Jones, S., Free-
man, W. B., and Falloon, I. R. H. (1984). Fami-
ly factors in schizophrenic relapse: A California
replication of the British research on expressed
emotion. Archives of General Psychiatry, 41,
1169-1177.
Waxler, N. (1977). Is outcome for schizophrenia
better in non-industrialized societies? The case
of Sri Lanka. Journal of Nervous and Mental
Disorders, 167, 144-158.
Weaver, J. L. (1973). Mexican-American health
care behavior: A critical review of the litera-
ture. Social Science Quarterly, 54, 85-102.
World Health Organization (1979). Schizophre-
nia: An international follow-up study. Chiche-
ster: Wiley.
4. Do Relatives Express Expressed Emotion?*
A. M. Strachan, M. J. Goldstein, and D. J. Miklowitz
Introduction
Recent research has shown that family in-
tervention programs, in combination with
neuroleptic medication, can have dramat-
ic effects on the course of schizophrenia in
terms of both psychiatric and social func-
tioning (Goldstein, Rodnick, Evans, May,
and Steinberg, 1978; this volume, Chaps.
12-14). All these programs have focused
on treating and supporting the family as a
unit, although it has not always been clear
what important aspects of family function-
ing these programs affect. However, as sug-
gested by Goldstein and Strachan (Chap.
Early Research on EE and Interaction
Two early papers describe the development
of a method for assessing the expression of
emotion (Rutter and Brown, 1966; Brown
and Rutter, 1966), which was later to be-
come the Camberwell Family Interview.
The original study was of 30 families con-
taining at least one child in which one of the
parents had been admitted either to an out-
patient psychiatric clinic or to the hospital.
For each family, there was 2- to 3-h inter-
view with the patient, a 3- to 4-h interview
with the spouse, and (for 18 families only) a
joint interview with the husband and wife
together, which lasted about an hour. The
individual interviews focused on the impact
of the psychiatric problems and on family
relationships. The joint interviews focused
on the family's utilization of medical and
social services. At each interview, two
people made independent ratings of a num-
ber of dimensions of emotional expression,
including warmth and critical remarks. Rat-
ings were made of attitudes toward the
15), the way in which emotional attitudes
are communicated and problems are solved
may be crucial factors, with expressed emo-
tion (EE) as a key index of such processes.
This chapter focuses on the interactional
correlates of EE, a measure based on rat-
ings of an interview with a relative. Such
ratings have been correlated in a number of
studies with direct behavior between the
patient and a relative, providing measures
of the congruence and incongruence be-
tween behavior in these two situations.
spouse in the individual interviews and be-
havior toward the spouse in the joint inter-
views.
Results for these 18 wives and 18 hus-
bands showed that criticisms of their spous-
es were much less likely to be expressed in
the joint than in the individual interview.
Only 13 people of 36 made one or more criti-
cism during the whole hour. This was proba-
bly because of the presence of their spouse
and because the topic of conversation was a
• The UCLA research studies were supported by
grants MH08744, MH30911, and MH14584 from
the US National Institute of Mental Health and
by a grant from the MacArthur Foundation
(K830902) to support a Network on Risk and
Protective Factors in the Major Mental Dis-
orders. The authors are deeply indebted to Ian
Falloon and Julian Leff for their warm spirit of
collaboration and to Jeri Doane, Jeanette Nor-
ton, Karen Snyder, Keith Valone, and Sibyl
Zaden for their important contributions to vari-
ous phases of this research.
52 A. M. Strachan et al.
relatively neutral one which would tend to sible that attitudes elicited on the CFI are
pull for criticism of external agencies rather imperfectly correlated with interactional
than of each other. behavior and that the predictive validity of
However, it was found that when there the Camberwell-derived EE attitudes in
were any criticisms in the joint interview, anyone study depends upon the degree of
there were nearly always (11 of 13 cases) 10 attitude-behavior correlation observed in
or more criticisms in the individual inter- that particular sample. Surprisingly, despite
view. Conversely, if there was a complete the obvious nature of this issue, it has rarely
absence of criticisms in the joint interview, been addressed in the accumulating litera-
there were usually (21 of 23 cases) fewer ture on EE. A correspondence between
than 10 criticisms in the individual inter- EE-type behaviors in direct family interac-
view. This was a highly significant differ- tion and subsequent clinical outcome would
ence (P < 0.(01). Thus, although the base- provide further support for the hypothesis
rate of critical remarks in the two situation that the affective climate of the family is rel-
was very different, the correspondence be- evant to the course of schizophrenia.
tween the two measures suggested some In our research program, we decided that
cross-situational validity to the rating of at- it was necessary to demonstrate that affec-
titude. The Pearson correlation between tive attitudes expressed by a relative in an
the two measures of criticism was 0.51. For interview corresponded with the manner in
warmth, the correlation was even higher, which that relative spoke when face-to-face
0.68. with the patient. To do this, two things were
The exciting and provocative findings of needed: a standardized situation to evalu-
Brown, Birley, and Wing (1972) and ate interactional behavior and a coding sys-
Vaughn and Leff (1976) stimulated a spate tem that reflected the behavior analogues
of replication studies designed to evaluate of EE attitudes. For the standardized situa-
whether the predicitve validity of the EE tion, we used the direct interaction task
measure holds up cross-nationally and used previously in our prospective longitudi-
cross-culturally (e.g., Chaps. 5, 7, this vol- nal study (Goldstein, Judd, Rodnick, Al-
ume; Vaughn, Snyder, Jones, Freeman, kire, and Gould, 1968) to assess family in-
and Falloon, 1984). While these efforts are teraction in a laboratory setting. In this
very understandable, they have focused modified version, family members were
upon a particular methodology for assess- asked to discuss two emotionally loaded
ing EE, the CFI, and lost sight of the origi- family problems derived from a prior inter-
nal model underlying this research. The ori- view. One problem was initiated by the pa-
ginal hypothesis of the Brown et al. studies tient and the other by a relative. These dis-
was that certain behavior expressed by rela- cussions were recorded in the absence of an
tives toward a recently released schizophre- experimenter.
nic patient increases the stress level within This procedure differs substantially from
the home and lowers the patient's threshold that used by Rutter and Brown (1966).
for relapse. The interactional behaviors First, the interactions are much shorter,
which were identified as important were being closer to 10 min than 1 h. Secondly,
rather specific styles of negative affective no raters are present: family members are
communication within the family. left alone to discuss the topic. Thirdly, the
However, the CFI is not a measure of in- topic is chosen to be one of immediate per-
trafamilial affective communication; in- sonal relevance to the participants. Where-
stead, it measures attitudes toward a patient as Rutter and Brown had couples discuss
expressed within the context of an interview the relatively neutral topic of their use of
with a mental health professional. These at- medical and social services, we had relatives
titudes mayor may not be expressed in face- discuss a much more emotionally charged
to-face interaction with the patient. We issue.
know, from years of research in social psy- For a coding scheme, we used the affec-
chology, that there is rarely isomorphism tive style (AS) coding system developed by
between verbally expressed attitudes and Doane (Doane, West, Goldstein, Rodnick,
overt interactional behavior. Thus, it is pos- and Jones, 1981) which was designed to
 Do Relatives Express Expressed Emotion?
measure affective attitudes expressed to-
ward a relative during a direct interaction
task. Some of the codes in this system over-
lap with the EE dimensions of criticism and
emotional overinvolvement. This system
was used to code the interactional data in
our sample of schizophrenic patients and
their parents. The codes in the AS system
are presented in Table 1.
Table 1. Negative Affective Style Codes
Criticism
1. Personal criticism
Unnecessary or overly harsh modifiers and/or
negative reference to broad classes of behav-
ior and/or negative evaluation of the child's
character or nature
2. Benign criticism
Mild, circumscribed, matter-of-fact, directed
toward specific incidents or sets of behaviors
Guilt induction
Conveys child is to blame or at fault for some
negative event and parent has been distressed
or upset by the event
Intrusiveness
Parent implies knowledge of child's thoughts,
feelings, or motives without basis for such
knowledge
1. Critical intrusiveness
Contains a harsh, critical attribution of intent
2. Neutral intrusiveness
Neutral quality; refers to child's emotional
states, ideas, preferences, etc.
Note. From "Parental Communication Deviance
and Affective Style: Predictors of Subsequent
Schizophrenia Spectrum Disorders in Vulnerable
53
Adolescents" by J. A. Doane, K. L. West, M. J.
Goldstein, E. H. Rodnick, andJ. E. Jones, 1981,
Archives of General Psychiatry, 38, pp. 679 - 685.
Although this scheme was originally de-
veloped to capture aspects of family process
which the research literature suggested had
a negative impact, it can be seen that some
of these codes are analogous to the EE mea-
sures. Thus, benign and personal criticisms
are analogous to critical comments. The in-
terpersonal analogue of overinvolvement is
reflected in the intrusion code in which one
person speaks as though they were the ex-
pert on the patient's thoughts, inner states,
and motives. The guilt induction and critical
intrusiveness codes have elements of both
criticism and emotional overinvolvement.
A series of three studies has examined
the relationship between EE and direct in-
teraction. The first study (Valone, Norton,
Goldstein, and Doane, 1983) examined 52
families in the UCLA Family Project, a lon-
gitudinal prospective study. These families
had all come to an outpatient clinic for
problems with an adolescent offspring.
Families were selected in which both natur-
al parents were living with the adolescent
and in which the offspring was disturbed but
nonpsychotic. Both dyadic and triadic data
were available. In all, 28 mothers and 34
fathers were classified as low EE, whereas
24 mothers and 18 fathers were placed in
the high EE category, mostly on the basis of
a highly critical attitude. It was found that
high EE parents expressed more benign cri-
ticisms and more harsh criticisms toward
their offspring than low EE parents when

Table 2. Benign and Harsh Criticisms Expressed to Patients by Dual High, Mixed, and Dual Low EE
Parental Pairs

Benign Criticism Harsh Criticism

Group n Mean (SD) Mean (SD)
a. Dual high EE 11 12.36 (7.55) 7.45 (7.01)

b.MixedEE 20 8.05 (5.32) 2.20 (2.98)

c.DuallowEE 21 4.95 (5.95) 1.43 (1.78)

Significant differences (P < 0.05): benign criticism, c < b, c < a; harsh criticism, a> b, a > c

Note: From "Parental Expressed Emotion and Affective Style in an Adolescent Sample at Risk for
Schizophrenia Spectrum Disorders" by K. Valone, J. P. Norton, M. J. Goldstein, and J. A. Doane,
1983, Journal of Abnormal Psychology, 92, pp. 399-407.
54 A. M. Strachan et a1.
data from both dyads and triads were sum-
med. When dyadic and triadic data were
examined separately, it was found that high
EE mothers made more benign criticisms
than low EE mothers in the dyads, whereas
there was no difference for fathers. There
were trends for both mothers and fathers to
use more benign criticisms in the triad.
Next, the authors examined the synergist-
ic influence of parents on each other's ex-
pression of criticism. They divided the fami-
lies into dual low EE (both parents low EE),
mixed EE (one parent high, the other low),
and dual high EE (both high EE). They
found that the three parental EE profile
groups differed significantly in the expres-
sion of both benign and harsh criticism (see
Table 2). More specifically, data analyses
suggested that the high EE parent in mixed
EE parental pairs is especially prone to ex-
press benign criticisms toward the adoles-
cent. However, the presence of the low EE
spouse in the mixed EE families appears to
inhibit the frequent expression of harsh criti-
cism by the high EE parent. These data
show that the sharpest predictor of critical
behaviors from EE attitudes occurred when
the EE profile of the parental pair was
examined together, as would be predicted
by family-systems theorists.
Unfortunately, in this study there was an
almost complete absence of emotionally
overinvolved parents so that it was not pos-
sible to study the behavioral correlates of
such an attitude. This may have been be-
cause the sample was not one of schizophre-
nics but of problem adolescents who may
elicit critical responses from parents but
whose protective responses are in the nor-
mal range. The next two studies to be de-
scribed overcame these drawbacks by study-
ing relatives of schizophrenics, some of
whom had emotionally overinvolved atti-
tudes.
Miklowitz, Goldstein, Falloon, and Doa-
ne (1984) assessed the relationship between
EE attitudes and direct interactional be-
havior with a sample of schizophrenic pa-
tients and their parents from 42 families. A
count was made of the total number of nega-

10

c-
en
Q)
E
8
Critical statements Intrusive statements

--
Q)
til 6
en
0
....
Q)
..c 4
E
:::J
C
c 2
co
Q)
~

0
LowEE High EE- High EE- High EE- Low EE High EE- High EE- High EE-
Critical Overinvolved Critical and Critical Overinvolved Critical and
EOI EOI
(n = 7) (n = 14) (n = 5) (n = 4) (n = 7) (n = 14) (n = 5) (n = 4)

Expressed emotion subgroup

Fig. 1. Mean number of critical and intrusive statements in direct interaction among high and low EE
parental subgroups
Note. From "Interactional Correlates of Expressed Emotion in the Families of Schizophrenics" by
D. J. Miklowitz, M. J. Goldstein, I. R. H. Falloon, and J. A. Doane, 1984, British Journal of Psychia-
try, 144, pp. 482-487.
 Do Relatives Express Expressed Emotion?
tive AS statements (sum of criticisms and
intrusive statements) expressed by the par-
ents in triadic discussion between the pa-
tient and his or her parents. The high EE
relatives expressed a significantly greater
number of negative AS statements than low
EE parents with means of 9.6 and 5.4, re-
spectively (P < 0.002).
However, since high EE relatives attain
this status by way of diverse attitudinal pat-
terns (high on criticism, high on criticism
and emotional overinvolvement, and high
on emotional overinvolvement only), we
investigated whether these subtypes, with
the high EE group, expressed different styl-
es of AS as well. Figure 1 presents the fre-
quency of criticism and intrusive statements
with the low EE and the different subtypes
of high EE. We can see that there is notable
group differentiation such that the subtypes
of high EE attitudes correspond with dif-
ferent patterns of interactional behavior
55
(criticism with critical EE attitudes and in-
trusiveness with emotional overinvolve-
ment). These data are also significant
(P < 0.05).
The third study in this series (Strachan,
Leff, Goldstein, Doane, and Burtt, 1985)
differed in three ways from the Miklowitz
et al. (1984) study. First, the study was
done in Britain, so that a cross-national re-
plication could be performed. Secondly,
the sample of schizophrenics was mostly of
recent onset. Thirdly, interactional behav-
ior was assessed from dyadic interactions
between patient and each relative separat-
ely. Thus, a direct comparison of EE atti-
tudes with AS behaviors could be conduct-
ed without the possibly confounding pres-
ence of another relative.
As in the Miklowitz study, it was found
that high EE relatives expressed a signifi-
cantly greater number of negative AS state-
ments than low EE relatives, with means of

10

Critical statements Intrusive statements

-
en
c
(1)
8

--
E
(1)
6

-
co
en
0
....
(1)
4
.0
E
::l
C
2
c
co
(1)
~
0
Low EE High EE- High EE- High EE- Low EE High EE- High EE- High EE-
Critical Overinvolved Critical and Critical Overinvolved Critical and
EOI EOI
(n = 28) (n = 11) (n = 16) (n = 7) (n = 28) (n = 11) (n = 16) (n = 7)

Expressed emotion subgroup

Fig. 2. Mean number of critical and intrusive statements in direct interaction among high and low EE
parental subgroups
Note. From "Emotional Attitudes and Direct Communication in the Families of Schizophrenics: A
Cross-National Replication" by A. M. Strachan, J. P. Leff, M. J. Goldstein, J. A. Doane, and
C. Burtt, 1985, manuscript submitted for publication.
56 A. M. Strachan et al.
11.6 and 4.7, respectively (P < 0.02). Fur-
thermore, when the high EE relatives were
subdivided into those who were critical,
those who were emotionally overinvolved
(EOI), and those who were both critical
and EOI, a similar pattern to the Miklowitz
data was obtained (see Figure 2). However,
it was found in this sample that a critical at-
titude was associated with both criticisms
and intrusions, whereas in the more chronic
American sample a critical attitude was
more uniquely associated with critical be-
havior. A study is in progress to assess
whether different types of neutral intru-
sions are associated with critical as com-
pared with overinvolved attitudes.
Another interesting finding in Strachan
et al. 's work was that in dyads where the rel-
ative was high EE both the relative and the
patient used more words than in dyads
where the relative was low EE. These fin-
dings suggest that in high EE dyads both
participants are actively involved in the in-
teraction, rather than the relative being
dominant and the patient submissive. These
findings are similar to those reported by
Kuipers, Sturgeon, Berkowitz, and Leff
(1983) who observed relatives and patients
talking to an interviewer and found that
high EE relatives spent 57% more time
talking than low EE relatives, a percentage
which is very close to the 61 % higher word
count observed here. Further, they found
that the amount of silence was significantly
longerin low EE pairs than in high EE pairs.
Strangely enough, none of these studies
showed that low EE relatives made signifi-
cantly more supportive statements during
the interactions, although there were trends
in the predicted direction. This may have
been because the discussions focused
around emotionally charged issues and elic-
ited critical and intrusive statements. Dis-
Patient Attributes Associated with EE Status
We mentioned earlier that Brown et al.
(1972) indicated that certain patient char-
acteristics were associated with both high
EE attitudes and subsequent relapse as
well. However, the patient's contribution
to the affective climate of the family follow-
ing hospitalization has received little atten-
cussions around more neutral issues may
elicit greater differences between relatives
in their rate of use of supporting statements.
This thesis is supported by data from Green-
wald, Kornblith, and Hopkins (1980). They
applied the marital interaction coding sys-
tem to samples of interaction between schi-
zophrenic patients and their relatives. The
discussion topics were more neutral than in
the other studies. They found no differ-
ences between low and high EE relatives on
their use of negative or neutral codes, but
low EE relatives showed more positive be-
haviors (n = 25, t = 2.64; P < 0.05) such
as approval, agreement, smiling, and hu-
mor. This suggests that low EE relatives
have the potential to be more supportive
than high EE relatives in some situations.
Overall, these studies suggest that inter-
actions with high EE relatives can be char-
acterized as calmer, less verbal, with more
pauses, and more positive support.
Although there is a reasonable level of
comparability between attitudes expressed
in the Camberwell and affective style ex-
pressed interactionally, the distributions
revealed that the correspondence is far
from perfect. The majority of high EE par-
ents, defined as critical on the Camberwell,
are indeed expressive of criticism but others
express none at all. Thus, some high EE
parents are congruent across situations and
others appear to dampen or inhibit these
high EE attitudes when in direct confronta-
tion with the patient. Currently, we are in-
vestigating whether patients living with
high EE individuals who express these nega-
tive affective attitudes when interacting
directly with the patient are more relapse-
prone than those patients with relatives
who have the negative attitudes but do not
express them directly to the patient.
tion since. If high EE attitudes are associat-
ed with relapse, is it not possible that this is
due to a correlation with a third variable,
such as the severity or quality of patient
symptomatology at the time of hospital dis-
charge? Leff (1976) reported negative data
in this regard when psychiatrists' ratings of
 Do Relatives Express Expressed Emotion?
the severity of symptoms at discharge were
examined.
A similar analysis of symptom ratings, of
the sample from the Falloon study (Miko-
witz, Goldstein, and Falloon, 1983), re-
vealed parallel findings to those reported by
Leff. There were no significant differences
between patients from high or low EE
homes on the Present State Examination
(PSE; Wing, Cooper, and Sartorius, 1974)
rated during the most acute phase of the
current episode, on the Brief Psychiatric
Rating Scale (BPRS; Overall and Gorham,
1962) obtained shortly before discharge,
nor in the number of prior hospitalizations.
There were also no differences in ratings of
premorbid adjustment on the UCLA social
attainment scale (Goldstein, 1978). The lat-
ter is particularly important as premorbid
adjustment is an established prognostic in-
dicator, and it is important to demonstrate
its independence from EE attitudes.
Miklowitz et al. (1983) did find, however,
that patients from one high EE subgroup,
the emotionally overinvolved subsample,
had poor premorbid adjustment and high
levels of residual symptomatology at dis-
charge. However, patients whose high EE
relatives were defined by highly critical at-
titudes were indiscriminable from patients
from low EE families in terms of their level
of residual symptomatology and premorbid
adjustment.
Symptoms and permorbid adjustment do
not exhaust the personal qualities of a schi-
zophrenic individual that can arouse high
EE feelings in a relative. Possibly some
other more subtle aspects of the schizophre-
nic's behavior arouse critical or overin-
volved feeling in a close relative. Currently,
we are investigating whether there are cer-
tain key behaviors emitted by the schizo-
phrenic during the direct interaction task
which trigger criticisms or intrusiveness by a
parent.
In a study currently underway, a coding
system (the coping style scale) was applied
to the same transcripts which had previous-
ly been scored for affective style. This scor-
ing was done by raters blind to the AS scor-
ing done previously and to the parental EE
57
level. The coping style codes covered three
broad categories of self-statements express-
ed by a patient: positive, negative, and auto-
nomous. Positive self-statements refer to
statements of worth or pride in some activi-
ty or attribute of the patient. Negative self-
statements are self-critical remarks, and
autonomous statements reflect assertions
by the patient to follow some self-selected
path of action.
Data analyzed so far have revealed that
patients from high and low EE homes differ
markedly in the rate of negative and posi-
tive self-statements. As hypothesized, pa-
tients from high EE homes express a signifi-
cantly higher number of negative self-state-
ments and a lower number of positive self-
statements than patients from low EE
homes (P < 0.001). Possibly, the sense of
ineptitude expressed by these negative self-
statements may be particularly infuriating
to relatives who see them as one more sign
of the frustrating negative symptoms of the
disorder.
To date, our analyses of parent and pa-
tient interactive behaviors have been car-
ried out independently. Yet, to understand
better the patterns of intrafamilial trans-
actions in families varying in EE attitudes,
future research must integrate these two
levels of data (see also Chap. 7). Thus, cur-
rent research in our laboratory is designed
to further our understanding of how patient
behaviors trigger relatives' emotional re-
sponses and, conversely, to identify those
behaviors emitted by relatives which trigger
negative social reaction by the patient.
Thus, sequential analyses are being used to
determine whether patient's self-denigrat-
ing remarks trigger parental criticism and
intrusiveness or whether they represent re-
actions to these parental behaviors. A com-
prehensive analysis of the manner in which
family relationships have an impact on the
course of schizophrenia requires an appre-
ciation of the subtle reciprocities in these
families and the manner in which the level
of tension within the home escalates or de-
escalates because of such interactive pro-
cesses.
58 A. M. Strachan et al.
References
Brown, G. W., Birley, J. L. T., and Wing, J. F.
(1972). Influence of family life on the course of
schizophrenic disorders: A replication. British
Journal of Psychiatry, 121,241-258.
Brown, G. W., and Rutter, M. (1966). The
measurement of family activities and relation-
ships: A methodological study. Human Rela-
tions, 19,241-263.
Doane, J. A., West, K. L., Goldstein, M. J.,
Rodnick,E. H., andJones,J. E. (1981). Paren-
tal communication deviance and affective style:
Predictors of subsequent schizophrenia spec-
trum disorders in vulnerable adolescents.
Archives of General Psychiatry, 38, 679-685.
Goldstein, M. J. (1978). Further data concerning
the relation between premorbid adjustment
and paranoid symptomatology. Schizophrenia
Bulletin, 4, 236-243.
Goldstein, M. J., Judd, L. L., Rodnick, E. H.,
Alkire, A., and Gould, E. (1968). A method
for studying social influence and coping pat-
terns within families of disturbed adolescents.
Journal of Nervous and Mental Disease, 147,
233-25l.
Goldstein, M. J., Rodnick, E. H., Evans, J. R.,
May, P. R. A., and Steinberg, M. R. (1978).
Drug and family therapy in the aftercare of
acute schizophrenics. Archives of General Psy-
chiatry, 35,1169-1177.
Greenwald, D. P., Kornblith, S. J., and Hop-
kins, J. (1980). Behavioral Assessment of Fami-
lies of Schizophrenics: A Validity Study of Ex-
pressed Emotion. Paper presented at the 14th
Annual Convention of the Association for the
Advancement of Behavior Therapy, New
York, November.
Kuipers, L., Sturgeon, D., Berkowitz, R., and
Leff, J. (1983). Characteristics of expressed
emotion: Its relationship to speech and look-
ing in schizophrenic patients and their rela-
tives. British Journal of Clinical Psychology,
22,257-264.
Leff, J. P. (1976). Schizophrenia and sensitivity
to the family environment. Schizophrenia Bul-
letin, 2, 566-574.
Leff, J. P., and Vaughn, C. E. (1981). The role of
maintenance therapy and relatives' expressed
emotion in relapse of schizophrenia: A two-
year follow-up. British Journal of Psychiatry,
139, 102-104.
Miklowitz, D. J., Goldstein, M. J., and Falloon,
I. R. H. (1983). Premorbid and symptomatic
characteristics of schizophrenics from families
with high and low level of expressed emotion.
Journal of Abnormal Psychology, 92, 359-367.
Miklowitz, D. J., Goldstein, M. J., Falloon, I. R.
H., and Doane, J. A. (1984). Interactional cor-
relates of expressed emotion in the families of
schizophrenics. British Journal of Psychiatry,
144,482-487.
Overall, J. E., and Gorham, D. R. (1962). The
brief psychiatric rating scale. Psychological
Report, 10, 799.
Rutter, M., and Brown, G. W. (1966). The relia-
bility and validity of measures of family life and
relationships in families containing a psychia-
tric patient. Social Psychiatry, 1, 38-53.
Strachan, A. M., Leff, L. P., Goldstein, M. J.,
Doane, J. A., and Burtt, C. (1985). Emotional
attitudes and direct communication in the fami-
lies of schizophrenics: A cross-national replica-
tion. Manuscript submitted for publication.
Valone, K., Norton, J. P., Goldstein, M. J., and
Doane, J. A. (1983). Parental expressed emo-
tion and affective style in an adolescent sample
at risk for schizophrenia spectrum disorders.
Journal of Abnormal Psychology, 92, 399-407.
Vaughn, C. E., and Leff, J. P. (1976a). The mea-
surement of expressed emotion in the families
of psychiatric patients. British Journal of Clini-
cal and Social Psychology, 15, 157-165.
Vaughn, C. E., and Leff, J. P. (1976b). The in-
fluence of family and social factors on the
course of psychiatric illness: A comparison of
schizophrenic and depressed neurotic patients.
British Journal of Psychiatry, 129, 125-137.
Vaughn, C. E., Snyder, K. S., Jones, S., Free-
man, W. B., and Falloon, I. R. H. (1984).
Family factors in schizophrenic relapse: A Cali-
fornia replication of the British research on Ex-
pressed Emotion. Archives of General Psychi-
atry,41,1169-1177.
Wing, J. K., Cooper, J. E., and Sartorius, N.
(1974). The description and classification of
psychiatric symptoms: An instruction manual
for the PSE and CATEGO system. London:
Cambridge University Press.
5. Short-Term Relapse in Young Schizophrenics:
Can It Be Predicted and Affected by Family (CFI) ,
Patient, and Treatment Variables?
An Experimental Study*
B. Dulz and I. Hand
The Hamburg Schizophrenia and CFI Project
The Hamburg Schizophrenia and CFI Pro-
ject evolved from an attempt by various aca-
demic and nonacademic members of the
Psychiatric University Clinic to develop
more reliable and effective means for pre-
dictive assessment and treatment of young-
er, mainly first- or second-breakdown schi-
zophrenic patients at high risk for a chronic
course of illness. We decided to work with
younger schizophrenics as there had been
much less systematic treatment research
performed with them than the more chronic
conditions.
The development of our research design
(Hand and Kottgen; details in Hand and
Gross, 1980) was strongly influenced by the
following sources:
1. The results published concerning ex-
perience with the Camberwell Family
Interview (CFI) (extensive literature re-
views by Hooley, Chap. 2, and Vaughn,
Chap. 8, this volume) and direct support
from C. Vaughn and K. Snyder regard-
ing the training of raters
2. The frank and stimulating presentations
and discussions of the schizophrenia re-
search groups participating in the 1979
conference of the Society for Psycho-
therapy Research in Oxford, England
3. The results of the three major long-term
studies in German-speaking countries
(Bleuler, 1972; Ciompi and Miiller,
1976; Huber et aI., 1979) on the course
of schizophrenia over decades
4. The continuous flow of information
from the ongoing research work of R.
Liberman and his colleagues in Camaril-
lo and Los Angeles.
Within the context of our study, we con-
ducted the first and so far only attempt at
replication of the Anglo-American CFI stu-
dies in a German-speaking culture. At the
same time, with our diagnostic and treat-
ment interventions, we wanted to investi-
gate what the "meaning" of the published
results and of our own CFI results could be
for a conceptualization of the findings that
until then had been purely empirical. Much
dispute had occured and still seems to be
going on about whether or not the CFI liter-
ature implicitly "blames" the relatives for
the patients' relapse and illness. Research-
ers investigating expressed emotion (EE)
felt themselves misunderstood by these cri-
tics. Yet, they themselves had not made
clear whether high EE-related behaviors of
relatives were supposed to "cause" relapse
of psychosis or vice versa. Does the assump-
tion of a behavioral feedback loop really
answer this question?
We shall first summarize the main aims of
this project and some general data on out-
come before turning to our detailed ana-
lyses of relapse and the variables that pos-
sibly contribute to relapse.
Replication of Previous CFI Results and
Direct Evaluation of the CFI
The main intentions of the following inves-
tigations were to replicate previous CFI re-
sults and to directly evaluate the CFI.
Replication of previous CFI results con-
sisted of:
* This project was supported by the Deutsche
Forschungsgemeinschaft, Sonderforschungsbe-
reich 115, Project B 23, University of Hamburg
(main study) and by the Charles Hosie Founda-
tion, Hamburg (pilot study).
60 B. Dulz and I. Hand
Identification of the proportions of high
EE relatives in younger schizophrenic
patients in a German-speaking culture.
Comparison of relapse rates in patients
from high EE and low EE families.
Comparison of relapse rates in patients
from high EE families who either did or
did not receive regular medication and
who did or did not participate in group
therapy.
Direct evaluation of the CFI consisted of:
Assessment of the stability of initially
high and initially low EE ratings over
time (no such investigation had been
published at the start of this project). If
high EE ratings (in relatives) at the time
of the patient's discharge do predict re-
lapse (in patients) 9 months later, one
ought to expect some kind of causal in-
teraction and, hence, stability of high
EE ratings over this period of time.
Comparison of global EE ratings over
time with ratings on the EE subscales of
emotional overinvolvement (EOI) , criti-
cism, and hostility. All published studies
attribute a high EE "total" score to a rel-
ative as soon as she or he has reached
high ratings on at least one of the EE
subscales. Yet, there is no published evi-
dence that these subscales are directly
related to a basic, unifying concept. If
ratings over time on these subscales
were not correlated or went into oppo-
site directions, this would not necessarily
question the clinical usefullness of the
EE total score as a "melting pot" for dif-
ferent relapse-related variables but it
would certainly question its "meaning"
as a construct.
Comparison of reepeated assessments
over time of EE ratings and nuclear
symptoms (according to the Present
State Examination, PSE). Only such an
analysis of correlations over time can
produce empirical evidence for a con-
ceptualization of the CFI that may stop
the arguments about the "meaning" of
CFI results obtained in the "classic" way
(EE at discharge, relapse at the 9-
months follOW-Up).
Comparison of EE ratings and self-rat-
ings of multiple neurotic symptomatolo-
gy and neurotic traits over time in (a) the
relatives (e.g., are obsessional relatives
more critical than nonobsessional?) and
in (b) the patients (do specific neurotic
symptoms or traits in patients specifical-
ly correlate with high scores in EE scales
of their relatives?).
Indirect Evaluation of the eFI:
Group Therapy for Patients and Relatives
Separately
According to the CFI literature in the late
1970s, "immunization by neuroleptic medi-
cation" (of the patient) and reduction of
face-to-face contact between the patient
and high EE family members (less than
35 hlweek seemed to be powerful means of
protecting vulnerable patients from re-
lapse-inducing, high EE-related behaviors
of their relatives. (It ought to be remem-
bered here that this causal attribution was
implicit in most of the studies available until
the late 1970s.) These interventions seemed
easy to administer and were the most fre-
quently applied.
Surprisingly, most of the CFI-related tre-
atment studies in those years employed
long-term neuroleptic medication and only
short-term psychosocial treatment (often
only on an inpatient basis). Assuming that
family and social learning factors have a
major impact on the course of schizophre-
nia, one would rather have expected the re-
verse proportion of both treatment ingre-
dients - medication for patients as intermit-
tent crisis intervention and psychosocial
treatment as the long-term stabilizer.
As far as the psychosocial treatments
themselves are concerned, they did not
seem to systematically use interventions
derived from information gathered with the
CFI.
In recent years, there has been a major
change in the content of CFI-related treat-
ment programs. A variety of approaches to
directly affect negative family interactions
are now being investigated: (a) social skills
training groups for patients (Liberman,
Wallace, Vaughn, and Snyder, 1980; Wal-
lace, Vaughn, and Snyder, 1982), (b) "psy-
choeducational" groups for relatives, in
combination with family sessions (Ander-
son, Hogarty, and Reiss, 1980), (c) rela-
tives' groups, with a main focus on CFI-
derived themes to achieve reduction of high
 Short-Term Relapse in Young Schizophrenics
EE-related attitudes and behaviors (Berko-
witz, Kuipers, Eberlein-Vries, and Leff,
1981), (d) mixed relatives' groups with high
and low EE members in an attempt to em-
ploy low EE relatives as co-therapists in
teaching skills for coping with schizophre-
nia to family members with high EE ratings
(Leff, Kuipers, and Berkowitz, 1979), and
(e) treatment of the single family unit in its
natural environment emphasizing CFI-re-
lated themes (Falloon, Boyd, and McGill,
1984).
In our design, we (Hand and K6ttgen, in
Hand and Gross, 1980) decided to try yet
another approach, indirect family therapy:
The families were treated in separate
groups for patients and their relatives with
the same therapists treating those groups
into which the members of each single
family had been separated. In half of these
families, the patients received "high-inten-
sity" (HI) treatment weekly group sessions)
and their relatives a corresponding "low-
intensity" (U) treatment (monthly group
sessions); in the other half, the reverse ap-
plication of group treatments was made.
Treatment was entirely on on outpatient
basis and was meant to go on for 2 years.
CFI-derived information provided the main
themes, both in the patient's and relatives'
groups.
It was hoped that a comparison of the
outcome in the patient-centered group
therapy would allow evaluation of the fol-
lowing questions within one single study:
Can high EE ratings be directly affected by
intensive group treatment of relatives, and
how does this affect patients?
Can patients and their course of illness be
directly changed by intensive treatment,
and how does this affect high EE ratings in
their relatives?
Are separate group treatments for patients
and their relatives a helpful and economical
means of treating families with problematic
or traumatic interactions?
The main treatment aims originally were
to (a) reduce (interactional) vulnerability in
the patient, (b) reduce harmful (interac-
tional) events in daily life, in particular from
high EE-related attitudes and behaviors in
relatives, (c) reduce patients' behaviors that
might cause or maintain high EE behaviors
in their relatives, (d) increase coping and
61
other skills in patients and relatives, and (e)
offer training in recognizing the occurrence
of schizophrenic symptoms such as an ex-
pression of the illness, an attempted coping
response for more severe (e.g., social) de-
ficits, and an "interactional maneuver."
Additionally, basic information about
the illness of schizophrenia was given. In
contrast to the "psychoeducational" ap-
proach, the delivery of "knowledge" to pa-
tients and relatives was tried in a much less
directive manner, spaced over much longer
periods of time, with smaller amounts of in-
formation to be digested in a given session,
and with more room for the patients to in-
fluence the structure and content of the
treatment session.
The treatment plan is described in some
detail by K6ttgen, S6nnichsen, Mollen-
hauer, and Jurth (1984). This final strategy
was developed on an eclectic-psychodynam-
ic basis derived from the cooperation and
close mutual supervision of the therapist
dyads (for each treatment group) composed
of therapists from the project and from the
hospital. It was not possible to develop the
originally intended, structured behavioral
therapy program. It was also not possible to
continue the project for the originally plan-
ned 6-years period.
General Design and Overall Outcomes
One-hundred and twenty patients were re-
ferred from two hospitals based on admis-
sion interviews as part of the project. Crite-
ria for inclusion were: nuclear symptoms
(PSE) exhibited during the month before
admission; age, 18-30 years; totallength of
previous inpatient treatments no longer
than 1 year; maximum number of previous
hospital admissions, 3; key relative willing
to participate in assessments and, if offered,
treatment. The results of the application of
these selection criteria are shown in Fig-
ure 1. Only 13% of the originally referred
sample eventually received treatment.
Originally, 52 patients and their 79 rela-
tives were included in the project (further
reduction of participants, cf. p. 64). An at-
tempt was made to secure complex assess-
ments of patients and relatives: with self-ra-
ting scales on psychotic and mUltiple neurot-
ic symptomatology, personality, social
62 B. Dulz and I. Hand

1. Admission interviews
1 120 1
Excluded
22
tor" other reasons"

2.PSE Excluded
46
PSE

3. CFI

4. Design 29 23
(49reL) (30 reL)

EE + EE-

15 EE + 14 22
1 EE-
(16 tam.; 29 reL) (21 reL) (29 reL)

EE + EE-

Patient Relatives
groups groups

Fig. 1. Selection of patients and relatives.

 Short-Term Relapse in Young Schizophrenics
skills and deficits, and socioeconomic de-
velopment. Structured interviews were
done with the PSE, CFI, and a relapse que-
stionnaire. Except for the PSE and CFI, an
attempt was made to perform most assess-
ments with both the patients and relatives.
Assessment intervals, except for relapse
(first follow-up after 9 months), were plan-
ned as usual: during admission and 6, 18,
and 30 months after discharge. The CFI was
assessed with the abbreviated version by
Vaughn and Leff. A family was rated as
high EE when at least one relative scored 4
or 5 on EOI and made six or more critical
remarks. One project rater (C. K6ttgen)
received CFI training with C. Vaughn and
K. Snyder. She then trained two other rat-
ers in the project. For 12 families, interrater
reliability between her and one of the ad-
ditionally trained raters was reported to be
high (r = 0.95).
Of the 16 patients who received treat-
ment, only 3 were female. Of the 26 rela-
tives, 14 were mothers, 10 fathers, and 2
siblings. All relatives of high EE families
were offered participation in the treatment
program (if they were low EE themselves,
they are called low EE co-relatives in the
subsequent analyses). Allocation to high-
or low-intensity treatment groups was
mainly influenced by the sequence in which
patients were referred to the project. Be-
cause of the loss of families due to the vari-
ous selection steps, it often took months
from identification ofthe first "suitable" pa-
Experimental Results
We shall start our analyses of predictor vari-
ables for the course of schizophrenia (re-
lapse) with the family variables from the
CFI, to be followed by the patient variable
"nuclear symptoms" (NS) and the treat-
ment variables "medication" and "group
treatments." Finally, results from a combi-
nation of two to five variables for predic-
tions of relapse will be presented.
Family Variables
Before investigating the CFI variables in
detail, we will first present results with the
"classic" CFI analysis: assessment of re-
63
tient and relative to the last to form a group.
The patients had a mean age of 23
(18-30), their relatives of 53 years. Two-
thirds of the relatives interviewed were
mothers, and two-thirds of the parents were
still married.
The results for patients and relatives with
most of these assessment instruments are
reported in detail by K6ttgen et al. (1984) as
well as a direct comparison with the previ-
ous studies in London (Vaughn and Leff,
1976) and Los Angeles (Vaughn et al.,
1984). As differences may be important for
a discussion of our results, some are sum-
marized here (LO = London; LA = Los
Angeles; HH = Hamburg):
Percentage of families where both parents
were interviewed:
LO 27; LA 52; HH 38
Percentage of parents in total sample:
LO 46; LA 94; HH 85
Percentage of high EE families:
LO 57; LA 67; HH 56
Percentage of male patients:
LO 41; LA 77; HH 65
Percentage of first admissions:
LO 57; LA 17; HH 69 (first and second ad-
missions together: 92!) Percentage of re-
lapse in EE+: EE- families:
LO 48:6; LA 56:17; HH 41:57!
The EE and relapse results in our own study
here differ slightly as we - unlike K6ttgen
et al., but similar to the L.A. study - did
not include the patients with "persisting
symptoms" in all the analyses.
lapses at the 9-month follow-up and their
correlation with single EE ratings at the
time of the patients' discharge. This will be
folowed by a series of analyses with repeat-
ed assessments of EE variables (at dis-
charge and the 6-month follow-up, CFI-
FU), eventually again related to relapse
rates (at the 9-month follow-up, REL-FU).
"Classic" EE Assessment and Relapse
Relapse Rates at the 9-Month Follow-up:
Control and Treatment Groups Combined
The course of illness in our patient sample is
64 B. Dulz and I. Hand
summarized in Table 1. We formed then
three categories of "no relapse," "relapse,"
and "special course': As the operationaliza-
tion of relapse in the literature on schizo-
phrenia is still a largely unresolved prob-
lem, we decided to use criteria similar to
those described by Vaughn et al. (1982).
The six criteria which constitute the three
main courses of illness in Table 1 are:
la: No psychotic symptomatology since
discharge from the hospital; Ib: psychotic
symptomatology without deterioration in
the quality of daily life, maximum duration
of 1 week, with or without hospital admis-
sion; lc: intermittent, short-term psychotic
symptomatology without deterioration of
the quality of daily life.
2a: Psychotic symptomatology without
nuclear symptoms, for longer than 1 week,
without hospital admission;
2b: nuclear symptoms in addition to 2a.
3a: Readmission to the hospital for more
than 1 week, but no nuclear symptoms; 3b:
3a with nuclear symptoms.
4: Persisting symptoms with deteriora-
tion of the everyday quality of life.
5: Hospitalization for more than 50% of
the follow-up time.
6: Suicide.
Relapse was defined with regard to its in-
dividual as well as social relevance. Hospi-
tal readmission of less than 1 week is inter-
preted as one means of preventing relapse
in a short-term psychotic crisis. Likewise,
intermittent short-term psychotic sympto-
Table 1. Relapse rates at the 9-month follow-up: control and treatment groups combined.
Relapse
category n %
1a 16 33
1b,c 5 10
2a,b 3 6
3a,b 13 27
4 6 13
5 3 6
6 2 4
48
matology was not defined as relapse when
the private and (or) professional life of the
patient was not affected by the sympto-
matology.
Table 1 does not include the data of four
patients (which are included in the publica-
tion by Kottgen et al., 1984), as these pa-
tients had moved away and information had
come only from their relatives. The remain-
ing 48 patients include those 5 from cate-
gories 5 and 6 ("special course") who will be
excluded from all those subsequent analy-
ses that rely on follow-up data (resulting in
a n of 43). Category 5 was reached by three
an n of 43). Category 5 was reached by three
female patients who had not been discharg-
ed from the hospital during the REL-FU.
The two patients in category 6 had commit-
ted suicide while on the waiting list for their
assigned project group.
The overall outcome of approximately
50% relapse in young schizophrenics during
the first 9 months after discharge is in ac-
cordance with previous reports in the litera-
ture.
EE at Discharge and Relapse at the
9-Month Follow-up
Of all our patients from the relapse cate-
gories 1-5 (n = 46; cf. Table 1), 25 (54%)
had relapses. The relapses occured in 14 of
the 29 (48%) patients with high EE rela-
tives, but - contrary to all previous publica-
tions - even in 11 ofthe 17 (65% ) from low
EE families.
n %
21 43 No relapse
22 46 Relapse
5 10 Special course
48
 Short-Term Relapse in Young Schizophrenics
Course of EE Ratings over Time
In all subsequent analyses and figures with
EE results, we shall first report the "total"
EE score (EE), followed by the results on
the subscales of emotional overinvolve-
ment (EOI) and criticism (CRIT). The sub-
scale of hostility is not included as relatives
never reached the cutoff point for a high
score. Simultaneously high ratings on EOI
and CRIT at the first assessment were only
observed in 4 of 63 relatives; one remained
high at FU only in CRIT, two scored low on
both scales, and the fourth did not partici-
pate in the FU.
Course of EE Ratings over Time
Treatment and Control Groups Combined
The results on all scales together are shown
in Table 2: 63 relatives of 46 patients partici-
pated in admission interviews as well as in
CFI-FU. Of these 63,46 relatives came from
high EE families. These 46 included 13 low
EE co-relatives from families with at least
one high EE relative. Therefore, the EE+
ratings in Table 2 are from a total of 33 rela-
tives (46 minus 13), whereas the EE+ ra-
tings are from 30 relatives (17 from the low
EE control group and 13 from low EE co-
relatives of high EE families). This differen-
tiation cannot be made in both subscales be-
Table 2. Course of EE ratings over time: treatment and control groups combined.
2nd CFI
EEscales
+ -
EE+ 17 16
EE- 5 25
EOI+ 7 8
EOI- 2 46
CRIT+ 10 12
CRIT- 8 33
65
cause any of the EE+ patients could score
initially low on one of the two subscales.
On the EE score, 16 of 33 initially high
EE relatives scored low at FU, whereas
only 5 of 30 initially low EE relatives
changed to high ratings at FU. Only 8 of all
these relatives were spouses, 7 of whom
scored low on both occasions.
Results on the total EE score resembled a
labile, state-like quality of high EE and a
rather stable, trait-like quality of initially
low EE ratings. The EOI and CRIT scales
showed almost identical developments of
ratings over time, with a particularly high
stability of low EOI ratings.
Course of EE Ratings over Time:
Control Groups
Whereas in the treatment groups all rela-
tives participated in both CFI ratings, the
control groups, in particular the one with
low EE relatives, showed a much lower re-
test compliance: 12 of 29 (41 %) of the low
EE relatives and 3 of 21 (14%) of the high
EE relatives refused to participate in the
second assessment. Understandably, retest
compliance is higher in treated than in un-
treated persons, but it does not appear to be
as obvious as to why the dropout rate was so
much higher in the low EE relatives.
33
30 EE
total score
63
15
EE
48 emotional
overinvolvement
63
22
EE
41 criticism
63
66 B. Dulz and I. Hand
Table 3. Natural course ofEE ratings over time: control group (high and low EE groups separately).
2nd CFI
EEscales
+ -
EE+ 6 7
[EE-] 2 3
EE- 2 15
EOH 2 2
EOI- 1 13
EOI- 0 17
CRIT+ 3 6
CRIT- 2 7
CRIT- 2 15
Table 3 shows the results for the high EE
and the low EE control group separately.
The low EE co-relatives of high EE families
are now separated (the separation being in-
dicated by brackets around EE- ratings)
from the low EE relatives in the low EE
control group. We introduced this addition-
al subgroup, as we wanted to investigate
whether low EE co-relatives resemble more
their high EE counterparts or the low EE
control group in the course of ratings (cf.
p.74).
At CFI-FU, 7 of 13 initially high EE rela-
tives scored low, whereas only 2 of the ini-
tially 17 low EE relatives had changed in the
reverse direction. Results in the untreated
control groups thus resembled those from
the total sample. The same holds true for
the EOI ratings. The scale for CRIT even
showed a decrease of high scores in two-
thirds of the subjects (without treatment).
Of the five low EE co-relatives, two
changed to high ratings.
Course of High EE Ratings over Time:
Treatment and Control Groups
Direct comparison of treated (n = 20) and
untreated (n = 13) initially high EE rela-
13 HighEE
5 control grp.
EE
total score
17 LowEE
controlgrp.
4 HighEE
14 control grp. EE
emotional
17 LowEE overinvolvement
controlgrp.
9 HighEE
9 control grp.
EE
17 LowEE criticism
control grp.
tives on all three scales surprisingly did not
show a specific treatment effect (in both
conditions, some 50% of initially high scor-
es were low at FU, whereas low scores
showed little change). Paradoxically, the
CRIT ratings even seemed to indicate a
beneficial control group effect. In the low
EE co-relatives, two of eight became high
in the treatment groups compared with two
of five in the control groups (Table 4).
Comparison of high-intensity groups for
relatives (HI rel.), with 23-26 treatment
sessions offered to participants, and low-in-
tensity groups for relatives (LI rel.), with
8-24 sessions offered, showed no differ-
ences between groups, and both resembled
the results reported before on the total EE
scores.
However, as Table 5 indicates, HI rel.
groups seemed to specifically affect ratings
in both subscales; while initially high EOI
ratings decreased in this group, initially low
CRIT ratings increased (the LI rel. group
did not show this effect).
Course of EE Ratings and Relapse
Since EE ratings at discharge did not prove
 Short-Term Relapse in Young Schizophrenics 67
Table 4. Course of high EE ratings over time: treatment and control groups.

2ndCFI
Treatment groups Control groups

EEscales EEscales
+ - + -
EE+ 11 9 20 6 7 13 EE
total score
[EE-j 2 6 8 2 3 5
28 18 46

EOI+ 5 6 11 2 2 4 EE
emotional
EOI- 1 16 17 1 13 14 overinvolvement
28 18 46

CRIT+ 7 6 13 3 6 9
EE
CRIT- 4 11 15 2 7 11 criticism

28 18 46

to be useful predictors of relapse in our investigate EE ratings and relapse in high

study, we shall now investigate whether re- EE families (n = 46 relatives), both in treat-
peated EE ratings can increase the predic- ment and control groups (again including
tive power of this instrument. We shall first the low EE co-relatives separately under

Table 5. Course of high EE ratings over time: effects of high-intensity vs low-intensity relatives
groups.

2nd CFI
Treatment groups Control groups
(LI pat. groups) (HI pat. groups)

EEscales EEscales
+ - + -

EE+ 5 5 10 6 4 10 EE
3 1 4 5 total score
[EE-j 1 2
13 15 28

EOI+ 2 5 7 3 1 4 EE
emotional
EOI- 1 5 6 0 11 11 overinvolvement
13 15 28

CRIT+ 4 3 7 3 3 6
EE
CRIT- 3 3 6 1 8 9 criticism

13 15 28
68 B. Dulz and I. Hand
the total EE score), then in low EE families no relapse, but two-thirds changed to low
(n = 17 relatives). even when patients relapsed!
Results from High EE Families (Treatment Do such diverse developments of the
and Control Groups Combined): three EE scales over time question the va-
lidity of the (most commonly used) total
As already mentioned, some 50% of the score?
high EE relatives changed to low EE total
scores at FU, independently of relapses in Results from Low EE Families
patients. Initially low EE co-relatives re- (Control Group):
mained the same only when patients did not Of the 17 relatives in the low EE control
relapse; three of four initially low EE co- group (10 with a relapsed patient), 15 re-
relatives of the patients who relapsed mained the same at FU on all three scales.
changed to high EE. It is difficult to inter- The other two relatives, both with a re-
pret this result, as the initially low EE rela- lapsed patient, remained low in EOI, but
tives in the control group in the vast majori- became high in CRIT.
ty remained low, even when patients re-
lapsed (Table 6).
In contrast to high total scores, high EOI Patient Variables
scores showed a positive correlation be-
tween relapse and their development over Nuclear Symptoms and Relapse
time. Four of six initially high EOI relatives All seven patients with nuclear symptoms
remained the same at FU when patients re- (NS) at discharge (two) or at discharge and
lapsed, as compared with four of ten, when FU (five) "relapsed" during FU, five of
patients had suffered no relapse. Low EOI them with persisting symptoms (relapse cat-
ratings remained the same, independently egory 5).
of relapse. Of the 36 patients without NS at dis-
Results became even more inconclusive charge, 15 (42%) relapsed. Of the 21 pa-
with the CRIT ratings: high ratings changed tients without NS at discharge and no re-
to low in about 50% when patients showed lapse, only 1 developed NS at FU.

Table 6. Course of EE ratings over time: high EE families and relapse.

2ndCFI
High EE relatives of patients

Without relapse With relapse

EEscales EEscales
+ - + -

EE+ 10 9 19 6 7 13 EE
total score
[EE-] 1 9 10 3 1 4
29 17 46

EOI+ 4 6 10 4 2 6 EE
emotional
EOI- 1 18 19 2 9 11 overinvolvement
29 17 46

CRIT+ 6 6 12 3 6 9
EE
CRIT- 3 14 17 2 6 8 criticism
29 17 46
 Short-Term Relapse in Young Schizophrenics 69
Of the 22 patients with relapse, about taken into consideration. In fact, 9 of 12 pa-
one-third (n = 8) showed no NS at dis- tients who relapsed in spite of CNM com-
charge and FD, the second third (n = 7) NS plained about side effects, whereas only 6 of
only at FD, and only the last third (n = 7) 11 who received CNM and did not relapse
NS at discharge or on both assessments. expressed the same complaint.

Treatment Variables Group Therapy and Relapse

Continuous Neuroleptic Medication
and Relapse
Continuous neuroleptic medication (CNM)
was defined as a neuroleptic dosage com-
parable to 100 mg Haldol Decanoat in
monthly intervals for at least over half of
the follow-up time (the majority of patients
received depot neuroleptics, e.g., halo-
peridol decanoate, independently of the
project).
Although CNM compliance seemed to be
much higher in patients from high EE fami-
lies (17 of 26 = 65 %) - among those partic-
ularly high in the treatment groups - than
from low EE families (6 of 17 = 35%) me-
dication did not seem to affect relapse rates
specifically. Fifty percent of the patients
with and without medication relapsed
(Table 8).
Medication may show specific effects
when their side effects are additionally
Of the patients from high EE families, 7 of
13 (54%) relapsed in the control and 5 of 14
(36%) in the treatment groups. We shall
now look at both treatment conditions sep-
arately.
High-intensity group treatment (HI pat.)
was given to eight patients in two groups.
One of them only had the chance to attend
9 sessions, the others were offered between
28 and 43 sessions. Five of these patients re-
mained without relapse. They as well as two
of the three who relapsed attended more
than 75% ofthe treatment sessions offered.
Low-intensity treatment (LI pat.) was
given to another eight patients in two
groups. Of these, one refused treatment
and another one could not participate for
other reasons (both relapsed). Of the re-
maining six, who were offered between 8
and 14 sessions, two relapsed. These two at-
tended less than 50% of the sessions offered,
but similarly so did three of the four who did

Table 7. Treatment intensity and relapse.

Group treatments
HI pat. group LI pat. group
(LI rei group) (HIrel group)

Relapse No relapse
+ - + -
i::'"<) >75% 2 5 0 1
§ i::'"
co
<) ·3co 50%-75% 0 0 0 0
!:i ~
0.. <50% 1 0 2 3
;:::
0
....
OJ) 3 5 8 2 4 6
..c::
......
.~ >75% 1 3 0 2
<) '"
<)
c..> .~
I:: co 50%-75% 2 2 2 0
.~
V
0.. p::: <50% 0 0 0 2
8
0
U 3 5 8 2 4 6
70 B. Dulz and I. Hand
not relapse (cf. Table 7, upper section). These
short-term relapse data do not show better
results for high-intensity treatments (with
higher appointment compliance of patients)
than for low-intensity treatments (with also
lower appointment compliance of patients).
As the relatives of LI pat. groups them-
selves were offered HI reI. groups., with
regard to family problems, this may have
been a compensation: Every single family
unit in both patient groups received the
same total amount of treatment.
If HI reI. groups were to have such a com-
pensatory effect, this could only be ex-
pected from high compliance of the rela-
tives. As data in Table 7 (lower section)
show, this was not the case.
Whereas most patients in HI pat. groups
also showed optimum compliance and
those in LI pat. groups showed low com-
pliance (i.e., they intensified by their com-
pliance the separation of treatment condi-
tions by design), the relatives did not show
the same response pattern. We found that
in the HI reI. groups the four relatives of pa-
tients who did not relapse were equally
distributed to both extreme pools of com-
pliance, whereas the two relatives of re-
lapsed patients showed medium compli-
ance. In the LI reI. groups, no relative at-
tended less than 50% of the sessions of-
fered! Four of the eight relatives reached
medium, the other four optimum compli-
ance.
With regard to the single family units, it
became clear that HI paULI reI. treatment
produced by far the best compliance of
both family members in the treatment of-
Table 8. Relapse: family and treatment variables.
I EE ratings
Patients with CNM
Treatment Control
groups groups
+ + - n
+ 5 2 5 12
- 6 4 1 11
11 6 6 23
fered. It appears that patients' compliance
with treatment motivated their relatives to
participate regularly rather than the other
way around.
Multivariate Prediction of Relapse
As we have seen so far, the single variables
investigated in most analyses separated the
total patient sample into two or three al-
most equally sized subsamples - with the
exception of the presence of NS "at dis-
charge." Can the combination ofthese var-
iables enhance their predictive power?
Relapse: Family and Treatment Variables
Neither CNM nor group therapy alone spe-
cifically affected relapse rates. If we regard
both together, a third variable, high EE,
implicitly (by design) affects the outcome.
A look at the data of patients from high
EE families only (treatment and control
groups combined) reveals that 15 of them
had suffered no relapse, 9 of whom had
been in treatment groups. Of the nine with
and the six without group therapy, one-
third in each condition had received no me-
dication. Of the patients in group therapy,
there were only five who (together with one
relative) attended more than 50% of the
sessions offered, whereas four did not (cf.
Table 8).
In the treatment groups, a combination
of EE+, CNM, and group therapy (Table
8) only prevented relapse in 6 of 11 patients.
This result seems to be even worse than for
patients from high EE families with CNM,
Patients without CNM
Treatment Control
groups groups
+ + - n Total
2 2 6 10 22
3 2 5 10 21
5 4 11 20 43
 Short-Term Relapse in Young Schizophrenics
but without additional group therapy; in
this group, four of six patients did not re-
lapse. In the small subsamples of patients
from high EE families without CNM, group
treatment was not revealed to have exerted
a specific effect compared with no treat-
ment at all.
In the patients from low EE families,
CNM did not show any "protective" effect
regarding relapse rates; five of these six pa-
tients relapsed; on the other hand, only 6 of
11 patients without medication relapsed.
Relapse: Family, Treatment, and Patient
Variables
If neither family nor treatment variables
71
alone or in combination enhance predicta-
bility of relapse, what then can additonal
application of a patient variable ("no NS at
discharge" , no NSd) bring about?
No NSd and no CNM led to relapse in 9 of
19 and no NSd with CNM in 6 of 17 patients.
Of the five patients with no NSd and CNM
without side effects (SE) (ct. p. 69), only
one relapsed. In all following analyses, we
shall only use CNM without SE as the me-
dication variable, as this outcome as well as
the analysis of side effects tentatively sup-
port the clinical impression of their rele-
vance.
Relapse
+ -
+ 4 11 15

- 21 10 31

25 21 46

All groups
Relapse / Relapse

+ - + -

+ 2 10 12 + 2 1 3

- 12 5 17 - 9 5 14

14 15 29 11 6 17

,/ All EE+ groups EE- control group

Relapse Relapse

+ - + -

+ 0 6 6 + 2 4 6

- 7 3 10 - 5 2 7

7 9 16 7 6 13

/ All therapy groups ~ EE+ control group

Relapse Relapse

+ - + -

+ 0 5 5 + 0 1 1

- 3 0 3 - 4 3 7

3 5 8 4 4 8
HI pat. therapy group LI pat. therapy group

Fig. 2. Prediction of relapse by Family Compliance Index (FCI) for all groups.
72 B. Dulz and I. Hand
Eleven patients had no NSd and group
therapy, and only three relapsed. Six pa-
tients had no NSd, CNM without SE, and
group therapy, and none of them relapsed.
However, a fourth and possibly even a fifth
variable may have influenced this outcome.
These six patients were from a total sample
of seven, of whom two relatives participated
regularly in group treatment; also, five of
these six patients attended the HI pat.
groups. Obviously, introduction of the pa-
tient variable increased plurivariate pred-
ictability of relapse.
Finally, we would like to suggest a multi-
variante "Family Compliance Index" (FCI)
for prediction of the course of illness from
the compliance behavior of patients and
relatives with regard to the treatment of-
fered. This FCI consists of the following six
variables, four of which have to be present
in families with group therapy offered and
three in those without such an offer to
achieve a "high" rating:
1. Regular outpatient psychiatric treat-
ment after discharge.
Summary and Discussion
Overall Results: Prediction of Relapse
Of 46 mainly (92%) first- and second-
breakdown schizophrenic patients, 46%
(56%, including those with persisting symp-
toms) had relapsed at the 9-month follow-
up. Relapses occurred in 58% of the high
(total n = 29) and 65% of the low (total n =
17) EE families.
Prediction of relapse with any single of
the following family or treatment variables
appeared impossible: EE in single or
repeated assessments, continuous neuro-
leptic medication (CNM), and group treat-
ment for patients and for relatives. With or
without any of these variables, relapse rates
were about 50%. The apparent, but statis-
tically nonsignificant tendency for a treat-
ment effect (36% relapse in treated as com-
pared with 54 % in untreated high EE fami-
lies) disappeared when HI and LI pat.
groups were compared: the HI pat./LI reI.
families showed much higher treatment
compliance (with group sessions and med-
2. Regular, high-potency neuroleptic med-
ication at least during half of the obser-
vation time.
3. No side effects of neuroleptic medica-
tion.
4. Regular participation in patients'
groups.
5. Regular participation in relatives'
groups.
6. If group treatment is not offered, use of
alternative treatment modalities (thera-
peutic hostels, etc.).
A high FCI was shown by 6 ofthe 16 high
EE families in the treatment groups, and
none of the respective patients relapsed. Of
the 13 high EE families from the control
group 6 showed a high FCI; two of the re-
spective patients relapsed.
Of the 19 low EE families, only 3 showed
a high FCI - again indicating the specificity
of low EE families. Two of these three pa-
tients were among the 11 who suffered re-
lapse.
The results of the systematic application
of FCI on our total patient sample and its
subgroups is shown in Figure 2.
ication) than LI pat.IHI reI. families; yet,
relapse rates were identical. Within HI and
LI pat. groups, the low and high compliance
patients again showed almost identical re-
lapse rates. It should be remembered here
that these results refer to the effects of
"being under treatment," not to treatment
outcome! The patient variable of nuclear
symptoms (NS) had equally low predictive
power when there was a negative rating at
discharge. However, persisting NS during
inpatient treatment invariably were cor-
related with relapse.
Prediction by a combination of variables
prooved hardly more satisfactory: a com-
bination of relatives' and treatment variables
- EE, CNM, and group therapy - did not
result in lower relapse rates. In contrast, the
results appeared odd, as group therapy did
not seem to have an additional effect to
CNM in patients from high EE families,
and in those from low EE families (who, by
design, did not receive group treatment)
 Short-Term Relapse in Young Schizophrenics 73
CNM even seemed to make outcome worse The results do not support most of the im-
than no treatment at all. plicit assumptions in CFI research: High
Only the additional application of the pa- ratings decreased to low on all scales in
tient variable of NS (no NS at discharge, no most analyses in about 50% of the subjects
NSd) and the modification ofCNM to CNM - independently of relapse rates or treat-
without side effects (SE) seems to increase ments! They seemed to measure a state-like
predictability of relapse: of the six patients quality, the state-related variables remain-
who showed this combination of variables ing unidentified. Low ratings in the vast ma-
(EE+, CNM without SE, group treatment, jority remained stable (trait-like) over time
and no NSd), none relapsed. However, five on all three scales, with only about 10%
of these patients were also in the HI pat. changing to high scores.
groups; and of all six patients, both parents In comparison with high EE families, the
participated in relatives' groups (which only low EE families showed very distinct re-
happened in 7 of all 16 families with group sults: a lower compliance with FU appoint-
treatment). ments (59% as compared with 86% in high
From these results, we construed a Family EE control groups); a much lower FCI
Compliance Index (FCI) with six variables. (only 3 of 17 scored high, and this without
A high FCI indicated a low relapse rate in relapse prevention, 6 of 16 high EE families
high EE families, both with and without tre- with and 6 of 13 without group therapy
atment. It did not do so in low EE families! scored high, with low to zero relapses); fi-
This kind of plurivariate prediction of re- nally, patients from low EE families
lapse is only of limited practical value. So showed a higher relapse rate and lower
far, the CFI has been used as a predictor at compliance with medication (the latter
the beginning of outpatient aftercare (hospi- hardly being the cause of the first). Yet, as
tal discharge), whereas the FCI relies much K6ttgen et al. (1984) described, low and
more on (compliance) variables during out- high EE families did not differ in socio-
patient aftercare. Any prediction of an event economic variables.
that includes variables more closely related To complicate matters even further,
(in time and content) to this event is, of some of the more puzzling results with EE
course, bound to achieve better results - in subsamples are summarized. The two
with lower generalizability. Further, it can- subscales EOI and CRIT showed a very
not be concluded from such an assessment similar state-like instability over time (like
(FCI) whether a high score with a low re- the total EE scores) when ratings were ini-
lapse rate is a consequence of a treatment tially high; initially low ratings remained
variable (e.g., HI pat. groups as a "nuclear" low, particularly in EO!. Thus, there is no
treatment, around which the other inter- support for the assumption that drops in
ventions can be established) or whether an high EE total scores are due to drops in
as yet undetected "family health factor" CRIT rather than EOI (d. Chap. 8, this
(e.g., benign atmosphere, independent of volume).
EE) enabled all family members to respond On the CRIT scale, more high ratings
to any help offered - this health factor pos- changed to low in the control than in the
sibly being in itself sufficient enough to pre- treatment groups, and HI pat. groups even
vent relapse. Our results with the FCI in seemed to increase initially low ratings
treated and untreated high EE families (while reducing high CRIT, and even more
seem to support the latter hypotheses. high EOI). When patients developed no re-
lapse, some 50% of high ratings changed to
Specific Results: Direct and Indirect
low - but when patients relapsed, 66% be-
came low.
Evaluation of the CFI
On the EOI scale, there was a more direct
This is the only study so far that has tried to correlation of high ratings with relapse, the
systematically assess the course of ratings ratings remaining high in families with and
on three EE variables (EE, EO!, CRIT) turning low in families without relapse. Ini-
simultaneously and in relation to CNM and tially low ratings remained the same re-
CFI-derived group treatment. gardless of relapse. It appears difficult to
74 B. Dulz and I. Hand
subsume these divergent results under one
CFI concept.
Ratings of the low EE co-relatives (of
high EE families) differed from the low EE
control group, as they - like the high rat-
ings of their high EE family members -
seemed to correlate with relapse: three of
the four whose patients relapsed turned
high in their EE.
These small numbers certainly do not ques-
tion the experimentally supported assump-
tion by K6ttgen et al. (1984) that the vast
majority of low EE families of young schi-
zophrenics may actually be high EE (if in-
stead of one at least two family members
are interviewed). However, taking into ac-
count all the specific outcomes in low EE
families in this study, the overall impression
is that low EE families are very special and
not just "hidden" high EE families.
Our relapse rates in low EE families seem
to support notions that this group of fami-
lies is inhomogeneous and includes a
"burnt-out" type (cf. Chap. 8, this volume),
with learned apathy and indifference to-
ward stressful life events. Too little positive
stimulation may occur for the patients of
these families (cf. Chap. 6, this volume).
Additionally, one may have to think of a
subgroup of "hypocritical" noncritical rel-
atives, who verbally cope and understand
and nonverbally send opposite messages.
Conclusions
The prediction of relapse and derivation of
optimal treatments for young schizophrenic
patients remains an unresolved problem.
EE results in this study raise rather than
answer questions. Almost none of the cur-
rent assumptions in the CFI literature could
be supported by our data.
Reasons for our contradicting results
may be manifold: K6ttgen et al. (1984) have
described in detail differences between our
patients' and relatives' samples and those in
previous studies with regard to duration of
illness, age, sex, and type and number of
References
Anderson, C., Hogarty, G., and Reiss, D.
(1981). The psychoeducational family treat-
ment of schizophrenics. In M. Goldstein (Ed.)
relatives interviewed. A particularly im-
portant variable from previous studies
seemed to be the amount of face-to-face
contact between patient and relatives (>35
hlweek increasing relapse rates). Again,
this could not be reproduced in our study: in
the high EE families, some 50% had more
than 35 h contact; yet, around 40% relapsed
each in the high- and low-contact environ-
ments. In the low EE families, 77% had low
contact, but nevertheless 69% of them re-
lapsed.
Could there be a relevant amount of er-
raneous CFI ratings, as there was only an
interrater reliability test between the
"properly" trained CFI rater in the project
and one of the two other raters she had
trained herself? This appears unlikely be-
cause of the very close cooperation between
all three raters throughout the whole pro-
ject.
Could it then have mattered that 65% of
all and 81 % of the patients treated were
male, whereas two-thirds of the relatives
interviewed were mothers and all CFI inter-
viewers and main therapists (in the thera-
pist dyads) were female? Finally, if differ-
ences in duration of illness were the decisive
factor for differences in EE and relapse re-
sults (i.e., treatment effects), then it would
no longer appear clinically meaningful to
give first- and second-breakdown schizo-
phrenics this diagnosis.
It follows that, at least for young "schizo-
phrenics" , more studies are needed with re-
peated assessments over time of all varia-
bles regarded relevant, and more attention
should be paid to low EE families.
The rather short duration of treatment
and FU, the loss of the majority of families
by the selection criteria, and compliance
problems with assessment appointments
and treatments severely restrict the con-
clusions that can be drawn for many of the
questions investigated. Unfortunately, pre-
vious studies with other outcomes and con-
clusions were hardly less affected by these
problems.
New developments in interventions with fami-
lies of schizophrenics. San Francisco: Joessey-
Bass.
 Short-Term Relapse in Young Schizophrenics
Berkowitz, R., Kuipers, L., Eberlein-Fries, R.,
and Leff, J. (1981). Lowering expressed emo-
tion in relatives of schizophrenics. In M. Gold-
stein (Ed.) New developments in interventions
with families of schizophrenics. San Francisco:
Jossey-Bass.
Bleuler, M. (1972). Die schizophrenen Geistes-
storungen. Stuttgart: Thieme.
Ciompi, L., and Muller, C. (1976). Lebensweg
und Alter der Schizophrenen. Berlin, Heidel-
berg, New York: Springer.
Falloon, I., Boyd, J., and McGill, C. (1984).
Family case of schizophrenia. A problem-solv-
ing approach to the treatment of mental illness.
New York: Guilford.
Hand, I. and Gross, J. (1980). Diagnostik und
Therapie von hochgradig ruckfallgefiihrdeten,
fruh erkrankten schizophrenen Patienten.
Finanzierungsantrag SFB 115, 1981-1983 Uni-
versitiit Hamburg.
Huber, G., Gross, G., and Schuttler, R. (1979).
Schizophrenie. Berlin, Heidelberg, New York:
Springer.
Kottgen, C., Sonnichsen, I., Mollenhauer, K.,
and Jurth, R. (1984). Results of the Hamburg
Cambcrwell Family Interview Study, I-III.
International Journal of Family Psychiatry, 5:
61-94.
75
Leff, J., Kuipers, L., and Berkowitz, R. (1983).
Intervention in families of schizophrenics and
its effect on relapse rate. In W. McFarlane
(Ed.) Family therapy in schizophrenia. New
York: Guilford.
Liberman, R., Wallace, c., Vaughn, c., and
Snyder, K. (1980). Social and family factors in
the course of schizophrenia. Toward an inter-
personal problem solving therapy for schizo-
phrenics and their families. In: Strauss, J.,
et al. (Eds.) The psychotherapy of schizophre-
nics. New York: Plenum.
Vauhn, c., and Leff, J. (1976). The influence of
family and social factors on the course of psy-
chiatric illness. A comparison of schizophrenia
and depressed neurotic patients. British Jour-
nal of Psychiatry 129: 125 -137.
Vaughn, c., Snyder, K., Simon, J., Freeman,
W., and Falloon W. (1984). Family factors in
schizophrenia relapse. Replication in Cali-
fornia of British research on expressed emo-
tion. Archive of General Psychiatry 41:
1169-1177.
Wallace, C. (1982). The social skills training pro-
ject of the Mental Health Clinical Research
Center for the study of schizophrenia. In Cur-
ran, J., Monti, P. (Eds.) Social skills training: a
practical handbook for assessment and treat-
ment. New York, Guilford.
76 B. Dulz and I. Hand
Comment on Chap. 5*
C. Vaughn high persisting symptoms appeared to be
less responsive to social influence than
The Hamburg investigators admit to being other patients, perhaps because of a dif-
puzzled by their own results, which appear ferent type of schizophrenia, more biologi-
to refute the main findings of all EE-based cally based. Yet, the Hamburg investiga-
studies to date. None of the factors previ- tors included patients with high persisting
ously identified as offering a measure of symptoms in the "relapsed" group. Thus,
protection against schizophrenic relapse the allocation of patients to "relapse" and
did so in this study: neither regular medica- "no relapse" categories differed considera-
tion, nor relatives' low EE, nor family ther- bly from past practice. The broader out-
apy. Particularly surprising is the high rate come criteria employed in Hamburg are not
of relapse in the low EE group, irrespective necessarily inferior to a criterion of out-
of medication compliance. In previous stu- come which is narrowly defined, but they
dies, the rate of relapses for low EE patients are different and should not be viewed as
on regular medication has been negligible. equivalent. Meaningful comparisons with
These contradictory conclusions clearly previous results cannot be made in the ab-
warrant additional consideration and com- sence of comparable criteria.
ment. A second major difference concerns an
That this was a well-intentioned and care- unreported but crucial aspect of the Ham-
fully executed study is not in doubt. How- burg study: the fact that a majority of pa-
ever, there were major differences in the tients, and low EE patients in particular,
design and methodology of the Hamburg did not live at home throughout the follow-
study, vis-a-vis previous EE studies, which up period. Indeed, low EE patients were
made it less of a "replication" than one much more likely than high EE patients to
would have wished. These may be sum- live away from parents and to have low
marized as follows: face-to-face contact (Kottgen, 1984, per-
1. Differences in the relapse criteria used sonal communication). The failure of the
to define outcome Hamburg investigators to either acknowl-
2. Unreported differences in the propor- edge this fact or to allow for it in the collec-
tion of patients who were living contin- tion and analysis of data casts doubt on
uously at home during the follow-up pe- many of the conclusions concerning low EE
riod families and relapse. Results from the Cali-
3. Unreported differences in the proce- fornia EE study emphasized the impor-
dures used to allocate families to high tance of the "at-home" factor. In the Cali-
EE and low EE groups. fornia study, unlike earlier British studies, a
A fuller account of these differences may sizeable minority of the total sample lived
explain at least in part the discrepant results away from home for some part of the fol-
obtained. low-up period. Because the impact of the
Firstly, the Hamburg relapse criteria dif- key relatives might be "diluted" in these
fered in several important respects from cases (depending also on patterns of con-
those employed in earlier EE studies, de- tact), it was thought essential to consider
spite the author's claims to the contrary. the results for the at-home group separately
Past EE studies have used changes in spe-
cifically schizophrenic symptomatology to
define relapse, without regard to days spent * A copy of Chap. 5 was not available to Christine
in the hospital or quality of social function- Vaughn when she wrote her contribution to this
ing in the follow-up period. Furthermore, book. As results seem to question previous CFI
publications with regard to their generalizability
patients who never recovered from the key to young schizophrenics, we greatly appreciate
admission episode were excluded from sub- that she has accepted the invitation for her special
sequent analyses of the relationship be- comments. It was agreed beforehand that the
tween relatives' EE scores and their own re- authors would not comment upon these com-
lapse (Vaughn et al. 1984). Patients with ments in this volume.
 Short-Term Relapse in Young Schizophrenics
from the rest of the sample. As reported
elsewhere (Vaughn, Snyder, Jones, Free-
man, and Falloon, 1984), the relationship
between EE and relapse was predictably
strongest for patients living at home con-
tinuously, although there was still a signifi-
cant relationship for the sample as a whole.
However, the significant relationship disap-
peared if one considered only the patients
who lived away for some part of the follow-
up period. High EE patients were more
likely to remain well if away from home,
while low EE patients were more likely to
relapse. The interpretation of the low EE
data was that the benign influence of the
low EE relative was diminished if the pa-
tient lived away from home, with other re-
lationships and other stresses (including life
events) assuming increased importance.
These results have important implications
for the Hamburg study. Given the high pro-
portion of low EE patients living away from
home and exposed to various nonfamilial
influences, it seems quite unjustified to sug-
gest that their high relapse rates may be due
to relatives who are "burnt-out" or "hypo-
critical"!
A third major difference was the Ham-
burg decision to allocate to the low EE
group patients from households in which a
key relative who was not interviewed was
described by an interviewed low EE relative
as being critical/hostile or overinvolved
(Kottgen, 1984, personal communication).
This is not appropriate, even ifthe noninter-
viewed relative was in low contact with the
patient. In previous EE studies, such cases
would have been dropped because of insuf-
References
Vaughn, C. E., Snyder, K. S., Jones, S., Free-
man, W. B., and Falloon, I. R. H. (1984).
Family factors in schizophrenics relapse: Re-
77
ficient information. The households in que-
stion might well be high EE, and certainly
should not be treated as low EE. Such
misallocations are likely to have contrib-
uted to the increased rate of "relapse" in the
Hamburg low EE group.
In summary, there are a number of im-
portant differences in design and methodol-
ogy between the Hamburg study and previ-
ous EE studies which call into question its
status as a "replication" study. It is neces-
sary to look no further than these differ-
ences for possible explanations of discre-
pant results regarding EE and relapse.
Other puzzling findings, such as the failure
of medication to protect patients against re-
lapse and the absence of a specific effect for
family therapy, warrant a closer examina-
tion, which is beyond the scope of this com-
mentarv. However, the low levels of com-
pliance with family treatment may be parti-
ally explained by the fact that many patients
(especially low EE patients) lived away
from home, as well as by the clinicians' que-
stionable decision to delay family therapy
until the time when a treatment "group"
was complete. An approach in which pa-
tients and relatives may wait weeks or even
months for treatment to begin does not
seem to provide an optimal test of family
therapy.
Finally, it is not an implicit assumption of
the CFI research that EE levels are constant
over time. This issue is discussed at some
length elsewhere in this book (Chap. 8). In
most respects, the Hamburg results are con-
sistent with earlier findings concerning the
stability of EE.
plication in California of British research on ex-
pressed emotion. Archives of General Psychia-
try, 41,1169-1177.
6. Emotional Atmosphere in Families of Schizo-
phrenic Outpatients: Relevance of a Practice-
Oriented Assessment Instrument
G. Buchkremer, H. Schulze-Monking, L. Lewandowski, and C. Wittgen

Introduction
The Camberwell Family Interview (CFI),
the measure used to assess "expressed emo-
tion" (EE), can hardly be used in routine
psychiatric work since the time required for
relatives and raters with this test instrument
is very great (about 4-5 h per rating). In ad-
dition, the amount of effort necessary to ar-
rive at a sufficiently high interrater reliabili-
ty during rater training is hardly possible
under the given conditions.
For psychiatric practice, the question
arises as to whether the results of Vaughn
and Leff (1976 a) obtained with the CFI can
be replicated using a simpler measurement
instrument. An appropriate survey instru-
ment, the Munster Familiy Interview
(Munsteraner Familienbogen, MFB), is de-
scribed in this paper. This interview not
only provides a measure of the emotions ex-
pressed by the individual family members,
but also registers the entire emotional fami-
ly atmosphere (EFA) and might therefore
be an alternative measure of the EE con-
struct.

Method
Munster Family Interview (MFB)
The MFB is a practicable survey instrument
which requires both less rater training and
expenditure of time than the CFI. By means
of this instrument, nine qualities of feeling
(EFA factors) in the emotional family at-
mosphere are surveyed (Table 1).

Table 1. Emotional Family Atmosphere (EFA) Factors.

Hostility Statements of aggressive attitude toward patient, impatience

Criticism Comments of nonagreement with behavior or attributes of the patient
Overprotectiveness Tendency to take care of patient in overly concerned way, belief that
patient cannot take care of himself
Rejection Statements against the patient as a person, lack of any kind of empathy
Resignation Hopelessness that patient will get better, no impulse to improve
situation, pity
Acceptance Relative compliance with illness, patient with his problems accepted,
constructive criticism
Warmth Empathy, attachment, expression of affection
Indifference No interest at all in patient's matters, no inclination to support,
no empathy, laissez-faire attitude
Predominance Tendency to influence and control patient
80 G. Buchkremer et al.
The rating is performed by one or two
trained raters in the context of a home visit
by a semistructured interview correspond-
ing to the CFI. The interview is conducted
with the whole family, i.e., the patient as
well as all relatives together. The evalua-
tion covers directly expressed feelings, re-
ports on feelings and behaviors from the re-
cent past, as well as other forms of emotion-
al expression such as the sound ofthe voice,
mimicry, gestures, and bearing. A rating on
a four-point scale is carried out with regard
to each EFA factor by raters who have been
trained in this procedure (it takes approxi-
mately 3 days to arrive at a sufficient inter-
rater agreement).
The test subjects on whom the MFB was
used and validated were from a research
project on relapse prevention in schizo-
phrenic outpatients (cf. Buchkremer and
Fiedler 1982). Twenty-four patients were
diagnosed as schizophrenic according to
ICD reference criteria. They had at least
two acute psychotic manifestations and
were checked in addition by an independent
expert with regard to the diagnosis of schi-
zophrenia. In addition, the last acute psy-
chotic exacerbation or the last hospitaliza-
tion had to have occured more than 6
Results
Reliability
Ten advanced psychology students were
trained for 24 h in the rating of the EFA fac-
tors. The training was carried out using
videotapes of family interactions. A loss of
information had to be accepted as the fami-
ly atmosphere can be experienced more viv-
idly during the home visit than on a video-
tape. In this way, the EFA factor indiffer-
ence and warmth were hard to rate since
their rating is especially based on nonverbal
behavior which is difficult to assess from
videotapes.
Ratings of the EFA factors were done in-
dividually by each rater, and overall 18 fam-
ily members were rated. To test reliability a
dichotomy was used, contrasting a moder-
ate or high rating with a mild or "no evi-
dence" rating. The results are shown in
Table 2.
months previously. The patients were thus
not receiving inpatient or partial inpatient
treatment at the time of the study. All of
them were living in their family.
All family members, including the pa-
tient, were rated during a home visit with
regard to their emotional expressiveness
(EFA factors). A mean value was calculat-
ed for each emotional factor based upon the
values of all the relatives present at the
home visit. These mean values were used as
indicators of the emotional atmosphere of
the whole family. In this way, a distortion is
possible as when, for example, a highly crit-
ical mother was present with an uncritical
father or siblings.
Rating in the context of a home visit had
the further advantage that the interactions
between the family members and the pa-
tient could be observed in their natural sur-
roundings. In addition, the home visit
enabled the establishment of a strong and
buoyant relationship between the relatives
and the therapist which was helpful in sub-
sequent therapy and motivated the relatives
to make themselves available for catamnes-
tic investigations (Buchkremer and Lewan-
dowski 1984).
For the EFA factors of criticism, over-
protectiveness, and rejection, a good de-
gree of reliability resulted, and for the EFA
factors of hostility, resignation, predomi-
nance, and acceptance, a moderate degree
of reliability. Low Kappa coefficients were
calculated for the EFA factors of indiffer-
ence and warmth. For the reasons men-
tioned above, they were difficult to rate and
had only a low frequency of moderate or
high degree rating but a good agreement.
(In a study which is currently in progress, a
reliability check will be carried out in the
context of home visits.)
Validity
Different aspects of the validity of the MFB
will be investigated below. According to the
results of Vaughn and Leff (1976 b), criti-
cism, and to a small extent, emotional over-
 Emotional Atmosphere in Families of Schizophrenic Outpatients 81
Table 2. Interrater Reliability of the EFA Factors (n = 18).

Kappa Agreement!

EFAfactors Mean SD Median Range Mean

Criticism 0.69 0.23 0.73 0.18-1.0 0.89
Hostility 0.41 0.35 0.44 0.00-1.0 0.93
Overprotectiveness 0.75 0.12 0.73 0.45-1.0 0.90
Rejection 0.66 0.20 0.64 0.31-1.0 0.94
Resignation 0.51 0.22 0.46 0.17-1.0 0.87
Indifference 0.09 0.04 0.00 -0.08-1.0 0.94
Acceptance 0.40 0.20 0.37 0.05-0.82 0.81
Predominance 0.56 0.15 0.56 0.28-0.85 0.80
Warmth 0.14 0.26 0.00 -0.15-1.0 0.83
1 Percentage of agreement between two raters.

involvement, were especially predictive of lapse criteria. A high measure of criticism

relapse over the course of 1 year. The ques-
tion therefore arises as to whether these
same EFA factors were equally valid re-
lapse predictors.
To answer this question, three different
and mutually exclusive relapse definitions
were chosen:
1. Rehospitalization (more than 2 days) or
more than 5 days in a day clinic
2. Outpatient treatment: sick leave due to
psychotic symptoms, recommencement
of neuroleptics, doubling of the dose of
neuroleptics
3. Subjectively experienced relapse
A subjectively experienced relapse was
the term used to designate a crisis situation
experienced by the patient, which could be
mastered without the aid of the psychiatrist.
The results (Table 3) indicate that the
EFA factors of criticism and indifference
display significant correlations with the re-
by the relatives is accompanied by an espe-
cially great danger of relapse in the patient.
According to the results of Brown et al.
(1972) and of Vaughn and Leff (1976 b),
emotional overinvolvement would also
have had to lead to the expectation of a
raised risk of relapse. The English word
overinvolvement, i.e., emotional overin-
volvement, was translated in the definition
of the MFB factor with the term "overpro-
tectiveness." A change in meaning thereby
arose which might possibly explain the dif-
ference between the result of this study and
that of the English authors. With regard to
the MFB factor, a prognostic validity is thus
not to be assumed. The EFA factor of in-
difference, surprisingly, showed several
positive correlations with relapse criteria.
These criteria were met in the course of 1
year after therapy.

Table 3. Phi Correlation Coefficients Between EFA Factors and Relapse on I-Year Follow-up
(n = 23)

Relapse Rehospitali- Psychiatric outpatient treatment Subjectively

zation Sick leave due Recommence- Doubling of experienced
EFAfactors to psychotic mentof the dose of relapse
symptoms neuroleptics neuroleptics

Criticism 0.44* 0.48** 0.38* 0.14 0.44*

Indifference 0.73** 0.57** 0.06 0.39* 0.35***
* P < 0.05; ** P < 0.01; *** P < 0.001
82 G. Buchkremer et al.
Table 4. Rank Correlation Coefficients (Spearman) Between EFA Factors and Psychopathological
Symptom Areas (AMDP) (n = 22-24)

Impairment
Attention Compul- Delusion Hallu- Ego Affective
EFA disorder sion cination disorder disorder
factors

Criticism 0.38* 0.15 0.28*** 0.06 0.35* 0.35*

Overprotec-
tiveness 0.16 0.14 0.11 0.28*** 0.06 0.Q7
Rejection 0.09 0.10 0.52** 0.13 0.Q7 0.21
Indifference 0.42* -0.13 -0.04 0.32*** 0.08 0.10
Warmth -0.33*** 0.28*** -0.Q7 -0.46* -0.09 -0.Q1
Acceptance -0.26 0.10 -0.14 -0.44* -0.16 0.02
* P < 0.05; ** P < 0.01; *** P < 0.001

Emotional Family Atmosphere and
Psychopathology of the Patient
Marked involvement of the relatives in the
patient'slife can be understood as a possible
reaction of the relatives to a schizophrenic
disorder (coping strategies). The correla-
tion between the EFA and the psychopa-
thological status of the patient will there-
fore be investigated next.
The psychopathological status of the pa-
tient was scored by third-person rating
using the AMDP system (a German psychi-
atric rating system; AMDP 1982) and by
self-rating using the Frankfurt syptom in-
ventory (FBF; Siillwold 1977).
Results of the Third-Person Rating
(AMPD)
The EFA factor of criticism displayed sig-
nificant correlations with three areas of psy-
chopathological disorders (attention, ego,
and emotional disorders). The EFA factor
of indifference, which (like criticism) was
found to have prognostic value , also showed
a correlation with disorders of attention.
The paranoid hallucinatory syndrome
was not correlated with the EFA factors of
criticism or indifference. However, in the
delusions there was a significant correlation
with the EFA factor of rejection. Hallucina-
tions correlated with low levels of warmth
and acceptance.
The following investigation is intended to
clarify whether there is a correlation be-
tween the cognitive disorders rated by the
patients themselves and the EFA.
The patients (n = 24) answered the short-
ened form of the FBF by which the subjec-
tive experience of thought and perception
processes can be surveyed (Siillwold, 1977,
1983). The shortened form of the FBF as-
sesses three factors.
Factor A: Disturbance of perception and
movement activity
Factor B: Disturbance of thought and
speech
Factor C: Anxiety and coping reactions

Table 5. Rank Correlation Coefficients (Spearman) Between FBF Scores and EFA Factors (n = 24)

EFA
FBF FBFscore FBFfactorA FBF factor B FBFfactorC

Criticism 0.30*** 0.25 0.28*** 0.36*

Indifference 0.60** 0.56** 0.59** 0.37*
* P < 0.05; ** P < 0.01; *** P < 0.001
 Emotional Atmosphere in Families of Schizophrenic Outpatients
Results of the Self-rating (FBF)
There were statistically significant correla-
tions between the EFA factors of criticism
and indifference and the FBF factors. In
particular, indifference did not only corre-
late in the overall score of the FBF but also
significantly with the individual factors
below the 1% level.
The results appear to indicate that schizo-
phrenia-linked disorders of cognitive
thought and perception processes (factors
83
A and B) are significantly correlated with
"indifference" of the relatives. Anxiety and
coping reactions of the patients (factor C),
on the other hand, are correlated with criti-
cism and indifference of the relatives.
The validation of the MFB with other cli-
nical or prognostic criteria failed to yield
any results which were significant in the sta-
tistical sense. In particular, there were no
significant correlations between the MFB
and the social status and disease course of
the patients.

Discussion
In contrast to the CFI, the MFB proved to teria of doubling the neuroleptic medica-
be a practical instrument for rating the en- tion indicates that the more chronic patients
tire EFA in the presence of all family mem- who are undergoing chronic neuroleptic
bers and of the patient. The results indicat- treatment are likely targets of indifference.
ed a satisfactory criterion validity. The in- In patients with a mainly critical family at-
terrater reliability was sufficient. mosphere, a long period of neuroleptic
With regard to prognostic criteria, the re- treatment was rarer. In these patients, there
sults of the MFB agreed with previous CFI was a more frequent fresh prescription of
research. As noted with the CFI, relapse neuroleptics in cases of crisis. Furthermore,
was associated with criticism expressed by they also experienced a subjective feeling of
the relatives. The present study confirms relapse significantly more often.
the results of Brown et al. (1972) as well as With some caution, these results can be
Vaughn and Leff (1976 b). interpreted in the following way: A critical
Surprisingly, an indifferent attitude of the family atmosphere is too great a strain for
relatives with regard to the patient likewise many patients, above all when they are
indicated a raised risk of relapse of the pa- under chronic neuroleptic treatment so that
tient. One reason for this appears to be that their vulnerability threshold is exceeded
the MFB assesses not only the family mem- and a relapse occurs. A high measure of in-
bers who are interacting most intensively difference in the family indicates that self-
with the patient but also the remaining rel- help potentials are absent in these families
atives. Evidently, both an emotionally in- and in these patients so that pychiatric in-
tense as well as a low-tension family atmos- terventions (e.g., rehospitalizations) be-
phere can entail a raised risk of relapse. come necessary to a greater extent in times
Great significance is to be accorded to of crisis.
this finding since it indicates that the "tight- In the investigation of the connection be-
rope walk" between overstimulation and tween the symptoms of the patients and the
understimulation as described by Wing EFA, significant correlations were revealed
(1976) also applies to the EFA of schizo- with "criticism" and the psychopathological
phrenic patients. Thus, not only emotional symptom areas of attention, ego, and emo-
overinvolvement in the families of schizo- tional disorders. In addition, disorders of
phrenic patients is to be avoided, but also a attention also correlated with indifference
lack of involvement is to be prevented. This of the relatives.
means that family therapies which empha- Analogous findings resulted when the ex-
size more reserve or more indifference by tent of the disorders of cognitive processes
the relatives toward their unwell family rated by the patients themselves as well as
member must be viewed with caution. their coping reactions were related to the
The finding that an indifferent family at- EFA.
mosphere is associated with the relapse cri- These findings indicate a reciprocity of
84 G. Buchkremer et al.
the correlations between the EFA and the
psychopathological status of the patients.
The correlation between the EFA and the
emotional expression behavior of the pa-
tients could already be demonstrated in an
earlier investigation of Brown et al.
(1972), Vaughn and Leff (1976 b), and
Miklowitz et al. (1983), who reported no as-
sociation between clinical symptoms on ad-
References
AMPD. (1982). Das AMPD-System. Manualzur
Dokumentation psychiatrischer Befunde (4th
ed.). Berlin Heidelberg New York: Springer.
[English translation: Guy, W., and Ban, T. A.
(Eds.) The AMDP system, Berlin, Heidelberg,
New York: Springer (1982)].
Brown, G. W., Birley, J. L., and Wing, J. K.
(1972). Influence offamily life on the course of
schizophrenic disorders: A replication. British
Journal of Psychiatry, 121,241-58.
Buchkremer, G., and Fiedler, P. (1982). Ange-
hOrigentherapie bei schizophrenen Patienten.
In Helmchen, H., Linden, M., and Ruger, U.
(Eds.), Psychotherapie in der Psychiatrie. Ber-
lin, Heidelberg, New York: Springer.
Buchkremer, G., and Lewandowski, L. (1984).
Therapeutische Gruppenarbeit mit AngehOri-
gen schizophrener Patienten. In: C. Anger-
meyer, and A. Rinzuen (Eds.), Die AngehOri-
gengruppe. Stuttgart: Enke.
Buchkremer, G., Lewandowski, L., and Fiedler,
P. (1982). Emotionale Interaktionsmuster in
Familien schizophrener Patienten. Eine Unter-
mission and EE status. This is probably be-
cause our patients were examined 6 months
after their discharge from the hospital. The-
refore, the range of symptoms from few to
severe was broader than in the investiga-
tions of the authors mentioned above, who
examined patients on admission, when they
normally all have severe symptoms.
suchung zum EE-Index. Aktuelles Forum Psy-
cho (Suppl. I). (pp. 9-10). Erlangen: Perimed.
Mikiowitz, D., J., Goldstein, M. J., and Falloon,
J. R. H. (1983). Premorbid and symptomatic
characteristics of schizophrenics from families
with high and low levels of expressed emotions.
Journal of Abnormal Psychology, 92,
359-367.
Siillwold, L. (1977). Symptome schizophrener
Erkrankungen. Uncharakteristische Basissto-
rungen. Berlin, Heidelberg, New York: Sprin-
ger.
Siillwold, L. (1983). Schizophrenie. Stuttgart:
Kohlhammer.
Vaughn, C., and Leff, J. (1976 a). The measure-
ment of expressed emotion in the families of
psychiatric patients. British Journal of Social
and Clinical Psychology, 15, 157-165.
Vaughn, C., and Leff, J. (1976 b). The influence
of family and social factors on the course of psy-
chiatric illness: A comparison of schizophrenic
and depressed neurotic patients. British Jour-
nal of Psychiatry, 129, 125 -137.
7. The Marriages and Interaction Patterns
of Depressed Patients and Their Spouses:
Comparison of High and Low EE Dyads
J. M. Hooley and K. Hahlweg
Introduction
Although we can now be relatively confi-
dent that there is indeed a reliable associa-
tion between family levels of expressed
emotion (EE) and relapse rates in schizo-
phrenic and depressed patients, a word of
caution is perhaps appropriate. Despite the
encouraging results of recent intervention
work, we are still some distance from a full
understanding of what EE actually is and
how it might operate to influence the pro-
cess of relapse. Discussions of the content of
interviews conducted with high and low EE
relatives, such as the one provided by
Vaughn in Chap. 8, are of clear value in in-
creasing our knowledge of the emotional
responses of families and in enhancing our
understanding of how they cope with a psy-
chiatrically ill individual.
A second approach to understanding EE
concerns systematic investigation of the be-
havior of high and low EE relatives, in an
effort to learn more about what goes on
when family members are involved in direct
interactions with an ill patient. Despite the
importance of this avenue of research, only
in very recent years have investigators
turned their attention toward the study of
the behavioral correlates of high and low
EE attitudes.
Clearly though, if high EE relatives have
a negative influence on the course of a pa-
tient's illness, it is likely that some differ-
ences between high and low EE families
must exist to trigger the relapse process.
Indeed, this is an assumption implicit in all
EE work. At the present time, however, we
have relatively little information about the
types of behavior or behavior patterns as-
sociated with high and low EE attitudes.
The EE literature assumes that a critical at-
titude expressed during an interview has a
corresponding behavioral analogue in na-
turalistic patient-relative interactions. Only
in recent years has the validity of this as-
sumption actually been subjected to empiri-
cal scrutiny (see Chap. 4). Results of a study
published in 1983 by Valone, Norton,
Goldstein, and Doane first suggested that
there might indeed be differences between
the interactions of families which varied ac-
cording to EE level. Using a sample of 52
disturbed but non psychotic adolescents
and their families, these investigators were
able to demonstrate that in affectively
charged situations, parents who expressed
high EE attitudes made significantly more
criticisms during face-to-face interaction
with their offspring than parents who were
low in EE.1
Valone et al.'s results are important be-
cause they indicate that a high EE attitude
might be associated with high levels of criti-
cal behavior toward a patient. Additional
but rather weak support for behavioral dif-
ferences between high and low EE relatives
also comes from a study conducted by Kui-
1 While such data are encouraging, two minor
methodological issues should be noted. First, the
subjects used in the study were not adult psychia-
tric patients (the population on which all previous
EE research has focused) but were disturbed and
possibly "preschizophrenic" adolescents.
Second, the assessment of levels of EE was made
by rating the UCLA parental interview and not
the CFI which is the standard EE assessment
instrument. How much effect the use of a dif-
ferent interview has on the ratings of EE is not
clear, but nevertheless this is a feature of the
study which should be borne in mind in any inter-
pretation of the data.
86 J. M. Hooley and K. Hahlweg
pers, Sturgeon, Berkowitz, and Leff
(1983), which investigated rates of talking
and durations of looking in high and low EE
relatives. Although no differences between
patients were apparent, high EE relatives
did spend more time talking to patients and
less time looking at them than their low EE
counterparts. Such findings are consistent
with the view of Berkowitz, Kuipers, Eber-
lein-Vries, and Leff (1981) that low EE rel-
atives are better listeners.
To date, perhaps the best study of the in-
teractional correlates of EE in the families
of schizophrenics is that of Miklowitz and
his coUegues (Miklowitz, Goldstein, Fal-
loon, and Doane, 1984). In this investiga-
tion, not only did high EE parents differ
from low EE parents - the former showing
more emotional negativity toward their off-
spring during face-to-face interactions -
but within the high EE group differences
were also found between high EE critical
and high EE overinvolved parents. Over-
involved parents were much more likely to
use neutral-intrusive statements than high
EE critical parents. Thus, the pattern of in-
teraction exhibited in schizophrenic fami-
lies may differ not only according to the EE
level of the relatives, but also according to
whether there is marked evidence of emo-
tional overinvolvement.
All three studies of the behavioral corre-
lates of EE described above were con-
ducted using schizophrenic (or "preschizo-
phrenic") patients and their families. The
behavioral data from these samples, togeth-
er with the interview data presented by
Vaughn in Chap. 8, clearly suggest that all is
not well within high EE families and imply
Subjects
The data presented in this chapter are based
on a total sample of 44 patients and their
spouses, all of whom were recruited while
the patients were in a hospital suffering
from an episode of major depressive dis-
order. Since not all of these couples agreed
to be videotaped, however, behavioral data
are available on only 30 dyads. Demo-
graphic data and data concerning EE levels
and marital satisfaction ratings are based on
the full sample. Comparisons between the
that some real difficulties do exist within the
relationships of patients and high EE family
members. The physiological data referred
to earlier (see Chap. 2) also lend further
support to this view. Since there is now evi-
dence that high family levels of EE are asso-
ciated with elevated rates of relapse in de-
pressed patients (see Chap. 2), it is becom-
ing clear that EE influences are not specific
to schizophrenia. Information about the in-
terpersonal relationships and interaction
patterns of depressed individuals and their
relatives is thus important and worthy of
discussion here because of its potential val-
ue for furthering our understanding of the
nature of EE. Since no data currently exist
which examine such issues in schizophrenic
samples, a chapter on depressed patients
has consequently been included in this
book.
One principal difference between schizo-
phrenics and depressed patients, however,
is that the latter are much more likely to be
married. Their key relative is thus most
likely to be a spouse. This provides an ideal
opportunity for the assessment of the rela-
tionship between the patient and relative
using psychometrically sound inventories of
marital satisfaction and for examining the
association between marital harmony and
levels of EE. This chapter will focus on the
marital satisfaction levels reported by high
and low EE dyads2 within a depressed sam-
ple. In addition, data will be presented con-
cerning the interaction styles typical of such
groups. Finally, some remarks relevant to
intervention work with high EE families
will be made, based on the results describ-
ed.
videotaped subgroup and the full sample
(see Hooley, 1984) revealed that there was
no tendency for the 30 videotaped couples
2 Throughout this chapter, the term high EE cou-
ple or high EE dyad will be used to refer to a part-
nership involving a depressed patient and a high
EE spouse. The term low EE couple will be used
in a similar way to denote the relationship be-
tween a depressed patient and his/her low EE
spouse.

Table 1. Demographic Characteristics of Patients and Spouses Contributing Data to the Study
Group
Patients (n = 44)
Age
DAS
Beck
Social class
Children
Length of marriage (years)
Spouses (n = 44)
Age
DAS
Beck
to be in any way untypical of the total re-
search sample in their EE levels or their
level of reported marital satisfaction.
All participating patients were recruited
Inclusion Criteria
To be accepted into the study, patients were
required to meet the criteria of the Present
State Examination (PSE; Wing, Cooper,
and Sartorius, 1974) and be cases of depres-
sion at ID level 5 or above. In addition, it
was also necessary for each patient to meet
the Research Diagnostic Criteria (Spitzer,
Endicott, and Robbins, 1978) for major de-
pressive disorder. Beck Depression Inven-
Assessment of EE and Videotaping
Once it had been established that a particu-
lar patient met the criteria for inclusion into
the study, the spouse was interviewed either
at home or during a visit to the hospital, us-
ing the Camberwell Family Interview. This
interview, used to assess family levels of
EE, is fully described in Chap. 2. Then,
some days after the interviews had been
completed, couples who consented were
videotaped while talking together for a pe-
riod of 15 min. The Strodtbeck revealed dif-
Coding the Videotaped Interaction
The behavior of each couple videotaped
was coded using the Kategoriensystem fur
The Marriages and Interaction Patterns 87
M SD
48.2 11.7
103.6 17.8
26.0 11.0
19.6 11.1
2.3 1.1
21.8 13.0
48.8 11.9
106.0 16.6
5.6 3.9
from three psychiatric hospitals in or near
Oxford, England and, as can be seen from
Table 1, are fairly representative of a mid-
dle-aged, middle-class population.
tory (BDI) scores (Beck, Ward, Mendel-
sohn, Mock, and Erbaugh, 1961) confirm
that the patients recruited were generally
severely depressed, their mean BDI score
being 26.0. In contrast, their spouses
showed the low BDI scores typical of non-
depressed individuals (mean BDI for
spouses = 5.6).
ferences technique (Strodtbeck, 1951) was
used to generate items for discussion. Vid-
eotaping was carried out in the hospital
setting except in the case of one couple
who, due to the patient's refusal to remain
in the hospital, had to be videotaped at
home. During this recording, all efforts
were made to ensure that conditions were
kept as similar as possible to those used in
the hospital setting.
partnerschaftliche Interaktion (KPI). This
system, developed by researchers at the
88 J. M. Hooley and K. Hahlweg
Max Planck Institute of Psychiatry in
Munich, is expressly designed to handle
marital interaction data and has been de-
monstrated to have good reliability and va-
lidity. Although the KPI contains 12 dif-
ferent verbal codes, for the purposes of the
analyses to be described below, all behav-
iors were collapsed into their superordinate
LevelsofEE
Spouses were assigned to the high or the
low EE groups by a median split of the num-
ber of critical comments made. Thus,
spouses were placed into the high EE group
if they made seven or more critical com-
ments; those who made six comments or
less were categorized as being low in EE.
Since none of the spouses for whom video-
High and Low EE Spouses
A high EE relative may be one who makes
critical comments, emotionally over-
involved remarks, or both. Emotional over-
involvement in spouses, however, is much
less common than it is in parents. In the
present sample, this is reflected by the fact
that of the 44 spouses interviewed, only 3
showed levels of emotional overinvolve-
ment high enough to classify them as high
EE on the basis of this index alone. More-
over, two of these spouses were also highly
critical and thus would have been placed
into the high EE group solely on the basis of
Demographic and Clinical Findings
Planned t comparisons performed on the
high and low EE groupings derived from
the median split revealed several important
differences. As Table 2 shows, high EE
spouses were younger than low EE spouses
(mean age for high EE group = 44.4; mean
age for low EE group 53.3: t = -2.56, P <
0.02) and had been married for less time
(t = -2.56, P < 0.02). High EE spouses
were also rated as less warm than low EE
spouses (CFI rating of 1.8 vs 3.8: t = -4.93,
P < 0.001) and more hostile (t = 3.84, P <
0.001). The higherlevel of hostility found in
categories and classified as either positive,
negative, or neutral. Cohen's Kappa was
0.81. Full details of the coding system, its
psychometric properties, and its method of
application, can be found in Hahlweg, Reis-
ner, Kohli, Vollmer, Schindler, and
Revenstorf (1984).
tape data were available actually made ex-
actly six critical comments, the median split
resulted in high and low EE groups near-
ly identical to those recognized by Brown,
Birley, and Wing (1972) and Vaughn and
Leff (1976) and found to be significantly as-
sociated with relapse rates in schizophrenic
populations.
their critical comments. The third emotion-
ally overinvolved spouse, however, was a
man who made no critical remarks at all
while discussing with his depressed wife. His
wife unfortunately proved unable to com-
plete more than a few minutes of videotap-
ing, and this couple did not therefore con-
tribute any behavioral data to the study. All
high EE spouses on whom videotaped data
are available and whose interactions are de-
scribed in this chapter are thus critical indi-
viduals.
high EE spouses is not in itself very remark-
able since all hostile comments are, by def-
inition, also critical. Thus, individuals mak-
ing larger numbers of critical comments are
more likely to make some comments which
are also hostile, and this is reflected in the
strong positive correlation which exists be-
tween the two variables (r = 0.53, n = 43,
P < 0.001).
Differences between high and low EE
spouses were also foundf when levels of
marital adjustment were examined. Using
scores obtained on the Dyadic Adjust-
 The Marriages and Interaction Patterns 89
Table 2. Comparison of High and Low EE Spouses

HighEE LowEE
Measure (n = 24) (n = 18) P
Age M 44.4 53.3 <0.02
SD 9.5 12.8
Length of marriage M 17.7 25.6 <0.05
SD 10.4 14.6
Children M 2.3 2.4 NS
SD 0.9 1.5
Social class M 17.9 21.7 NS
SD 11.2 11.1
BDIofspouse M 5.7 5.1 NS
SD 4.4 3.1
BDlofpatient M 25.9 26.2 NS
SD 11.1 11.3
Warmth" M 1.9 3.9 <0.001
SD 1.4 1.1
Hostility" M 1.6 0.4 <0.001
SD 1.1 0.9
Positive remarks" M 1.7 2.6 NS
SD 2.3 1.7
D AS of spouse M 99.2 112.4 <0.02
SD 16.0 13.7
DAS of patient M 97.4 111.6 <0.05
SD 12.5 20.7
"CFI rating.

ment Scale (DAS), a self-report inventory
of marital satisfaction developed by Spanier
in 1976, high EE spouses reported being
less satisfied with their marriages than low
EE spouses. This is reflected by their lower
DAS scores (99.2 vs 112.4, t = -2.52, P <
0.02).
Some variables, on the other hand, did
not discriminate between spouses differing
in EE levels. These variables included so-
cial class, number of children, and also the
BDI scores of both patients and spouses.
The findings for BDI scores are interesting
since they suggest that the intensity of the
depression reported by the patient is not it-
self directly related to the number of criti-
cisms made by the spouse. Moreover, the
level of the depressed mood reported by
spouses also did not differ according to EE
level. Thus, high EE spouses do not appear
to be critical simply because they have more
severely depressed partners or because they
themselves feel more depressed than low
EE spouses, nor does having a highly criti-
cal spouse seem to influence the level of de-
pressed mood reported by the patient. The
correlation between the BDI score of the
patient and the number of critical remarks
made by the relative was 0.14 (NS). The
correlation between the BDI score of the
spouse and the number of critical comments
he or she makes is even lower (r = -0.05,
NS). The amount of criticism made by the
spouse thus seems to be independent of
both the BDI score of that spouse and ofthe
BDI score of their patient-partner.

Marital Satisfaction Levels

In contrast to the results for BDI scores de-
scribed above, the quality of the marital re-
lationship does seem to be an important
variable and one which is associated with
the frequency of critical remarks made. It is
also a variable related to the frequency of
critical remarks received. Just as high EE
spouses reported higher levels of marital
distress than low EE spouses, patients liv-
ing with high EE partners also reported
90 J. M. Hooley and K. Hahlweg
being more unhappily married than pa-
tients living with low EE partners (DAS
scores: 97.4 vs 111.6, t = 2.35, P < 0.05).
The number of critical comments made (in
the cases of spouses) or received (in the case
of patients) was significantly associated
with the reported DAS score - high levels
of criticism being related to low levels of
marital satisfaction. In patients, there was a
correlation of -0.47 between criticism re-
ceived from the spouse and DAS score
(n = 34, P = 0.005) while in spouses the
correlation between DAS score and criti-
cism given was -0.32 (n = 34, P < 0.05,
onetailed).
Although several differences exist be-
tween high and low EE spouses, the fact
that there is a significant age difference be-
tween two groups presents a problem. Do
the differences described truly reflect genu-
ine differences between high and low EE
spouses or are such differences simply arti-
facts which have arisen as a consequence of
the age difference between the groups? To
answer this question, it is necessary to per-
form the same statistical analyses on high
and low EE groups which do not differ in
age. To do this, the youngest ofthe high EE
spouses and the oldest of the low EE
spouses were systematically removed from
the groups in a pairwise fashion until a t test
revealed that no significant age difference
between the two EE groups remained. Da-
ta from these age-equivalent groups were
then reanalyzed and the differences be-
Interaction Patterns of High and Low EE Dyads
In addition to the marital and demographic
data described above, data concerning the
interaction patterns of 30 couples were also
obtained by means of videotaped observa-
tions. These took place while the couple
was left alone to discuss issues on which
they held differing opinions. All interac-
tions lasted 15 min. The behavior of the
couple was subsequently coded into posi-
tive, negative, or neutral categories. Inter-
actions of high and low EE dyads were then
compared. Examples of verbal behavior
rated as positive are agreements, self-dis-
closures, and accepting statements, while
negative behavior is comprised primarily of
tween high and low EE spouses investigated
a second time. The repeated analyses re-
vealed that the only variable which did not
discriminate as before was duration of mar-
riage - a finding which is hardly suprising
given the close association between this and
age. All other variables, however, still
showed differences between the high and
the low EE spouses. Thus, when the age dif-
ference between the groups is reduced to
nonsignificant levels by the systematic
removal of selected individuals from those
groups, high EE spouses were still rated as
being less warm and more hostile and still
reported higher levels of marital distress
than low EE spouses.
Such conclusions are further supported
by the fact that the number of critical com-
ments made by the spouse was not signifi-
cantly correlated with age (r = -0.24, n =
43, P> 0.1) and by comparisons made us-
ing data from the 30 spouses for whom vid-
eotape data were available (see Hooley,
1984). Within this subsample, the high and
low EE spouses did not differ significantly
in age, nor did they differ in the number of
years they had been married. However, as
found in the larger sample, differences did
exist between high and low EE spouses for
warmth, hostility, and marital satisfaction.
In view of this, it seems reasonable to sug-
gest that such variables do reflect genuine
differences between high and low EE
spouses and cannot be explained simply by
reference to age.
criticisms, disagreements, and justifica-
tions.
To facilitate comparison between the
high and low EE dyads, cumulative records
of each of the 30 interactions were pro-
duced using a pen plotter. Figure 1 provides
an example of the type of plot produced by
this method. In this plot, which shows data
from a low EE couple, the patient's behav-
ior is represented by the solid black line,
while data from the spouse are shown by the
broken line.
A positive behavior causes the pen plot-
ter to move upward while a negative behav-
ior results in the pen moving down. Any
 The Marriages and Interaction Patterns 91
LEE MAR2
10

60

50

40

30

20

10

-10 Patient

- 20 Spouse

- 30 Change
0 10 20 30 40 50 60 10 80 90 100 110 120 130 140
Fig. 1. Low EE dyad: Positive interaction pattern.

HEE ST01
10

60

50

40

30

20

10

-10 Patient

- 20 Spouse

- 30 Change
0 10 20 30 40 50 60 10 80 90 100 110 120 130 140

Fig. 2. High EE dyad: Negative interaction pattern.

92 J. M. Hooley and K. Hahlweg
HEE : JOB2
10

60

50

40

30

20

10

-10 ---Patient

- 20 ---Spouse

- 30 Change
0 10 20 30 40 50 60 10 80 90 100 110 120 130 140
Fig. 3. High EE dayd: Asymmetric interaction pattern.

HEE GRA2
10

60

50

40

30

20

10

-10 Patient

- 20 Spouse

30 Change
- ~---.--.---.---.---.---.---r---r-~~~---.---.--~~~
o 10 20 30 40 50 60 10 80 90 100 110 120 130 140
Fig. 4. High EE dyad: Neutral interaction pattern.

Table 3. Relationship Between EE and Interaction Pattern
EE Positive
Low 9(30.0%)
High 5(16.7%)
Total: 14
x = 6.5 (P < 0.01)
Percentage agreement = 73.3%
neutral behavior is denoted by a line paral-
lel with the central base line (broken),
which is the neutral position, falling as it
does between the positive and negative
sides of the ordinate. The total number of
utterances is shown on the abscissa, while
the ordinate represents the extent of the de-
viation from the neutral position. Thus, the
couple in this figure shows a strongly posi-
tive interaction pattern. The plot indicates
that both patient and spouse begin positive-
ly and generally remain positive throughout
the interaction. By contrast, the couple
plotted in Figure 2 shows a very different in-
teraction style. In this dyad, positive behav-
ior is much less consistent and, as time pro-
gresses, the pattern of the interaction grad-
ually becomes more negative.
The latter type of interaction pattern is
typical of high EE dyads in this study. Whe-
reas low EE dyads generally show a clear-
cut positive pattern, high EE couples have
more varied, but nevertheless nonpositive,
styles of interaction. Thus, the couple
whose data are shown in Figure 3 have an
asymmetric pattern, while those shown in
Figure 4 are generally neutral in their
behavior.
The interaction patterns of high and low
EE dyads are so different that it is actually
possible to identify couples reliably as high
or low in EE on the basis of their interaction
styles. When couples were grouped visually
by the second author (who was blind to the
The Marriages and Interaction Patterns 93
Interaction pattern
Neutral, negative,
or asymmetric Total
3(10.0%) 12
13(43.3%) 18
16 30
EE levels of the spouses), by focusing on
their overall style, with couples who showed
a clear and sustained positive pattern being
separated from those who exhibited a neu-
tral, negative, or asymmetric pattern, 73.3 %
of dyads were found to be correctly identi-
fied as high or low in EE. As can be seen
from Table 3, the pattern most typical of
low EE dyads is one of continuous positive
exchange. This is shown in 9 of the 12 low
EE couples. By contrast, such a positive
pattern is only found in 5 of the 18 high EE
dyads (Chi-square = 6.5, P < 0.01). Within
such dyads, interaction patterns tend to be
negative, neutral, or asymmetric, with hus-
band and wife showing independent behav-
ior profiles.
t Tests also reveal the positive interaction
pattern to be associated with lower levels of
criticism. While spouses whose dyadic in-
teraction pattern is nonpositive (i.e., nega-
tive, neutral, or asymmetric) make a mean
number of 10.8 critical comments about
their partners during an interview, those
whose dyadic interaction plots reveal a po-
sitive pattern make only 5.7 critical remarks
(t = 2.6, P < 0.02). The cumulative inter-
action plots also validate the Dyadic Adjust-
ment Scale, with positive patterns being-
more indicative of marital satisfaction and
nonpositive patterns being more typical of
marital distress (DAS scores: 113.5 vs 96.9,
t = 3.2, P < 0.005).
94 J. M. Hooley and K. Hahlweg
Discussion
The data presented in this chapter suggest
that EE is a measure of important aspects of
the interpersonal relationships of depressed
patients and their spouses. While it is not
yet clear how relevant these findings are
for schizophrenic patients, both partners
within high EE depressed relationships re-
port being less happy with their marriages
than individuals involved in low EE rela-
tionships. Moreover, for both patients and
spouses, high EE levels are correlated with
low levels of marital satisfaction, although
for patients the magnitude of the correla-
tion is much larger. Those patients who live
with the most critical spouses are the most
unhappy with their marriages.
Behavioral observations of the interac-
tions which take place within high and low
EE dyads also confirm these self-report da-
ta. Whereas the behavioral-interaction pro-
files of low EE dyads typically reflect a con-
tinuous exchange of positive verbal com-
munication, those generated by high EE
couples are much more variable. They are
also much less likely to be positive. Inter-
action patterns of high EE dyads typically
involve neutral, negative, or asymmetric
exchanges.
Rather importantly, cumulative plots
which record the positive, negative, and
neutral aspects of interactions appear to be
a reliable way of differentiating high from
low EE dyads. Such a discrimination can be
easily made visually and, suprisingly, is best
made by focusing on the presence or ab-
sence of positive behavior patterns. This is a
particularly interesting finding since, on the
basis of previous studies of EE, it might
have been predicted that a clear-cut nega-
tive interaction style would have been the
most likely differentiating high and low EE
dyads. Thus, while the data concerning
marital satisfaction levels validate the CFI
and support the view that high EE relation-
ships are generally more troubled than low
EE relationships, the interaction plots sug-
gest that one manifestation of these inter-
personal problems involves the lack of po-
sitive communicatory behavior.
Clearly, however, the interaction pat-
terns described represent a relatively crude
way of depicting the nature of high and low
EE interactions and, focusing as they do on
global measures of positive, negative, and
neutral behavior, fail to capture the real es-
sence of such relationships. In subsequent
publications, we will present more detailed
data concerning the specific behaviors
which both differentiate high and low EE
individuals and discriminate between de-
pressed patients interacting with them. The
very important issue of the sequencing of
these behaviors will also be addressed with
reference to positive and negative escala-
tion patterns.
Returning to the data presented here,
however, the crucial question is why high
EE dyads interact more negatively (and less
positively) than low EE dyads. Any inter-
ventions made within high EE families
must necessarily be predicated on assump-
tions about the answer to this question.
While a high EE attitude may be reflective
of underlying personality variables in some
individuals, in others it may arise as a con-
sequence of continued exposure to particu-
lar types of behavior in patients. In still
other instances, high EE may be the pro-
duct of the interaction of both these factors,
with certain types of symptoms (or constel-
lations of symptoms) eliciting a high EE re-
sponse in certain types of individuals.
While the results of this investigation
highlight the absence of mutual exchanges
of positive communication within high EE
couples, they also demonstrate the diversity
of interactional styles within this group as a
whole. Most low EE dyads show quite con-
sistent patterns of positive behavior over
the course of a face-to-face interaction. In
contrast, the interaction styles of high EE
dyads reveal much more variability and are
sometimes negative, sometimes neutral,
and occasionally asymmetric or positive.
Whether this diversity is a reflection of un-
derlaying differences in the etiology of the
high EE response is not yet clear. If this is
the case, however, interaction plots such as
these may provide valuable information to
researchers and clinicians concerning the
level and type of intervention most appro-
priate for any particular dyad. Although we
are still some distance from being able to
tailor our interventions to individual fami-

lies, detailed data concerning families who
fail to respond to conventional forms of EE
modification as well as on those who do de-
rive therapeutic benefit is clearly a neces-
sary first step. EE is a subtle and complex
phenomenon most likely to be determined
in a variety of ways. While the data pre-
References
Beck, A. T., Ward, C. H., Mendelsohn, M.,
Mock, J., and Erbaugh, J. (1961). An invento-
ry for measuring depression. Archives of Gen-
eral Psychiatry, 4, 561-571.
Berkowitz, R., Kuipers, L., Eberlein-Vries, R.,
and Leff, J. (1981). Lowering expressed emo-
tion in relatives of schizophrenics. In M. J.
Goldstein (Ed.), New developments in inter-
ventions with families of schizophrenics. Lon-
don: Jossey-Bass.
Brown, G. W., Birley, J. L. T., and Wing, J. K.
(1972). Influence of family life on the course of
schizophrenic disorders: A replication. British
Journal of Psychiatry, 121,241-258.
Hahlweg, K., Reisner, L., Kohli, G., Vollmer,
M., Schindler, L., and Revenstorf, D. (1984).
Development and validity of a new system to
analyse interpersonal communication (KPI:
Kategoriensystem fUr partnerschaftliche Inter-
aktion). In K. Hahlweg and N. S. Jacobson
(Eds.), Marital interaction: Analysis and mo-
dification. New York: Guilford.
Hooley, J. M. (1984). Criticism and depression.
Unpublished D. Phil. thesis, University of Ox-
ford.
Kuipers, L., Sturgeon, D., Berkowitz, R., and
Leff, J. (1983). Characteristics of expressed
emotion: Its relationship to speech and looking
in schizophrenic patients and their relatives.
British Journal of Clinical Psychology, 22 (4),
257-264.
Miklowitz, D. J., Goldstein, M. J., Falloon, I. R.
The Marriages and Interaction Patterns 95
sen ted in this chapter cannot provide us
with any simple answers about the optimum
method of intervention for high EE fami-
lies, they do provide us with a basis for ask-
ing more refined and informed questions
about the directions subsequent interven-
tion research might take.
H., and Doane, J. A. (1984). Interactional cor-
relates of expressed emotion in the families of
schizophrenics. British Journal of Psychiatry,
144,482-487.
Spanier, G. B. (1976). Measuring dyadic adjust-
ment: New scales for assessing the quality of
marriage and similar dyads. Journal of Mar-
riage and the Family 38,15-28.
Spitzer, R. L., Endicott, J., and Robbins, E.
(1978). Research Diagnostic Criteria for a se-
lected group of functional disorders. New
York: Biometric Research Unit, New York
State Psychiatric Institute.
Strodtbeck, F. L. (1951). Husband wife interac-
tion over revealed differences. American
Sociological Review, 16,468-473.
Valone, K., Norton, J. P., Goldstein, M. J., and
Doane, J. A. (1983). Parental expressed emo-
tion and affective style in an adolescent sample
at risk for schizophrenia spectrum disorders.
Journal of Abnormal Psychology, 92 (4),
399-407.
Vaughn, C. E., and Leff, J. P. (1976). The in-
fluence of family and social factors on the
course of psychiatric illness. British Journal of
Psychiatry, 129, 125 -137.
Wing, J. K., Cooper, J. E., and Sartorius, N.
(1974). The description of psychiatric symp-
toms: An introduction manual for the PSE and
cat ego system. London: Cambridge University
Press.
8. Patterns of Emotional Response
in the Families of Schizophrenic Patients
C. E. Vaughn
Introduction
In Chap. 2, Hooley succinctly summarizes
the history of expressed emotion (EE) re-
search and its principal findings to date. She
reviews the considerable body of evidence
which suggests that the EE index has pre-
dictive validity where relapse patterns in
schizophrenia are concerned and refers to
the content and concurrent validity studies
which offer further proof that EE is concep-
tually sound. However, she also acknowl-
edge the existence of unresolved questions
and issues raised by the EE investigations.
The intention of this chapter is to take up
some of these questions and issues and ex-
plore them in the light of unpublished data
from relatives' interviews - rich sources of
information concerning patterns of emo-
tional response and family interaction in the
months preceding a psychiatric admission.
The material to be presented suggests some
directions for further research and tentative
guidelines for clinical interventions in the
future.
A recurrent criticism of the EE research
is that it has failed to indicate precisely what
the concept measures, how it manifests it-
self, or how it operates to influence the vul-
nerable individual. While it appears to tap
qualities of family life which can either pro-
tect patients against relapse or make relapse
more likely by the exacerbations of symp-
toms, the process by which this happens has
remained unclear.
Hints as to what might be going on ap-
peared in the early epidemiological studies
of schizophrenia conducted by Brown and
his group. This work began as a search for
environmental explanations of why some
patients relapsed after discharge (despite
the protection afforded by antipsychotic
medication) and others did not. The origins
of schizophrenia were assumed to be bio-
logical; there was considerable scepticism
regarding some theorists' views of the fami-
ly's role in the etiology of the illness. Never-
theless, a series of anomalous findings early
on in the research suggested that influences
from within the home might affect the
course of a schizophrenic illness. One such
anomaly was the link between relapse and
type of living group; another was the fact
that male patients apparently did well even
if unemployed as long as the mother went
out to work (Brown, Carstairs, and Top-
ping, 1958; Brown, 1959). Both anomalies
could be explained, the investigators rea-
soned, if patients were reacting to close ties
because of a sensitivity related to the dis-
ease process. Theories of sensory overload
and emotional overarousal in schizophrenia
perhaps were relevant. It therefore seemed
appropriate to try to identify aspects of fam-
ily life which might contribute to a better or
worse psychiatric outcome. There was no
suggestion that the qualities or characteris-
tics to be identified were necessarily deviant
or that they were unique to the relatives of
schizophrenic patients. This important
point is emphasized by Brown (1985) in a
recent account of the path to the "discov-
ery" of EE. He recalls:
It seemed important that the occasional presence
of deeply disturbed or unusual relationships be-
tween parent and patient should not be allowed
to dominate our thinking. If I had any hunch
about what was going on, it was that it involved
something a good deal less fundamental, indeed
everyday. Therefore one way forward would be
to develop an instrument capable of recording
the range of feelings and emotions to be found in
98 C. E. Vaughn
ordinary families. Indeed, the family instrument
used to record "expressed emotion" was not de-
veloped with the families of schizophrenic pa-
tients and it did not occur to me that there was
anything amiss in this.
Brown and his colleagues' initial attempts
to identify features in the home environ-
ment predictive of relapse produced hun-
dreds of rating scales concerning different
aspects of family life. Of all these scales,
only three showed independent relation-
ships with the course of schizophrenia:
relative's critical comments, hostility, and
emotional overinvolvement. These three
scales together comprise the index of EE,
which essentially is an indicator of a rela-
tive's negative affect or intrusive overcon-
cern toward the patient. Two additional EE
scales, warmth and positive remarks, are
measures of positive affect which undoubt-
edly contribute to the emotional atmos-
phere in the home. They were not included
in the EE index, however, because oftheir
complex interrelationships with the other
EE scales and with relapse. For example,
high warmth toward the patient is associa-
ted with a good outcome only when the rel-
atives score low on emotional overinvolve-
ment (EOI). Where high warmth is found
in combination with high EOI, relapse is
very likely.
The EE index is rated on the basis of both
feelings expressed during the inverview and
reported behavior outside it. The Camber-
well Family Interview (CFI) was designed
to elicit two kinds of information: objective
information concerning life in the house-
hold in the months leading up to the pa-
tient's admission and subjective information
concerning the relative's attitudes and feel-
ings toward the patient and the illness.
Brown believed that if the relative were
asked in a neutral fashion about quite spe-
cific details of the illness and related events,
information about feelings would naturally
follow. A number of different interviewing
techniques help to ensure that this happens.
Questions about the occurrence of patient
behaviors cover many different aspects of a
particular behavior: onset, severity, fre-
quency, social contact, reactions of every-
one who witnessed the behavior. Relatives
are asked to give examples of the last time a
behavior occurred and how they coped at
the time. During the course of the interview
the focus shifts from symptom behaviors to
more general patterns of family interaction,
such as the occurrence of irritability and
quarreling and the amount and intensity of
contact between the patient and other fam-
ily members. There are specific questions
too about the quality of the relationship be-
tween patient and relative: Can you get
close to him? Have your feelings for him
changed since the trouble came on? Addi-
tional questions concern the impact of the
illness: What difference has this hospitaliza-
tion made to you and the family? What has
been the most disturbing aspect of the trou-
ble for you? Although the interviewing pro-
cedure is governed by rigorous guidelines,
the approach is a flexible one and the tone
that of an informal conversation rather than
an interrogation. The result almost always
is a rich picture of family life during the pre-
admission period. The quality of informa-
tion obtained in most instances provides
quite a good idea of relatives' characteristic
responses in a time of crisis and their re-
ported impact on the patient, as well as
more general patterns of family interaction.
While EE researchers generally have ac-
cepted relatives' accounts as valid, until re-
cently the assumption of veracity was based
primarily on the usefulness of the EE index
as a predictor of symptomatic relapse. An
element of doubt lingered: in the absence of
more direct evidence, one could not be sure
that a relative's interview behavior or re-
ported reactions necessarily reflected what
went on in the patient's presence. How-
ever, recent independent investigations of
direct interactions between patient and rel-
ative have confirmed the impressions glean-
ed from relatives' interviews, suggesting
that in a majority of cases there is a corres-
pondence between these different sources
of information (Miklowitz, Goldstein, Fal-
loon, and Doane, 1984). Reassured by this
knowledge, let us now consider some of the
ways in which low EE and high EE house-
holds differ. The discussion of qualitative
data which follows is based on interviews
with several hundred relatives of patients in
England, America, and Australia.

EE as an Indicator of the Pre-illness Relationship
A detailed analysis of the interview data for
78 relatives in the 1976 Vaughn and Leff
study established a link between the con-
tent of relatives' criticism and the quality
of the pre-illness relationship, for both schi-
zophrenic and depressed samples of pa-
tients (Leff and Vaughn, 1985). For this
analysis, a distinction was made between
critical remarks concerning symptom
behaviors which first appeared in the con-
text of an illness episode (for example, de-
lusional behaviors, changes in levels of ac-
tivity or irritability) and critical remarks
about more general, enduring personality
traits of the patient. Results revealed that
for both diagnostic groups a major deter-
minant of a relative's current response to
the patient and the illness was the way in
which patient and relative got along before
the present illness episode.
If the pre-illness relationship was good,
the relative at interview tended to either
make no critical remarks or to confine criti-
cism to comments about florid symptom
behaviors. A critical response seemed to
depend less on the degree of the patient's
disturbance than on the relative's own per-
sonality. Relatives who described them-
selves as generally tolerant and easygoing
persons tended to be noncritical. Those
who admitted to being tense or moody
Distinguishing Characteristics of High EE and Low EE Relatives
Measures of severity of psychopathology at
admission consistently fail to distinguish be-
tween patients from low EE and high EE
families: high EE patients are not more ill
than their low EE counterparts, by any cri-
terion adopted (Brown, Birley, and Wing,
1972; Vaughn and Leff, 1976; Vaughn, Sny-
der, Jones, Freeman, and Falloon, 1984).
Miklowitz, Goldstein, and Falloon, (1983)
demonstrated that low EE and high EE
schizophrenic patients were indistinguisha-
ble on measures of premorbid adjustment
and residual symptomatology after dis-
charge. The differences lie not in the pa-
tients, but in the relatives - low EE and
high EE relatives differ markedly in their
response to a patient's objective condition.
Patterns of Emotional Response 99
when stressed, however, frequently found
the strain of coping with someone who was
psychiatrically ill intolerable. Feelings of
frustration and helplessness commonly re-
sulted in expressions of anger toward the
patient. However, these relatives' criti-
cisms invariably were directed at specific
symptom behaviors for which the patient
was not held responsible.
Where there appeared to be a poor rela-
tionship between patient and relative prior
to the onset of the illness episode, critical
remarks were almost exclusively about
long-standing behaviors. For both the schiz-
ophrenic and depressed groups, approxi-
mately 70% of all critical remarks fell into
this category. Criticism tended to center on
aspects of the relationship with which the
relative was markedly dissatisfied, partic-
ularly communication and the amount of af-
fection and interest shown by the patient.
The relatives concerned did not identify
these behaviors as manifestations of illness,
but considered them to be integral charac-
teristics of the patient's personality.
Thus, it seems that the kind of remarks
made by a relative at a crisis point (time of
admission) can provide important clues as
to how patient and relative get along in nor-
mal circumstances.
Throughout the several decades of EE
research, investigators have been impres-
sed by the ways in which differences in lev-
els of EE tend to reflect differences in
thresholds of tolerance, attitudes toward
the legitimacy of the illness, and styles of
coping. It has been possible to identify a
number of characteristics which tend to
distinguish relatives who score highly on
criticism or emotional overinvolvement
from those who do not, as judged by their
behavior in the interview and reported re-
actions.
The first characteristic concerns the rela-
tive's respect for the patient's relationship
needs, which may vary from patient to pa-
tient, and by diagnosis. For example, many
100 C. E. Vaughn
schizophrenic patients find mhmate rela-
tionships difficult and prefer to maintain a
certain social distance even when well. By
contrast, depressed patients may be socially
demanding and seek repeated expression of
caring and concern from persons around
them. Low EE relatives "attend" to the pa-
tient more; they are sensitive to requests
for either privacy and autonomy (schizo-
phrenia) or increased social support (de-
pression).
Additional characteristics which tend to
distinguish low EE and high EE relatives
are the relative's attitude toward the legiti-
macy of the illness and the relative's level of
expectations for patient functioning. These
two characteristics are linked in predictable
ways. The low EE view that the illness is
genuine tends to result in lowered expecta-
tions: the patient is unwell; therefore, one
must make allowances. Low EE relatives
consistently show greater empathy and
make more efforts to understand the pa-
tient's ordeal than do high EE relatives.
While they are objective enough to recog-
nize that certain symptom behaviors are
signs of illness, they can accept that certain
peculiar ideas or experiences might be real
enough for the patient, and this realization
also contributes to an attitude of tolerance.
These attitudinal differences, it should be
noted, tend to evolve in the absence of any
professional counseling.
Finally, low EE and high EE relatives dif-
fer in their emotional reactions to the ill-
ness. This may seem a tautological state-
ment, but it is one which warrants further
examination. If one compares the respec-
tive accounts of parents who are discrepant
for EE - one highly critical, one not - the
same behavior will be described in very dif-
ferent ways. The less critical relative will
tend to give a dispassionate, objective de-
scription of a particular behavior by the pa-
tient, even when that behavior is undesira-
ble. For example, when asked about his
son's personal hygiene (Does he look after
himself all right?), one father simply an-
swered: "Not good. He doesn't want to
shave, wash, that sort of thing." The lan-
guage used by the patient's mother in re-
sponse to the same question was much less
neutral: "That's a very sore point with me
... He goes beyond being a bit grubby. He
doesn't shave, doesn't wash his hair, never
changes his underwear - it's quite revolt-
ing!" This reply goes beyond mere descrip-
tion; the strength ofthe negative reaction is
considerable.
In another household, both parents de-
clared the son's playing of loud rock music
to be a source of family tension. As with the
preceding example, however, parental
reactions differed in content and in tone.
The father reported low-keyed attempts to
intervene: "If he played loud music, I might
just say 'Tum the music down, Kev, the
whole street doesn't need to hear. '" The
mother's response was one of anger and
acute distress: "Music, when he's on the
downgrade, would nearly drive you crazy.
He'd play the same records over and over
and over again, until you felt if I hear that
one more time I'll scream!"
Low EE relatives frequently do express
dissatisfaction with a patient's behavior,
but the dissatisfaction is muted; it tends to
take the form of statements such as "I wish
he weren't like that." Their opinions lack
the emotional intensity of remarks made by
high EE relatives. The use of words like
"revolting" and "drive you crazy" reflect
strong disapproval. Indeed, the high EE
relative often responds as if the behavior in
question were a personal affront. Given
such strong reactions, confrontations with
the patient are very likely.
Clinical intervention programs suggest
that many high EE relatives have low prob-
lem-solving skills, a conclusion supported
by relatives' own accounts. High EE rela-
tives will freely acknowledge their inability
to react calmly and nonintrusively. Even
when attempts to change the patient's un-
desirable behaviors repeatedly fail, the rel-
ative will persevere with the same ap-
proaches. Predictably, a failure to adapt to
the changing circumstances created by the
illness tends to exacerbate the situation and
to increase levels of tension, frustration,
and irritability generally.
One high EE father vividly described his
inability to cope with his daughter's "mad"
talk when ill. His response was one of ex-
treme intolerance. In this particular case,
the relative recognized that the illness was
genuine and acknowledged the need to
show more understanding, yet no amount
 Patterns of Emotional Response 101
of insight facilitated the process of adjust- kill him. The father managed to disarm his
ment: son of a knife through gentle persuasion:
I had no technique for dealing with her. I was at "Have a cup of tea, son, and let's talk about
a loss. My wife is much better talking to her when this." He did not contact the hospital until
she's like this, but I find it very difficult ... When the following morning. Admission did
she gets a bit more and more insane, I could never prove to be necessary, but a major social
- having been brought up as an engineer, used to
crisis was averted. When asked how he
examining things according to the rules of science
- I could never, ever think it worthwhile trying to managed to remain so composed, the father
talk to my daughter. It was such absolute non- replied: "It didn' worry me, I thought if he
sense! I couldn't stand it, couldn't adapt, couldn't goes for me I'll put him in his place. I wasn't
develop a method for dealing with her. But my too upset, as the wife had a breakdown
wife did. once. I do take things as they come, I sup-
Low EE relatives are more adaptable. Over pose."
time and usually by a process of trial and er- Contrary to some interpretations of the
ror - if it works, keep it up; if it doesn't, try EE data, low EE relatives do attempt to
something else - they gradually develop exercise social control over the patient, but
techniques for dealing with difficult behav- in less intrusive and confrontational ways
iors which minimize their impact. The pro- than their high EE counterparts. They also
cess by which repertoires of successful cop- exert considerable self-control. While some
ing techniques are built up is worthy of low EE relatives describe themselves as
more systematic investigation. What exists naturally easygoing, as in the above exam-
at the moment is a wealth of anecdotal ma- ple, a calm response in a crisis does not ne-
terial pertaining to relatives' efforts to come cessarily indicate a lack of emotion. Feel-
to terms with the disturbing changes ings of distress and dissatisfaction can be
brought by the illness, some of which may considerable. Frequently, there is evidence
be permanent. that low EE relatives have learned to con-
Particularly noteworthy are the excep- trol their emotions over many years, deal-
tionally calm and self-contained reactions ing not only with periodic crises but also
by some low EE relatives to extremely agi- with the everyday problems of living with a
tated or bizarre patient behaviors. Only very person who is vulnerable to episodes of psy-
occasionally does this response style seem chiatric illness. One mother described her
inappropriate, overcontrolled, or evidence trial-and-error efforts to discover what wor-
of some denial process at work. In most ked best:
cases, these relatives do seem concerned, I realized that if I got upset, exploded, it was bad
but they are not unduly anxious. Many dis- for Martin. It had a more calming effect on him if
play an admirable ability to defuse a crisis. I kept very calm. So I let many things pass. I
In such circumstances, they are likely to wouldn't make an issue. I became more permis-
have a calming effect not only on the patient sive in order to survive. You just do the best you
(an impression supported by the biological can each day. My husband and I are pretty happy-
data reviewed by Hooley) but also on other go-lucks. We don't let many things bother us, we
family members. At interview they offer try to keep a sense of humor. That's essential! In
cool, even humorous accounts of their ef- any event, we're not the same people we were
forts to deal with the patient. One low EE when we started out.
father recounted a preadmission incident in It is this quality ofJlexibility which, perhaps
which his son woke him up one night with more than any other characteristic, distin-
an announcement of a God-sent mission to guishes low EE from high EE relatives.

Impact of Different Response Styles on the Schizophrenic Patient

Relatives' accounts indicate that schizo- seldom avoided relatives who were low-
phrenic patients react to their widely vary- keyed and nonintrusive. They tended to
ing behaviors in very different ways. In the stay in the same room even if disinclined to
British and California EE studies, patients converse; they showed no inclination to re-
102 C. E. Vaughn
treat. In high EE households, however, then it's over with. But I think that's what my son
social withdrawal was commonplace. Often doesn't like. He doesn't like to hear the scream-
there seemed to be a clear link between a ing and the hollering. He doesn't like being con-
stressful stimulus and protective with- fronted, he doesn't like people to become upset
with him. It does not good at all to nag, it just
drawal by the patient. For example, one makes him more irritable.
young man divided his time between the
households of his divorced parents. In the These very similar comments, made by dif-
months after going to live in his mother's ferent relatives, were echoed again and
conflict-ridden high EE household, he again in interviews.
spent more and more time by himself, with- Thus, relatives' reported observation
drawing to his room or going for long solita- support psychophysiological findings (cf.
ry walks. During family rows, he ceased to Chap. 2) which suggest that high EE rela-
speak at all; he would cower in a corner of tives are more emotionally arousing to the
the room and put his hands over his ears as patient. Why should this be so? Research
if to shut out the noise. His mother com- into the precise mechanisms by which EE
plained that her attempts to "reach" him operates to influence the return of schizo-
seemed to make him more ill. Indeed, the phrenic symptoms is just beginning, but the
patient relapsed 3 times during a 12-month interview data provide some clues. Let us
period. It is notable that he showed no signs assume that persons prone to episodes of
of withdrawal, nor did he relapse, during schizophrenia have an underlying psycho-
the periods when he lived with his father in biological vulnerability which makes it dif-
less tense circumstances. ficult for them to process complex stimuli
A second kind of reaction by the patient and handle social relationships. Let us ac-
to overstimulation or stress was agitated ag- cept, also, that issues of control and influ-
gression. There were repeated instances in ence frequently preoccupy such persons. It
which a critical or intrusive action by the follows that certain environments may pro-
relative would provoke an irritable or vio- duce high and sustained levels of arousal in
lent outburst from the patient: schizophrenic patients which make them
Would he ever get excited or agitated? more likely to become ill again. By the rela-
That would be if someone was too much pressing tives' own accounts, these enVIronments
him. are characterized by discord and/or intru-
He likes the quiet, to be talked to softly; he sive attempts on the part of the relative to
doesn't like it when you try to press him. It makes advise, complain, or merely get through to
him very nervous.
Would he ever get irritable? Only if you aggra-
the patient. Whether or not these overtures
vated him. That would make him very angry, are well-intentioned, they are likely to be
very upset. He wouldn't say anything, but he'd viewed as threatening by persons vulnera-
get very aggressive ... He would kick chairs or ble to episodes of schizophrenia and may
throw something, to take his frustration out. precipitate a return or exacerbation of flor-
I scream and holler and get it out of my system, id symptoms.

Premorbid Personality, Relative's EE, and Relapse:

The Issue of Vulnerability
The content analysis of critical comments
made in the 1976 study revealed an associa-
tion between the patient's previous person-
ality and the relative's emotional response,
suggesting that in at least a proportion of
cases the relative was responding to ab-
normality in the patient. In such cases, the
relative's reaction may be entirely under-
standable. how can one be sure that the link
between a relative's EE and symptomatic
relapse is not due to another intervening
variable thought to be associated with a
poor prognosis, for example, a poor prem-
orbid personality?
Results of the EE studies to date clarify
the ways in which different variables may
interact to influence the outcome of any
given patient. The schizophrenia literature
suggests that clinical variables and aspects
of a patient's psychiatric history are not lin-

ked in any obvious way to the course of an
individual patient's illness. The EE findings
suggest why this is so for patients living with
relatives. Patients showing few of the clini-
cal signs associated with a poor prognosis
are likely to remain well whether or not
they return to a high EE or low EE home.
This finding is consistent with results of
other studies in which certain criteria such
as acuteness of onset or marital status are
associated with a good prognosis or remis-
sion in schizophrenia. Where the clinical in-
dicators are unfavorable, however, the
quality of the postdischarge environment is
highly important for the patient's clinical
outcome. Thus, an emotionally more neu-
tral setting, such as a low EE home, exerts
the greatest protective influence on those
patients who are constitutionally the most
vulnerable. In three different studies of
schizophrenia, patients with poor prem-
orbid personalities who returned to low EE
relatives almost always stayed well over 9-
Determinants of EE
At a given point in time, determinants of
the relative's degree of EE are likely to in-
clude the patient's behavior (degree of
disturbance and extent of secondary handi-
cap); the relative's perceived ability to cope
and related problem-solving skills; idiosyn-
cratic personality factors; the amount and
quality of information available concerning
the illness; external stresses such as unem-
EE: State or Trait?
An assumption throughout the EE investi-
gations is that a high degree of EE on one
occasion "is a measure of the relative's pro-
pensity to react in that way to that particular
patient, even though other factors may be
needed to precipitate this." (Brown et aI.,
1972, p. 246). Thus, the EE level at the time
of key admission represents a potential to
respond in a characteristic manner, partic-
ularly at a time of crisis for the family. This
tendency does not preclude fluctuations in
EE levels over time, however. These are
most likely to occur in the proportion of
cases in which the relative's emotional re-
Patterns of Emotional Response 103
month follow-up period. Those who re-
turned to high EE relatives usually re-
lapsed. The particular vulnerability of un-
married schizophrenic men again was high-
lighted by the analysis of differential pat-
terns of parental response. Although schi-
zophrenic sons and daughters were equally
likely to be found in high EE homes, the
sons were significantly more likely than the
daughters to return to homes featuring pa-
rental conflict, according to the same three
studies (Brown et aI., 1972; Vaughn and
Leff, 1976; Vaughn et aI., 1984). At the
same time, schizophrenic sons were less
likely than schizophrenic daughters to be
protected by the combined effects of main-
tenance phenothiazine therapy and reduced
contact with relatives, both of which can
modify the effects of an excessively stimul-
ating environment. These findings explain
at least in part their greater chances of de-
veloping a chronic or relapsing condition
after an initial attack of schizophrenia.
ployment; and cross-cultural differences in
societal responses to the illness. The last of
these factors is discussed by Jenkins et aI.,
in Chap. 3. With the exception of the pa-
tient's behavior, none of these influences
on EE has been studied in a systematic fas-
hion, but relatives' accounts suggest that
each is of some importance and warrants
further investigation.
sponse appears to be a direct reaction to the
patient's behavior when ill. (Brown's esti-
mate of 35% is in accord with the figures
which emerged from the content analysis of
criticism in the 1976 Vaughn and Leff
study). In such cases, one would expect a
good patient recovery with full remission of
symptoms to be followed by a lowering of a
relative's EE, particularly the number of
critical comments. This is precisely what
happened in follow-up interviews conduct-
ed by Brown et al. (1972) within a year of a
patient's discharge from the hospital. The
number of criticisms made by relatives had
104 C. E. Vaughn
decreased considerably; the greatest reduc-
tion in number occured with respect to pa-
tients who had markedly improved. Thus,
the Hamburg results regarding the appar-
ent instability of EE (Chap. 5); are not in-
consistent with earlier findings. Other in-
dices of expressed emotion, such as emo-
tional overinvolvement (EOI), showed a
less marked tendency to diminish over
time. These differential findings are con-
sistent with clinical impressions of thera-
pists who have attempted to modify levels
of EE. Apparently, it is easier to lower high
levels of criticism than to alter patterns of
marked EOI, which by their nature tend to
Exceptional Response Styles: "Schizophrenogenic" Highs and
"Burnt-out" Lows
Because the main focus of EE investiga-;
tions has been on schizophrenia, the risk re-
mains that the components of the EE index
will be perceived to be specific to families of
that diagnostic group. It is worth reiterating
that they are not. Criticism, hostility, and
emotional overinvolvement may feature in
any relationship, and they have been shown
to occur across a wide range of conditions
and cultures. Furthermore, while the dis-
tinguishing characteristics of low EE and
high EE relatives may be readily identified,
th!!re is considerable variation of response
within the dichotomized categories. Each
family is different. Relatives obtaining simi-
lar scores on key EE scales do not neces-
sarily resemble each other, nor does any
one relative display all the "markers" of
being high or low EE. Trainees in the use of
the Camberwell Family Interview and the
EE scales frequently express surprise at the
ordinariness of many high EE respondents,
particularly when the patient has a diagno-
sis of schizophrenia. Yet the majority of rel-
atives interviewed do seem to be "normal,"
whatever that word may signify nowadays.
Reactions of extreme distress or anger are
often understandable, given the nature of
the disorder and its sometimes grievous im-
pact on family life. In the words of one rela-
tive, "It's all a very sad business, very sad."
A high EE response by the relative usually
is notable not for its own sake, but for its ap-
parent impact on the vulnerable patient.
be deep-rooted and hence less amenable to
clinical intervention.
In both the London (1972) and Hamburg
follow-up studies, changes from low EE at
admission to high EE at follow-up were un-
common. However, one can imagine cir-
cumstances in which this direction of
change might occur: for example, the de-
velopment of an increasingly pessimistic
view of the long-term prognosis when a pa-
tient fails to recover completely after a
single acute episode. In the absence of ap-
propriate counseling or support, a relative
might easily become more anxious or more
critical, and EE levels rise accordingly.
Nevertheless, a minority of relatives of
schizophrenic patients - perhaps 15%, in
the author's experience - do display ex-
treme or bizarre response styles reminis-
cent of the "schizphrenogenic" stereotypes
described in the past by Fromm-Reich-
mann, Laing, Lidz, and others. These rela-
tives are extraordinary individuals by any
criteria, especially memorable for the inter-
viewer because of the manner in which their
"stories" are told. They tend to be mothers
and almost always score highly on EO!.
They mayor may not be highly critical as
well. It is not uncommon, however, for
these relatives to score highly on all the key
EE scales. The result can be a classic "dou-
ble-bind" style of response in which signs
of hostility toward the patient alternate
with signs of warmth, and critical and posi-
tive remarks are juxtaposed in the same
passage. When there also is evidence of
communication deviance, the difficulties
multiply for the interviewer/rater. One
wonders whether the resultant sense of con-
fusion may not approximate, to some de-
gree, the confusion experienced by the vul-
nerable patient exposed to such conflicting
messages. Almost certainly these high EE
relatives pose problems of a special order
for the clinician hoping to influence the
course of the illness by family intervention.
There also exists a subgroup of low EE
relatives who do not fit the characteristic
low EE profile, i.e., who are not benign,

tolerant, mature individuals living in stable
households. The atypical low EE relative
has emerged fairly recently, in studies of
large inner-city conurbations whose resi-
dents are economically and socially de-
prived. In such families, a low EE response
may reflect apathy and indifference rather
than calm concern. These families tend to
be subjected to multiple stresses; they feel
they have little or no control over the course
Conclusion: Some Implications for Clinical Intervention
In their efforts to cope with a serious psych-
iatric illness, most relatives claim to have
received little or no assistance from mental
health professionals. By their own ac-
counts, the nature of the advice and infor-
mation provided has altered little over the
years - the amount is negligible and the
quality poor, in general. In the absence of
professional guidelines, families tend to
manage on their own, more or less well.
The EE data suggest that mental health
education is a basic imperative for all pa-
tients and their families. For both ethical
and pragmatic reasons, they deserve to
know more about the disorder, medication
and its side effects, and other related issues.
In the 1976 British study, only half of the
relatives interviewed considered the pa-
tient's schizophrenic illness to be a mental
problem. Moreover, in all the schizophre-
nia studies to date, highly critical relatives
were significantly more likely than other
relatives to take an unsympathetic view of
the illness, i.e., to blame the patient for his
symptoms and hold him responsible for not
exercising control over them. It is essential
that the persons most directly concerned -
the patient and those around the patient -
are involved in the treatment process from
the first psychiatric contact. If they are not,
then "administrative" solutions (such as the
routine prescription of medication) to the
problems associated with relapse and read-
mission are likely to fail. Ideally, mental
health education for psychiatric consumers
should become as commonplace as anten-
atal care for pregnant women or counseling
for persons who have diabetes. The analogy
between schizophrenia and diabetes is in
Patterns of Emotional Response 105
of their lives. For such persons, the identi-
fied patient's schizophrenia represents just
another of a host of seemingly insoluble
problems. The positive characteristics usual
in low EE relatives are lacking. To date,
these "burnt-out" relatives are few in num-
ber, but they may feature more prominently
in future studies, particularly of popula-
tions for whom rates of unemployment are
high.
some respects an apt one; both present
management problems which can benefit
from a constructive, problem-solving ap-
proach.
The particular qualities and coping abili-
ties of many low EE relatives deserve to be
noted. The more positive and constructive
elements of the low EE response style re-
present behaviors which can be modeled in
intervention programs with high EE fami-
lies. Yet low EE relatives should not be
neglected in such programs, for at least two
reasons. Firstly, they too need advice, in-
formation, and support. In the rush to iden-
tify "high-risk" households, there is a dan-
ger that the needs of those who appear able
to cope will be overlooked. Secondly, the
EE studies to date have been concerned
with one measure of outcome only: symp-
tomatic relapse. Other measures of out-
come, such as presence or severity of sec-
ondary handicaps, were not attempted.
There is some evidence that when a low EE
relative has no expectations for a schizo-
phrenic patient and exerts no pressure to
perform, there may actually be an increase
in negative symptoms and higher levels of
social impairment. Thus, home environ-
ment which is benign by one criterion may
not be so by another.
The existence of atypical high EE and
low EE response styles further emphasizes
the need to look beyond simplistic labels in
efforts to identify families who might bene-
fit from direct clinical intervention. An EE
index, or any other "risk" indicator, is of
very limited use when it comes to planning
interventions. What are desirable - follow-
ing on from mental health education for all
106 C. E. Vaughn
- are treatment and rehabilitation pro-
grams which are based on comprehensive
References
Brown, G. W. (1959). Experiences of discharged
chronic schizophrenic mental hospital patients
in various types of living group. Millbank Me-
morial Fund Quarterly, 37, 105 - 13l.
Brown, G. W. (1985). The discovery of "ex-
pressed emotion": Induction or deduction? In
J. Leff and C. Vaughn (Eds.), Expressed emo-
tion in families: Its significance for mental ill-
ness. New York: Guilford.
Brown, G. W., Birley, J. L. T., and Wing, J. K.
(1972). Influence of family life on the course of
schizophrenic disorders: A replication. British
Journal of Psychiatry, 121, 241-258.
Brown, G. W., Carstairs, G. M., and Topping,
G. G. (1958). The post-hospital adjustment of
chronic mental patients. Lancet, 2, 685-689.
Leff, J. P., and Vaughn, C. E. (1985). Expressed
emotion in families: Its significance for mental
illness. New York: Guilford.
Miklowitz, D. J., Goldstein, M. J., and Falloon,
assessments of individual families' differing
assets and needs.
I. R. H. (1983). Premorbid and symptomatic
characteristics of schizophrenics from families
with high and low levels of expressed emotion.
Journal of Abnormal Psychology, 92,359- 367.
Miklowitz, D. J., Goldstein, M. J., Falloon, I. R.
H., and Doane, J. A. (1984). Interactional cor-
relates of expressed emotion in the families of
schizophrenics. British Journal of Psychiatry,
144,482-487.
Vaughn, C. E., and Leff, J. P. (1976). The in-
fluence of family and social factors on the
course of psychiatric illness: A comparison of
schizophrenic and depressed neurotic patients.
British Journal of Psychiatry, 129, 125-137.
Vaughn, C. E., Snyder, K. S., Jones, S., Free-
man, W. B., and Falloon, I. R. H. (1984).
Family factors in schizophrenic relapse: Replic-
ation in California of British research on ex-
pressed emotion. Archives of General Psychi-
atry, 41,1169-1177.
B. Modification of the Course of Schizophrenia
by Family Interventions
9. Working with Families of Acute Psychotics:
Problems for Research and Reconsideration
L. C. Wynne
My observations of acute psychotics and
their families in recent years lead me to pro-
pose that educational programs and other
forms of intervention with the families of
these patients need to be given much more
distinctive consideration. Although much
that we have learned about educational
programs with families of chronic schizo-
phrenics is relevant, the programs with
families of acute, first-admission psychotics
should be significantly different.
My current thinking about these pro-
blems has arisen since 1978 when Dr. Tho-
mas Gift and I, and our associates in Ro-
chester, began to participate in the WHO
Study of the Determinants of Outcome of
Servere Mental Disorder (DOS). Patients
selected for this study differed from those
with whom I had worked in earlier years at
NIMH and in Rochester. In the Interna-
tional Pilot Study of Schizophrenia (IPSS)
(WHO, 1973), in which I participated at
NIMH, patients were selected only with the
very broad stipulation that the onset was to
be within 5 years of admission. The catch-
ment area from which the patients were
selected was broad and loosely defined. In
no sense could the study be regarded as one
of epidemiologic incidence, even treated
incidence. The DOS selection criteria were
much more rigorously defined by limiting
the sample to a specific, geographically de-
fined catchment area, with the expectation
that each first-admission patient in that
catchment area would be consecutively in-
cluded if he or she met screening require-
ments for functional psychosis. Further-
more, the first outpatient psychiatric con-
tact for psychotic symptoms had to occur
within the previous 3 months. Although the
primary focus was on schizophrenia, such
acute psychotic patients obviously cannot
be definitively diagnosed when first seen.
Hence, all functional psychoses were pro-
visionally included except clear-cut manics
or psychotic depressives. Patients with a
history of use of street drugs or alcohol
could be included provisionally if there was
evidence that these substances were not the
primary or major cause of the episode. In
Rochester, we fortunately have a psychia-
tric care system in which essentially all acute
psychotics in our catchment area have been
funneled by families, doctors, and com-
munity agents (such as the police) into the
Emergency Department (ED) ofthe Strong
Memorial Hospital. The ED, then, is the
established point of first contact for essen-
tially all of these patients in this catchment
area.
In sharp contrast to my experience in
Washington and to prior studies in Roches-
ter with chronic patients and their families,
these first-admission acutely psychotic pa-
tients and their families in the DOS present-
ed a new array of issues and problems.
First, the patient regularly defied diagnostic
classification, even with DSM-I1I (1980).
As Fox (1981, p. 63) has said in a British
perspective on DSM-III:
The least satisfactory aspect of the DSM-III
treatment of the psychoses is the classification of
those acute and episodic disorders with features
of both schizophrenia and affective disorder. By
assigning most such disorders to schizophrenia or
affective disorder, DSM-III may obscure the un-
certainty surrounding such patients, and diminish
the possibility of further clarifying their corre-
lates.
Indeed, most of these patients do have stri-
king affective features, in addition to the
presence of hallucinations, delusions, or in-
coherence that is used to define psychosis
operationally. Although DSM-III is inade-
quate for classification of acute psychoses,
110 L. C. Wynne
the British approach, building upon Present
State Examination (PSE) data and
CATEGO (Wing, Cooper, and Sartorius,
1974), is, in my view, even more unsatisfac-
tory for these patients. If one faces these
diagnostic dilemmas squarely, educational
programs with the patient's families need to
acknowledge this uncertainty and recognize
that the relatively explicit diagnostic criteria
and treatment and outcome data used in
programs with the families of chronic schi-
zophrenics simply are not available for
acute psychoses.
Second, the very fact of the acuteness of
the episode, and especially the fact that this
is the first such episode, clearly makes for a
very different impact upon the family com-
pared with the experience of families who
have lived through several episodes and re-
hospitalizations over many months or
years.
Third, the way that the families reorgan-
ize themselves internally and in relation to
the support network evolves over time for
the families of the chronic patients in a way
that is structurally quite different than for
families of first-episode psychotic patients.
Thus, the ease and appropriateness offam-
ily changes inevitably differs for families of
recurrent or chronic versus first-admission,
acutely psychotic patients. I shall return to
each of these points later, together with
some preliminary comments on family in-
tervention programs with the families of
acute patients.
Misdiagnoses of patients inevitably pro-
duce misunderstanding of the family's ap-
propriate contricution to the patient's
treatment. Research psychiatrists typically
do not see acute psychotics until after a psy-
chiatric resident or attending psychiatrist
has first interviewed, diagnosed, and treat-
ed the patient. Initially in Rochester, Dr.
Gift and I usually were able to see the acute
psychotics quickly, within 24-48 h after ad-
mission. Even so, we were the second or
third, not the first, to do so, and at that time
we did not determine the treatment. We
were impressed by the frequency with
which important florid symptoms had al-
ready disappeared or had been trans-
formed, either by medication or by the hos-
pital admission process. Most importantly,
the administration of neuroleptic medica-
tion within hours of arrival in the ED left us
uncertain about our diagnostic assessment
and dissatisfied as to whether our standard
criteria were consistently relevant. There-
fore, in 1981, we successfully changed the
admission procedure so that we now are
called directly by a nurse in the ED about
any patient broadly meeting our selection
criteria. We ourselves see the patient in the
emergency setting before any other psychi-
atrist has done so. We then follow up with
control over our own program of clinical
care and treatment within the framework of
the WHO research program and do not
need to "wash out" the often confusing ef-
fects of various medications.
While in the ED, we are able to meet per-
sonally with the family members and to be-
gin immediately the process of support and
education at a time when the family is
highly receptive and grateful for attention.
This process also facilitates the task of in-
teresting the family members a little later in
participating in the Camberwell Family In-
terview (CFI) (Vaughn and Leff, 1976) and
our other family research procedures.
Although I have been using the PSE since
1965, when it was in an early edition, I have
become concerned, especially in. our work
with acute psychotics, by the inadequacy of
this instrument for a broad spectrum of
acute psychoses. First, the PSE and the
CATEGO system associated with it rely
heavily on Schneiderian first-rank symp-
toms. These symptoms formerly were as-
sumed to be quite diagnostic of schizophre-
nia and were incorporated into the DSM- III
with this viewpoint still in mind. However,
the analysis of diagnostic data, by Carpen-
ter and Strauss (1974), Brockington, Ken-
dell, and Leff (1978), and others, have fully
documented the inadequacy of the Schnei-
derian symptoms for diagnosing schizo-
phrenia. This is much more of a problem
with acute psychotics than with chronic pa-
tients because the Schneiderian features are
found in most patients, schizophrenic or
not, when they are floridly psychotic, espe-
cially those with marked affective features.
A second problem with the PSE is the
lack of items relevant to affective psycho-
ses. We had had to use supplementary items
to cover these areas in our Rochester re-
search.
 Working with Families of Acute Psychotics
Third, most acute psychotics have vivid
paranoid delusions that are strongly weight-
ed on the PSE to produce a CATEGO diag-
nosis of paranoid schizophrenia. Since par-
anoid delusions (and hallucinations) are
prominent in mania and even hypomania,
these CATEGO diagnoses are often in er-
ror. Helzer, Brockington, and Kendell
(1981) have found the CATEGO wanting
as a predictor of outcome.
At the same time, not only the CATE-
GO, but also other diagnostic classifica-
tions, excepts in the Scandinavian classifica-
tion system, clearly underestimate the like-
lihood that a diagnosis of reactive psychosis
or psychogenic psychosis will be made. The
category of schizophreniform disorder in
DSM-III requires the characteristic
(Schneiderian) symptoms for schizophre-
nia, but at the same time fails to allow for
the distinctively different symptoms, most-
ly affective, and the differing course and
precipitants that are found in the "reactive
psychoses," "psychogenic psychoses," and
"benign schizophreniform psychoses," as
they have been labeled in Denmark (WeI-
ner and Stromgren, 1958; McCabe, 1975).
By now it is commonplace to recognize
that American psychiatrists of 10-20 years
ago diagnosed schizophrenia loosely and
much more frequently than did the British.
As a turnabout, affective psychoses may
now be diagnosed more frequently in Ame-
rica than they are in Britain. Although
these changes in diagnostic habits may lead
to gains in the long run, confusion about the
acute picture remains, especially when eval-
uated from hospital records. First-admis-
sion records of patients who are now chron-
ic and of family members hospitalized in the
past are of dubious value for establishing a
valid family history. Many American rec-
ords fail to mention the affective status of
manics when delusions or hallucinations
have been reported so that the affective
aspect of the clinical status is often unclear.
Another shift has occurred through the
belated recognition that manic-depressive
illness often begins in adolescence and
young adulthood. In the past, manic-de-
pressive illness was ruled out in many clinics
if the patient was less than 25 years old.
These younger manics, especially males,
often present with a rather different picture
111
than do older manics. Recent, unpublished
observations by L. Crabtree (personal com-
munication, 1982) in Philadelphia (which
are concordant with my own observations)
suggest that many acting-out adolescents
and young adults with impulsive, assaul-
tive, and often disorganized behavior are in
fact mancis. They are subdued by neuro-
leptics but otherwise do poorly with these
drugs, although they respond well to lithi-
um. Some are further benefited, especially
if they have had a developmental disorder,
in focusing their attention through use of
tricyclic antidepressants, which have a
paradoxical effect in this age range similar
to amphetamines in the treatment of hyper-
active children.
An additional problem for diagnosis and
intervention arises because some of the
acute psychoses arise from a baseline of
severe personality disorders. Here a very
important distinction will need to be made
between those who have had a schizo typal
personality disorder, usually with a poor
premorbid adjustment, and those with a
"borderline" personality disorder, which
often appears to be part of the affective
spectrum. The latter patients are more apt
to have transient psychotic episodes than
are the schizotypal and schizoid patients.
A still further complication to the di-
agnostic problem with acute psychotics con-
cerns the effects of street drugs. There do
seem to be marked differences in the sex
distribution of acute psychotics, which may
reflect in many settings the greater use of
substance abuse in males. Recent evidence
suggests that not only marijuana but also
PCP and LSD become bound in the tissues
through chemical mechanisms and are re-
leased to produce recurrent flashbacks that
clinically are often indistinguishable from
functional, acute psychotic episodes. Taka-
hashi (1982) has told me that Japanese data
suggest that psychotic relapses may occur
on this basis, without additional ingestion
of street drugs, for as long as 2 years; he is
quite convinced that this phenomenon may
account for the remitting and relapsing
qualities for some of the allegedly schizo-
phrenic psychoses.
The nature of psychoeducational pro-
grams with families will of course need to
vary considerably with the nature of the ill-
112 L. C. Wynne
ness of the identified patient. Whether or
not all of the points that I have raised above
about diagnosis are confirmed by further
research, it seems clear that a definitive di-
agnostic assessment will remain difficult in
early stages of acute psychoses and that an
educational program oriented to just one
disorder (schizophrenia) may be misleading
for the family.
In years past, experimental psychologists
have given considerable emphasis to the
concept of process schizophrenia versus re-
active schizophrenia. Many data have sup-
ported the idea that process schizophrenics,
with poor premorbid personality, insidious
onset, and poor outcome, have core psy-
chological deficits, particularly in the at-
tentional sphere. For many years, Strom-
gren (1965) and other Scandinavians have
recommended that the so-called good prog-
nosis schizophrenics, or what Vaillant
(1964) calls the remitting schizophrenics, be
more fully separated from the nuclear or
nonremitting schizophrenics. This view-
point is now being explored in relation to
ventricular enlargement found in CT scans
of different categories of patients (Wein-
berger, De Lisi, Perman, Targum, and
Wyatt, 1982). Strauss and Carpenter (1974)
have shown that poor premorbid adjust-
ment is associated with poor postmorbid
adjustment, particularly in the social and
occupational sphere. It appears that the
poor premorbid and prodromal symptoms,
such as social withdrawal, selectively pre-
dispose to later negative or deficit symp-
toms after a florid episode. What have been
called the "characteristic," florid symptoms
of schizophrenia are actually not so charac-
teristic because they also are found in other
florid psychoses. The longitudinal se-
quences of florid symptoms preceded by,
and followed by, deficit symptoms make a
more definitive basis for the diagnosis of
schizophrenia.
My present view is that it would be desir-
able to spell out still more than does DSM-
III the following limitations for the di-
agnosis of schizophrenia: The florid (posi-
tive) symptoms remit and relapse, but the
deficit (negative) symptoms persist with
only minor improvements and are inade-
quately documented in DSM-III. From the
standpoint of the deficit symptoms, which I
believe are primary in schizophrenia,
though admittedly difficult to delimit, schi-
zophrenia is a nonremitting disorder. It is,
of course, useful to forestall relapse of the
florid symptoms, but in the home and com-
munity, the deficit symptoms are most
troublesome. This includes what Wynne
and Singer (1963) have called "amorphous"
forms of communication. Special attention
to the issues of building compensatory so-
cial skills, in the family setting and else-
where, is important with these families.
However, with the families of acute psycho-
tics, only those with a long prodome and
poor premorbid history need to have this
emphasis.
Diagnostic distinctions of the kind I am
making here should facilitate more focused
interventions with both patient and family.
Most importantly, I believe that we must
take seriously this distinction between nu-
clear schizophrenia, as I have just described
it, and good prognosis schizophrenia, which
may be better regarded as a form of affec-
tive disorder or, as McCabe (1975) has sug-
gested, as a third major type of disorder dif-
fering from both schizophrenia and manic-
depressive psychosis. If my gloomy view of
the outcome of nuclear schizophrenia is
justified, much of what is included in pre-
sent-day family educational programs, such
as those of Anderson, Hogarty, and Reiss
(1980), will still be quite appropriate. On
the other hand, programs with acute psy-
chotics who mayor may not be on their way
to nuclear schizophrenia will need to be
planned with much more discretion and
with a different orientation. It is crucial to
consider this diagnostic distinction even in
the present day when more emphasis is
given to affective psychoses because the
acute phase of illness is still inadequately
understood and is quickly masked in most
settings by large doses of neuroleptics. It is
indeed rare to find a setting in which even a
week of diagnostic study with nonmedic-
ated care is available for acute psychotics,
despite the fact that Carpenter, McGlas-
han, and Strauss (1977) clearly showed that
such care is possible for even longer pe-
riods.
Beyond the immediate diagnosis, the
heavy use of neuroleptics also leads to mis-
interpretation of deficit symptoms. All too
 Working with Families of Acute Psychotics
often, the deficit symptoms that follow a
florid episode are viewed as requiring larger
doses and longer duration of medication.
Many patients who have been heavily me-
dicated for months or years, show side ef-
fects that blur with the deficit symptoms of
the illness. I wish it were true that this kind
of problem only occurs with therapists who
are inexperienced or incompetent, but I
fear that the problem is, in fact, far more
widespread.
What are the details of family interven-
tion and educational programs that can be
useful with the families of acute psychotics?
My suggestions here are highly preliminary
and tentative, and they can by no means
cope with all of the complexities that I have
noted above.
1. The interventions with both patient
and family need to be carried out most in a
phase-oriented manner (Wynne, 1983), be-
ginning at the first psychiatric contact,
when the family typically brings the patient
to the psychiatric facility. Perlmutter (1983)
at the Strong Memorial Hospital in Roches-
ter collected survey data from the ED staff
for 1000 consecutive patients. The data spe-
cified whether or not family members came
with the patient and whether the patient's
central presenting problem was about a
family difficulty; 700 of the 1000 patients
were thus involved with their families at the
time of the emergency contact. Dismaying-
ly, relatively few families were seen at that
time. As I have indicated, when the families
are brought into the picture and viewed as
assets and resources for both the ongoing
clinical care and for future planning with
the patient, their subsequent cooperation is
much more likely. As Scott (1976) has
shown, a psychological closing off by the
family from the patient is greatly reduced if
the family is contacted within 72 h of admis-
sion and brought into the treatment plan-
ning.
2. A crisis intervention approach with the
family is appropriate in the subsequent
days. Surely, acute onset of psychosis in a
family member is experienced as a crisis. As
in other crisis intervention, the family needs
time to talk over and to ventilate their feel-
ings of perplexity, frustration, and anger.
The CFI often serves a useful therapeutic
purpose of this kind, even though it is ex-
113
plicitly designed to be a neutral research
procedure. Indeed, a good therapist at that
crisis stage would handle the situation in a
very similar manner. Increasingly in Roch-
ester, we have been finding that the CFI can
usefully be integrated into the clinical pro-
gram and should not be labeled and put out
of sight as data only for the researcher.
3. After preliminary meetings with the
relatives, without the patient present,
meetings with the whole family (including
the patient) are nearly always useful, at the
very least to take stock of their future to-
gether. Some family therapists have recom-
mended against having patients meet with
their families during the so-called florid
phase, perhaps because they fear that this
will evoke high expressed emotion from the
family (Vaughn and Leff, 1976). If the in-
terviewer takes a firm, structuring position
and adapts the length of the family meeting
to the functioning of those present, I have
found that these difficulties, when they do
occur, are quite manageable. Such meetings
can be quite helpful in assessing the family
relationships along dimensions (especially
structured) that are not so fully evident
from individual interviews or test situa-
tions. However, I do not automatically con-
tinue with conjoint interviews when assess-
ment indicates this seems unworkable or
unrealistic.
4. The concept of psycho educational
workshops similar to those used with fami-
lies of chronic schizophrenics need to be
modified for families of acute psychotics. I
am intrigued by a recent innovation by
Crabtree (personal communication, 1982)
in which he introduced five successive even-
ing workshops, a week apart, with parents
of four to five families, each having a psy-
chotic offspring in the 18- to 25-year-old
range. They spend one session each on schi-
zophrenia, affective psychoses, borderline
problems, street drugs and therapeutic me-
dication, and developmental-life cycle is-
sues. This explicity educational program
leads, as in the experience of most others
who have carried out educational pro-
grams, into frequent requests for more ex-
tended family therapy, which may be on a
group or on an individual basis.
Thus far, I have seen these families one at
a time, rather than in multiple-family
114 L. C. Wynne
groups. I label several meetings as educa-
tional and informational, set apart as not
having explicit therapeutic intent. How-
ever, I now regard education as a funda-
mental component of all the family therapy
that I do. An advantage in Crabtree's ap-
proach of discussing a series of disorders,
rather than the one disorder presumed to be
diagnostic for an identified patient, is that
the diagnostic complexity is conveyed at a
sensible, understandable level, which the
families seem to value highly and do not fall
so easily into demands for premature cer-
tainty.
5. Later follow-up will vary tremendously
with the diverse problems, and with their
course. Even when the diagnosis later is nu-
clear schizophrenia, early intervention of
this kind continues to prove helpful in eas-
ing the burden on the family as a unit, and,
in tum, helping the family to realize its po-
sitive capabilities in facilitating treatment.
My recent experience strongly indicates
that the willingness of families during and
after a first admission to be interested in ex-
tended programs of this kind is quite dif-
ferent from that of families of chronic pa-
tients. Although the work with families of
chronic patients may not be continuous, in-
termittent contact with family members,
sometimes about a new problem, some-
References
Anderson, C. M., Hogarty, G. E., and Reiss, D.
J. (1980). Family treatment of adult schizo-
phrenic patients: A psychoeducational ap-
proach. Schizophrenia Bulletin 6: 290-505.
Brockington, I. F., Kendell, R. E., and Left, J.
P. (1978). Definitions of schizophrenia: Con-
cordance and prediction of outcome. Psycho-
logical Medicine 8: 387-398.
Carpenter, W. T., McGlashan, T. H., and
Strauss, J. S. (1977). The treatment of acute
schizophrenia without drugs: An investigation
of some current assumptions. American Jour-
nal of Psychiatry 134: 14-20.
Carpenter, W. T., and Strauss, J. S. (1974).
Cross-cultural evaluation of Schneider's first-
rank symptoms of schizophrenia: A report
from the International Pilot Study of Schizo-
phrenia. American Journal of Psychiatry 131:
682-687.
DSM-II1. (1980). Diagnostic and Statistical
Manual of mental disorders (3rd ed.). Wash-
times to make sure that the staff member
was available, is useful. With the families of
acute psychotics, compared with the fami-
lies of chronic patients, the kinds of ques-
tions that are raised and the kind of anxiety
and alarm are quite different. The families
of chronic patients are often focused on
themselves. They are inclined to blame
themselves or, more usually, feel blamed
and neglected by the psychiatrists, all too
often with justification. In contrast, the
families of first-admission patients are more
inclined to look for "the problem" in pos-
sible marijuana usage, bad companions, or
in some life event, and only rarely wonder
whether something within the family may
be relevant. Nevertheless, these families of
acute psychotics are in a crisis; they are dis-
tressed; and they are responsive to those
who will listen to them and who will re-
spond with information and guidance. A
major task for the future in the field of pre-
vention lies in developing early support and
educational programs for the families of
acute psychotics. These programs require
careful design and are not so easily organiz-
ed in a standard format as with the families
of chronic schizophrenics. Nevertheless,
the payoff in preventive success may be in-
deed high.
ington, D. C.: American Psychiatric Associa-
tion.
Fox, H. A. (1981). The DSM-II1 concept of schi-
zophrenia. British Journal of Psychiatry 138:
60-63.
Helzer, J. E., Brockington, I. F., and Kendell,
R. E. (1981). Predictive validity of DSM-II1
and Feighner definitions of schizophrenia: A
comparison with research diagnostic criteria
and CATEGO. Archives of General Psychia-
try 38: 791-797.
McCabe, M. S. (1975). Reactive psychoses: A
clinical and genetic investigation. Acta Psychi-
atrica Scandinavica, Supplement 259: 13-133.
Perlmutter, R. A. (1983). Family involvement in
psychiatric emergencies. Hospital and Com-
munity Psychiatry 34: 255-257.
Scott, R. D. (1976). Closure in family relation-
ships and the first official diagnosis. In J. J(Ilr-
stad and E. Ugelstad (Eds.), Schizophrenia 75,
pp. 265-281. Oslo: Universitetsforlaget.
 Working with Families of Acute Psychotics
Strauss, J. S., and Carpenter, W. T. (1974). The
prediction of outcome in schizophrenia: II. Re-
lationships between predictor and outcomc
variables: A report from the WHO Intcrnat-
ional Pilot Study of Schizophrenia. Archives of
General Psychiatry 31: 37-42.
Stromgren, E. (1965). Schizophreniform psycho-
ses. Acta Psychiatrica Scandinavica 41:
483-489.
Takahashi, R. (1982). Personal communication.
Vaillant, G. E. (1964). An historical review of
the remitting schizophrenias. Journal of Nerv-
ous and Mental Disease 138: 48-56.
Vaughn, C. E., and Leff, J. P. (1976). The in-
fluence of family and social factors on the
course of psychiatric illness: A comparison of
schizophrenic and depressed neurotic patients.
British Journal of Psychiatry 129: 125-137.
Weinberger, D. R., DeLisi, L. E., Perman, G.
P., Targum, S., and Wyatt, R. 1. (1982). Com-
puted tomography in schizophreniform dis-
order and other acute psychiatric disorders.
115
Archives of General Psychiatry 39: 778-783.
Weiner, J., and Stromgren, E. (1958). Clinical
and genetic studies on benign schizophreni-
form psychoses based on a follow-up. Acta Psy-
chiatrica et Neurologica Scandinavica 33:
377-399.
World Health Organization. (1973). The Inter-
national Pilot Study of Schizophrenia, Vol. 1.
Geneva: World Health Organization.
Wing, J. K., Cooper, J. E., and Sartorius, N.
(1974). The measurement and classification of
psychiatric symptoms. Cambridge: Cambridge
University Press.
Wynne, L. C. (1983). A phase-oriented approach
to treatment with schizophrenics and their
families. In W. R. McFarlane (Ed.), Family
therapy in schizophrenia, pp. 251-265. New
York: Guilford.
Wynne, L. c., and Singer, M. T. (1963). Thought
disorder and family relations of schizophrenics:
II. A classification of forms of thinking. Ar-
chives of General Psychiatry 9: 199-206.
10. Family Education as a Component
of Extended Family-Oriented Treatment
Programs for Schizophrenia
L. J. Cozolino and 1\1. J. Goldstein
Introduction
The desire for knowledge in relatives of
schizophrenic patients has stimuled the de-
velopment of long-term intervention pro-
grams which include an initial phase of fam-
ily education. The present paper reviews
four family intervention programs designed
for these relatives, which were summarized
in a recent volume (Goldstein, 1981). These
programs were developed in the context of
clinical research studies headed by Falloon
and Liberman in Los Angeles, Leff and his
associates in England, and Anderson and
Hogarty in Pittsburgh. Each program con-
tains a component specifically designed to
transmit some basic information about schi-
zophrenia and to answer questions concern-
ing diagnosis, etiology, prognosis, and the
relatives' role in the recovery process.
Family and patient education programs
did not begin with these newer family inter-
vention programs but have been developed
during the last 20 years throughout the Uni-
ted States, Canada, and England in both
medical and community settings. They have
grown with the support of physicians, thera-
pists, community workers, and family mem-
bers to serve the needs of relatives and to
explore ways of getting information and
support to a wide audience. Information is
delivered in many forms, ranging from
books, pamphlets, and daylong courses to
being transmitted as part of ongoing family
therapy programs or mutual support or-
ganizations. The family education pro-
grams analyzed in this paper have drawn
heavily on this prior work.
The four family education programs out-
lined in this paper are part of long-term
family intervention programs established
for research purposes in medical settings. In
beginning to evaluate family education, it
seems appropriate to start with such pro-
grams because of the wealth of information
that is available from the broader research
project. This context allows us to have an
accurate understanding of the information
made available and the type of information
and treatment received by the patient and
provides an opportunity for direct assess-
ment of the interpretation and utilization of
information and practical management ad-
vice.
While all of these programs attempt to
delineate a family education phase from a
long-term intervention phase, it is impor-
tant to note that long-term interventions
have the opportunity to continue the edu-
cational process throughout the treatment
periods which range from 9 weeks with a 3-
month follow-up (Snyder and Liberman,
1981) to 2 years (Falloon, Boyd, McGill,
Strang, and Moss, 1981). Thus, it is some-
what arbitrary to delineate an "educational
component" in isolation from the rest of the
program. In all of these programs, the edu-
cational component can be interpreted as a
seeding process or as an opportunity to be-
gin with a common language and set of as-
sumptions rather than expecting to transmit
information which will be understood and
used by itself. In subsequent group meet-
ings, Berkowitz, Kuipers, Eberlein-Fries,
and Leff (1981) found that although many
participants seemed to need to hear the in-
formation given during the educational
component a number of times, it did seem
" ... to have served the purpose ... of help-
ing relatives formulate questions that they
may not have been able to ask before."
(p. 38). All program descriptions stress the
118 L. J. Cozolino and M. J. Goldstein
importance of exercising a great deal of
flexibility in both the educational and long-
term intervention phases because of the
individual needs of particular families.
Because this is a relatively new move-
ment, systematic assessment of the most ef-
Comparison of the Four Programs
To appreciate similarities and differences in
the family education component in each of
the four programs, it is necessary to have
some overview of the total, extended pro-
gram in which each is embedded. There-
fore, in this section we will first present a
brief summary of each of the programs
which presents the underlying rationale and
a short description of each of the different
phases of the program. The brief summa-
ries are based on program descriptions pro-
vided by these investigators in the Gold-
stein (1981) monograph on this subject.
LefT, Berkowitz, and Knipers
This program is the most specifically fo-
cused toward lowering expressed emotion
(EE) in the home and reducing the amount
of direct contact between the patient and
the family. The first component of the pro-
gram is the joint interview, which takes
place as soon as possible after the patient's
discharge from the hospital. This interview
consists of a meeting with the family, pa-
tient, a psychologist, and a psychiatrist to
discuss ways of reducing contact and to cov-
er specific areas of conflict. There may be
one or more of these interviews during the
next 9 months taking any form which is
helpful to the particular family.
Mental health education was introduced
on the basis of research which suggested
that high EE families tended to attribute
the patient's behavior to deliberate acts
while low EE families attributed the pa-
tient's behavior to the illness. It was felt that
the more information the family had about
schizophrenia, the better they would be
able to cope with the patient and to keep
down their level of EE.
The group meetings for relatives were de-
signed to bring together low and high EE
families and to encourage high EE families
ficacious methods of transmitting informa-
tion to the families has yet to be done. It is
hoped that this comparison will serve as a
starting point in the delineation of research
questions to be addressed in the future de-
velopment of family education programs.
to engage in behavior modeled for them by
the low EE group members. This was a dif-
ficult process, and the groups tended to take
on the qualities of "support, catharsis, en-
couragement, and interest." Overinvolved
relatives were also encouraged to find inter-
ests in areas other than the patient. Mem-
bers often discussed information about
schizophrenia, management issues, and
their feeling about the entire process.
Liberman and Snyder
Based on the premise of an interrelation-
ship between EE and psychophysiological
arousal, Liberman and Snyder hypothesize
that relapse in a biologically vulnerable in-
dividual is a result of a balance between life
stressors and problem-solving skills. Too
much environmental stress (for example,
an overly critical or involved family) and/or
too few coping skills may lead to relapse.
Thus, the emphasis here, as in the Leff et al.
program, is on improving interactive coping
strategies used by family members and pa-
tients, instead of on emphasizing reduction
of contact.
Thus, families in which most members
were rated as high in EE were admitted to
the program. The first two sessions were
family and patient education, which focus-
ed on information about schizophrenia and
its effects on both the patient and the fami-
ly. Practical information on the use of re-
sources in the hospital and community were
also discussed.
The final seven sessions took the form of
joint therapy for relatives and patients, de-
signed to address specific communication
and problem-solving skills which could as-
sist the family during the recovery period.
 Family Education as a Component of Treatment
Anderson, Hogarty, and Reiss
Similar to the Liberman and Snyder pro-
gram in theoretical assumptions, Anderson
et al. base their model on research in
arousal/attention dysfunction and deviant
family communication as well as on re-
search on EE.
Anderson et al. introduce certain social
work concepts, stressing a period of con-
necting with the family, which the other
programs do not emphasize. Anderson
et al. also refer to their daylong educational
component as a survival skills workshop for
relatives, which underscores their emphasis
on pragmatic, day-to-day issues of coping
and management. Their overall goal is to
increase the predictability and stability of
the home environment, to increase adher-
ence to medication, and to provide support
for the family, as well as the patient,
through the recovery process.
Like the Berkowitz et al. project, there is
a strong emphasis on the use of support
groups and on the expansion of the family
and the patient's social support network.
Their interventions focus on present prob-
lems during family therapy, postponing
long-term issues until after the immediate
crisis of the schizophrenic episode is resolv-
ed. While family therapy in this program
does not appear to be explicitly behavioral
in orientation, it does seem to be pragmatic
and problem focused. Communication
skills and problem-solving skills are pre-
sented in a less structured manner than in
the more behaviorally oriented programs.
When, Where, and to Whom is Family Education Provided?
Next, we will focus upon the specific educa-
tional unit, variously termed mental educa-
tion, family education, or survival skills
workshop, in which family members, some-
times with and sometimes without the pa-
tient present, are offered information about
schizophrenia within a didactic format.
Table 1 presents a schematic of each of
these educational components which pre-
sents details concerning the lengths and
number of sessions, the location, member
119
Falloon, McGill, and Boyd
Two initial home sessions comprise the
family education component of treatment
which take place with the patient present.
Interaction and sharing are encouraged be-
tween the patient and their family. Thera-
pists often return to education later during
therapy as the need arises.
Family education and therapy are con-
ducted in the home, which increases atten-
dance and is thought to improve the gener-
alizability of behavioral interventions. The-
rapists are on 24-h call.
This program is the most behaviorally
oriented program reviewed here. Assess-
ment begins with a behavioral analysis of
each individual and then of the family
group. The interviewer attempts, through
this process, to define a number of core
communication and problem-solving defi-
cits which become the focus of the com-
munication and problem-solving training.
Consistent with behavioral theory, assess-
ment is carried on throughout the process,
and the specific behavioral interventions
can be subsequently modified to meet new
or changing needs.
The Falloon et al. program is perhaps the
most extensive of all the approaches in
terms of actual family involvement and
length of treatment, with 40 family sessions
being held over a 2-year period. Sessions
are conducted in the home for the first 9
months and are either continued through
the next 15 months or the family has the
option of participating in community-based
family groups. Monthly medication evalua-
tions are conducted with the patient.
of families present, whether the patient is
present or not, and the style of presenta-
tion.
It can be seen that the number of sessions
ranged from one to three and that they var-
ied from relatively brief sessions to the all-
day format of Anderson's group (Ander-
son, Hogarty, and Reiss, 1981). Mostly,
these programs begin while the patient is in
the hospital or shortly thereafter when the
patient has been stabilized on medication as
120 L. J. Cozolino and M. J. Goldstein
Table 1. Overview of Psychoeducational Programs

Research No. of Length of Location No. of Patient Timing of Style of

group sessions sessions families present education presentation
present
Berkowitz 2 No set Hospital 1 No After Brief textis read to
etal. (1981) length or home patient participants and
discharge given to them to
take home, ques-
tionsare
answered
Snyder 2 2h Storefront 3 patient Yes During hospi- 3 co-leaders jointly
and Liber- and their talization, lead 1st h and then
man (1981) families most after a separate into in-
few weeks dividual therapy
groups, structured
syllabus, class-
room format -
handouts and
homework
assignments
Anderson, 1 5h Hospital 3-4fam- No Early in treat- Seminar structure
Hogarty, ilies ment,most - lecturelhand-
and Reiss patients still in outs
(1981) hospital
Falloon, 2or3 Phh Home 1 family Yes When patient At-home discus-
Boyd, is (re )stabiliz- sion where family
McGill, ed on medica- members are
Strang, and tion asked to share
Moss their experiences
(1981) and the patient is
encouraged to
discuss personal
and interpersonal
experiences

in the Falloon et al. program (Falloon,
Boyd, McGill, Strang, and Moss, 1981).
The programs do differ in one important
factor and that is whether the patient is pre-
sent or not at these educational settings. In
two of the programs they are and in two
they are not. One ofthe key issues for future
investigation is whether this particular com-
ponent of a family program is inhibited or
facilitated by patient participation. This is
not as simple an issue to resolve as it might
seem as the programs outlined in Table 1
vary considerably in important parameters
which would affect the value of patient par-
ticipation. For example, the Anderson
et al. family education program (survival
skills) begins shortly after the patient is ad-
mitted to the hospital and still very psychotic.
Participation of the patient at this point may
not be fortuitous, and family members may
need an opportunity to consider matters in
a less emotionally charged context. In the
Falloon project, the patient has been dis-
charged and stabilized on medication for
4-6 weeks before family education is car-
ried out in the home with only the patient
and significant relatives present. Is it the
level of remission, the home context, or the
focus on the single family unit that deter-
mines the patient's participation? Some of
the projects use a group format while
others, such as Leff et al. and Falloon et aI.,
meet with an individual family unit. In
what way do these contextual factors affect
the impact of a family education unit, if at
all? Further research is needed to clarify
 Family Education as a Component of Treatment
whether the timing, group composition, or
site (home vs clinic) affect the impact of a
family education program.
Note that those programs that involve
relatives only do not appear to provide sep-
arate education for the patient. It remains
to be demonstrated whether separate pro-
grams targeted specifically toward patients
will enhance the impact of the programs
directed at the relatives.
To date, none of these family education
units have been prepared for mass media
distribution to be utilized in other settings.
Clearly this represents a distinct possibility
as there has been sufficient experience with
the administration of these programs that
scripts could be created and put on video-
Content of the Various Family Education Programs
The informational contents of the family
education component across the four pro-
grams have a great many similarities. There
are many common features in that they all
deal with (a) definitions of a schizophrenic
disorder, (b) a description of the various
symptoms, (c) present theories about etiol-
ogy and prognosis, and (d) treatment re-
commendations concerning pharmacother-
apy and family impact. The overlap be-
tween programs is not due to chance as a
number of investigators worked together in
London at one time (Leff, Liberman, and
Falloon), and Hogarty had contact with
these investigators in the course of develop-
ing his program with his colleague, Carol
Anderson. While the nature of the informa-
tion across programs is relatively constant,
the manner and detail in which it is present-
ed varies a great deal. The Leff et al. project
keeps factual information to a minimum,
while the Falloon and Liberman programs
contain a great deal of factual information
which is presented at a fairly sophisticated
level. The Anderson program lies some-
where in the middle, presenting factual in-
formation but balancing it with a good deal
of phenomenological examples of the ex-
perience of both the patient and the rela-
tives.
Due to differences in terminology and
emphasis, it was difficult to compile a sim-
ple table which would convey the similari-
121
tape. An issue not clarified in Table 1 is that
in each of the treatment programs the fami-
ly education program was delivered, in
part, by one of the persons who was involv-
ed in the longer term family program. Thus,
it was not simply a case of a mental health
professional imparting information and
withdrawing from future contact from the
family. Any future research on the impact
of these family education programs needs
to consider not only the issue of live vs
videotape format, but whether the person
or persons who provide the information are
likely to be involved in the subsequent sup-
port provided to the families who receive
the family education.
ties and differences of information given
across programs. Instead we felt it would be
more informative to offer to the reader a
general outline of information given in the
areas of diagnosis, symptomatology, etiol-
ogy, course, and treatment and to include
short passages from the various syllabi to
give a flavor of each program. If the reader
is interested in developing a program at her/
his own facility, the authors strongly sug-
gest contacting these programs for more
complete syllabi and information.
Please note that while passages are taken
from the syllabi or written materials offered
from these programs, at the time of writing
this report the syllabus from the Anderson
et al. program was not available, and pas-
sages are taken from a videotape of their
survival skills workshop.
Diagnosis
All of the programs begin with a discussion
of the diagnosis of schizophrenia, which
usually includes common misconceptions
(i.e., schizophrenia does not mean split
personality, most schizophrenics are not
violent), a basic idea of the statistical occur-
rence of schizophrenia, and phenomeno-
logical reports of patients who have been
through an experience ofthe illness.
A major part of the discussion of diagno-
sis is to demonstrate that the bizarre behav-
122 L. J. Cozolino and M. J. Goldstein
ior they have been experiencing is actually a
part of a "known" medical entity and that
this experience is shared by many people
throughout the world. Another goal is to re-
conceptualize the patient's deviant behav-
ior as a reflection of his mental disorder
rather than as "bad" or "willful" behavior.
For example, in the syllabus of the Leff
et al. family education program, it is stated
that:
· .. Schizophrenia is an illness. If affects people in
many ways. The difficulty is that the sort of ex-
periences it gives rise to seem completely real to
the people suffering from it. For instance, some-
one who hears voices may talk back to them be-
cause he thinks they are voices of people who are
actually there ... Because the patient cannot usu-
ally explain what is happening in his mind, it is
not always easy for other people, even those who
live with him like yourselves, to realize that many
of the odd or upsetting things he does are caused
by the illness ... Schizophrenia is not a rare ill-
ness ... it can affect anyone.
The positive impact of such clear definitions
of the schizophrenic disorders can be seen
in the following quotes from the relatives
cited in the preface of this volume whose re-
actions to a psychoeducational program
were sought:
I don't think the lay person knows very much
about what schizophrenia means, so it was a big
educational, a big learning experience just to
learn what schizophrenia was.
Symptoms
A discussion of symptoms usually follows
diagnosis, which continues the process of
normalization. Perhaps for the first time,
relatives become aware that others have
witnessed bizarre behavior such as their rel-
ative has been exhibiting. The Falloon and
the Liberman programs go through the var-
ious symptoms, giving definitions and ex-
amples as seen below:

Besides attempting to "normalize" their ex- Falloon et al. (1980) Liberman et al. (1979)
perience to some degree, information Symptoms Symptoms
about diagnosis allows relatives to be better Disturbances of Thought broad-
thinking casting
able to communicate with professionals
Delusions Thought with-
they may encounter both in and outside of drawal
the program. Perhaps the biggest initial ob- Hallucinations Thought insertion
stacle in explaining the diagnosis of schizo- Incoherent talk Paranoid
phrenia is overcoming preconceptions delusions
based on movies and news reports and to Inappropriate Social with-
allay the fears of the patient and their fami- feelings drawal
lies based upon these misconceptions: Bizarre habits Work behavior
Self-care
· .. Schizophrenia means a disintegration of the Feelings
personality where a person finds it difficult decid- (flat affect)
ing what is real and what is not real. It is a little Thoughts
like having dreams when you are wide awake ... Conversation
We describe schizophrenia as a major mental ill- and speech
ness because it has a great effect on nearly every
aspect of a person's life. Everything that is im- While Leff and his group describe the above
portant in our lives may be affected by schizo- symptoms, their descriptions are less pa-
phrenia. The symptoms of schizophrenia are not tient centered and more focused on the in-
identical for every person, but they nearly always teraction between the patient and their rel-
produce a handicap in many aspects of everyday atives, as seen in the following excerpt:
living. (Taken from "What is Schizophrenia", Schizophrenia can also affect feelings .. . You
Falloon, McGill, and Boyd, 1980). may have noticed that he doesn't seem to care for
· .. a person suffering from schizophrenia has a you as he did before, or show his love for you in
different experience of reality than other people. the same way. There may be fewer and fewer
This means that the person with schizophrenia times when you can really talk to each other, and
sometimes relates to persons and events in a man- you may sometimes wonder whether he still feels
ner that is different enough from what most anything for you at all. He can't help this, be-
people usually expect that it can be said that the cause his usual feelings have been swamped by
schizophrenic has a split with reality. (Liberman the illness and he has become very wrapped up in
et al. 1979). himself.
 Family Education as a Component of Treatment
Leff also describes: tions to the person with schizophrenia - is crucial
in the development and outcome of schizophre-
1. Instability of mood where the patient is at
times "miserable or desperate" and others nia.
when he is "very excitable and overactive"
2. Social isolation
3. Reduced energy
4. Apathy
5. Personal hygiene issues
Some of the programs go into an explora-
tion of the phenomenology of a schizophre-
nic psychosis. In the Hogarty/Anderson
survival skills workshop, excerpts of first-
person report of the inner experience of a
psychotic break are read aloud and hand-
outs of such accounts are also provided to
relatives. In the Falloon sessions, which are
carried out in the home with the patient pre-
sent, both patient and relatives are asked to
share their experience of the psychotic bre-
akdown. In both instances, it is reported
that relatives report more sympathetic atti-
tudes toward the patient after such a session
as they appreciate that their relative's expe-
riences and behavior are not as unique as
they formerly believed.
Etiology and Course
Although the etiology and course of schizo-
phrenia may be the aspect of the illness
which is least understood, it is perhaps the
most important for the attitudes and morale
of the relative. Counteracting guilt con-
cerning the cause of the illness and balanc-
ing realistic expectations for the future with
a sense of hope is an important component
of these programs. All four programs are
grounded in a diathesis stress model of
schizophrenia, and they describe the genet-
ic as well as the environmental factors
which may contribute to the development
of schizophrenia. To the question "What is
the cause of schizophrenia?" Liberman et
al. explain:
It is probably a better idea to think of the causes
of schizophrenia rather than a single cause. And
it is also important to think of factors which can
make the symptoms and life functioning better or
worse in a person who already has schizophrenia.
Since medication can improve the symptoms of
schizophrenia, it is likely that brain chemistry
changes are part of the problem. On the other
hand since psychotherapies and behavior thera-
pies which focus on the patient's environment
also produce therapeutic change, it is felt that the
environment ... especially other people's reac-
123
Thus, an important task of the information-
al content is to shift focus from past fears
and blaming to a realistic and positive view
for the recovery process, involvement in the
program, and maintenance of the well-
being of all family members.
The problem of how to handle theories
concerning genetic factors is frequently ad-
dressed by developers of family education
programs. The Liberman handout states,
"Many people want to know if schizophre-
nia comes from parents, they want to know
if they have to worry that someone else in
the family will get it. The heredity of schizo-
phrenia is a complicated business, but it is
clear that it sometimes runs in families and
sometimes does not." In other programs,
the genetic data are dealt with more expli-
citly. Any presentation of genetic theories
is likely to stimulate considerable discussion
of risk factors for other offspring as well as
revive family arguments concerning whose
side of the family passed on the "bad"
genes. Therefore, it is a topic that is covered
in most family education programs but
needs to be handled with great care and sen-
sitivity.
Similarly, with regard to the stress side of
the stress-diathesis theories covered in
these programs, theories concerning the
family etiology of schizophrenia are either
played down or presented as unproven.
Stress, particularly during the post-dis-
charge, aftercare period, is presented as a
trigger of relapse, and family conflict is
identified as a particularly potent stressor
during that period. Since a number of the
family education units are part of more ex-
tended aftercare programs targeted spe-
cifically for high EE family units, the im-
portance of reducing criticism and emotion-
al overinvolvement is emphasized as a
mechanism for reducing family stress.
While hope must be maintained, realistic
expectations must also be supported. "Pre-
paration" for relapse may serve as a stress
inoculation as well as a sensitizer to a wor-
sening of symptomatology before another
episode.
In the Leff et al. syllabus, the following is
stated:
124 L. J. Cozolino and M. J. Goldstein
Some people only have one attack of schizophre-
nia; they recover from this and never have anoth-
er. Others, luckily only a small number, do not
respond to treatment at all. However, most pa-
tients although they recover from the attack are
likely to have other attacks. These may occur
within weeks of recovery, or may happen years
later. During further attacks new kinds of odd
behavior can appear, but often the same pattern
will repeat itself.
Although all of the family education pro-
grams present material about etiology, they
vary greatly in the detail covered. For ex-
ample in the handout Some Plain Facts
about Schizophrenia, distributed by the
Liberman group as part of the family educa-
tion program, the section on causes of schi-
zophrenia states, "It is clear that schizo-
phrenia involves a chemical problem in the
brain. This is a problem that the person is
born with and it's no one's fault." Contrast
this brief explanation with the extensive
presentations by Hogarty, which covers
some brief facts of neuroanatomy, neutro-
transmitter theory, the dopamine hypo-
thesis, and the relationship ofthis model to
a hyperarousal theory of cognitive dysfunc-
tion on schizophrenia. Hogarty states:
What you see on the board are two things that
look like a star fish ... this is my feeble attempt to
represent a neuron ... a neuron doesn't look that
big, in your brain right now there are the better
part of ten billion neurons, and they are tightly
packed in ... I'm going to tell you a little bit about
how the neuron works ... The center of the neu-
ron is the cell body and those little branches
which come out are ... called dendrites, and this
is a very important part ofthe cell body, the axon,
these are the nerve fibers ... There are messages
that go through the body ... These ten billion
neurons are connected but they are not physically
wired together .. , and the message travels
through a very complex process ... Neuroanato-
mists refer to the message coming in as the af-
ferent, they are affecting the cell and this would
be efferent ... (the spaces between cell) are cal-
led synapses. After a description of dopamine
transmission, Prof. Hogarty continued: My guess
is that everyone in this room has a family member
that has been ill who is receiving some form of
medication which we call antipsychotic medica-
tion or antischizophrenic medication. You may
hear them referred to as tranquilizers. It turns out
that they don't tranquilize people, they operate
on this (dopamine transmission) process ... They
put back the chemical balance. Those molecules
look like the dopamine and they take its place,
they block the crevice ... They'll compete and
substitute with the dopamine and sneak-in there.
The effect is to tone down the strength of the mes-
sage, not as much can get through. This whole
process of being excited and distracted and too
much coming in at one time. This is a chemical
way of enhancing the ... toning down process in
the cell which is not working at the moment. It
helps the body tone down and get equilibrium
back. It's the tune up - we're playing a trick on
the receptors with medication. That's all we're
doing.
A key issue that is still unresolved is what
relatives and/or patients retain from the in-
formation provided. In some instances, lit-
tle of a technical nature is retained; in other
cases, a key concept is retained which is
found helpful. Consider the following
quote from one of the relatives described in
the preface of this book who was sharing
her experience of family education:
What helped was when you explained what hap-
pens when there is a chemical imbalance or what
is it, a dopamine flow? What happens inside their
head, I always had the idea that he could control
his symptoms, but we found that day that they
really cannot control it. It is something that you
have no control over. It helped a greal deal, it
really did, to know that he had no grip over real-
ity.
Or as another parent stated:
The number one point is that it is a chemical im-
balance like sugar diabetes that can be treated
with medication. He (the lecturer) went on to ex-
plain how this chemical imbalance bombarded
with stimuli on the receiving end of the synapse
that made all kinds of sounds and sights magni-
fied - so the medication would screen out the
dopamine that was bombarding your senses and
eliminated some of the frightening sounds or sce-
nes or colors and they were brought down to nor-
mal perception and became more realistic.
Even when retention is not great, relatives
often express appreciation that mental
health professionals share with them the
latest knowledge in this field. It is the be-
ginning of an alliance, a building process
than can facilitate relatives' investment in
the longer and more demanding phases of a
sustained aftercare treatment program.
Difference in detail are also evident with
regard to information provided about prog-
nosis. Some programs are extremely ex-
plicit in providing figures about relapse po-
tential while others are more vague One
advantage of providing explicit figures
about relapse is that they provide a natural
 Family Education as a Component of Treatment 125
segue into material concerning the protec- Maintenance drug
tive value of antipsychotic medication, a therapy
point emphasized in all four programs. Relapses rates and
medication
Symptoms likely
Treatment to improve
Symptoms likely not
All four educational components divide to improve
their treatment emphasis between pharma- Reasons patients
cological and psychosocial interventions re- stop taking
flecting the synthesis between these two medication
orientations, which until recently have Getting the right
dose
often been seen in opposition. The goal of Role of the family
this information is to increase drug com- in sustaining drug
pliance and stress the importance of both therapy
chemical and environmental components of
the illness.
Falloon et al. summarize their sessions on
medication thusly:
Medication 1. Regular tablet taking is the mainstay of
The Leff program describes medication: treatment of schizophrenia.
2. Major tranquilizers are very effective
These play an important part in the treatment of medicine for the treatment of schizo-
schizophrenia. They help to stop the voices in the
patient's head, they make him less anxious and phrenia.
restless and help him to think more clearly. They 3. In low doses they also protect a person
protect him against stresses coming from his own from a relapse of symptoms.
experiences and his everyday life. The effects 4. Side effects are usually mild and can be
cannot always be seen straight away. Some pa- coped with.
tients are not given tablets but are put on injec- 5. Street drugs make schizophrenia worse.
tions. These have the same effect as tablets but For the Liberman and Falloon projects, the
can be given less often. two sessions of education are followed by
interactive problem solving and communi-
In contrast to this minimal explanation of cation skills training, the actual details of
medication, the Falloon and Liberman pro- which are not described during education.
grams offer in-depth explanations of the The Leff and Anderson educational com-
range of medication and the benefits and ponents, however, include advice as to
risks of drugs used to combat side effects. what the families can do to assist the patient
The Leff et al. syllabus contains a single and themselves during the recovery pro-
page of drug information, while the Fal- cess. Their advice centers around reduction
loon, McGill, and Boyd syllabus is ten of contact, lowering EE, gaining support
pages long. Some of the psychopharmacol- from other families, and taking care of
ogical material provided by these investiga- yourself (the relative).
tors is summarized below. From the Leff et al. syllabus:
Falloon Liberman
Biochemical theory of Types of drugs The best thing for your to do in this situation is
schizophrenia Neuroleptics firstly, not to spend so much time with him so that
Types of medication Mood stabilizers you don't get on each other's nerves. It is impor-
Benefits of medication Antidepressants tant that the patient leads as independent a life
Reduced relapse Stimulants as possible ... If you have to be together a lot of
rates Types of neuroleptics the time, the best thing to do is not to shout or cri-
Relapse worse and effects ticize or get too involved.
without medication Generic names
Dosages Trade names The Anderson et al. psychosocial treatment
Warning signals Equivalent doses emphasizes both patient management and
Side effects Who is likely to concern for the relatives. The topics em-
Drugs/alcohol improve phasized are as follows:
126 L. J. Cozolino and M. J. Goldstein
Management
Benign indifference
Realistic expectations
Normalization of family life
"Setting limits"
Concern For Self
It may be a long process
Take care of yourself
Mutual support between families
Families asked to share their
gains with other families new to the project
Written suggestions as to how to
start a parent's group are handed out as well
as information about parent of adult
schizophrenic (PAS) groups
With these foci, the Anderson et al. family
education session remains true to its title of
survival skills workshop.
Throughout the educational component,
each program maintains its basic perspec-
tive. While they are all balanced with a mix-
ture of didactic and experiential materials,
the Falloon and Liberman programs take a
more information-oriented stance, which is
separated from the rest of the treatment
program. The Leff and Anderson programs
on the other hand seem more directed to-
Implications for Future Research on Family Education
Clearly, families need and desire informa-
tion and assistance. However, a number of
important issues concerning format, deliv-
ery, and content require more research to
optimize the impact of a family education
program.
These programs described above, being
experimental in nature, incorporate many
nonspecific factors, such as novelty and
enthusiasm, which are difficult to quantify.
It is important, however, to compare pro-
grams in a way that provides information
about which components se.em most useful
and in which combinations. There are some
specific issues which call for discussion and
investigation as program technology con-
tinues to evolve and expand.
The programs outlined previously have a
great many similarities, which is their most
striking feature upon first inspection. Their
common goals are to (a) transmit informa-
tion, (b) lead to a reduction of EE in the
home, (c) supply a rationale for the medic-
ward understanding the experience of the
illness and contain information focused on
the long recovery period.
The differences in detail offered and
manner of presentation may, in some cases,
be more an artifact of the authors presenta-
tion in their syllabi than of actual differ-
ences in their programs. A great deal of
flexibility is necessary to accomodate the
wide range of needs and interests of the
various families of which these syllabi are
merely a general format for presentation.
Thus, a program with a very detailed sylla-
bus may omit detailed information to an
upset or low functioning family, while an-
other program with little technical informa-
tion offers the opportunity for relatives with
a greater need for detailed information to
have their questions answered.
It would seem that assessing the needs
and abilities of the participants and match-
ing information to these needs is crucial. In
program development, it may be best to
have the ability and facilities to use these
various modes of offering information
about schizophrenia so that a wide range of
needs could and would be met.
ation and thus increase compliance, (d) re-
duce family guilt concerning their role in
the cause of illness, (e) replace unrealistic
expectations with more accurate estimates
of the length and nature of the recovery pe-
riod, and (f) give practical tips concerning
management issues. While there are differ-
ences in the emphasis and focus of informa-
tion given to families during education, the
basic theoretical positions are relatively
similar. Grounded in a diathesis-stress no-
tion of etiology and theories of arousal and
attention dysfunction, these programs have
a somewhat different emphasis from other
educational programs based on orthomo-
lecular hypothesis, for instance. Thus, in
analyzing family reaction to this informa-
tion, it must be held in mind that their reac-
tions may be related to particular packages
of information and not to information about
schizophrenia in general.
One salient aspect of family education is
the timing of the educational component in
 Family Education as a Component of Treatment
the context of the long-term intervention.
Both the Berkowitz et al. and the Anderson
et al. programs have a phase prior to educa-
tion. While somewhat different in empha-
sis, they can serve to both assess the level of
intellectual and emotional functioning of
the relatives and to set the stage for the edu-
cational component. In discerning the po-
tential position and negative effects of fam-
ily education, it may be important to ask
whether this preeducational connection
should be considered optimal or whether it
is for many a necessary prerequisite for a
positive educational experience. Variations
in the response to family education empha-
sizes the need for other tailoring informa-
tion in the educational component or for
doing remedial work with the family in pre-
paration for a structured group presenta-
tion.
Another important issue of timing re-
volves around at what point in the course of
the patient's illness should family interven-
tion and education begin. The Anderson
et al. and Liberman and Snyder programs
most often begin during hospitalization
while the Falloon et al. group waits from 4
to 8 weeks after patient discharge. The rela-
tive merits of the different perspectives on
this issue of timing seems to be an important
point of discussion and further investiga-
tion. The composition of the groups is also
relevant to the issue of flexibility in con-
veying certain information. A group for-
mat, as used by Anderson, as opposed to
presentations to individual families, allows
little adjustment to the individual needs of
participants. The combined group/individ-
ual approach of Snyder and Liberman is an
interesting compromise allowing general
information dissemination followed by fo-
cused discussion. Responses from families
in such programs will allow assessment of
the importance of presenting information
specific to the ability level of the particular
family members and specific to the particu-
lar ill relative. The usefulness of general in-
formation regarding the disorder presented
when the patient is first returning home can
also be examined. Group presentation
might inhibit relatives who are embarrassed
in a group or have difficulty in expressing
themselves in front of strangers. This
inhibition may be especially salient when
127
discussing issues concerning the mental ill-
ness of a relative. On the other hand, the
group format may allow support and mutual
sharing of feelings to develop across several
families and may reduce the pressure felt by
any single family. The general applicability
of the group format becomes even more
crucial when it is in the context of the ex-
tended treatment, as in the Berkowitz et al.
program. The relative effectiveness of mul-
tiple- vs single-family educational programs
has yet to be assessed.
Whether family education takes place in
the home or at the treatment facility is an-
other difference among these programs.
The added intimacy, increased information
about natural family context, possible
higher generalizability of treatment, and
the lowered rate of missed sessions of home
visits are balanced by the lower cost and
greater time efficiency for the professional
of having meetings in the hospital. Com-
partive data on the effects of the location of
family education may yield information
concerning the relative effectiveness of the
different venues as well as possible sugges-
tions about which families tend to benefit
most from which settings. Home visits may
prove to be especially important when
treating families of certain cultural and so-
cioeconomic groups.
Patient reaction to family education has
yet to be systematically assessed as well as
the differential effects of having the patient
involved or uninvolved with the family edu-
cation. Possible positive effects of patient
involvement, such as the added information
obtained by the patient, the patient's ability
to give direct feedback to the family, and
avoidance of the patient's reactions to being
"left out," may be counterbalanced by the
relatives' needs to express their feelings and
ask questions without the patient present.
Comparative patient feedback between the
programs which include and exclude pa-
tients from family education should be eval-
uated and considered in future program de-
velopment.
Consideration of informational content,
differences in intellectual level, level of
stress, and the individual needs of the parti-
cipants suggests investigation in the areas
of: (a) the need for an impact of technical
information, (b) the relative placement in
128 L. J. Cozolino and M. J. Goldstein
time and ordering in importance oftheoret-
ical and practical information, (c) the eval-
uation of what is retained and what is for-
gotten and, (d) the possible negative side
effects of information concerning medica-
tion, diagnosis, the extended recovery pe-
riod, and overall prognosis. Family reac-
tions to information can be assessed in a
variety of ways in subsequent therapy ses-
sions, tests of information retention, and
open-ended questions concerning their in-
References
Anderson, C. M., Hogarty, G., and Reiss, D. J.
(1981). The psychoeducational family treat-
ment of schizophrenia. In M. J. Goldstein
(Ed.), New developments in interventions with
families of schizophrenics (pp. 79-94). San
Francisco: Jossey-Bass.
Berkowitz, R., Kuipers, L., Eberlein-Fries, R.,
and Leff, J. (1981). Lowering expressed emo-
tion in relatives of schizophrenic. In M. J.
Goldstein (Ed.), New developments in inter-
ventions with families of schizophrenics (pp.
27-48). San Francisco: Jossey-Bass.
Falloon, I., McGill, c., and Boyd, J. (1980).
Treatment of Schizophrenia - Part II: Medica-
tion. Unpublished Manuscript.
Falloon, I., McGill, c., and Boyd, J. (1980).
What is schizophrenia? Unpublished Manu-
script.
Falloon, I., Boyd, J. L., McGill, C., Strang, J.,
terpretations and reactions to various
pieces of data. Continued contact allows
more direct observation and assessment of
the impact and interpretation of the infor-
mation by family education.
While these research questions have
evolved in the context of long-term inter-
vention programs for schizophrenic pa-
tients and their relatives, many of the same
issues apply to any educational program of-
fered to relatives of a mentally ill person.
and Moss, H. (1981). Family management
training in the community care of schizophre-
nia. In M. J. Goldstein (Ed.), New develop-
ments in interventions with families of schizo-
phrenics (pp. 61-78). San Francisco: Jossey-
Bass.
Goldstein, M. J. (Ed.) (1981). New develop-
ments in interventions with families of schizo-
phrenics. San Francisco: Jossey-Bass.
Liberman, R. P., Aitchison, R. A., and Falloon,
I. (1979). Family therapy in schizophrenia: Syl-
labus for therapists. Unpublished Manuscript.
Snyder, K., and Liberman, R. (1981). Family as-
sesment and intervention with schizophrenics
at risk for relapse. In M. J. Goldstein (Ed.),
New developments in interventions with famil-
ies of schizophrenics (pp. 49-60). San Fran-
cisco: Jossey-Bass.
11. Pilot Study of the Impact of a
Family Education Program on Relatives
of Recent-Onset Schizophrenic Patients
L. J. Cozolino and K. Nuechterlein
Introduction
It is assumed that advances in the technolo-
gy of family education gained from broad
psychosocial research programs with schi-
zophrenic patients and their relatives (this
volume, Chap 10, Goldstein, 1981) will be
adapted and utilized in a variety of settings.
Hospitals, professional training centers,
community mental health centers as well as
relative and patient support groups all de-
sire new and more impactful ways of de-
livering information to their members. Spe-
cific analysis ofthe educational components
of these long-term treatment programs may
be especially important for programs in
hospitals or community settings which can
offer family education but may not have the
resources to conduct a full-scale family the-
rapy program. In theory, educational com-
Pilot Study
As a first approach to this and more general
issues of family education, ten relatives
from six different families with a schizo-
phrenic member were interviewed. All rel-
atives participated in a family education
program in the context of an extended indi-
vidual patient aftercare treatment program.
All schizophrenic patients in this sample
were enrolled in the Developmental Pro-
cess in Schizophrenic Relapse Project (P. I. ,
Keith Nuechterlein, Ph.D.) of the UCLA
Clinical Research Center for the Study of
Schizophrenia. The patients were all young,
recent-onset cases who were relatively early
in their contact with the mental health sys-
tem. The patients received medication and
individual and group therapy, while the
families were given three family education
ponent analyses may allow program coor-
dinators to direct available resources in
ways which will benefit the most clients.
However, there may be a tendency to enth-
usiastically adopt family education, ab-
stracting it from broader intervention pro-
grams, without understanding that the fam-
ily's experience of education may be very
different when such material is presented as
a self-contained unit. Thus, as the technol-
ogy of family education develops in the con-
text of broader intervention programs, par-
allel research focused on family response to
education in the absence of family therapy
will increase our knowledge of the effects of
family education programs in different the-
rapeutic or support contexts.
sessions near the beginning of the outpa-
tient program within 1-2 months of hospi-
tal discharge. Family members met with
two UCLA Aftercare Clinic staff members,
usually on 3 consecutive weeks. They were
given information about schizophrenia,
medication, and family stress management
and discussed particular problem areas they
were experiencing. The information given
to these families was adapted by Nuechter-
lein and Snyder from the programs of the
earlier Liberman and Snyder (1981) project
and the ongoing Anderson, Hogarty, and
Reiss (1981) program. The information in-
cluded the sections on symptoms, possible
causal factors, and medication from the
Liberman and Snyder outline and those on
management of stress and communication
130 L. J. Cozolino and K. Nuechterlein
from the Anderson et al. outline. The fami-
ly components of these programs were de-
liberately excluded because one focus of the
project was a medication-controlled exa-
mination of familial, behavioral, and psy-
chophysiological predictors of the course of
schizophrenic disorders and their predictive
interrelationships. Thus, the family educa-
tion program was purposely not an intensive
attempt to change levels of expressed emo-
tion (EE) or other familial factors and can
serve as an example of the presentation of
family education without family therapy.
These relatives were generally the same
age and level of education as samples taken
by Hatfield (1981), and none were partic-
ipating in mutual support family groups in
the community. Thus, these people were
generally isolated from others who shared
their difficulties and, as a result of the ge-
ography of Los Angeles, most lived a good
distance from UCLA.
A semistructured consumer satisfaction
interview was conducted at their homes by
the senior author. The interview form is
presented in Appendix A. While the format
was somewhat structured, open-ended que-
stions allowed families a wide range of re-
sponse. Relatives were first asked about the
information they possessed prior to the ses-
sions, where they obtained this informa-
tion, and how they reacted to it. Focusing
on the content of the program, they were
then asked if they remembered specific con-
tent in the areas of diagnosis, cause, the ef-
Preliminary Results
General Reactions
While there were numerous suggestions for
improvement, there was also consistent
praise for the program. Most felt that they
had finally found an ally after terrifying and
frustrating experiences with police, hospital
staff members, and doctors. While the in-
terview was fairly structured, relatives took
over for the most part and seemed to have a
strong need to describe the details of their
family's difficulties. The interview had a
cathartic function for many as they vented
concerns and frustrations. We would often
have to wait between 15 and 30 min before
fects of stress, symptoms, the family role in
rehabilitation, family causes, medication,
and prognosis. If content was remembered
in each area, the usefulness and their reac-
tion (i.e., if they found the information
comforting or disturbing) were discussed.
Ratings of usefulness of information
were done on 8-point Likert scales where 1
was most useful and 8 was not useful at all.
Families were also asked to state what in-
formation they would have liked to have re-
ceived which they were not given. Overall
program ratings were taken as well as rat-
ings of specific management tips they re-
membered being told. Interviews were con-
ducted in the home because it was felt that
home interviews would afford families
more freedom of expression than the hospi-
tal environment. Families were assured of
confidentiality and told that no feedback
would be given to the clinic staff except
after it was combined with data from other
families. The interviewer (L.J.C.) was in-
troduced to the family by a telephone call
from the family's case manager and describ-
ed as an independent researcher who had
the support of the aftercare program but
was in no way part of the treatment pro-
gram. This was done due to our concern
that they would feel that negative feedback
might jeopardize their relative's treatment.
The interview was conducted an average of
8.2 months after the last family education
session with a range from 5 months to 1
year.
beginning the interviews, listening to step-
by-step detailed accounts of their relative's
decompensation, their search for assist-
ance, or, more often, a description of some
current difficulty. Relatives also used us as
a resource, asking for books, telephone
numbers of family organizations, and refer-
rals for therapists.
Five of ten subjects remembered being
given practical tips, such as calling the
emergency number for assistance in a crisis,
reducing contact, and not pressuring the pa-
tient. Of all the information remembered,
these tips were rated as the most useful. Of
the eight content areas, information con-
 Pilot Study of the Impact of a Family Education Program 131
cerning medication was the most consist- I don't think it bothered us, we had already been
ently remembered and thought to be useful. bothered previously. (How?) When you go to the
hospital every day for a month and you see no
Relatives rated the data they were given change in your son - that bothers you ... so we
concerning cause and prognosis especially had already been bothered before. No it (the in-
unhelpful, describing it as too general and formation during the program) didn't bother us
vague. Especially in these areas, they felt at all.
that general information about schizophre-
nia was of little use as they went to the ses- On the other hand, some relatives who
sions expecting information specifically came expecting a rapid recovery period
about their ill relative. were distressed by the information that re-
It was found that the numerical ratings of covery would likely occur over a period of 1
relative usefulness had little meaning for year or 2. This raises an important issue
most of the categories beyond practical tips concerning the role of family education in
and medication because of the wide variab- fostering hope as well as instilling realistic
ility and the different mindsets with which expectations for the recovery process. A
the relatives were answering them (because delicate balance may need to be struck to
of this and in light of the small sample size at keep families actively and positively en-
the time of writing this report, statistical gaged in the treatment process. The impor-
evaluation has been omitted). Some rela- tance of maintaining this balance supports
tives took the stance that all of the informa- the notion of assessing the relative's expect-
tion was useful in contrast to total igno- ations prior to family education.
rance:
Anything is useful, you see, these questions are Did You Find the Information Pitched
really slanted from our context because - useful? at the Right Level for You?
Well God! Tell me anything - anything is useful.
I want to know! We have no backbone of infor- Another point of interest for us was wheth-
mation. er relatives found the information overly
technical, too elementary, or pitched at the
Others who really wanted extensive specific right level. None found it to be too techni-
information concerning their relative tend- cal, while those desiring more specific infor-
ed to rate all but the specific tips and infor- mation rated it as too elementary. While, in
mation on medication as too general to be general, relatives liked having the technical
useful: information about the possible role of neu-
We felt it was too general and a waste of time - it ral transmission abnormalities, "Its com-
got to the point where we didn't want to go back. forting to know that its something electri-
cal," their versions of what they were told
Was There Any Information Which You often caught us off guard:
Found Disturbing? Schizophrenia is like a tear in the brain - too
much chemical in the brain.
One goal of these interviews was to deter- There is a spot here (points to the center of the
mine whether any of the information dis- top of his head), she is missing this particular
turbed the family in any way. We thought thing and the medication can kind of work on that
that hearing the diagnosis of schizophrenia, area ... I can see that the medication has not
finding out about a prolonged recovery pe- made the gap any better.
riod, or hearing of the possible side effects And there are always the diehards even
of the medication might be particularly dis- after family education: "Schizophrenia is a
turbing during the family educational pro- split personality, isn't it?"
gram. In general, relatives reported that
they were so disturbed by their relative's
Did the Program Change the Way You Felt
behavior and the things they heard from po-
About Your Relative's Treatment?
lice and hospital staff prior to the educa-
tional program that they did not find the in- When asking relatives about the positive ef-
formation in the program to be particularly fects of family education, they found it diffi-
disturbing. One father stated: cult to separate the effects of the education
132 L. J. Cozolino and K. Nuechterlein
from the broader treatment program. Con-
cerning the sessions, one mother stated that
" ... they made me feel much more secure
about the treatment. It also made me feel
that there was someone I could call." Thus,
the flesh and blood presence of a concerned
professional seemed at least as important as
the information. The information, while
poorly remembered in many cases, repre-
sented a gesture of the establishing of a life-
line, which was extremely positive. Rela-
tives unanimously expressed their gratitude
and positive feelings toward the good inten-
tions of the staff for offering their time.
None had ever been really given this sort of
attention before. A mother, when asked
whether she found any of the information
comforting, keyed on the word comforting
and stated:
Everything was a feeling of relief because we felt
so isolated - once you get talking with other
people who deal with these problems and that
there was some place to tum for help - that's
what gave me relief - not so much the informa-
tion ... this program was a godsend.
Another mother was comforted by the fact
that other forms of treatment were explain-
ed to her as well as the one offered by the
clinic. She said it made her feel as though
she wasn't "up against an institution."
General Issues Concerning Family Education
Certain issues appeared consistently
throughout the interview process which
have specific ramifications for family edu-
cation both inside and outside the context
of longer-term family treatment.
Information
Let's Get Specific!
It became clear to us during the course of
the interviews that enthusiasm for impart-
ing theoretical information to the relatives
of schizophrenic patients has some poten-
tial drawbacks which should be discussed.
Schizophrenia was in no way a matter of
sheer intellectual curiosity to these families,
they needed help, and they needed it fast.
They came expecting specific information
about their relative and were somewhat
Would You Be Interested in Participating
in Relatives's Groups?
None of these ten relatives were participat-
ing in support groups, and only one expres-
sed even an interest in attending one ("but
only if it was close"). One said she was
"anti-group" and another was quite upset
by the thought, saying:
I don't want groups, I don't go for groups, I never
did go for groups, I don't want it ... I find that I'm
just not ready to go swap stories with other
people. I don't feel that that is going to help me in
any way ... or just to know misery loves company
. .. NoWay!
While previous research has addressed the
needs of members of relatives' groups,
there is probably a still larger population of
relatives who have never participated in
groups of any kind. It is interesting to con-
sider that the needs of these less social, less
interactive (if we can call them that) rela-
tives, may differ in significant ways from
those who will join groups and involve their
broader social network in the process of re-
covery. At the very least, it suggests that the
form of support and treatment that is view-
ed as most desirables for families with schiz-
ophrenic members may differ from one
population to the next.
confused and distanced by information
which was of "general" interest. One mo-
ther stated:
I don't remember the discussion very well at all.
At the time you're so involved with your own par-
ticular person's symptoms ... I don't think a
whole lot of things sink in that is general informa-
tion - at least it didn't for me. I knew what R.'s
symptoms were.
Generally, relatives felt that it was difficult
or impossible to translate general informa-
tion into specifically useful behaviors with-
out the help of the doctor. It seems that at
the beginning, their needs may be more
specific to management issues and, as the
pressures associated with severe symp-
tomatology subside, perhaps a cognitive
understanding of the illness becomes more
important. The need for management ad-
 Pilot Study of the Impact of a Family Education Program
vice relating to their relative seems most im-
portant. This need was sometimes not met
in the UCLA educational program because
the deliverers of the family education com-
ponent were not the patients' therapists and
were not necessarily deeply familiar with
the individual case. Thus, they had only a
limited amount of specific information that
had been provided by the therapist. Famil-
ies expressed a need for the doctors treating
their relative to be present at the sessions.
The illness of a family member does not
occur in isolation of other problems and a
specific history. While specific skills are
needed, they are interpreted and imple-
mented through the interpersonal and his-
torical context of the family. One mother
stated her confusion:
In four sessions one can't fill up a lifetime on what
had happened prior, so they go only by the time
that she had gotten ill. How do you separate that
part of your life from the area she's in now? ...
you can't and they did.
It seems difficult for relatives to separate
this part of their lives from their historical
difficulties and conflicts. Without feedback,
it is impossible to monitor how the informa-
tion is implemented and if it at all resembles
your intended results.
As part of their concern for the family,
parents are now vigilant for symptoms in
their other children and want specific infor-
mation concerning them: "They (the psy-
chiatrists) asked about our son (the pa-
tient's brother) and you start to worry about
him too. If it's hereditary ... what to say
that it won't." Besides specific information
concerning the possibility of illness in their
other children, parents requested informa-
tion concerning their children's interac-
tions, especially competition. There were
also instances of siblings concerned about
their own health: "The psychiatrist said to
my other son that there is a 30% chance that
this will happen to him." (This occurs de-
spite the presentation of accurate figures
for risk of illness).
Thus, there was much support for very
specific information about an array of issues
particular to a patient and the individual
family. Given the choice, these relatives
would have much preferred information
tailored to them, and most could have done
without general explanations.
133
Format
When Is Our Next Meeting?
Besides wanting specific advice, relatives
felt that follow-ups were essential. The in-
formation and management advice appear-
ed self-explanatory until they tried to im-
plement it at home. One father described
his experience:
You think, "That sounds easy ... I can do that"
but when it comes right down to it lots of times
you can't. We have the leaflet they gave us -
don't be sarcastic and all that stuff and that's fine
... it was useful ... but is wasn't very effective.
The things they said were the right things to do if
you could do them.
Parents expressed a need for the program to
be given over a longer period of time for
them to test out the management advice
they are given, to have an opportunity to
see what sort of reactions the patient will
have to medication, and to see how the re-
covery process will go. Specific difficulties
were consistently reported in the areas of
keeping a balance of discipline and freedom
for the patient especially when it involved
keeping the level of emotion low. One
mother stated: "How do I keep it low-
keyed? I'm caught between fights (between
brothers). This is a natural conflict - I'm
fighting off fights." In giving general man-
agement advice without extensive follow-
up, one risks establishing a failure experi-
ence which may add to an already high level
of guilt, frustration, and stress. One might
hypothesize that a management failure with
"expert" advice may add to feelings of pes-
simism and hopelessness far outweighing
those which would occur in a trial-and-error
approach by relatives. Relatives were sup-
portive of a program which would be less
concentrated and spread out over a longer
period of time so that they could get advice
with their specific problems as they arose.
One relative expressed this quite well when
she said:
Parent interviews should be spaced out rather
than all at the start cause you just don't know
what you need to know and need to remember
when it happens at the beginning when you're
upset anyway ... As things progress, they should
give you more as you need to know it.
134 L. J. Cozolino and K. Nuechterlein
Establishing Source Credibility aud
Coordiuatiou of Treatmeut
Who Is in Charge?
The UCLA Aftercare Clinic program, like
the other programs reviewed in this vol-
ume, has a multidisciplinary staff. The re-
sult is that patients and their families come
into contact with a variety of different kinds
of professionals, specifically psychiatrists,
psychologists, social workers, and psychia-
tric nurses. A greater understanding and ac-
ceptance of a medical illness model coupled
with the observable potency of pharmacol-
ogical interventions seemed to lead families
to express a greater need to have interactive
feedback with the physician rather than
with other staff members. This need is also
supported in that most individuals under-
stand the medical context and interventions
from personal experience while compara-
tively few have had psychotherapeutic con-
tacts. One father stated: "Didn't under-
stand what the treatment was ... except for
medication. Everytime he came over here
(this was the one family interviewed at the
clinic) he was filling out some question-
naire, ... doing group therapy, ... I still
don't know what's going on." The notion of
combining psychosocial and psychophar-
macological therapies seemed beyond their
understanding (despite education empha-
sizing this combination), while the magical
presence and power of the physician was
held onto tenaciously. This perspective is
clearly described by one mother:
They should set an appointment with the doctors
to talk to the parents even if it is for five or ten
minutes - it would calm them considerably ...
otherwise you just don't know, you have all sorts
of fantasies about what you think they are doing
and even though there is a psychologist telling
you, you don't trust those people like you do his
actual physician. Ten minutes with the doctor is
worth five hours with the psychologist - because
you feel you are at the horse's mouth.
Instances were also noted when a specific
problem of management was difficult to
label as either psychological or medical.
Relatives were often unsure whether a new
set of symptoms was related to medication
or a function of a "disturbed state of mind."
At these times, family members were at a
loss as to which staff member to contact.
These data seem to support Anderson
et al. 's use of a family ombudsman to co-
ordinate treatment, represent family con-
cerns to various staff members, and direct
them to the appropriate professional for a
particular difficulty.
Where Are They When You Need Them?
All relatives stressed the importance ofhav-
ing someone to call in emergencies, espe-
cially because things seem to go wrong late
at night and on holidays. Support in emer-
gency situations was seen as one of the most
important needs and most appreciated
aspects of the Aftercare Clinic program,
which includes a 24-h emergency phone
number.
Relatives were unanimous concerning
the importance of feedback both to and
from staff during the treatment process.
Most wanted some form of regular informa-
tion concerning process in therapy and pro-
gnostic updates when possible. They also
expressed a need for feedback during the
implementation of management advice
given to them during the family educational
program.
All relatives expressed a special need for
feedback to and from the physician during
periods when the patient is adjusting to new
medication. They reported that the side ef-
fects were quite disturbing, and there was a
need for periodic reassurance as well as
feedback to the doctors concerning the pa-
tient's symptoms.
We were surprised to find that relatives
consistently mentioned that they did not
feel that the patient presented an accurate
picture of their "real" behavior to the doc-
tors and wanted to have input in this way.
One mother was quite adamant that a short-
coming of the program was that:
... more effort isn't made to get from the parents
what their opinions and observations are as to the
patient's behavior. I don't think the patients are
honest with the doctors - they ought to check
with the parents - the patient doesn't give an ob-
jective view of what his behavior is.
This poses difficulty in the areas of parental
overinvolvement as well as the maintenance
of integrity and confidentiality of individual
 Pilot Study of the Impact of a Family Education Program
patient treatment. Patient reaction to pa-
rental involvement in their therapy may
yield important information in this area as
well as a greater understanding of the rela-
tionship between overinvolvement and pa-
tient treatment. Another issue which family
education without subsequent family thera-
py might raise is the creation of positive ex-
pectancies for participation in the ongoing
treatment on the part of the parents. As a
result of their initial contacts, they may feel
more isolated than if they were excluded
from the beginning. The UCLA program
includes informal follow-ups by the individ-
ual therapists. Perhaps initial family educa-
tion necessarily implies formal, scheduled
follow-ups with parents to keep them sup-
portive of the therapeutic process.
Who Can I Believe?
Family members in this sample described
the individuals who presented the educa-
tional materials as "nice social worker
types", "well intentioned", "sweet", and as
"really having their hearts in their work".
Most, however, said that they were expec-
ting an "authority" (i.e., physician) and
were thus somewhat skeptical about the
credibility of the information. Another fac-
tor to consider is that most of the informa-
tion is presented as tentative theory (e.g.,
the dopamine model) rather than fact,
which is quite unlike the assertive confi-
dence with which their doctors have di-
agnosed and treated most of their illnesses.
Thus, it is easy to see how the lack of certain-
ty in the presentation of information con-
cerning issues of diagnosis, etiology, and
prognosis could result in negative attribu-
tions concerning the expertise of the in-
structor rather than appreciation for the
comparative complexity of schizophrenia
and the actual state of knowledge relative to
most commonly treated medical difficul-
ties.
This feedback suggests that an initial ap-
pearance by the physician could be impor-
tant in decentralizing power and describing
the roles and expertise of the various
professionals. The task at the outset may be
to redirect a lifetime of thinking of illness
from a medical perspective to an under-
standing of the interaction between psycho-
135
logical and biological factors and the quali-
tative differences between an illness such as
schizophrenia and most well-defined me-
dical conditions.
What Can I Believe?
Most information in educational programs
is presented as theory and expressed in ten-
tative probabilities rather than delivered
with certainty. The notable exceptions to
this rule are the positive aspects of drugs
and the lack of evidence supporting a family
cause for schizophrenia. While this infor-
mation appeared initially comforting to rel-
atives, later some began to speculate: "If
they are so unsure about all the other
things, how can they be sure about this?"
Another trend among mothers was initial
relief upon hearing that they were not to
blame, followed by a return of guilt feelings
during the stress of the slow recovery peri-
od. This growing doubt also took the form
of fear concerning the long-term side effects
of the medications, which seemed to be
fueled by the constant media barrage about
the delayed effects of all kinds of drugs,
pesticides, and chemicals. This data may
point to a need for an evaluation of the
amount of certainty with which the infor-
mation is presented as well as the need for
"booster shot" classes and follow-ups.
Assessing the Recipients of
Family Education
Research has demonstrated differences in
learning as a function of anxiety level, pro-
longed stress, prior learning, and expecta-
tions. Recipients of family education are af-
fected by all of these factors. They are re-
ceiving information about an emotionally
charged issue which has caused them a great
deal of personal distress; they are also
asked to make specific changes in their
behavior. Thus, benefiting from family ed-
ucation not only involves one's cognitive
skills and affective state but also personal
history and interpersonal and communica-
tion skills.
Most relatives enter family education
with a great deal of misinformation, with a
high level of stress, and uneasy as to what
will happen next. They also have a wide
136 L. J. Cozolino and K. Nuechterlein
variety of expectations concerning the ses-
sions and what will take place. The great
variability in this group of relatives led us to
speculate that there may be information
about some specific aspect of family func-
tioning which could suggest variations that
could optimize information presentation
and format, if assessed prior to the pro-
gram. Five areas which stand out from the
interview materials are coping style, previ-
ous information, level of anxiety, attitudes
toward treatment, and need for a balance of
hope and realism.
Coping Style
In our experience with this small group of
relatives, we found that different coping
styles seem to be operating with respect to
information and involvement in the recov-
ery process. A coper/avoider phenomenon
(Goldstein, 1973) seemed to be present in
many family members that is similar to the
integrator/sealing over distinction noted
among recovering schizophrenic patients
(McGlashan, Levy, and Carpenter, 1975).
For some, the need to know and be involved
was very strong, which led them to ask nu-
merous questions during family education
and continue the process by reading and tal-
king to others. They seemed to gain some
sense of control or mastery over the process
by wanting to know as much as possible.
This feeling was expressed by four of ten of
the relatives in statements such as:
Ignorance is frightening - the more you know,
the more you feel secure.
I'd like to know, that's what I'd like to do. I'd
really like to know, I'd like to know what the pos-
sibilities are, what we can do, anything that we
can help with, whatever, I want my daughter to
be a well girl.
These individuals were the most impressed
with information concerning biochemical
theory and the action of the medication.
The more precise or scientific the better.
These were often the individuals who also
wanted the most specific information con-
cerning their involvement in the treatment
process: "I wish they could have said to me
you're doing this wrong, you're doing that
wrong and this is what you should do."
Others seemed nearly indifferent to the
information and somewhat annoyed that
we are causing them to recall their involve-
ment in the program. These individuals see-
med to want to carry on life as usual and
were most pleased with the program as an
auxilary caretaker: "It was good just to have
someone to tum your troubles over to."
These individuals also seemed to remember
least about the program, even though they
were interviewed within a shorter amount
of time since receiving family education
than other relatives who remembered much
more. If this is a reliable phenomenon, it
may offer some clues about how to tailor
family education and treatment to address
this issue. A full complement of references
and detailed information may be made
available for copers while avoiders may
need clear, step-by-step instructions for
emergencies that they can take home and
use when needed.
Previous Information
Everyone had their horror story concerning
what they had been told by doctors about
their relative's illness before becoming part
of the Aftercare Clinic program. For ex-
ample:
Doctor: Your daughter is a psychotic schizo-
phrenic.
Mother: Does that mean my daughter is insane?
Doctor: Yes.
Father: At that point she (the mother) became a
basketcase.
After 5 days in the hospital, all of A ... 's doctors
got together and came to the conclusion that he
was schizophrenic. They felt that he was an in-
curable schizophrenic ... then I got mad.
Beside these experiences specific to their
relative, some other connotations of the
word "schizophrenic" were not shared. In
interviewing one husband and wife, we
asked if the fact that the recovery process
would be a couple of years was disturbing to
them. The husband said "yes" because he
had heard from someone that it would be a
matter of months; the wife, who had taken
a psychology course 20 or so years ago, sta-
ted that that made her quite happy because
she had remembered it to be a totally in-
curable illness. This example underscores
the discrepancies between family members'
prior knowledge and the importance of
establishing an information base line prior
 Pilot Study of the Impact of a Family Education Program 137
to the family educational process. This may helpful in understanding and supporting the
facilitate the goals of reeducating and pro- patient through the treatment process.
viding comfort to these families.
Balancing Hope and Realism
Level of Stress One of the expressed purposes of family ed-
ucation is to reduce pressure on the patient
It is clear that families are under a great deal by revamping relatives' unrealistic expect-
of stress during or immediately after the ations for a quick recovery. Relatives are
time of their relative's hospitalization, often anxious for the patient to get back to
which is when most family educational pro- work or school and to just be "normal"
grams (in the context of long-term treat- again. In most programs, the recovery peri-
ment) are given. While this has been re- od is described as at least a 1- to 2-year pro-
peated numerous times by many profes- cess with possible severe setbacks without
sionals, it was especially meaningful to hear proper adherence to the medication sched-
it expressed by relatives. When we asked ule. While this is important, there seems to
one mother about the most important be an element of danger in being too realis-
points in the program, she described a need tic with relatives right from the beginning.
to state the information slowly, clearly, and One mother expressed her distress by stat-
repeatedly. She dramatized the process of ing:
giving the relative the clinic number by tak- I would have liked more positive than negative
ing a card and pressing it firmly against her hope. The whole picture they presented was so
palm, saying: "Make sure people have the bleak - they didn't give me much hope. I think
clinic number - it's like talking to a child that when you're talking to a parent at that time,
because parents are very distraught." to sit and tell the parent that it's at least 3 years.
If he stops the medication, he may regress a
Perhaps, at least initially, our goals in
whole year and never come back. If I wasn't so
transmitting information should be quite positive I would have said, that's it - I have an
modest and subordinate to the needs of rel- idiot son for the rest of my life.
atives to build a bond of trust with the staff.
Assessing the amount of trauma or shock a While all relatives interviewed stressed the
relative is experiencing may give important need to be told the truth concerning the ill-
direction about how to proceed in the be- ness, most stressed the importance of en-
ginning stages of the program. This might couragement during the initial phases of the
take on the role of venting, dealing with treatment process. While they found the
personal issues of the relatives, or just shar- existence of the program reassuring, they
ing coffee. During the interviewing expe- found the realities of the extended recovery
rience, we found that relatives had a very period unsettling. Families enter education
high amount of emotion still locked into with a spectrum of expectations concerning
these issues 5 months to 1 year later. schizophrenia drawing parallels to every-
thing from a fever delirium to an incurable
Attitudes Toward Treatment brain disease. Assessment ofthese assump-
tions, expectations, and their level of
It may prove helpful to assess initially the distress prior to education may give facilita-
relative's attitudes toward the treatment tors more insight into how to balance hope
process and their general perspective to- with realistic expectations for a particular
ward medication, hospitals, and therapy to family. Some balance which would relieve
see if any remedial education or public rela- pressure from the patient but would also
tions work needs to be done. Some family encourage the relatives would be optimal.
members called themselves "anti-psychia-
try" while others seemed to have an almost
religious commitment to medication and
the medical model. Few had experience
with verbal psychotherapies, and some edu-
cation in this area may have been quite
138 L. J. Cozolino and K. Nuechterlein
Summary
Preliminary evaluation of these initial fami- into specific behaviors at home. There
ly interviews suggests specific and nonspe- appears to be a possibility of discourage-
cific effects of family education and raises ment and an increase in feelings of pessi-
many possibilities for investigating the rela- mism and guilt when advice is unsuccessful-
tive effectiveness of various approaches. ly implemented. Parents reported concern
This feedback from relatives present sug- about the possibility of their other children
gestions for education in a broader treat- becoming ill and requested more informa-
ment context, as well as pointing to possible tion in this area. Further research may sug-
drawbacks in presenting family education gest that especially in education programs
without adequate follow-up. without familiy therapy, equipping rela-
For the most part, relatives interviewed tives with survival skills should have priori-
thus far have found most of the theoretical ty over and precede the presentation of
information interesting but too general to general information.
be of much practical use. The exception to The consensus regarding format was that
this was information concerning the med- the need for follow-up and feedback is es-
ication and its side effects. Relatives seem- sential especially in the implementation of
ed to be most appreciative of practical management advice. Relatives suggested
management advice at least during the giving the information slowly, repeatedly,
initial phase of the patient's reintegration and simply, especially at first. The more
into the home and community. The sessions spread out the sessions, the better the op-
seemed to have the overall positive effect of portunity to implement and monitor the ef-
creating the feeling of a lifeline between fects of specific management tips and then
families and the treatment program. Rela- receive feedback. They also reported
tives reported a greater support for keeping changes in symptomatology over the course
their relative on medication as a result of of treatment and the need for updated ad-
the sessions. They were not disturbed by vice as new difficulties arose. Thus, given
the information except where the news of a limited resources, relatives seemed to sug-
1- to 2-year recovery period was introduced gest a slow pace with long intervals, as op-
to relatives who expected a rapid recovery. posed to a concentrated program when the
They found the information pitched at the patient is first released. From this perspec-
proper level and were particularly impres- tive, family education in the absence of on-
sed by biochemical explanations of neural gong support would not be of adequate as-
transmission and the effects of medication. sistance. A free-standing educational pro-
Only one of ten relatives thus far have been gram lacks the all important ability to give
willing to participate in family support ongoing feedback concerning practical ad-
groups. This points to the existence of a vice and lacks a way to monitor how the in-
large population of relatives of the mentally formation is interpreted and utilized. Be-
ill who are not participating in relatives' yond this, there may be the risk of creating
groups and thus may not be represented in positive expectancies of long-term assist-
research of family needs (Hatfield, 1979, ance with initial family education, which
1981). Research with these populations may lead families to be more discouraged
may suggest some important differences in with the passage of time. This is speculation
the uses of social support and the willing- on our parts and needs to be assessed.
ness to participate in group treatment. Many changes in the conceptualization
In the area of information, relatives stres- and treatment of schizophrenia have result-
sed the priority of specific management ad- ed in confusion for relatives. Information
vice over theoretical information. At least they have from the past and that which they
initially, survival skills superceded curiosity are now receiving is often vastly contradict-
in their minds. The more concrete and pa- ory. Add to this the mUltidisciplinary staf-
tient specific the information, the more use- fing of therapy programs, the strength of
ful it was found to be. Relatives reported the medical illness model in their minds,
great difficulty in translating general advice and the potency of the pharmacological in-
 Pilot Study of the Impact of a Family Education Program
terventions and you have relatives straining
to get to speak to the physician for some
"truth". This is accentuated by the fact that
most of the information is presented as
theory instead of fact. Thus, establishing
source credibility and coordination of treat-
ment stands out as a major obstacle to the
family's understanding and integration into
the treatment process. For families to gain
the full benefit available from individuals
with various expertise, it may be that edu-
cation about the treatment program and the
decentralization of power should be one of
the first orders of business. In line with this,
the question of who should present infor-
mation becomes an important area of in-
vestigation.
In interacting with these families, the
wide variability of anxiety levels, expecta-
tions, and prior learning seemed to be the
most salient factors in determining their re-
action to family education. The possibility
of assessing the recipients of family educa-
tion to maximize utility and learning could
be a possible area of investigation as pro-
gram technology develops. Differences in
coping style, previous information, level of
stress, and attitudes toward treatment, if as-
sessed, may allow a tailoring of information
to best fit the needs of a particular relative
or family. Another important variable
seems to be a proper balance of hope and
realism as relatives brace themselves for the
recovery process. This may be the most im-
portant role of family education and the
best reason to properly assess the relative's
expectations and attributions prior to giving
them information. Broader interventions
which have contact with relatives prior to
family education are able to assess the psy-
chological state, intellectual capacity, and
other important factors which allow tailor-
ing of presentations to match the recipients.
Group programs do not allow this flexibi-
lity, but broader intervention programs
have the opportunity to clear up misunder-
standings and deal directly with family re-
actions. Group programs which stand on
their own have no means of assessment and
no opportunity to make adjustments
without follow-ups.
The notion of choosing the content and
format based on the strengths and weak-
139
nesses of students is done in most educa-
tional settings. What we are asking of rela-
tives is far more than remembering infor-
mation in a usual classroom sense. We are
requesting specific changes in lifetime
behavior patterns in the context of an ex-
tremely stressful experience with a handful
of tentatively held theories. In light of this,
it seems that tailoring the information and
manner of presentation to account for what
relatives are capable of understanding and
effectively using at the time may be a fruit-
ful area of future research.
While many factors have led to the de-
velopment of family education and therapy
in the treatment of schizophrenia, the nega-
tive effects of emotional distress and ten-
sion on patient recovery have been among
the most important. As researchers focus-
ing on schizophrenia, we measure program
success as a function of lowered relapse
rates or symptomatic relief of the patient.
The well-being of the family is usually ap-
proached only from the perspective of its
positive or negative effects on the identified
patient. It is clear that for most families, a
psychotic break in one of their members is a
major life stressor. The amount of stress
and disruption may vary from the aunt who
lives in the next town to the husband who is
the primary supporter of the family. Be-
sides the great interpersonal strain, a rela-
tive's breakdown may result in major
changes, such as the modification and/or
dismissal of long-term plans and goals, ad-
ded economic burdens or total loss of in-
come, marital pressures, separation or
divorce, decreasing outside contact and
possible isolation, and a consistently stress-
ful home environment. Thus, the families
as well as the patients seem to be "at risk"
for emotional difficulties. As educational
and therapeutic interventions expand to in-
clude the family, it may be that the meas-
urement of program utility and success
should also expand to include the coping
and well-being of other family members.
Perhaps by focusing only on the patient we
are missing the other half of the value and
overall cost effectiveness of family inter-
ventions as a primary prevention strategy
for physical and psychological illness for
relatives.
140 L. J. Cozolino and K. Nuechterlein
References
Anderson, C. M., Hogarty, G., and Reiss, D. J.
(1981). The psychoeducational family treat-
ment of schizophrenia. In M. J. Goldstein
(Ed.), New developments in interventions with
families of schizophrenics. San Francisco: Jos-
sey-Bass.
Goldstein, M. J. (1973). Individual differences in
responses to stress. American Journal of Com-
munityPsychology,1 (2),113-117.
Goldstein, M. J. (1981). New developments in
interventions with families of schizophrenics.
San Francisco: Josey-Bass.
Hatfield, A. B. (1979). The family as partner in
the treatment of mental illness. Hospital and
Community Psychiatry, 30, 338-340.
Hatfield, A. B. (1981). Coping effectiveness in
families of the mentally ill: An exploratory
study. Journal of Psychiatric Treatment and
Evaluation, 3,11-19.
McGlashan, T. T., Levy, S. T., and Carpenter,
W. T. (1975). Integration and sealing over.
Archives of General Psychiatry, 32,
1269-1272.
Snyder, K., and Liberman, R. (1981). Family
assessment and intervention with schizophre-
nics at risk for relapse. In M. J. Goldstein
(Ed.), New developments in interventions with
families of schizophrenics, San Francisco: Jos-
sey-Bass, p. 49-60.
Appendix A
Family Interview
Do you recall the session(s) at UCLA (or at your
home) in which you were given information
about the causes and treatment of your relative's
mental illness?
I would like to ask you some questions con-
cerning your experiences during these sessions.
This is not a test in any way! It is a consumer sur-
vey designed to find out whether or not these ses-
sions served your needs. All of this information is
confidential and will not be shared with clinic
staff except after it is combined with responses
from others who have participated in the pro-
gram.
Previous Knowledge
1. Before we discuss the sessions, I would like
to get an idea of how much you had learned
about your relative's illness before you par-
ticipated in this program. How would you
rate the amount of information you had
prior to these sessions?
A Great Deal
1. .... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
2. What prior information did you find useful?
3. What prior information did you find not so
useful?
4. Where did you obtain this information?
Previous doctors _ _
Books _ _
Friends _ _
Others __________________________
5. Was any of the information you received be-
fore you participated in the session(s) dis-
turbing to you?
6. (If yes), what information did you find dis-
turbing?
7. Why do you feel that it was disturbing to
you?
8. Was any of the information you received be-
fore you participated in the session(s) com-
forting or reassuring to you?
9. (If yes), what information did you find com-
forting or' reassuring?
10. Why do you feel that it was comforting or
reassuring to you?
UCLA Sessions/General Impressions
11. Generally, what do you recall being told
about the nature of your relative's disorder
during the family education sessions?
12. Overall, did you find this information useful?
Yes No
1..... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
13. What information stands out in your mind as
being particularly useful in some fashion?
14. What kinds of information stand out in your
mind as not so useful?
15. Overall, did the information you were given
during the sessions seem to agree with your
experience with your relative?
Completely In large part About half
Somewhat Not at all
Nothing
16. In what ways did it seem to agree?
17. In what ways did it disagree?
18. Overall, did the information tend to agree
with the information you knew or believed
prior to the sessions?
Completely In large part About half
Somewhat Not at all
19. What stands out in your mind in this regard?
20. Were any of the things you were told disturb-
ing to you?
21. (If yes), what were they and how were they
disturbing?
22. Overall, do you feel the information was pit-
ched at the right level for you?
Too elementary Right Too technical
1. .... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
23. What stands out in your mind as being too
elementary?
24. What do you remember as being too tech-
142 Appendix A
nical?
25. What things do you remember that seemed
just right?
26. As a result of the program do you feel that
you have a greater or lesser understanding of
your relative's illness?
Greater understanding Lesser understanding
1. .... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
27. Overall, what weren't you told about that
you would have liked to have known?
Diagnosis
28. Many sessions include a discussion of diagno-
sis, you know, words like schizophrenia,
schizophreniform, psychotic reactions, or
other mental disorders. Do you recall di-
agnosis being discussed during your ses-
sion(s)?
29. What stands out in your mind in this regard?
30. Do you find that this information is useful to
you?
Useful Not useful
1 ..... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
31. In what ways do you find this information
useful?
32. Are there ways in which you feel it is not use-
ful?
33. Was there anything in the discussion of di-
agnosis which you found comforting or re-
assuring?
34. (If yes), what did you find comforting or re-
assuring and why?
35. Was there anything in the discussion of di-
agnosis which you found disturbing?
36. (If yes), what did you find disturbing and
why?
37. Was there anything about diagnosis that you
would have liked to have been told that
wasn't included in the sessions?
Etiology
38. Often, the cause of mental illness are dis-
cussed in these sessions, do you recall this
being discussed?
39. What stands out in your mind in this regard?
40. Did you find this information useful?
Useful Not useful
1 ..... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
41. What information did you find most useful?
42. What information did you find least useful?
43. Did you find any of the information concern-
ing the causes of mental illness disturbing?
44. (If yes) , what did you find disturbing and why
do you feel it disturbed you?
45. Did you find any of the information concern-
ing the causes of mental illness comforting or
reassuring?
46. (If yes), what did you find comforting and
reassuring and why?
47. What weren't you told about the causes of
mental illness that you would have liked to
have been told?
Effects of Stress
48. Do you recall being told anything about the
effects that stress may have on mental ill-
ness?
49. What stands out in your mind in this regard?
50. How useful have you found this information
to be?
Useful Not useful
1.. ... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
51. What information did you find most useful?
52. What information did you find to be not so
useful?
53. Did you find any of the information comfort-
ing or reassuring?
54. (If yes), how was it comforting and reassur-
ing and why do you think that it had this ef-
fect?
55. Did you find any of this information disturb-
ing?
56. (If yes), how was it disturbing and why do
you think that it was disturbing to you?
57. What weren't you told about the effects of
stress that you would have liked to have
known?
Symptoms
58. Do you recall during the sessions discussing
the various symptoms your relative may be
experiencing?
59. What stands out in your mind in this regard?
60. Did you find the information useful?
Useful Not useful
1. .... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
61. What did you find to be most useful?
62. What information was least useful?
63. Was any of the information concerning
symptoms disturbing to you?
64. (If yes), what did you find disturbing and
why?
65. Was any of the information concerning
symptoms comforting or reassuring to you?
66. (If yes), what did you find comforting and
reassuring and why?
67. Are there things which weren't discussed
that you would have liked to know more
about?

Family Role in Rehabilitation
68. Many sessions include advice to relatives as
to how they can best help the patient during
the recovery process. Did your sessions con-
tain any information in this area?
69. What stands out in your mind in this area?
70. Have you found this information to be use-
ful?
Useful
1. .... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
71. What do you recall as being most useful?
72. What information seems to be least useful?
73. Was any of this information comforting or
reassuring to you?
74. (If yes), what did you find comforting or re-
assuring and why do you feel that it had this
effect on you?
75. Was any of the information disturbing to
you?
76. (If yes), what was disturbing and why do you
feel that it was disturbing to you?
77. Are there issues concerning the family's role
in rehabilitation that you would have liked to
discuss that were overlooked during the ses-
sions?
Family Etiology
78. Many times when a family member is stric-
ken with a psychological illness, close relati-
ves may feel that they are in some way to
blame or that there may have been some-
thing that they could have done to prevent
the illness. Did you experience any of these
feelings when your relative became ill?
79. What kinds of feelings did you have in this
regard?
80. Was this discussed in the sessions at all?
81. Was anything said that was disturbing to
you?
82. (If yes), what was disturbing to you and why?
83. Was anything said that was reassuring or
comforting to you?
84. (If yes), what was comforting and why do feel
it was comforting to you?
85. What do you remember being said about
these issues?
86. As a result of your attending the sessions,
have your thoughts and/or feelings changed
in regard to this issue?
Medication
87. The role of medication in the treatment pro-
cess is often a topic of discussion. Do you re-
call medication being discussed?
88. What sorts of information stand out in your
mind in this regard?
Appendix A 143
89. Did you find this information useful?
Useful Not useful
1.. ... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
90. What do you recall as being useful?
91. What do you recall as being not so useful?
92. Do you recall being told about the possible
side effects of the medication?
93. Did knowing about the possible side effects
Not useful
help you keep your relative on medication?
94. Was there anything discussed concerning
medication which was disturbing to you?
95. (If yes), what did you find disturbing and why
do you feel that it disturbed you?
96. Was there anything discussed concerning
medication which was comforting or reas-
suring to you?
97. (If yes), what did you find comforting or re-
assuring and why do you feel that it was reas-
suring to you?
98. What weren't you told about medication that
you would have liked to have known?
Prognosis/Recovery Period
99. One session may have included advice con-
cerning your relative's convalescence or
recovery period. What do you remember in
this regard?
100. Was this information useful?
Useful Not useful
1 ..... 2 ..... 3 ..... 4 ..... 5 ..... 6 ..... 7 ..... 8
101. Was this information comforting or reas-
suring in any way?
102. (If yes), what did you find comforting and
reassuring and why?
103. How did you react to this information, was
it disturbing to you in any way?
104. (If yes), what did you find disturbing and
why?
105. Was there any discussion concerning your
relative's future?
106. What do you recall being discussed?
107. How did you react to this information, was
it disturbing to you in any way?
108. (If yes), what did you find disturbing and
why?
109. How useful did you find information con-
cerning your relative's future?
Useful Not useful
1.. ... 2 ..... 3 .... .4 ..... 5 ..... 6 ..... 7 ..... 8
110. Were there areas of discussion concerning
the future that you would have liked to
cover that were not discussed?
Practical Tips (Survival Skills)
111. Often the sessions include practical tips in
144 Appendix A
day-to-day living with the patient. Do you 117. In what way?
remember any of these tips? 118. Did participation in the educational pro-
112. Which ones stand out in your mind? gram affect the way you feel about your rel-
113. Please rate each tip after description ative's treatment?
Useful Not useful 119. In what way?
1. .... 2 ..... 3.... .4 ..... 5 ..... 6..... 7 ..... 8

114. What things have you learned about how to
deal with your relative's illness which might
be useful to other families?
115. Were there practical sorts of advice that you
weren't given that you would have liked to
have gotten?
Other Participants
120. Were there members from other families
with you when you went through the pro-
gram?
121. How many others were there?

Relating to Doctors, Clinic Staff,

Other Professionals, etc.
116. Did participation in the educational pro-
gram affect the way you feel about the after-
care clinic program or staff?
12. Psycho educational Family Therapy
C. M. Anderson
Introduction
In recent years, several models of family
treatment have been developed which at-
tempt to include an educational component
as part of a larger program of intervention
for schizophrenic patients and their famil-
ies. These models can be generally classi-
fied as psychoeducational family therapies.
They aim to keep patients out of hospitals
and return them to productive life in the
community by teaching patients and famil-
ies about the illness and its management,
treating the family as a resource rather than
as a stress, focusing on concrete problem
solving and specific helping behaviors for
coping with stress, and working in tandem
with medication programs. These models
emphasize the likelihood of a biological
basis of the illness and the negative impact
of stress on patient ability to function. The
underlying assumption of the interventions
used in psychoeducational models is that
giving families information about the na-
ture of the illness, along with specific sug-
gestions for coping with it effectively, can
decrease tIie intensity and conflict inherent
in family life, and thus decrease the likeli-
hood that patients will relapse.
There seem to be a number of reasons
why these models are attracting so much
attention today. First, major changes have
occurred in patterns of mental health care
for severely disturbed patients. Policies
supporting deinstitutionalization have cau-
sed minimally functioning patients to be re-
turned to the community after very brief
hospitalizations. Consequently, families
have become long-term primary care agents
for very disturbed patients. Many of these
patients do not cooperate with plans for af-
tercare treatment, failing to attend at all or
dropping out after only a few visits. It has
become important, therefore, to develop
treatment programs which would help fam-
ilies to perform this caretaking role, pro-
grams which are not dependent solely upon
the motivation of patients.
Second, recent British studies of expres-
sed emotion (EE) have confirmed long held
clinical impressions that families have an
impact on the course of the patients' illness.
These studies demonstrate that measures of
high EE in family members is correlated
with high relapse rates for patients. (Brown
and Birley, 1968; Brown, Birley, and Wing
1972; Vaughn and Leff, 1976; Hooley,
Chap. 2). These findings have stimulated
increased interest in attempting to develop
ways of modifying family systems to de-
crease family intensity and thus the likeli-
hood of patient relapse.
Third, the first controlled study which in-
cluded an attempt to provide patients and
families with some understanding of the ill-
ness and ways to cope with had dramatic re-
sults. Michael Goldstein and his colleagues
from UCLA (Goldstein, Rodnick, Evans,
May, and Steinberg, 1978) designed a 6-
week program of family therapy in which,
following hospitalization, patients were
randomly assigned within a 2x2 design:
high- or low-dose drug therapy, and family
therapy or no therapy. The Goldstein
model of family intervention was brief, con-
crete, and problem focused. The goals of
the program were to identify the events that
were stressful to the patient and then to pre-
vent the occurrence of these events or to
mitigate their destructive impact. Family
sessions began by exploring the psychotic
experience, with the therapist helping the
146 C. M. Anderson
patient and the family to discuss the illness family therapy and high doses of medica-
and its symptoms. Out of these discussions, tion. In other words, this intervention pro-
the therapist helped the family and patient gram achieved its primary goal of helping
to identify and agree on two to three specif- patients to maintain themselves in the com-
ic stresses that were of particular current munity during the high-risk period immedi-
concern. Conflicts with significant others or ately following hospitalization through the
other stresses, that are viewed as potential- provision of only six sessions of family ther-
ly precipitating a psychosis were emphasiz- apy. Although the differences between the
ed. Although symptoms themselves were groups were less dramatic at long-term fol-
often labeled as stressors, attempts were low-up points (Goldstein and Kopeikin,
made to focus on the interpersonal conse- 1981), the impact of this program was dra-
quences of symptoms rather than symptoms matic enough to inspire increased optimism
perse. regarding attempts to intervene with fami-
Following the identification of stressors, lies of schizophrenic patients.
the therapist helped families to develop This paper will describe the psychoedu-
strategies to avoid these stresses or methods cational model developed in Pittsburg as a
of coping with them when they occurred. part of a larger research project, which is
With the therapists guidance, the need for both attempting to replicate the British
both the family and the patient to accomo- work on EE and attempting to investigate
date to the other was emphasized. Once the impact of various strategies of interven-
coping strategies were developed, the ther- tion in the aftercare of patients with schizo-
apist helped the patient and the family to phrenia.! The larger research project
implement them. When necessary, coping screens patients to insure they meet Re-
skills were developed using direct teaching, search Diagnostic Criteria for schiz-
coaching, and practice. When problems ophrenia, rates the level of EE of their
arose in the implementation of these new family members, and randomly assigns
coping strategies, obstacles to the imple- patients with high EE families to one of four
mentation process were analyzed and the treatment cells: family therapy, social skills
strategies were modified. Finally, thera- training, family therapy and social skills
pists helped families and patients to antici- training, or medication alone. Patients and
pate and plan how they would handle future their families are followed in these treat-
stress. ment modalities for 2 years with various
The results of this short-term program measures of individual and family function-
were extremely positive. After 6 weeks of ing taken at periodic intervals. This paper
treatment, the low medication-no family will briefly note the assumptions and goals
therapy group had a relapse rate of 24% of the family program, describe its compo-
while the high medication-family therapy nents and stages, present some preliminary
group had no relapses at all. After 6 findings about its effectiveness alone and in
months, the relapse rate increased to 48% combination with social skills training, and
for the low dose-no family therapy group, discuss some of the common elements of
but remained at 0% or those receiving this and four similar programs.

Assumptions
It is assumed that whatever the "cause" of
schizophrenia, patients with schizophrenia
appear to have a "core psychological defi-
cit," which appears to increase their vul-
nerability to internal and external stimuli
(Broen and Storms, 1966; Lang and Buss,
1965; Payne, Mattussek, and George, 1959;
Rabin, George, Doneson, and Jentons,
1979; Shakow, 1962; Silverman, 1972;
Tecce and Cole, 1976; Venables, 1964,
1978). This patient vulnerability is probably
exacerbated by the range of behaviors and
emotions that their families are likely to dis-
1 This research project, under the direction of
Gerard Hogarty, is partially funded by Grant No.
MH 30750 from the National Institute of Mental
Health.
 Psychoeducational Family Therapy 147
play. For instance, communication in fami- ture, and specific coping mechanisms of use
lies of schizophrenic patients often includes in dealing with a psychotic family member.
communicative behaviors that are vague, A highly structured and directive approach
unclear, amorphous, tangential, or lacking was used to increase the predictability and
in acknowledgment (Goldstein and Rod- stability of the family environment. An ed-
nick, 1975; Jacob, 1975; Jones, 1977; Jones ucational component was included to in-
et aI., 1977; Singer and Wynne, 1965,1966; crease family knowledge about the illness
Wynne, 1961). Whether or not these com- and confidence in coping with it, thus de-
munication patterns predate the illness or creasing family anxiety about the patient
contribute to its etiology, it would seem log- and their ability to react helpfully. The pro-
ical that patients who have problems con- gram has four basic overlapping phases,
trolling and processing stimuli would have separated here for the sake of clarity. Since
difficulties coping with these complicated the entire program is discussed in more de-
and confusing family communications. tail elsewhere (Anderson, Hogarty, and
Furthermore, when patients have be- Reiss, 1980), only the major points will be
come acutely psychotic, families usually re- stressed.
port feelings of anxiety, guilt, anger, and
sadness (Hatfield, 1978; Kreisman and Joy,
Phase I: Connecting with the Family
1974). These emotions are likely to increase
the intensity of family life, and, since the ill- Based on the assumption that no interven-
ness is a chronic one, this intensity is likely tion can succeed unless the family can hear
to increase over time as family members are or use it constructively, the first phase of
unable to find ways to help the patient. It treatment, which emphasizes the establish-
might be hypothesized that family members ment of an allliance with the family, begins
in such chronic crises would come to re- immediately after the patient's admission to
spond to patients in one of the two ways the hospital. Since all families begin the
Brown describes as components of high program during a serious crisis, and most
EE: becoming overinvolved, attempting to have had multiple unsuccessful contacts
constantly monitor and protect patients with other hospitals and professionals, spe-
from themselves or the environment, or be- cial attention is given to the creation of an
coming frustrated, angry, rejecting, and atmosphere which increases the family's re-
withdrawn from patients and treatment sys- ceptivity to treatment intervention. Phase I
tems. Either of these emotional responses interventions first involve joining the family
would appear to both decrease family abil- by eliciting their reactions to the patient's
ity to cope with patient behaviors and to be illness and to past attempts to cope with it,
problematic to a patient vulnerable to in- as well as eliciting their perceptions of their
tense stimuli. In summary then, these two own current needs and problems. Hope-
forces (the patients vulnerability and the fully, these discussions communicate to
turmoil of families) probably interact to the families that therapists care about what
patient's disadvantage in a spiraling man- they have been through, are not critical
ner; the patient's vulnerability to stimuli about how families have attempted to cope
causing symptoms which upset family mem- with patients, and genuinely want to know
bers, who in turn upset the patient, and so family ideas and views of what is helpful.
on. Once the family has begun to form a rela-
Because of this hypothesized relationship tionship with the therapist, the therapist is
between patient vulnerability and family established as the family's ombudsman or
anxiety or behaviors, this program of family representative in relationship to the hospi-
intervention was designed to accomplish tal system. Since the staff of inpatient facili-
two goals: (1) to decrease the patient's vul- ties are primarily involved with the patient
nerability to stimuli through a program of on a daily basis, it is easy for them to neglect
maintenance chemotherapy and (2) to de- families or fail to see family needs. The
crease the intensity of the family environ- creation of a family representative serves to
ment through a program of providing the balance this skewed perspective and pre-
family with support, information, struc- vent the alienation of families from the tre-
148 C. M. Anderson
atment team. Thus, the family ombudsman
keeps the family informed of ward decisions
about the patient, ensures the input offam-
ily concerns and needs into treatment plan-
ning, and provides the family with structure
and concrete help in coping with the illness
and the hospitalization. In this way, the the-
rapist also begins to mobilize the family's
concern and involvement into constructive
attempts to help themselves and the pa-
tient. By the end of the hospitalization
(which usually is less than 3-4 weeks), the
family, patient, and therapist arrive at a tre-
atment contract which roughly specifies the
goals, content, length, rules, and methods
of the aftercare family program. This pro-
gram, then, continues for 1-2 years after
patient discharge from the hospital.
Phase II: Survival Skills Workshop
Based on the assumption that people are
more anxious about what they do not un-
derstand, the survival skills workshop seeks
to provide the family with as much informa-
tion as possible about the nature of schizo-
phrenia. The education workshop was de-
signed as a multiple family enterprise to
simultaneously promote deisolation of the
family and desensitization about the subject
of mental illness. It is a daylong event at-
tended by all the members of four or five
families who are new to the program. (The
patient does not attend.) Every attempt is
made to encourage an informal atmosphere
in which families can question professionals
and interact with one another. The work-
shop is held as early in the treatment pro-
cess as possible because it also serves to
establish the basic themes of the entire fam-
ily program. The workshop focuses on the
following categories of information.
Information About the Illness
The most recent data about the phenome-
nology, onset, treatment, course, and out-
come of schizophrenic disorders is present-
ed in clear, understandable language.
Theories of etiology, ranging from genetic
and biochemical to family and cultural, are
explained. What is known about the pro-
gnosis of the illness is also outlined, as are
various methods of treatment: psycho-
therapy, pharmacology, megavitamins,
homeodialysis. Every attempt is made to
discriminate between the facts, theory, and
opinion about each ofthese issues. Because
medication compliance is viewed .as a cru-
cial component of the program, the impor-
tance of antipsychotic medication is given
special attention. Mechanisms of action,
possible negative side effects, and the use of
antiparkinsonian agents are explained and
statistics about the risk of relapse on and off
medication are shared. In particular, the
critical importance of family support for
and feedback about the medication pro-
gram is stressed.
Information About Management
of the Illness
Following the presentation of general facts
and theories, families are introduced to a
series of techniques for managing patients.
Based on the assumption that families can-
not accept staff suggestion unless they
genuinely believe that the staff know how
hard it has been to cope with this illness, this
discussion begins with a description of what
families have probably done over the years
that has not worked.
Following this description, families are
told that while there is no firm evidence that
families cause schizophrenia, there is rea-
son to believe that families have the power
to influence the course of the illness. They
are helped to see the need to create barriers
to overstimulation of the patient by estab-
lishing firm, clear, and approximate bound-
aries. Families are encouraged to set limits
on unacceptable behaviors, while allowing
patients to set their own pace toward recov-
ery. This theme is translated into specific
suggestions for responding to the patient's
fears, delusions, paranoid thoughts, obses-
sive rituals, or threats of violence. Finally,
families are strongly encouraged to avoid
centering their lives around patients. They
are asked to attend to their own needs, the
needs of other family members, and to mo-
bilize a social support network to maintain
their own ability to cope and survive over
time.

Phase DI: Reentry and Application
of Workshop Themes
Highly structured low-key individual family
sessions are held as soon as the acute phase
of the illness has been controlled sufficient-
ly to enable the patient to attend. Once pa-
tients have left the hospital, these sessions
occur once every 2-3 weeks. The interven-
tions of these sessions are based on the
themes established in the survival skills
workshop and relate largely to the reinfor-
cement of family boundaries and the gradu-
al resumption of responsibility by the pa-
tient.
Three kinds of boundaries are stressed on
an ongoing basis. The interpersonal bound-
aries between family members, the gener-
ational boundaries between parents and
offspring, and the family boundary with the
larger social community support system.
The first two types of boundaries are rein-
forced, largely by encouraging families to
establish clear expectations, rules, and
limit-setting processes. The third type of
boundary, that between the family and the
community, is minimized by stressing the
family's need for the development of a
support system beyond the nuclear family.
Over time, patients are gradually en-
couraged to assume more responsibility for
their lives and functioning. Initially, the en-
tire treatment focus is on the patient's sur-
vival outside of the hospital. As signs of life
begin to occur in patients, the sessions gra-
dually emphasize a return to effective work
and social functioning. This is initially ac-
complished by the assignment of small hou-
sehold tasks or tasks which involve a mini-
mal amount of socialization with outsiders.
Later, more ambitious tasks are assigned,
and families are encouraged to increase
Preliminary Results
Results for the 1st year of treatment using
this model have been encouraging. Among
treatment takers (n = 88), 19% ofthose re-
ceiving family therapy alone experienced a
psychotic relapse in the year following hos-
pital discharge. Twenty-one percent of
those receiving an individual behavioral
therapy (social skills training) relapsed, but
Psychoeducational Family Therapy 149
their expectations of patients. Since pro-
gress on these issues is exceedingly slow, a
great deal of support is given to family
members to enable them to tolerate inactiv-
ity, amotivation, and apathy.
Phase IV: Continued Treatment
or Disengagement
Once the goals for effective functioning
have been attained (and these goals differ
depending on the patient's abilities, the
length of impairment, and tolerance level of
families), the model calls for the family to
be presented with two possible options for
treatment: (1) more traditional family-
oriented treatment to resolve long-term
family conflicts of unfinished business or (2)
periodic supportive maintenance sessions
of gradually decreasing frequency.
The first three phases of this model of
family intervention do not offer families the
opportunity to deal with family issues and
problems that do not immediately relate to
the patient's survival in the community. In
fact, the model specifically discourages the
discussion of upsetting topics, such as mari-
tal discord, unresolved losses, or major
moves toward emancipation. Nevertheless,
some of these issues could interfere with the
ongoing growth and development of family
members, and once the crisis has passed,
family members are offered the opportuni-
ty to devote their energies to resolving
them. Unless it is thought that the issues
have direct and immediate impact on the
patient's progress, families are to be offered
a choice about this phase of treatment since
contracting to help the patient is. not
thought to give the right to unilaterally de-
termine the family's general goals or meth-
ods of attaining them.
no patient in the treatment cell which re-
ceived both family therapy and social skills
training experienced a relapse. In contrast,
nearly 36% of those receiving chemothera-
py and support relapsed. When all assigned
patients were studied (n = 102), only a sig-
nificant effect for family therapy could be
identified in lowering relapse. Most impor-
150 C. M. Anderson
tant, when patients who were entirely faith-
ful in adhering to their maintenance chemo-
therapy were included, a clear and signifi-
cant effect for family therapy was observed.
Analyses regarding the adjustment of these
patients are not complete as of this writing,
but in general, there is no evidence that
family therapy or social skills training are
simply maintaining poorly adjusted patients
in the community just to make the relapse
rates "look good". If anything, clinical
judgments regarding the "degree of illness"
and "change in clinical state" pre- and post-
treatment reveal a significant effect for fam-
ily therapy. (There is also a trend suggesting
better role performance for social skills pa-
tients as well.) Studies conducted at the
University of Southern California by Fal-
loon's group (Chap. 14) and in London by
Leff's group (Chap. 13) revealed that 6%
and 9% of family therapy-treated patients,
respectively, experienced a relapse by 9
months, but that 44% and 50% of controls
treated with drug and individual therapy
had relapsed by 9 months. Thus, there ap-
pears to be an increasingly broad validation
of the effectiveness of family approaches, at
least in the forestalling of a psychotic re-
lapse. This, in turn, increases the potential
for helping the patients to be integrated
within the communities.
We do not, however, wish to be intem-
perate in our claims. The results, at least
from our own study, have demonstrated
that these approaches forestall relapse but
in no way provide "prevention" per se. To
date, relapse rates continue to rise as pa-
tients near the end of their 2nd year of treat-
ment. An uncensured estimate of relapse
into the 2nd year of treatment reveals that
25% of family therapy alone patients, 35%
of social skills training alone patients, 22%
of the combined treatment patients, and
57% of drug-treated controls have ex-
perienced a relapse to date. Further, we
References
Anderson, C., Hogarty, G., and Reiss, D.
(1980). Family treatment of adult schizophren-
ic patients: A pscho-educational approach.
Schizophrenia Bulletin, 6 (3), 490-505.
Broen, W. E., and Storms, L. H. (1966). Lawful
disorganization: The process underlying a schi-
feel that investigators familiar with the EE
method and associated family therapists
(ourselves included) have succumbed to a
profound error in judgment in limiting
these modern interventions to high EE pa-
tients and their families. High EE, it seems,
has become the exclusive definition of high
risk in these studies. Admitting patients to
studies who come from high EE houses por-
tends to place under observation essentially
a group of unmarried males living in paren-
tal homes. In our own studies, this group
represents but a subsample of the "at risk"
schizophrenic population. Other samples,
including married and unmarried females,
are also at high risk for relapse, and we
believe these subgroups might also profit
from these interventions.
We surmise from the results we have
achieved so far that the primary effect of
family treatment is a prophylactic one and
that these experimentally treated patients
are not simply being maintained in the com-
munity in a more impaired condition. Al-
though both family treatment and social
skills training are useful, a limitation of the
social skills training approach seems to be
its inability to engage or maintain patients
who are actively psychotic at the time of
discharge. The family approach, however,
appears to buy time for patients, permitting
them to more fully recover by decreasing
family distress, educating them about the
illness, reducing their expectations and uni-
tary views of the illness, and otherwise in-
corporating new coping strategies. How-
ever, in facilities with long-term hospital-
izations or who otherwise discharge pa-
tients in a more clinically stable condition,
both family therapy and social skills trai-
ning might well be effective interventions in
significantly lowering schizophrenic re-
lapse. The combined treatment approach
provides an additional and important pro-
phylactic advantage.
zophrenic syndrome. Psychological Review,
73,265-279.
Brown, G. W., and Birley, J. L. T. (1968). Crises
and lifes change and the onset of schizophrenia.
Journal of Health and Social Behavior, 9,
203-214.

Brown, G. W., Birley, J. L. T., and Wing, J. H.
(1972). The influence of family life on the
course of schizophrenic disorders. A replica-
tion. British Journal of Psychology, 121,
241-258.
Goldstein, M., and Kopeikin, H. (1981). Short
and long term effects of combining drug and
family therapy. In M. J. Goldstein (Ed.), New
developments in interventions with families of
schizophrenics. San Francisco: Jossey-Bass.
Goldstein, M., and Rodnick, E. (1975). The
family's contribution to the etiology of schizo-
phrenia: Current status. Schizophrenia Bulle-
tin, 1, 14,48-63.
Goldstein, M. J., Rodnick, E. H., Evans, J. R.,
May, P. R., and Steinberg, M. (1978). Drug
and family therapy in the aftercare treatment of
acute schizophrenia. Archives of General Psy-
chiatry, 35 (10), 1169-1177.
Hatfield, A. B. (1978). Psychological costs of
schizophrenia to the family. Social Work, 23
(5),355-359.
Jacob, T. (1975). Family interaction in disturbed
and normal families: A methodological and
substantive review. Psychological Bulletin, 82,
33-65.
Jones, J. E. (1977). Patterns of transactional style
deviance in the TAT's of parents of schizophre-
nics. Family Process, 16,327-337.
Jones, J. E., Rodnick, E., Goldstein, M.,
McPherson, S., and West, K. (1977). Parental
transactional style deviance as a possible indi-
cator of risk for schizophrenia. Archives of
General Psychiatry, 34, 71-74.
Kreisman, D. E., and Joy, V. D. (1974). Family
response to the mental illness of a relative: A
review of the literature. Schizophrenia Bulle-
tin, 1 (10), 34-54.
Lang, P. J., and Buss, A. H. (1965). Psychologic-
al deficit in schizophrenia: Interference and ac-
tivation. Journal of Abnormal Psychology, 70,
77-106.
Leff, J., and Vaughn, C. (1980). The interaction
of life events and relatives' expressed emotion
in schizophrenia and depressiv neurosis. Brit-
ish Journal of Psychiatry, 135, 146-153.
Psychoeducational Family Therapy 151
Payne, R. W., Mattussek, P., and George, E. I.
(1959). An experimental study of schizophren-
ic thought disorder. Journal of Mental Science,
105,624-652.
Rabin, A. I., Doneson, S. L., and Jentons, R. L.
(1979). Studies of psychological functions in
schizophrenia. In: Bellak, L., (ed.). Disorders
ofthe Schizophrenic syndrome (pp. 181-231).
New York: Basic Books.
Shakow, D. (1962). Segmental set: A theory of
the formal psychological deficit in schizophre-
nia. Archives of General Psychiatry, 6, 1-17.
Silverman, J. (1972). Stimulus intensity modula-
tion and psychological disease. Psychopharma-
cologia, 24, 42-80.
Singer, M. T., and Wynne, L. C. (1965). Thought
disorder and family relations of schizophrenics.
Archives of General Psychiatry, 12, 187-212.
Singer, M. T., and Wynne, L. C. (1966). Com-
munication styles in parents of normals, neu-
rotics, and schizophrenics. Psychiatric Re-
search Reports, 20, 25-38.
Tecce, J. J., and Cole, J. O. (1976). The distrac-
tion-arousal hypothesis, CNV and schizophre-
nia. In: D. I. Mostofsky (ed.). Behavior control
and modification of physiological activity. Eng-
lewood Cliffs, NJ: Prentice-Hall.
Vaughn, C. E., and Leff, J. P. (1976). The in-
fluence of family and social factors on the
course of psychiatric illness. British Journal of
Psychiatry, 129, 125-137.
Venables, P. H. (1964). Input dysfunction in
schizophrenia. In B. A. Maher (ed.). Progress
in Experimental Personality Research (Vol. I,
pp. 1-47). New York: Academic Press.
Venables, P. H. (1978). Cognitive disorder. InJ.
K. Wing (ed.). Schizophrenia: Towards a new
synthesis (pp.117-137). New York: Academic
Press.
Wynne, L. (1961). The study of intra-familial ali-
gnments and splits in exploratory family thera-
py. In N. Ackerman, F. Beatmen, S. Sherman
(Eds.), Exploring the base for family therapy
(pp. 95-115). New York: Family Service As-
sociation of America.
13. Controlled Trial of Social Intervention in the
Families of Schizophrenic Patients
J. Leff, L. Kuipers, R. Berkowitz, R. Eberlein-Vries, and D. Sturgeon
Introduction
A robust association has been established
between the level of expressed emotion
(EE) shown by relatives and the outcome of
schizophrenia in patients living with them
(Brown, Monck, Carstairs, and Wing,
1962; Brown, Birley, and Wing, 1972;
Vaughn and Leff, 1976a; Leff and Vaughn,
1981). Relapse of schizophrenia is more
likely if patients live with relatives who are
excessively critical and/or overinvolved.
Such relatives are designated as high EE.
Two factors appeared to operate in a pro-
tective manner for patients living whith
high EE relatives. These were regular
maintenance therapy with neuroleptic
drugs and the establishment of a social dis-
tance between patient and relative, as mea-
sured by the amount of face-to-face contact
during a typical week. Furthermore, these
two factors seemed to be additive in their
protective effect. The evidence for these
conclusions was derived from an analysis of
pooled data from the two most recent stu-
dies (Brown et aI., 1972; Vaughn and Leff,
1976a). The relevant figure setting out
Method
Design
The problems to be surmounted in design-
ing a controlled trial of any social therapy
have been explored in detail by Leff (1981).
The double-blind placebo controlled drug
trial is a sound basis to start from, but a
number of modifications are necessitated
by the nature of social treatments. For ex-
ample, there is no placebo for a social treat-
ment. Any interaction between a patient
these results is reproduced here (Figure 1).
These issue of the direction of cause and ef-
fect in the relationship between relatives'
EE and patients' relapse has been ap-
proached by using statistical techniques to
define the role of possible mediating fac-
tors, such as patient's disturbed behavior.
The possibility that such behavior might
lead both to high EE in relatives and to re-
lapse in the patients was eliminated by a
variety of statistical approaches (Brown et
aI., 1972; Vaughn and Leff, 1976a). How-
ever, a causal relationship between high EE
and schizophrenic relapse and low EE and
patients remaining well can only be de-
monstrated convincingly by an experiment
in which relatives's EE is manipulated and
patients' relapse rate is monitored. The
same argument applies to the possible pro-
tective nature of low face-to-face contact,
which can only be established beyond doubt
by an experimental approach. This paper
reports an experiment directed at both
these issues.
and a staff member, however trivial, has
therapeutic potential, even a standardized
assessment procedure. We recognize that
there is no ideal solution to all the inherent
problems, but present our design as the op-
timum under the circumstances.
We chose to study patients at highest risk
of relapse, represented by subgroup 6 in
Figure 1. Virtually all these patients not
taking regular medication and living in high
contact with high EE relatives relapsed
154 J. Leff et al.
Subgroups

LowEEI3%

Face-to-face contact
! 1. On drugs

2 Not on drugs
12%

15%

!
Total group 3. On drugs 15%
<35h28%

} 4. Not on drugs 42%

!
HighEE51%
>35 h 69% 5. On drugs 53%

6. Not on drugs 92%

Fig. 1. Nine-month relapse rate of total group of 128 schizophrenic patients.

Note. From "The Influence of Family and and Social Factors on the Course of Psychiatric Illness. A
Comparison of Schizophrenic and Depressed Neurotic Patients" by C. E. Vaughn and J. P. Leff, 1976,
British Journal of Psychiatry, 129, pp. 125-137.

over the course of 9 months. In view of the
proven efficacy of maintenance neuro-
leptics in preventing relapse of schizophre-
nia (Leff and Wing, 1971; Hogarty and
Goldberg, 1973), we considered it unethical
to withhold such drugs from these high-risk
patients. Our strategy then was to ensure
that all patients entering the trial received
maintenance neuroleptics in the form of
long-acting injections, to eliminate prob-
lems with compliance. Patients were then
randomly assigned to an experimental
group, which received the social interven-
tion, and a control group, which received
routine outpatient care. In considering the
design, we were faced with the choice of
administering a nonspecific social treat-
ment of the control group or allowing their
responsible clinicians to treat them accord-
ing to their usual practice. Both procedures
have their drawbacks, and we chose the lat-
ter, recognizing that we would need to es-
tablish that the treatment experiences of ex-
perimental and control families were sub-
stantially different (cf. Orford and Ed-
wards, 1977).
Package of Social Interventions
It can be seen from Figure 1 that, provided
low social contact is genuinely protective,
reduction in face-to-face contact between
high EE relatives and patients on medica-
tion should reduce the relapse rate from
53% to 15%. This would be equivalent to
the effect on relapse rate of altering rela-
tives' EE from high to low (compare sub-
groups 1,2, and 3). Hence, the aims of so-
cial intervention were to reduce face-to-
face contact below the crucial level of 35 hi
week (Brown et aI., 1972) and/or to change
relatives' EE from high to low. A range of
strategies that might be expected to achieve
these aims has been set out by Leff (1976).
One possible design for our trial was to as-
sign experimental families randomly to a
variety of social interventions, for example,
an educational program, family sessions, or
behavioral modificaton. We decided that
this was too ambitious in terms of the num-
bers required and instead compiled a pack-
age of social interventions which was to be
applied flexibly according to the needs of
each family. Thus, we set out to compare
the relative effectiveness of this package
with routine outpatient care with the expec-
tation that if it did prove to be superior, the
individual elements could be evaluated sep-
arately in a subsequent trial.
The intervention basically consisted of
three elements: an education program, a
relatives' group, and family sessions.
Education Program
This consisted of four lectures on the etiol-
ogy, symptoms, course, and treatment and
management of schizophrenia. We put a lot
of thought into the writing of these lectures,
which incorporated our consensus opinions

on the above topics, couched in everyday
language with a strict avoidance of jargon.
The lectures were read out from a type-
script to each relative in his or her home.
Initially four visits were made, one for each
topic, but after a few relatives had been in-
structed in this way, we decided it would be
preferable to give two lectures at a time.
Following each lecture, we allowed unlimit-
ed time for the relative to ask questions.
This element in the package was in fact
evaluated separately by means of a knowl-
edge interview, administered before and
after the education program. This is pres-
ented in detail under "assessments."
Relatives' Group
This was originally conceived of as a way of
bringing together high EE and low EE rela-
tives. In addition to the family emotion stu-
dies, we had evidence from psychophysio-
logical studies (Tarrier, Vaughn, Lader,
and Leff, 1979; Sturgeon, Kuipers, Berko-
witz, Turpin, and Leff, 1981) that low EE
relatives had found way of coping with the
everyday problems of living with schizo-
phrenic patients that not only avoided dis-
tressing the patients but provided them with
active support.
It was our intention to use the group as
a means of altering the coping styles of high
EE relatives so as to resemble more closely
those of low EE relatives. Some of the
problems that relatives face were known
from Creer and Wing's (1975) study. Priest-
ley's experiences with a similar group of rel-
atives were also valuable (Priestley, 1979).
Given this information, the group was de-
liberately set up so that the therapists acted
as facilitators. Both high EE and low EE
relatives were encouraged to bring their
problems and their solutions to the meeting
and share them with others in a similar posi-
tion. The purpose of this was to enable
them to learn about coping strategies of
which they were unaware and finally to help
them try a different approach at home. The
focus of the group was thus on potential or
actual difficulties that relatives experienced
and not primarily on interpretations of the
relatives' own behavior. This latter was
more useful in discussions between the pro-
fessionals about the group process that oc-
Controlled Trial of Social Intervention 155
curred after each group meeting.
Other reasons for setting up the group in-
cluded the need to counter the sense of
isolation felt by many relatives of schizo-
phrenic patients, the possibility that the
group would allow relatives to discharge
emotions that would otherwise be directed
at the patients, and the fact that it was eco-
nomical of resources. The group was con-
structed to be flexible both in regard to
membership and attendance. Relatives
were inducted into the group as the patients
entered the study. The group met every 2
weeks in a large, comfortable room in the
Institute of Psychiatry for 1.5 h, and rela-
tives attended when they could. They were
encouraged to attend for the 9 months of
the study period, but were free to continue
thereafter if they wished, and a number car-
ried on coming for over 1 year. The group
had a maximum of seven members at any
one time and included two professionals for
most of the study, but latterly only one. Our
experience with the group is presented in
more detail elsewhere (Berkowitz, Kui-
pers, Eberlein-Fries, and Leff, 1981). Each
session was taperecorded, with permission
of the relatives, for future analysis.
Family Sessions
Because the relatives' group was not ap-
propriate for dealing with the whole range
of problems or for dynamic work and be-
cause patients were excluded from it, we
felt that it needed to be complemented by
sessions with the whole family. Conse-
quently, we saw each family in their own
home on a minimum of 1 and a maximum of
25 occasions. The sessions, which lasted
1 h, were conducted by two professionals, a
male psychiatrist and a female psychologist.
Where the key relative was a spouse, we
saw the patient and spouse together. Where
parents were involved, we saw the patient
with both parents if available, and adult sib-
lings were also encouraged to participate.
At the beginning of the study, the value
of family therapy was not well established,
particularly for the families of schizophreni-
ic patients. Consequently, we adopted a
pragmatic approach, guided by our princi-
pal aims of reducing EE and/or social con-
tact. Techniques used to achieve these aims
156 J. Leff et al.
Admission
'V
1st
Proce- knowledge Education knowledge Relatives'
dures PSE EE interview interview program interview
Experimental + +
High
Control + +
Experimental + +
Low
Control + +
_ _ _ _ _ _ _-'9 months..s_ _ _ _ _ __
Fig. 2. Flow chart of intervention program.
in family sessions varied from dynamic in-
terpretations to behavioral interventions.
Before each session, we discussed and for-
mulated our objectives, but we could be
diverted from these if other important is-
sues emerged in the course of the visit. Each
session was discussed afterward and a quite
detailed account written up.
In addition to these formal contacts,
some relatives phoned up one or other of us
on several occasions. We kept careful rec-
ords of all such contacts to calculate the
total amount of professional time taken up
by each family.
Assessments
All patients between the ages of 16 and 65
admitted to the Bethlem and Maudsley
Hospital, the Southwestern Hospital, and
the North Wing at St Pancras Hospital were
screened regularly. If they had lived with
relatives continuously for 3 months before
admission, resided within reasonable com-
muting distance from the hospital con-
cerned, and appeared from the case notes
to be suffering from a functional psychotic
illness, they were interviewed with the Pres-
ent State Examination (PSE) (Wing, Coo-
per, and Sartorius, 1974). The PSE data
were processed by the CATEGO program,
and those patients assigned a diagnosis of
schizophrenia were included in the trial if in
addition they spent more than 35 hlweek in
face-to-face contact with one or more rela-
tives. This was determined by constructing
a time budget of a typical week from an in-
terview with the patient, or failing that with
a relative (Brown et al., 1972).
Discharge
V
2nd 3rd
Family Follow-up knowledge
group sessions EEandPSE interview
+ + + + +
+ +
+ + + +
+ +
Having passed these screens, patients
then entered the study, and demographic
and historical data were collected with a
standardized schedule. Relatives in high
face-to-face contact with the patient (al-
most invariably a spouse or parents) were
interviewed with the abbreviated form of
the Camberwell Family Interview (Vaughn
and Leff, 1976b). Ratings of the four main
components of EE, criticism, hostility,
warmth, and overinvolvement, were made
from taperecordings of the interviews. Rel-
atives with a score of six or more critical
comments, any degree of hostility, or a rat-
ing of 3 or more on overinvolvement1 were
assigned to a high EE group. The further
assessments and procedures than ensued
are shown on a flow chart in Figure 2.
The knowledge interview was specially
constructed to test the effect of the educa-
tion program. It consisted of 21 questions
directly related to the basic information
given in the program. It was administered
before and after the program to experimen-
tal high EE and low EE relatives and at the
same interval to the control relatives. The
knowledge interview was given a third time
to all relatives at the end ofthe 9-month fol-
low-up.
Immediately following the first knowl-
edge interview, a joint interview was held
with the relative and patient in a studio.
This allowed us to make a videotape record-
ing of the interview, during which certain
1 The crucial level of overinvolvement was lower-
ed from 4 to 3 prior to the study in the light of
more extensive experience with this rather in-
frequent attitude.
 Controlled Trial of Social Intervention 157
psychophysiological measures of the pa- a more subjective judgment than the first
tient's respones were made. The data from kind of relapse and hence may be influ-
these interviews will be presented else- enced by the interviewer's bias. Unfortun-
where. ately, it was not possible to maintain blind-
The next step was to give the education ness to each family's status as experimental
program to the experimental high EE and or control because the research team was
low EE relatives. When this was com- both conducting the intervention and mak-
pleted, the relatives were invited to join the rng the assessments. This deficiency in the
relatives' group. At about this time, most design could only have been remedied by a
patients were close to discharge from the considerable increase in staff, which was
hospital. For all high EE families, the clin- not feasible. However, there were a num-
icians responsible for the patients were in- ber of safeguards against the influence of
formed of their high risk of relapse and bias. The key assessments, EE and the
were asked to prescribe long-acting neuro- PSE, were made by researchers who were
leptic medication if possible. It must be extremely experienced in the use of the
appreciated that patients remained under techniques and had been trained to a high
the care of their usual clinician. Our thera- level of interrater reliability.
peutic interventions with the experimental The assessment of EE was usually con-
families were conducted in parallel with the ducted independently by two raters. Where
patients' ordinary clinical care and de- they disagreed or where the rating was par-
pended on a close liaison with the clinical ticularly crucial, the recording was sent to
team. This was usually, but not invariably, a Dr Christine Vaughn, who rated it blindly.
successful arrangement. In addition, a representative selection of
Just before discharge, the patient's clin- eight follow-up interviews was rated blindly
ical state was assessed. If it was clear that by Dr R. Gosh, who had been trained in the
the psychosis had completely resolved, an assessment of EE. His reliability with the
informal assessment was carried out. How- definitive raters was calculated using the
ever, if there were any residual pychotic product moment correlation coefficient and
symptoms, a full PSE was conducted and was found to be 0.88 for critical comments
served as a baseline for any subsequent and 0.90 for overinvolvement. He agreed
change. Following discharge, virtually all with the definitive raters on the assignment
the patients attended outpatient clinics on a to high or low EE in seven of the eight
regular basis. A note was affixed to their cases. In the eighth case, he rated one ex-
case records asking the responsible clini- perimental mother as marginally high on
cian to contact the research team if there criticism whereas she had been rated as low
was any suggestion of an impending or by both definitive raters independently.
actual relapse. As a result, it was possible to With regard to the PSE ratings, these
see all but one of the patients who suffered were not done blindly in the case of a type 1
a recrudescence of symptoms at an early relapse. However, for possible type 2 re-
stage in their relapse. A full PSE was per- lapses, an independent rater, Dr Paul Beb-
formed at the time and a history of life bington, was recruited to make the assess-
events taken for the 3 months prior to re- ments blindly. He did not always concur
lapse (Brown and Birley, 1968). with the view of the patient's clinician that a
For the purposes of the research, relapse relapse had occurred.
was defined as a recurrence of schizophren-
ic symptoms as detected by the PSE, in pa-
Follow-up
tients who had been free of them at dis-
charge (type 1 relapse of Brown et al., All patients in the trial were followed up at
1972). For the small proportion of patients 9 months after discharge. If they had not al-
who still had active schizophrenic symp- ready relapsed, a PSE and life events his-
toms at discharge (two experimental, three tory were administered. The relative or rel-
control), relapse was defined as a marked atives were reinterviewed with the Camber-
increase in the number or intensity of symp- well Family Interview, if possible by the
toms (type II relapse). This clearly involves same person as on the first occasion, and
158 J. Leff et al.
EE was rated. The time budget was also
repeated. A history of medication taken by
the patient during the preceding 9 months
was obtained.
A 2-year follow-up has also been initia-
ted, the results of which will be presented
when it is completed.

Subjects
nificant difference between the two groups,
Characteristics of Patients
once more confirming the truly random
Over the course of 4 years, 49 patients pas- nature of patient assignment. In terms of
sed the screening criteria for the trial. The CATEGO classes, 11 patients in each
number is so small because our selection group were labeled S+ (schizophrenia),
procedures for high-risk patients eliminated while the remaining patient in each group
more than two of every three admissions for was classified as P+ (paranoid psychosis).
schizophrenia. Of those families ap-
proached, nine (18%) refused to take part
Characteristics of Relatives
in the study. Of the remainder, 24 lived with
high EE relatives and actually participated In the experimental group, six patients
in the trial. The randomization procedure, lived with a spouse and six with parents. In
using a table of random numbers, segregat- the control group, six lived with a spouse,
ed them into 12 control and 12 experimental five with parents, and one with her sister.
families. Relatives who were found to be Some patients in both groups lived with a
low EE from the Camberwell Family Inter- single parent, and in some households one
view were also randomly assigned to ex- of the parents was in low face-to-face con-
perimental and control groups for the pur- tact with the patient, in which case they
pose of assessing the effects of the educa-
tion program (see Figure 2). In addition,
the low EE experimental families were en- Table 1. Demographic and Historical Charac-
couraged to join the relatives' group, in ac- teristics of Patients
cord with the principal therapeutic aim of Experi- Control
identifying and using their coping skills as a mental
model.
Number 12 12
Male/female 6/6 6/6
Demographic and Historical Data Living group: parents 6 6
spouse/sibling 6 6
The experimental and control patients were Mean age 39 30
compared on a large number of demo- Education: CSE or above 4 4
graphic and historical features as shown in Ever married 6 7
Table 1. Except for unemployment, which Ever divorced or separated 1 2
was worse in the experimental group, none Mean number of children 0.8 1.3
of the items differed significantly between Drop from highest sociosexual
the two groups, endorsing the effectiveness achievement 3 2
ofthe randomization procedure. Occupation: nonmanual or
above 6 2
Mean length of unemployment
Clinical Data before admission in months 16.5 4.8'
Employed at admission 4 7
The PSE data were processed by the Abnormal premorbid
CATEGO program and a print-out ob- personality 8 5
tained of syndromes and of diagnostic clas- First admission 4 4
ses. The experimental and control patients Mean number of previous
were compared for the presence or absence admissions 2.3 1.2
of each of the 38 syndromes using Fisher's Age at first onset 33 27
exact test. For no syndrome was there a sig- • P<0.05.
 Controlled Trial of Social Intervention 1S9
Table 3. Overinvolvement Scores of Experimental and Control Relatives
Number of relatives Mean number of critical comments
with six or more
critical comments At9-month
initially Initially follow-up

Experimental 12 15.8 6.S t=3.37,P<0.00S

Control 8 12.0 10.7 NS
NS NS

Table 2. Critical Comments of Experimental and Control Relatives

Mean overinvolvement score

Number of relatives
scoring 3 or more At9-month
on overinvolvement Initially follow-up

Experimental 5 4.0 2.4 t=2.36,0.1 >P>O.OS

Control 6 4.0 3.7 NS
NS NS

were not interviewed for an EE assessment.
A total of 13 relatives in the experimental
group and 13 in the control group were
rated on EE. The mean number of critical
comments and the mean overinvolvement
scores for these individuals are shown in
Tables 2 and 3. There were no significant
differences between the groups on these
measures.

Results

Social Intervention
Two relatives did not attend the group at
all. One, a wife, attended for the first time
on the only occasion when no other rela-
tives turned up. She was given a one-to-one
session with the therapist and declined
further attendance or contact. She was the
only relative who did not receive the educa-
tion program. The other, a husband, was
managing a business which demanded his
presence at the time the group met. The re-
maining 11 high EE relatives attended the
group an average of 9.1 times ranging from
6- 21 times over the 9-month follow-up pe-
riod. The number of relatives present at
each session varied from two to seven, with
a mode of four.
Family sessions in the home wee held for
all families. The number of sessions ranged
from 1 to 25 with a mean of 5.6. However,
only one family had as many as 25 sessions.
In addition to the formal contacts of the
relatives' group and family sessions, rela-
tives were encouraged to phone members
of the team whenever they felt the need.
This facility was not abused as a total of 10
relatives used this method of contact be-
tween 2 and 19 times, the mean being 7.5
times.
Apart from our own efforts to work with
the relatives in the experimental group, the
clinical team responsible for the patient's
care provided some input in two cases. For
one married couple, the clinical team's so-
cial worker acted as a co-therapist with
one of the research psychologists, as he
wished to gain experience of mariti thera-
py. Another married couple received mari-
tal therapy from the clinical psychiatrist in
parallel with our own family sessions. We
160 J. Leff et al.
took pains to maintain a close liaison with
this therapist.
The therapeutic experiences of the con-
trol relatives were studied by consulting the
case notes of the patients as well as ques-
tioning the relatives at follow-up.
Treatment Experiences of Control
Relatives
In eight cases, no therapeutic help was
given to the relatives. In two cases, the rela-
tives themselves were seen by a psychiatrist
to receive treatment for overt symptoms.
One relative, the mother of a schizophrenic
girl, was discovered to suffer from schizo-
phrenia herself and was treated as an outpa-
tient with a neuroleptic drug. She attended
a total of 11 times in the course of 9 months,
but the focus of treatment was on her psych-
iatric condition rather than on her attitudes
toward her daughter. The other relative to
receive treatment was a husband who be-
came depressed and was seen as an outpa-
tient on several occasions over 2 months.
Two further families received help from
social workers. In one family, this took the
form of a single family session held during
the 9th month after the patient was dis-
charged with the aim of persuading him to
attend a day hospital. In the second family,
the wife of a schizophrenic patient was seen
every 2 weeks by a psychiatric social work-
er, and one family session was held during
the 9 months. The social worker wrote that
"my interviews with her have had the object
of providing her with some support through
helping her to share her feeling of responsi-
bility for maintaining her husband's pro-
gress." He gave her some advice on how to
handle her husband's behavior that stem-
med from his delusions. It is only in this
single case of the 12 controls that profes-
sional help to the relatives approximated
that received by the experimental families.
It is worth noting that in this control case
the wife's level of overinvolvement actually
rose during the follow-up period, although
her husband reduced his contact with her to
a low level.
FoUow-up Relatives
All relatives in the experimental group
went through the planned assessment pro-
cedures at the 9-month follow-up. 2-year
follow-up has been completed and will be
presented elsewhere (Leff, Kuipers, Ber-
kowitz, and Sturgeon, in press).
The control relatives were not as cooper-
ative as their experimental counterp!lrts,
probably because we offered them no help.
Two of the 13 originally interviewed re-
fused a second EE interview, and one, a
husband, had left his wife and was unob-
tainable. Thus the success rate for follow-
up was 77%.
The mean number of critical comments
and the overinvolvement scores obtained at
follow-up are shown in Tables 2 and 3. It is
evident that there has been a highly signifi-
cant reduction in criticism in the experi-
mental group over the 9-month period of
social intervention. Five experimental rela-
tives changed from high to low criticism,
while three others showed some decrease.
Only three critical relatives remained vir-
tually unchanged. By contrast, there was a
small and nonsignificant reduction in the
number of critical comments made by the
control relatives. This was entirely account-
ed for by two relatives, one a spouse and
one a sister, who changed from high to low
criticism. As neither relative had received
any therapeutic help from professionals,
this almost certainly represents a spontane-
ous alteration in attitude.
We can conclude from these data that our
package of social interventions met with
considerable success in ameliorating critical
attitudes in the experimental relatives. The
same cannot be said for emotional overin-
volvement. There was a drop in the mean
overinvolvement score for the experimental
relatives, as can be seen from Table 3, but
this failed to reach an acceptable level of
significance. In terms of individuals, two
relatives altered from high to low scores on
overinvolvement. One of these, the only
spouse to show overinvolvement in the ex-
perimental group, underwent a particularly
dramatic change from a score of 4 to zero.
Two others achieved a lower score, but did
not drop below 3, while one relative showed
no change. By contrast, none of the overin-
volved control relatives fell below 3 on this
scale, while one relative actually changed
from low to high during the follow-up period.

At follow-up, face-to-face contact had
fallen below 35 hlweek in six experimental
relatives, two being the parents of one pa-
tient. In one patient, this reduction resulted
from his getting a full-time job and in anoth-
er from attendance at a day hospital, while
the remaining three made alterations in the
ways they spent their leisure time. Low con-
tact was also found at follow-up in three
control families, which is not significantly
different from the proportion in experi-
mental families. However, in two of these,
the way in which this was achieved was dif-
ferent from that in the experimental group.
One patient's husband left her during the
follow-up period, while another patient
began spending most of his leisure time
alone in his room. The third patient de-
veloped more leisure interests which took
him out of the parental home.
As stated earlier, the aims of the social in-
tervention were to lower EE and/or face-to-
face contact below certain crucial levels.
Reduction of EE was achieved in five famil-
ies and reduction of contact also in five fam-
ilies, but there was some overlap, as can be
seen in Table 4.
As a consequence, one or both of our
aims were met in 9 of the 12 families with
follow-up completed to date. Hence, social
intervention was successful with 75% of
families.
FoUow-up Patients
In each group, all but one of the patients
were maintained on long-acting injections
Table 4. EE and Face-to-Face Contact at Follow-up in Experimental Group.
Patient code number Relative
1 Mother
1 Father
7 Husband
13 Wife
30 Wife
31 Wife
40 Wife
41 Mother
45 Husband
49 Mother
53 Mother
55 Mother
61 Mother
Controlled Trial of Social Intervention 161
of neuroleptics and received these regularly
as prescribed. One control patient with a
first episode of a paranoid psychosis was
prescribed oral trifluoperazine, which he
took religiously, while one experimental
patient could not tolerate the side effects of
injections and was transferred to oral med-
ication. She also adhered meticulously to
her drug regime.
During the 9-month follow-up, relapse
occurred in seven patients, six of them in
the control group. All relapse were type 1 in
nature. In six patients, the relapse took the
form of a florid episode, similar if not ident-
ical to the previous one. Three of these pa-
tients were readmitted to the hospital. The
seventh patient, a control case, had been
free of auditory hallucinations, which re-
turned abruptly. The relapse rate in the
control group was 50%, while the rate in the
total experimental group was 8% (exact
P = 0.032). Hence, the package of social
interventions produced a significant reduc-
tion in relapse rate of schizophrenia. Furth-
er light is thrown on the specificity of the
interventions by comparing the relapse rate
in the control group with that of the patients
in whose families we achieved one or both
of the aims of social intervention. In fact, in
these nine families not a single patient re-
lapsed, a difference from the relapse rate in
the control group of even greater signifi-
cance (exact P = 0.017).
The single experimental patient who re-
lapsed was living in high contact with her
mother, who was both highly critical and
EE Face-to-face contact
Low Low
Low Low
Low High
High High
Low Low
High Low
Low High
Low High
High Low
High Low
High High
High High
Low High
162 J. Leffetal.
extremely overinvolved. The patient was
burdened with two handicaps, since in addi-
tion to schizophrenia she also had Turner's
syndrome and was of conspicuously short
stature.
Discussion
The screening procedure for our trial was
calculated to select patients with schiz-
ophrenia who were at high risk of relapse.
The effectiveness of the procedure was sub-
stantiated by the relapse rate of 50% over 9
months in the control group, despite the
fact that they were all maintained regularly
on neuroleptic drugs. The significantly and
substantially lower relapse rate of 8% in the
experimental group provides evidence for
the therapeutic effectiveness of our pack-
age of social interventions. However, we
did not achieve our aims of lowering EE
and/or social contact in every family, but
only in three-quarters of them. It is notable
that in this smaller group of nine families no
patient relapsed. The single relapse of an
experimental patient occurred in one of the
three families in which neither EE nor so-
cial contact was lowered to the target level.
This shows that the achievement of one or
both of our stated aims was crucial for a
therapeutic effect.
The specificity of our therapeutic en-
deavor in totally preventing relapse of schi-
zophrenia in a high-risk group over 9
months provides unequivocal evidence
both for the causal influence of high EE rel-
atives on relapse and for the protective ef-
fect of lowered social contact between pa-
tient and relative. Our tentative interpreta-
tion of the data in Figure 1 (Vaughn and
Leff, 1976a) now takes on a definitive sta-
tus, all the more so because the relapse
rates predicted from the Figure were so
close to the actual rates (53% cf. 50%,15%
cf.8%).
It needs to be emphasized that our find-
ings not only constitute evidence for the
causal influence of relatives' EE on the
schizophrenic process in patients, but also
substantiate the effectiveness of a form of
social intervention with these families. To
date, the evidence for this has been some-
what sparse. Only the studies by Goldstein,
In the control group, as already mention-
ed, two relatives spontaneously changed
from high to low EE between the initial as-
sessment and the follow-up. In neither
household did the patient relapse.
Rodnick, Evans, May, and Steinberg
(1978), Hogarty, Goldberg, Schooler, and
Ulrich (1974) and Hogarty, Schooler,
Ulrich, Mussare, Ferro, and Herron (1979)
satisfy reasonably stringent scientific crite-
ria. Goldstein and his colleagues studied
schizophrenic patients in their first or se-
cond admission so that it is inevitable that
their average age, 23, was considerably
lower than that of our sample, 35. Their pa-
tients were discharged after an admission of
only 1-2 weeks on long-acting injections of
fluphenazine. The experimental subjects
received six sessions of crisis-oriented fami-
ly therapy, with the primary goal of helping
the patient and significant others use the
events of the psychosis to adjust to the ill-
ness. At a 6-month follow-up, no patient in
the drug plus family therapy group had re-
lapsed, whereas the relapse rate among pa-
tients who received the drug alone was 18%
(P = 0.04). The interpretation of this result
is slightly clouded by the fact that the thera-
py received by both groups was not stand-
ardized after the initial6-week trial period.
In the two studies of Hogarty and his col-
leagues, social therapies were also added to
maintenance neuroleptic drugs in a con-
trolled design. Their patients were more
like ours than were Goldstein's, being
mainly chronic and having a median of
three previous hospital admissions. In the
first study, the social therapy was a com-
bination of intensive social casework and
vocational rehabilitation counseling, while
in the second it comprised intensive individ-
ual and family social casework. In neither
study did the social intervention significant-
ly reduce the relapse rate when added to
maintenance drugs. However, it did appear
to confer an advantage on the patients in
the 2nd year of the follow-up, no relapse oc-
curring during this period among patients
receiving social therapy and drugs in both
studies.

Our result is thus closer to that of Gold-
stein et aI., although we were studying a
more chronic population in remission rath-
er than in the throes of a psychosis, and fur-
thermore our subjects were selected to re-
present a particularly high-risk group. This
difference is pointed up by comparing the
relapse rate of our control group over 9
months, 50%, with that of Goldstein's con-
trol patients over 6 months, 18%.
Our findings have important implica-
tions, both theoretical and practical. On the
theoretical side, now that the role of rela-
tives' emotional attitudes has been firmly
established in the causation of schizophren-
ic relapse, the question arises of whether
they operate in a similar manner prior to the
first episode. It is logically inadmissable to
extrapolate our findings backward in time
to before the first attack of illness since this
itself may dramatically alter relatives' at-
titudes. However, evidence exists linking
one emotional attitude we found to be in-
fluential, with maternal overprotectiveness
antedating the onset of schizophrenia. This
evidence, which derives from child guid-
ance clinic studies, was reviewed by Hirsch
and Leff (1975) who concluded that over-
protective attitudes were more commonly
shown by mothers of children who later
develop schizophrenia than mothers of con-
trol children. There is a clear link between
this early overprotectiveness and an aspect
of overinvolvement, as measured in the stu-
dies of EE. This suggests that the overpro-
tective component of overinvolvement
develops very early in the child's life, and
indeed we have anecdotal evidence of this
from several of our families. It is likely that
the long-standing nature of this component
of maternal overinvolvement accounts for
the lesser degree of success we had in alter-
ing this attitude when compared with our
effect on criticism. The single dramatic dis-
appearance of overinvolvement was not in a
mother but in a wife, who would of neces-
sity have held this attitude for a much short-
er period of time.
The other theoretical consideration is the
mechanism by which relatives' attitudes
produce a recurrence or exacerbation of
schizophrenic symptoms in the patients.
Previous studies have established that the
effect of high EE relatives on patients is
Controlled Trial of Social Intervention 163
detectable by monitoring the patients' auto-
nomic responses (Tarrier et aI., 1979; Stur-
geon et aI., 1981). An "arousal" hypothesis
has been proposed to accommodate these
findings and work has recently begun using
this paradigm to study cortical psycho-
physiology.
There are a number of practical issues,
the first of which concerns our failures in
the experimental group. With hindsight
could we have been more effective in work-
ing with the three families concerned, bear-
ing in mind our lack of experience with
these kinds of families when we began the
study? The answer for one family is possibly
yes. This was a marital couple where the
husband was the patient. They received
marital therapy from the clinical psychia-
trist in parallel with our sessions of family
therapy. It is possible that this diversifica-
tion of effort weakened the potency of the
two sets of therapists, and more would have
been achieved with a single set.
The other two families shared a number
of features in common: they both had multi-
ple problems, the patient was a single girl in
both, and the key relative was a mother who
was both excessively critical and overin-
volved. In one family, already, referred to,
the patient had Turner's syndrome in addi-
tion to schizophrenia, and the mother had
shown overinvolved attitudes to her almost
from birth. In the other, the mother was Ita-
lian and spoke broken English, and it was
doubtful, in retrospect, whether she took in
much of what went on in the relatives'
group and family sessions.
In the first of these two families, the pa-
tient relapsed after only 2 months at home
and before her mother had started to attend
the relatives' group. This was an unusual
delay and may have been symptomatic of
the mother's resistance to therapy. In fact,
we continued working with the family fol-
lowing the patient's relapse, and after a
year or more there are some indications of
change in the mother's attitudes. These will
be assessed at the 2-year follow-up. The
second of these two families, with the ita-
lian mother, we felt would probably have
resisted the efforts of the most experienced
therapist. However, someone who was
bilingual in English and Italian might well
have breached the communication barrier.
164 J. Leff et al.
The other two practical issues are closely
connected: which element in our package
was the most effective, and what should be
taught to professionals who intend to work
with the families of schizophrenic patients?
We cannot answer the first question fully
since we did not evaluate the individual ele-
ments of our package, apart from the edu-
cation program. In the absence of this infor-
mation' we have to assume that the rela-
tives' group and family sessions are of equal
importance. Working with families would
appear to require a longer training than the
running of relatives' groups, but it must be
appreciated that we had little experience in
either form of therapy with families of schi-
zophrenic patients when we started and yet
managed to achieve a high success rate in
meeting our aims. We consider that part of
the formula for success is the setting out of
clear aims from the beginning, and anyone
entering this area of clinical work should
have as a target the reduction of face-to-
face contact and of critical and overin-
volved attitudes in the relatives.
The clinician is faced with the problem
of deciding which families need this kind of
approach. Unfortunately, the training re-
quired to rate EE is a long and arduous pro-
cess so that the Camberwell Family Inter-
view cannot be incorporated routinely into
clinical assessment. However, most high
EE families are readily identified by clin-
icians, even those unfamiliar with the con-
cepts of criticism and overinvolvement.
Families with borderline levels on these sca-
les are likely to be missed in clinical prac-
tice, but the clinician should be alerted to
the possibility that relatives are high EE if
patients living with them require frequent
readmissions despite prophylactic medic-
ation.
The determination of the degree of face-
to-face contact is a simple procedure which
References
Berkowitz, R., Kuipers, L., Eberlein-Fries, R.,
and Leff, J. (1981). Lowering expressed emo-
tion in relatives of schizophrenics. In M. J.
Goldstein (Ed.). New Developments in Inter-
ventions with Families of Schizophrenics. Lon-
don: Jossey-Bass.
requires no special training and could read-
ily be introduced as part of the history tak-
ing. In choosing between the strategies of
attempting to reduce face-to-face contact or
to lower EE, it might be assumed that the
former was the easier and less time-con-
suming option. Indeed, we began the trial
with this explicit assumption. However, we
found that a number of families were very
reluctant to spend less time together and yet
responded to our attempts to reduce EE. In
other families, we felt that a reduction in
contact only became possible when the
emotional bonds tying them together had
been loosened. As it turned out, we were
equally successful in achieving both aims so
would recommend their joint adoption by
clinicians.
From the beginning, we have been sensi-
ble of the need to keep our interventions
within the scope of everyday clinical prac-
tice. We attach such great practical impor-
tance to the result of our trial that we see
our next task as the compilation of a train-
ing program that will make the necessary
skills available to social workers, clinical
psychologists, and psychiatric nurses.
Acknowledgment. We are grateful to the consult-
ants at the Bethlem and Maudsley Hospital, the
Southwestern Hospital, and the North Wing at St
Pancras Hospital who allowed their patients to
take part in this study. Our thanks are due to Dr.
Bebbington for carrying out PSE assessments on
some patients and to Dr. Vaughn and Dr. Ghosh
for conducting EE assessments on some relatives'
interviews.
An extended version of J. Leff, L. Kuipers, R.
Berkowitz, R. Eberlein-Vries, and D. Sturgeon,
"A Controlled Trial of Social Intervention in the
Families of Schizophrenic Patients". British
Journal of Psychiatry, 1982, 141, 121-134. Copy-
right 1982. The Royal College of Psychiatrists.
Reprinted by permission of authors and pub-
lisher.
Brown, G. W., and Birley, J. L. T. (1968). Crises
and life changes and the onset of schizophrenia.
Journal of Health and Social Behaviour, 9,
203-214.
Brown, G. W., Birley, J. L. T., and Wing, J. K.
(1972). Influence of family life on the course of

schizophrenic disorders: a replication. British
Journal of Psychiatry, 121,241-258.
Brown, G. W., Monck, E. M., Carstairs, G. M.,
and Wing, J. K. (1962). Influence of family life
on the course of schizophrenic illness. British
Journal of Preventive and Social Medicine, 16,
55-68.
Creer, c., and Wing, J. (1975). Living with a
schizophrenic patient. British Journal of Hos-
pital Medicine, 14,73-82.
Goldstein, M. J., Rodnick, E. H., Evans, J. R.,
May, P. R. A., and Steinberg, M. R. (1978).
Drug and family therapy in the aftercare of
acute schizophrenics. Archives of General Psy-
chiatry, 35, 1169-1177.
Hirsch, S. R., and Leff, J. P. (1975). Abnormali-
ties in parents of schizophrenics (Maudsley
Monograph No 22.). London: Oxford Univer-
sity Press.
Hogarty, G. E., and Goldberg, S. C. (1973).
Drug and sociotherapy in the aftercare of schi-
zophrenic patients. One-year relapse rates.
Archives of General Psychiatry, 28, 54-64.
Hogarty, G. E., Goldberg, S. C., Schooler, N.
R., and Ulrich, R. F. (1974). Drug and socio-
therapy in the aftercare of schizophrenic pa-
tients. II. Two-year relapse rates. Archives of
General Psychiatry, 31, 603-608.
Hogarty, G. E., Schooler, N. R., Ulrich, R. F.,
Mussare, F., Ferro, P., and Herron, E. (1979).
Fluphenazine and social therapy in the after-
care of schizophrenic patients. Archives of
General Psychiatry, 36, 1283 -1294.
Leff, J. P. (1976). Schizophrenia and sensitivity
to the family environment. Schizophrenia Bul-
letin, 2, 566-574.
Leff, J. P. (1981). Clinical and methodological
problems in interaction studies. In G. Tognoni,
C. Bellantuono, and M. Lader (Eds.), Epi-
demiological Impact of Psychotropic Drugs.
Amsterdam: Elsevier.
Controlled Trial of Social Intervention 165
Leff, J. P., Kuipers, L., Berkowitz, R., and Stur-
geon, D. (in press). A controlled trial of social
intervention in the families of schizohrenic pa-
tients: Two year follow-up. British Journal of
Psychiatry .
Leff, J. P., and Vaughn, C. (1981). The role of
maintenance therapy and relatives' expressed
emotion in relapse of schizophrenia: a two-year
follow-up. British Journal of Psychiatry, 139,
102-104.
Leff, J. P., and Wing, J. K. (1971). Trial of main-
tenance therapy in schizophrenia. British
Medical Journal, 3, 599-604.
Orford, J., and Edwards, G. (1977). Alcoholism
(Maudsley Monograph No 26). London:
Oxford University Press.
Priestley, D. (1979). Schizophrenia and the
family: II. In J. K. Wing and R. Olsen (Eds.),
Community Care for the Mentally Disabled.
London: Oxford University Press.
Sturgeon, D., Kuipers, L., Berkowitz, R., Tur-
pin, G., and Leff, J. (1981). Psychophysiologi-
cal responses of schizophrenic patients to high
and low expressed emotion relatives. British
Journal of Psychiatry, 138,40-45.
Tarrier, N., Vaughn, C. E., Lader, M. H., and
Leff, J. P. (1979). Bodily reactions to people
and events in schizophrenia. Archives of Gen-
eral Psychiatry, 36, 311-315.
Vaughn, C. E., and Leff, J. P. (1976a). The in-
fluence of family and social factors on the
course of psychiatric illness. A comparison of
schizophrenic and depressed neurotic patients.
British Journal of Psychiatry, 129, 125-127.
Vaughn, C. E. and Leff, J. P. (1976b). The mea-
surement of expressed emotion in families of
psychiatric patients. British Journal of Social
and Clinical Psychology, 15, 157-165.
Wing, J. K., Cooper, J. E., and Sartorius, N.
(1974). Measurement and classification of psy-
chiatric symptoms. Cambridge: Cambridge
University Press.
Appendix
Vignettes of mgh EE Families in Study
I = iDitiai assessment; F = foUow-up assessment;
C = critical comments; 0 = overinvolvement
score.
Experimental Families
Case 1 Mother I: Cll, 03; F: Cl, 02.
Stepfather I: C9; F: C2.
The patient, a man of 23 living with mother, step-
father, and a younger brother, suffered a first at-
tack of schizophrenia. Mother had always seen
him as the weaker of her two sons and appeared
to have been overinvolved with him virtually
from birth. He slept in her bed till the age of 15.
He was discharged from the hospital to a hostel,
but left after a few weeks to return home. How-
ever, he attended a day hospital so that contact
remained low. Family sessions included his par-
ents and brother and focused on the marital rela-
tionship, which was initially poor, but improved.
The patient was free of psychotic symptoms on
discharge and remained so throughout the 9
months.
Case 2 Husband I: C9; F: CO.
The patient, a married woman of 53, experienced
a second episode of schizophrenia. Her husband,
a retired professional and an intelligent, emo-
tionally inhibited man, was extremely intolerant
of her irrationality when she was ill. He attended
the group regularly, and in addition we conduct-
ed joint sessions in the home, but we were never
able to explore the emotional roots of his in-
tolerance. Nevertheless, his critical attitude had
disappeared at follow-up, and his wife remained
free of all symptoms. They continued to be in
high contact despite her joining day classes, and
his taking up a part-time teaching job.
Case 3 Wife I: C26; F: Cll.
The patient, a married man of 42, had been
admitted six times before the current episode of
schizophrenia. He had not been able to work in
the preceding 18 months. There were three chil-
dren, with the eldest of whom, a daughter, he had
had an incestuous relationship some years previ-
ously. This had come to light and he had served a
prison sentence. His wife was very intolerant of
his handicaps and extremely critical of his inabili-
ty to playa man's role in the family. She used the
group to vent a great deal of anger. Family ses-
sions were held, sometimes including the chil-
dren, but our attempts to persuade patient and
wife to spend less time together were unsuccess-
ful. Although she welcomed the opportunity to
go out in the evening, he both refused to ac-
company her and made her feel guilty about leav-
ing him. At follow-up, her criticism had consid-
erably abated, although not below the crucial
level, and they remained in high contact. Never-
theless, he remained free of psychotic symptoms,
although he continued to exhibit a significant de-
gree of depression.
Case 4 Wife I: C7, 01; F: C2, 01.
The patient, a married man of 47, was admitted
on the second occasion for an episode of paranoid
schizophrenia. He believed that the neighbors
downstairs were interfering with his apartment in
various ways, a belief that his wife shared, thus
exhibiting folie Ii deux. He had retired from his

job as a caterer 3 years earlier on account of phy-
sical ill health. One conjoint marital session was
held, following which his wife attended the group
on a single occasion, but as no other relatives ar-
rived that time, she was given an individual ses-
sion by the psychologist. She refused to come to
the group again. The patient lost his paranoid de-
lusions while in the hospital, and his wife's abnor-
mal beliefs also disappered. He remained well
and at follow-up was found to be in low contact as
he had increased the time spent going out on his
own.
Case 5 Wife I: C8; F: C9.
The patient, a married man of 55, was admitted
for the first time with an acute and florid attack of
schizophrenia. He had been unemployed for 3
years having lost his job as a chauffeur through a
drinking and driving offence. He had married his
wife when she was pregnant with another man's
child, and they had had four children of their
own. His wife was critical of him and showed little
warmth. She attended the group regularly, and
we held several conjoint marital sessions, but at
follow-up her critical attitudes were unaltered
and the marital relationship appeared to be
unchanged. However, the patient had taken on a
job as a security guard and because of the shifts he
worked was in low contact for 5 of every 6 weeks.
He remained free of psychotic symptoms
throughout the 9 months.
Case 6 Wife I: C3, 04; F: C1, 00.
The patient, a married man of 56, was admitted
for the tenth time with a psychotic illness. In the
past, his diagnosis had varied between schizo-
phrenia, mania, and schizoaffective psychosis.
On this occasion, he had first-rank symptoms as
well as grandiose delusions and was diagnosed by
CATEGO as S+. His wife had been a social
worker and gave up her job on marrying him 6
years previously. It was evident that she treated
him as though he was her client, and this was re-
flected in her overinvolvement with him. He had
given up his job as a commercial artist 7 years
previously and did clerical work part-time. His
wife attended the group regularly, and in addi-
tion, we held one conjoint marital session in
which the patient's feelings of inadequacy and his
wife's need to protect him were explored. This
appeared to have a significant effect as shortly
afterward she returned to full-time work. He con-
tinued to work part-time at home and remained
in high contact with his wife, but at follow-up her
overinvolvement had completely dissipated. The
patient remained well.
Appendix 167
Case 7 Mother I: C16, 01; F: C3, 02.
The patient, a divorced woman of 42, lived with
her elderly mother and was admitted for the first
time with an episode of schizophrenia. A younger
sister had suffered from schizophrenia for many
years and had eventually committed suicide. The
patient slept in the same bedroom as her mother,
who was very critical of her. The patient was also
physically disabled by chronic bronchitis and was
unable to work. Her mother attended the group
regularly, and sessions were also held in the
home. In particular, we attempted to persuade
the patient to move into a bedroom of her own
and to influence her mother to allow her to do
more for herself in the home. We made efforts to
get the patient to attend a day center, but she was
very resistant to this, so instead we arranged for
her to come to the hospital daily as a day patient.
At follow-up, she was spending much less time
with her mother, although still in high contact.
Her mother had become much less critical,
though was slightly more overinvolved and they
still slept in the same room. At discharge, the
patient was hearing voices continuously despite a
high dose of a phenothiazine. The voices gradual-
ly attenuated over the course of 9 months and
were much less insistent and frequent at follow-
up.
Case 8 Husband I: C26; F: C15.
The patient, a married woman of 44, was admit-
ted for the sixth time with an attack of schizo-
phrenia. She and her husband ran a hotel and
both abused alcohol. However, there was no
doubt of the schizophrenic nature of her illness.
He was exceedingly critical of her, and they often
had physical fights. Because of their work, the
husband could not attend the group, so instead
regular conjoint marital sessions were arranged
to fit in with the couple's working hours. These
were conducted by a female research psycholo-
gist and a male staff social worker. After a while,
it was decided that each therapist would see the
same-sex client on an individual basis in place of
the joint sessions. The couple owned a house in
addition to living in the hotel and were encour-
aged to spend more time in their house, where
they tended to occupy separate rooms. As a con-
sequence, at follow-up they were in low contact.
The husband's criticism had abated considerably
but was still well above the safe level. The patient
had no recurrence of symptoms during the 9
months.
168 Appendix
Case 9 Mother I: C41, 04; F: Cll, 03.
Father - low contact.
The patient, a youth of 20, lived with his parents
and two siblings. His father, an engineer, worked
long hours, and they spent little time together.
He was admitted with a first episode of schizo-
phrenia and was markedly overactive, running
away from the hospital on several occasions. His
mother was one of the most critical relatives in
the study and was also highly overinvolved. She
attended the group regularly, and we also held
several sessions with the patient and both his
parents. Although he recovered from the acute
episode, the patient continued to complain of
apathy and anhedonia throughout the follow-up
period. He improved very slowly, began to at-
tend the hospital as a day patient, and started
playing squash and taking up the piano again. At
follow-up, his mother showed a reduction in both
elements of EE, but not below the crucial levels.
However, the patient was found to be in low con-
tact with her, as well as with his father, and suf-
fered no return of the florid symptoms of his ill-
ness.
Case 10 Mother I: C15, 05; F: C14, 03.
The patient, a woman of 25, lived with her
mother and mother's boyfriend, her father hav-
ing committed suicide in a psychiatric institution.
The patient suffered from Turner's syndrome re-
sulting in dwarfism, as well as schizophrenia. She
was admitted for the third time. Her mother had
always been overprotective, and this had been
intensified with the advent of schizophrenia. Her
mother was also excessively critical. It proved dif-
ficult to get the mother to attend the group and
before she did so, the patient relapsed with a re-
currence of florid symptoms during a weekend
together at a holiday caravan. This occurred 2
months after discharge and was the only relapse
in the experimental group. Following this, her
mother began to attend the group regularly, and
we held several family sessions. At follow-up, the
mother's criticism was unchanged, though she
was considerably less overinvolved. However,
she and the patient were still in high contact.
Case 11 Mother I: C6, 04; F: C7, 04.
Father - low contact.
The patient, a woman of 23, lived with her par-
ents and a brother. Her mother was Italian and
spoke broken English, while the patient was
bilingual. Her father, a caretaker, spent long
hours away from the home. The patient was ad-
mitted for the third time with an acute episode of
schizophrenia. Her mother had few friends, took
the patient with her wherever she went, and
clearly treated her as her principal companion.
They slept in the same bed, while the father was
banished to another room. Her mother joined the
group and several family sessions were held, in-
cluding the brother. Communication in these ses-
sions were held, including the brother. Com-
munication in these sessions was very disrupted,
and we put a lot of effort into clarifying it. It pro-
ved very difficult to engage her father in the
therapeutic program. At follow-up, the mother's
attitudes were unaltered, and she remained in
high contact with the patient. Nevertheless, there
had been no relapse.
Case U Mother I: C27; F: C3, 02.
The patient, a single woman of 41, lived with her
mother. She was admitted for the second time
with an episode of schizophrenia. The patient's
mother was extremely critical of her, but she
spent all her leisure time with her. Her mother
received the education program and was a regu-
lar attender at the group. The group terminated
before the follow-up was due, and a research psy-
chologist saw the mother and daughter every 2
weeks for the rest of the follow-up period. This
patient was in the hospital for nearly 1 year and
appeared at times to be at risk of becoming in-
stitutionalized. At discharge, she was free from
florid symptoms but showed blunting of affect
and no spontaneous speech. She had been work-
ing prior to her admission, but was too handi-
capped to resume work and attended a day hospi-
tal three times a week and a day center twice a
week. Much of the therapeutic work was aimed at
enabling the patient to become more active, to
get up in the morning at a reasonable hour, and to
do some household tasks. In addition, our aim
was to help the mother to encourage her daughter
in these activities and to persevere with her ef-
forts. At the 9-month follow-up, the mother sho-
wed very little criticism although she was a little
more emotionally overinvolved. The patient ex-
perienced no recurrence of florid symptoms dur-
ing the 9-month follow-up period. Her affect re-
mained rather restricted, but she showed much
more spontaneous speech, had resumed some
hobbies such as knitting, helped with chores, and
in particular kept her own room clean and tidy.

Control Families
Case 1 Father I: C8; F: C14.
Mother -low EE.
The patient, a man of 29, lived with his parents.
He was unemployed and was in high contact with
his father, who had retired from his job as a te-
lephonist. The patient was admitted for the sixth
time with an episode of schizophrenia. On dis-
charge, he continued to hear voices, which per-
sisted unchanged throughout the follow-up per-
iod. No therapeutic work was done with the par-
ents, and the father's level of criticism had risen
at follow-up.
Case 2 Mother I: 03; F: 04.
Father - low contact.
The patient, a man of 30, lived with his parents.
He worked as a sales assistant but was in high
contact with his mother as they spent all their lei-
sure time together. His mother was overinvolved
with him, though not at all critical. He developed
a first attack of schizophrenia from which he re-
covered completely. No therapeutic work was
done with his parents, and he suffered a return of
his psychotic symptoms a few months later. At
follow-up, his mother's overinvolvement had
somewhat increased.
Case 3 Wife I: C6; F: CO.
The patient, a man of 33, lived with his wife and
two children. He was admitted for a second time
with an episode of schizophrenia. His wife's crit-
icism was just over the threshold. He recovered
completely with treatment and went back to his
job as a porter. The wife received no professional
help, but when assessed at follow-up was com-
pletely uncritical of her husband. They continued
to be in high contact and he remained well.
Case 4 Husband I: C8; F: unobtainable.
The patient, a woman of 48, lived with her hus-
band and the youngest oftheir eight children. She
was admitted with a second attack of schizophre-
nia, from which she made a full recovery. Her
husband was quite critical of her and had left
home for short periods in the past. Several
months after her discharge, he left her again, but
this time did not return. Some weeks later, she
suffered a recurrence of florid symptoms and was
readmitted.
Appendix 169
Case 5 Sister I: C6; F: C2.
The patient, a woman of24, was the only subject
in this study living with a sibling, her older sister.
She was admitted for the third time with an epi-
sode of schizophrenia, which completely re-
solved. Her sister was given no professional help,
but despite this was much less critical when seen
at follow-up. They remained in high contact, and
the patient experienced no return of her symp-
toms.
Case 6 Mother I: C13, 01; F: refused.
Father I: C38; F: C32.
The patient, a woman of 24, lived with her par-
ents and siblings. Both parents were highly criti-
cal of her, particularly her father. She was admit-
ted for the second time with schizophrenia and
responded well to treatment. She met and mar-
ried another patient while in the hospital and
went to live with him on discharge. However, this
was not a success, and after a while she returned
home. After a few days, her father had a furious
row with her and ordered her to sleep in a shed in
the garden. Sevcral days later, she suffered a
return of auditory hallucinations.
Case 7 Mother I: C5, 05; F: C?, 03.
Father - low contact.
The patient, a young woman of 21, lived with her
parents. Her father, a postman, lived virtually a
separate life from the rest of the family and was in
low contact with the patient. She was admitted
with a first attack of schizophrenia from which
she made a slow but complete recovery. When
her mother was seen for an EE assessment, it
became clear that she was psychiatrically ill and
she was referred to a psychiatrist. He decided that
she was also suffering from schizophrenia, which
had not been previously diagnosed. The mother
was treated for this as an outpatient with a
neuroleptic drug, but her attitudes to her
daughter were not focused on. The patient suf-
fered a recurrence of schizophrenia several
weeks after leaving the hospital. At follow-up,
her mother refused to allow the interview to be
recorded so that critical comments could not be
assessed. However, her overinvolvement with
her daughter was still found to be high, and they
remained in high contact.
170 Appendix
Case 8 Mother I: 05; F: 05.
The patient, a man of 31, lived with his mother.
He had never had a girlfriend, and he and his
mother appeared to be totally absorbed in each
other. He was admitted for the first time with an
episode of schizophrenia, from which he made a
full recovery. He had been unemployed for 1 year
before his admission, and he remained without a
job during the follow-up period. He was reluctant
to attend a day hospital and spent virtually the
whole day confined in their apartment with his
mother. He experienced a return of his symptoms
a few months after discharge. At follow-up, the
mother's overinvolvement was just as intense,
and they were in high contact. A single family in-
terview had been held to persuade him to attend
a day hospital.
Case 9 Husband I: C7, 04; F: refused.
The patient, a woman aged 37, lived with her hus-
band and three children. He was quite critical of
her, but also overinvolved, showing a degree of
morbid jealousy. She was admitted for the second
time with schizophrenia and during her admis-
sion asserted that she had had a sexual encounter
with another patient. The truth of this was un-
certain, but it exacerbated her husband's jeal-
ousy, and he could not get it out of his mind. He
became clinically depressed and received outpa-
tient treatment over the course of 2 months. On
discharge, the patient was still hearing voices, but
these gradually became less insistent, and she did
not suffer any exacerbation of her symptoms over
the 9 months. Her husband refused a follow-up
assessment, but in a brief interview the impres-
sion was gained that his attitudes were un-
changed.
Case 10 Mother I: 04; F: 03.
Father - low contact.
The patient, a youth of 20, lived with both par-
ents. He had never had a girlfriend and spent
much of his leisure time with his mother. He wor-
ked as a post-office clerk and returned to this job
on discharge. His admission was for a first epi-
sode of schizophrenia, and he made a full recov-
ery. No professional help was offered the
mother, whose overinvolvement was a little less
at follow-up, but still above the crucial level.
However, the patient had developed leisure acti-
vities of his own and was in low contact with his
mother. He remained well.
Case 11 Wife I: C8, 02; F: C8, 04.
The patient, a man aged 40, lived with his wife
and two children. She was quite critical and
somewhat overinvolved. He was admitted with a
second episode of schizophrenia, from which he
made an incomplete recovery, continuing to har-
bor a number of delusions when discharged. A
social worker saw his wife at regular 2-weekly
intervals throughout the 9 months, but at follow-
up her overinvolvement had actually intensified.
However, the patient had reduced his contact
with her by retiring to his room, and his conditon
did not worsen during the follow-up period.
Case 12 Husband I: C14, 03; F: C8, 03.
The patient, a woman of26, was admitted for the
second time with an episode of schizophrenia.
Unlike the first attack, this one immediately fol-
lowed the birth of a child and had a prominent
depressive component. Her husband was both
highly critical and overinvolved. Her illness ran a
protracted course, but she finally became free of
symptoms and went home. Her husband received
no professional help, and at follow-up was still
excessively critical and overinvolved. The patient
suffered a relapse with the same mixture of schiz-
ophrenic and depressive symptoms several
months after discharge.
14. Behavioral Family Therapy for Schizophrenia:
Clinical, Social, Family, and Economic Benefits
I. R. H. Falloon
Introduction
Behavioral family therapy represents a
clear departure from the focus on "ab-
normal" family interaction patterns that has
played a prominent role in the psycho-
dynamic and non behavioral family theories
of schizophrenia. Research and clinical
practice, especially with psychodynamic
and systemic methods, have sought to
examine the detrimental effects of family
transactions upon the family member who
is experiencing the symptoms of schizo-
phrenia (Fromm-Reichmann, 1948; Lidz,
Fleck, and Cornelison, 1965; Jackson and
Weakland, 1960; Wynne, Ryckoff, Day,
and Hirsch, 1958). Despite formulations to
the contrary, the paradoxical approaches of
the Milan group (Selvini-Palazzoli, Ce-
chin, Prata, and Boscolo, 1978) have tend-
ed to instigate change through the parents
of disturbed index patients, indirectly im-
plicating their behavior as detrimental. On
the other hand, the behavioral family thera-
py approach proposes that the family is the
basic unit for the promotion of the healthy
functioning of all its members, and it is a
crucial determinant in the recovery and re-
habilitation of those suffering from illness.
This natural care giving function of families
is considered to be the greatest resource
available to the mental health services.
For this reason, the behavioral family
therapy approach seeks first to evaluate the
healthy elements of family communication
in solving everyday problems before assist-
ing the family to enhance their problem-
solving efficiency. Minimal attention is
accorded to patently ineffective problem-
solving strategies. The aim is to build upon
pre-existing family strenghts, rather than
focus on manifest weaknesses. It is assumed
that in very situation each family member is
doing his or her best to cope with the stres-
sors that they are experiencing. Their abil-
ity to achieve ideal solutions and goals is
limited by a broad range of personal and en-
vironmental constraint, such as intelli-
gence, past experience, education, social
learning, interpersonal role patterns, and
finances. It is assumed that they will en-
deavor to fashion the most effective re-
sponses from all the resources within their
group. However, it is recognized that under
the added strain of having one (or more)
family member(s) suffering from a major
illness that the capacity for efficient prob-
lem solving will be eroded, sometimes to
the point where other individuals in the
family may decompensate and become phy-
sically or mentally ill under the burden.
This is particularly so when the symptoms
are persistent and disruptive to the usual
pattern of family interaction. Thus, it may
not be sufficient for the family of a chron-
ically ill person to function as an average
problem-solving unit; indeed, they may
need to achieve much greater than "nor-
mal" family functioning to preserve their
own health as well as restore the health of
the sick member. This paradigm of effective
problem solving in families alters, in a fun-
damental fashion, the way we assess family
functioning and subsequently formulate
our therapeutic interventions. The goal is
not merely to ameliorate symptoms through
reducing stress in the environment, but also
the enhance the functioning of each mem-
ber of the family unit.
172 I. R. H. Falloon
Behavioral Analysis of Family Functioning
The first step in behavioral family therapy
involves an assessment of the strenghts and
weaknesses of the family as a problem-solv-
ing unit. Particular note is made of the pre-
cise manner in which they tackle the prob-
lems of everyday life together, as well as
their responses to significant life events.
Life problems include the personal goals of
individual family members.
Initially, family members are interviewed
individually to establish their unique assets
and deficits, their personal life goals, and
the assistance they obtain from other family
members in their attempts to achieve these
goals. Their feelings and attitudes toward
other family members are discussed. The
Camberwell Family Interview (CFI) provi-
des a useful adjunct to this individual analy-
sis, although in its abbreviated form
(Vaughn and Leff, 1976a) it focuses on re-
lationships with the index family member
and his or her behavioral disturbance.
The second step in behavioral analysis of
a family unit involves an assessment of the
relationships between the various goals and
problems of individual family members. It
is clear that the aspirations of individual
family members may conflict at times with
the wishes and values of other family mem-
bers. An insecure mother may not want her
son or daughter to leave her and go to work
or a day center every day.
Strictly religious parents may be upset by
social activities that involve drinking or
sexual contact. Such conflicts with individ-
uals' goals for themselves and one another
may undermine progress unless they can be
resolved by effective problem solving.
Seemingly positive changes, from a thera-
pist's perspective, may be perceived dif-
ferently by the family as a whole. It is,
therefore, crucial that the therapist can for-
see these potential problems at the onset of
intervention and can help the family to seek
creative solutions at an early stage. This
analysis of how the family fits together as a
system is described as functional analysis by
behavioral therapists. It has much in com-
mon with the systemic analysis of ap-
proaches based on systems theory.
The third step involves observation of the
family having a problem-solving discussion.
This is the first time the family is brought
together as a group. Previous assessments
of individual family members may have re-
vealed reports of problem-solving efforts,
but such reports provide only a sketchy
view of family discussions from highly sub-
jective viewpoints. We have employed
several methods to obtain naturalistic ob-
servation of family problem-solving behav-
ior. Perhaps the best has been obtained
from observing the family in their home.
Occasionally this is not feasible, and clinic-
based analogues have been used. It is im-
portant to realize that families may not pro-
vide valid evidence of their functioning
within the constraints of the clinic and that
at least one home visit is crucial in their as-
sesment. The problem-solving test that we
have found most useful involves carefully
choosing a "hot issue" that concerns all
family members and leaving them to at-
tempt to resolve the problem within a 10-
min period. This time interval is often too
brief to allow detailed planning to be achie-
ved but provides a surprisingly good insight
into the structure and strategies the families
employ in their discussions. The strengths
and weaknesses of the family's communica-
tion and problem-solving skills can be relia-
bly coded for clinical and research analysis.
Once these steps have been undertaken,
often over several weeks, and up to 20 h of
intensive assessment, the behavioral family
therapist is in a position to formulate a
management plan that is uniquely structur-
ed to enhance the problem-solving efficien-
cy of each family unit. The current efforts of
families to cope with their problems and
achieve their goals are documented, and
these serve as the basis upon which to con-
struct increasingly effective problem solv-
ing. This base line behavioral analysis is
merely the initial phase of the program, and
throughout the course of treatment similar
assessment of family functioning continues,
enabling strategies to be modified as new
information is gathered. The therapist is
constantly addressing his or her own prob-
lem of "How can I assist this family to con-
duct their own problem-solving functions
more efficiently?"
 Behavioral Family Therapy for Schizophrenia
Treatment Program
The most important component of the tre-
atment program is the behavioral analysis.
The initial and continuous pinpointing of
specific deficits of family problem solving
enables the therapist to assist the family to
achieve gradually more fluent strategies for
coping with the wide variety of difficulties
they encounter, while at the same time con-
tinuing to promote their personal life goals.
Many families show considerable compe-
tence in problem solving, but through a lack
of understanding of the nature of schizo-
phrenia may seek solutions that provoke
excessive stress for the index patient. Other
families may have difficulties conducting a
calm discussion within their family group.
One person may assume dictatorial control
and prevent the flow of ideas and sugges-
tions from others; another may not listen to
what others are saying; in many cases, fam-
ilies do not sit down in a relaxed setting,
free from distractions to discuss important
issues; still other families who reach agree-
ment on what appears to be the optimal so-
lution to resolving a problem fail to imple-
ment that solution owing to inadequate
planning. There are no familial defects that
are specific to schizophrenia. However, the
unusual nature of the disorder and the
stress that ensues for patients and their
caregivers tend to provoke inefficiencies of
problem solving.
The major interventions employed in the
behavioral family therapy program can be
divided into four broad categories:
1. Education about schizophrenia
2. Communication skills training
3. Problem-solving skills training
4. Behavioral strategies for specific prob-
lems
Education About Schizophrenia
Although the mutual sharing of information
about the nature, course, and treatment of
schizophrenia is a continuous process, two
sessions at the onset of family therapy are
devoted to discussion about the diagnosis,
etiology, management, course, and drug
and psychosocial treatment of the disorder.
These initial sessions provide a rationale for
the subsequent family therapy and for the
173
continuing drug treatment that is carefully
monitored throughout. Index patients play
a crucial role in this education. They are en-
couraged to describe their own experiences
of schizophrenia and its management to
their families. Usually, this is the first oc-
casion on which the nature of the patient's
symptoms has been discussed among the
family. The patient assumes the "expert"
role, although at times other family mem-
bers may report similar episodes of schizo-
phrenic symptoms in the past or even the
present. These sessions are delayed until
the index patient is stabilized after an acute
episode and is usually able to discuss his ex-
periences in an objective manner.
A questionnaire is administered before
and after these initial sessions to assess the
amount of information acquired and to note
areas where further education may be ne-
cessary. Prior to this education, half of all
patients and family members knew the di-
agnosis of the patient's condition; after-
ward, 95% agreed with the primary diagno-
sis of schizophrenia. A 50% increase in
scores on a multiple-choice questionnaire
about schizophrenia has been observed
consistently (McGill, Falloon, Boyd, and
Wood-Siverio, 1983).
Communication Training
At times of crisis, ineffective patterns of
communication substantially impede cop-
ing efforts and contribute to the stress that
may trigger exacerbations of schizophrenia.
Effective communication is the basis for ef-
ficient problem solving. The major com-
ponents of family communication that are
assessed involve the ability to listen em-
pathically to one another when discussing
problem issues, mutual communication of
positive and negative feelings, and making
constructive requests for behavior change.
The nonverbal and verbal components of
interpersonal communication are addres-
sed, particularly where major deficits
detract from the clarity of such expression.
Improved communication is shaped
through skills-training procedures, where-
by brief segments of interaction are re-
hearsed repeatedly with instructions,
174 I. R. H. Falloon
modeling, social reinforcement, and per-
formance feedback from the therapist and
family members. Homework assignments
are employed so that family members con-
tinue to practice these skills outside the
therapy sessions.
No specific attempts are made to prevent
family members from expressing criticism,
hostility, rejection, or overinvolved com-
munication in their discussions. These pat-
terns of expression, directed toward the in-
dex patient, have been considered predic-
tive of symptom exacerbations (Vaughn
and Leff, 1976b). Nevertheless, effective
training of modes of communicating posi-
tive and negative feelings is likely to reduce
the frequency with which these potentially
destructive comments are made. Families
are encouraged to express their feelings of
anger, frustration, and guilt within a con-
structive problem-solving framework. In
this way, their negative expressions become
the initial steps in identifying and resolving
difficult problems, thereby relieving stress
in the family.
Problem-Solving Training
Families are encouraged to adopt a struc-
tured approach to their problem-solving
discussions. This involves arranging specific
times when the whole family can meet, free
of distractions, to discuss important issues.
The frequency of such family meetings,
their duration, and the choice of topic for
discussion is left to the discretion of the
family. However, they are advised to meet
at least once a week, to limit the time spent,
and to choose relatively straightforward
problem issues for their initial meetings.
The family therapy sessions examine the
progress of their family discussions and pro-
vide a workshop for enhancing their ef-
fectiveness. A six-step model of problem
solving is advocated as a guide for the fami-
ly meetings. The six steps include:
1. Identify a Specific Problem. This step
entails active listening to each person's de-
scription of the problem, seeking clarifica-
tion when necessary, and avoiding giving
premature advice or reassurance. The aim
is to be able to specify the problem in clear
operational terms that are readily under-
stood by all family members.
2. List Alternative Solutions. Once the
problem has been specified, a brain-storm-
ing approach is used to generate a list of po-
tential solutions. Discussion about the rela-
tive merits of each solution is postponed,
and both "good" and "bad" ideas are re-
corded. Reticent family members are en-
couraged to express their thoughts, and
every idea is acknowledged by adding it to
the list.
3. Discuss Merits of Each Solution. Each
solution is discussed in tum. The advan-
tages and disadvantages of each are high-
lighted.
4. Choose the Optimal Solution. The fam-
ily chooses that solution (or combination of
solutions) that best fits the needs of the
problem and the resources of the family.
5. Plan How to Implement the Solution.
Detailed plans are drawn up to map out the
steps through which the solution will be im-
plemented. Strategies for ensuring that the
solution is implemented smoothly are de-
vised. Major roadblocks are anticipated,
and ways of coping with these and other im-
portant consequences are discussed. At
times, families may rehearse the steps
among themselves in a dry run; often
merely talking through the procedures step-
by-step is sufficient.
6. Review All Efforts. Any genuine at-
tempt to implement the plan is acknowl-
edged and praised, even if a successful out-
come was not achieved. The family is en-
couraged to review their efforts and to use
all knowledge gained to construct a more
effective plan. The notion that successful
problem solving requires persistent and re-
peated efforts is clearly instilled.
The therapy sessions enable families to
practice their problem solving under the
supportive guidance of the therapist. Initi-
ally, and at times of crisis, he or she may
need to provide clear direction for the fami-
ly. Guidesheets that list all the steps of
problem solving are used during the ses-
sions and in family discussions at home.
One family member "chairs" the discus-
sion, recording notes on the sheets of each
problem-solving step. As soon as the family
begins to master the approach, therapist
involvement is withdrawn.
 Behavioral Family Therapy for Schizophrenia
Behavioral Strategies for Specific Problems
Few families have an adequate repertoire of
coping behavior to manage all the problems
that arise in the course of schizophrenia.
Although the problem-solving structure
helps facilitate creative resolution of diffi-
cult problems, there are times when ad-
ditional professional assistance may be em-
ployed. Behavioral strategies that have
been validated in research studies to assist
with symptoms of anxiety, depression, ob-
sessive-compulsive disorders, persistent de-
lusions or hallucinations, medication com-
pliance, or enhancing social and sexual inter-
action are taught within the family prob-
lem-solving context. In addition to teaching
methods of dealing more efficiently with
symptomatic states, families may learn a
broad range of general management skills,
such as operant reinforcement strategies,
contingency contracting, role playing limit
Controlled Outcome Study
To assess the efficacy of behavioral family
therapy as the core component of a family
management program for schizophrenia a
controlled outcome study was designed
(Falloon, Boyd, McGill, Razani, Moss, and
Gilderman, 1982; Falloon, 1985). It was not
considered feasible to compare the pro-
gram with a control treatment of drug ther-
apy alone because almost every patient with
schizophrenia who attended the psychiatric
outpatient service on this basis tended to
drop out unless more extensive therapeutic
contact was established. The most appro-
priate comparison appeared to be a patient-
oriented management program that pro-
vided a similar level of intensive support,
with optimal medication, comprehensive
care management, and crisis intervention,
including extensive family support, but
without the specific focus on enhancing
family problem-solving functions.
The study involved the random assign-
ment of 39 index patients and their families
to either family therapy of individual-
oriented therapy after they had been stabi-
lized on optimal doses of medication after
an acute episode of schizophrenia.
175
setting, and skill training.
The burden of management of severe,
persistently handicapped individuals can be
substantially reduced when families can
learn to make effective use of supportive
persons and agencies in their social net-
works.
At all times, the therapist remains clearly
aware that his role is that of a facilitator of
family problem solving and that the goal of
therapy is to enhance the family's own ef-
forts in the most efficient manner possible.
Thus, additional behavioral strategies are
not introduced until the families have failed
to achieve a satisfactory outcome with their
own attempted strategies.
A detailed manual of the behavioral
family therapy methods that have been de-
veloped for the community care of schizo-
phrenia are provided in Family Care of
Schizophrenia (Falloon, Boyd, and McGill,
1984).
The initial phase of treatment in the study
involved weekly sessions for 3 months,
tapering to biweekly sessions until 9
months, when once-monthly sessions were
conducted until the 2-year point. A total of
40 sessions were provided in this manner
over the 2 years.
Therapists were assigned a similar num-
ber of cases in each condition. They were
available to provide additional sessions at
all times throughout the study and covered
for emergency calls on a 24-h-a-day basis.
Case management by the same therapist
was continued throughout any periods of
hospital admission, day treatment, or resi-
dential care.
An extensive battery of dependent mea-
sures of clinical, social, and family morbi-
dity were administered at base line and 3,9,
and 24 months. In addition, monthly ratings
of clinical factors were made throughout
the study. Ongoing data regarding life
events and family coping were collected in
biweekly interviews with family members.
Information of potential relevance to the
monetary costs and benefits of treatment
was recorded throughout the study.
176 I. R. H. Falloon
Patient Selection
A syndrome diagnosis of schizophrenia or
delusional psychosis according to PSEI
CATEGO criteria (Wing, Cooper, and Sar-
torius, 1974) was an initial requirement for
selection. Most patients had symptoms of
"nuclear" schizophrenia, indicating the
presence of one or more of Schneider's
first-rank symptoms. One-third had expe-
rienced their first episode of schizophrenia
prior to entry to the study. Another one-
third were long-standing cases.
Age was restricted to between 18 and 45
years, and patients were required to be liv-
ing with (or in close daily contact with) at
least one natural parent.
At least one member of the family house-
hold was required to display high levels of
expressed emotion (EE) in the Camberwell
Family Interview or to display other clear
evidence of extreme family tension where
high EE was absent.
Three patients withdrew from treatment
prior to the end of the 1st month leaving 18
in each condition. These two randomly as-
signed groups were well matched on socio-
demographic and clinical variables. One
further patient withdrew after 10 months,
but continued aftercare in a neighborhood
clinic and was reassessed after 2 years.
Treatment Conditions
Certain aspects of treatment were identical
for all patients, regardless of the treatment
condition. Special efforts were made to
maintain patients on optimal doses of
neuroleptic medication throughout the 2-
year program. Owing to its relative lack of
Results
Thirty-six patients (18 in each condition)
completed the initial 9 months of intensive
management. In a few cases, pharmaco-
therapy was not optimal for periods, but in
every case this issue was resolved, obviating
the need for withdrawal of cases from the
data analysis.
Clinical Morbidity
Patients experienced a similar number of
severe side effects and the availability of
laboratory assay of plasma levels, chlor-
promazine was our drug of choice. Poor
compliance with oral preparations neces-
sitated the use of long-acting intramuscular
fluphenazine decanoate at times. Where
chlorpromazine was poorly tolerated, flu-
phenazine hydrochloride, haloperidol, and
thiothixene were used.
Pharmacotherapy was provided by two
psychiatrists and a clinical pharmacist at
monthly clinic appointments. A research
nurse collected blood samples for plasma
assays. The pharmacotherapists remained
blind to the psychosocial treatments pa-
tients received and conducted monthly clin-
ical assessments. On the basis of these as-
sessments, they continuously sought to
lower the dosage of neuroleptics to provide
the minimal dose that was sufficient to pre-
vent the exacerbation of symptoms of schiz-
ophrenia. When signs of imminent de-
terioration appeared, the pharmacothera-
pists increased the doses according to need.
Four 1-h sessions of rehabilitation coun-
seling were provided for each patient by an
experienced rehabilitation counselor. A
range of vocational and psychosocial re-
habilitation programs was made available
to patients in both conditions. Once again,
the counselor was blind to the treatment as-
signment.
Family therapy sessions during the first 9
months were conducted in the patients'
homes. Individual sessions were usually
conducted in the clinic. All missed sessions
were rescheduled so that each patient re-
ceived the same minimal amount of thera-
pist contact.
exacerbations of psychopathology in each
treatment condition during the first 9
months of management: a total of 40 epi-
sodes in the family condition and 43 with in-
dividual therapy. However, of the 18 epi-
sodes of schizophrenia experienced with
family management only, 3 were consider-
ed of major significance. Family cases had
10 further major episodes of nonschizo-
phrenic symptoms, thereby totaling 13
major exacerbations. This contrasted with
 Behavioral Family Therapy for Schizophrenia
130
120
100
80
60
40
20
Fig. 1. Exacerbations of all symp-
toms (0-9 months).
29 major episodes in the individual condi-
tion, 21 of which involved florid schizophre-
nia . Furthermore , the duration ofthe major
episodes that occurred in the family mana-
gement averaged less than half that of indi-
vidual management: 30 days vs 69.
Figure 1 summarizes the differences in
psychopathology between the two condi-
tions. Individual management maintained
symptom stability about half the time ,
whereas family management cases remain-
ed stable for about 80% of the time. These
trends were maintained during the less in-
tensive follow-up phase.
These clinical observations were sup-
ported by rating scale data provided by
blind raters and subjected to a repeated
measures analysis of covariance. Monthly
serial ratings of each patient's specific schiz-
ophrenic target symptoms were signif-
icantly lower for family management
(P < 0.003) throughout the 24 months (see
Figure 1). A similar trend was observed for
the BPRS "schizophrenic thought" factor
that combined ratings of thought disturb-
ance, delusions , and hallucinations
(P < 0.04). The greater stability with fami-
ly management was supported by a com-
177
i
DAYS
Other disorders
Anxiety
:: Depression
Schizophrenia
Individual Family
management management
parison between the coefficients of varia-
tion of the total BPRS scores of each patient
over the first 9 months. Family cases sho-
wed significantly lower coefficients of varia-
tion than individual cases (P < 0.009).
To date, the focus of our assessment has
been on florid symptoms. It is possible that
the lowered morbidity associated with fam-
ily management was accomplished at the
expense of an increase in the deficit , or neg-
ative, symptoms of schizophrenia. This did
not appear to occur. Indeed, there was a
trend for deficit symptoms to be reduced
more with family management. The BPRS
"withdrawal" factor, which was comprised
of the items "motor retardation," "blunted
affect," "emotional withdrawal," and
"grandiosity" (scaled negatively), favored
family management (P < 0.09) .
Qualitative blind ratings of psycho-
pathology using the PSE interview showed
no differences in the neurotic symptoms in
the two conditions. At 9 months, ten (56%)
of the family-managed cases were in full re-
mission from schizophrenic compared to
three (17%) of the individual group. At 2
years, 12 (67%) of the family cases showed
no evidence of schizophrenia, with 9 (50%)
178 I. R. H. Falloon
free from any psychiatric symptoms. In con-
trast, 14 (83%) of the individual cases sho-
wed persistence of schizophrenic or para-
noid symptoms.
Thus, there was substantial evidence that
family management was associated with in-
creased clinical stability related to all
sources of psychopathology. Furthermore,
there was a trend toward a gradual reduc-
tion of clinical morbidity to the extent that
half the patients who received family man-
agement were free of any signs of mental ill-
ness at the end of 2 years. As a result of this
reduction in clinical morbidity, substantial-
ly less hospital care was required. Over the
2 years, family cases averaged 4 days in the
hospital and individual cases 23 days.
Social Morbidity
The determinants of social functioning are
varied, not all associated with illness fac-
tors. Both the family and individual thera-
pies endeavored to enhance the social per-
formance of the index patient. This was at-
tempted through direct therapist-patient
goal setting and planning in the individual
condition and through teaching the family
as a whole how to use their problem-solving
discussions to address similar issues for the
index patient as well as other family mem-
bers.
In the initial phase of psychosocial treat-
ment, patients with good premorbid social
adjustment and schizophrenia of recent
onset tended to have less difficulty re-
establishing work and social activities.
Eight patients (44%) in each management
condition were paid for work during the
first 9 months. All had been working at the
onset of their index episode. Over 24
months, family cases were engaged in func-
tionally useful daily activities (i.e., full- or
part-time work, sheltered workshops, ed-
ucational or training programs) for an aver-
age of 12.6 months. This compared with 7.2
months for individual cases. During the 24
months prior to the study, patients subse-
quently assigned to family management had
been occupied in similar constructive activ-
ity for 7.0 months, and those assigned to in-
dividual management averaged 7.4 months.
Family management was, therefore, as-
sociated with a substantial increase in func-
tional performance; individual manage-
ment succeeded in maintaining previous
levels of functioning.
The results of rating scales of social func-
tioning showed a similar pattern, although
some deterioration of social adjustment was
evident during the first 9 months of individ-
ual management (Falloon, McGill, Boyd,
and Pederson, in press b). On the self-
report measure (SAS-SR) (Weissman et
al., 1978), an analysis of covariance, using
the base line scores as covariates, greater
overall social adjustment was reported with
family management (P < 0.02). Significant
advantages were seen for the family ap-
proach on scales of leisure activity (P <
0.01), and family relationships (P < 0.04).
Seven items from the SBAS social per-
formance scales were examined (Platt,
Weyman, Hirsch, and Hewett, 1980).
These were derived from blind assessor in-
terviews with key family members. Figure 2
summarizes these ratings of social perform-
ance as well as relatives' dissatisfaction with
the impairment they observed. Small, but
significant, gains in social performance
were evident in the first 9 months for family
management cases, whereas some overall
deterioration in social functioning was evi-
dent in half the individually managed pa-
tients at 9 months, with improvement
occurring in the 2nd year of the program.
An analysis of covariance, using the base
line scores as covariates, showed significant
between-group differences on the combin-
ed social performance score at 9 and 24
months (P < 0.02). The interaction be-
tween the groups over time was not signif-
icant.
The relatives' dissatisfaction with social
performance tended to parallel their re-
ports of impairment. During the first 9
months, family management appeared to
reduce dissatisfaction in several cases
where high levels of impairment remained.
This suggested that the family approach
may have assisted in instilling more tolerant
attitudes in the face of persisting handicaps.
Individual management relatives tended
toward less tolerant attitudes. The com-
bined dissatisfaction ratings over the 24
months significantly favored family man-
agement (P < 0.0002). It was possible that
more tolerant attitudes accounted for re-
 Behavioral Family Therapy for Schizophrenia 179

• f l .·· ·
% Family management Individual management
impairment dissatisfaction
100: fimljPairm*ent dissatisfaction
Household 5
tasks
...:;:;:;:;
~... ;:;:;:;:
.. ....
,
~~~~~ ~----~~~~~

o 9 mo 0 9 mo 0 9 mo o 9 mo

100~
oil •
Leisure
activity 50 *
o 9 mo o 9 mo o 9 mo o 9 mo

Work/study
activity

o 9 mo o 9 mo o 9 mo o 9 mo

Decision
making

.u:=.-**
o 9 mo
-----Lt.~~_
0 9 mo
I
0 9 mo

Friendliness/
affection

Everyday
conversation

.~'~I
o 9 mo o 9 mo o 9 mo

100 ~
ti
Relationships ~ ......
outside family s:~
....
o 9 mo o 9 mo o 9 mo o 9 mo
indicates proportion of family-treated patients significantly less
than individually treated, p < .01
* indicates p < .05, Fisher's Exact Test

Fig. 2. Social role impairment and parental dissatisfaction with index patients' performance.
180 I. R. H. Falloon
duced reporting of social impairment. The
consistency of clinician, patient, and rela-
tive reports makes this unlikely, as well as
the fact that the relatives reported continu-
ing impairment in those same areas of func-
tioning where their dissatisfaction had be-
come minimal. Therefore, despite the dif-
ficulties in assessing social functioning,
there appeared to be good evidence for the
superiority of family management in reduc-
ing the social morbidity associated with
schizophrenia.
Family Morbidity
We have already noted that family toler-
ance toward social deviance may have been
increased by the family approach. Further
evidence for benefits to the family members
who received behavioral family therapy in-
cluded significant reduction' in the distress
they expressed toward the patients' endur-
ing clinical and social morbidity and reduc-
tion in the overall burden they associated
with having the index patient living in the
household (Falloon, Doane, and Pederson,
in press c). Figure 3 summarizes the
changes in family burden in both manage-
ment conditions. A repeated measures anal-
100 %
Family management
50
o L..-_KU.
Baseline 9 Months 24 Months
100
Individual management
Baseline 9 Months 24 Months
ysis of covariance, using the base line scores
as covariates, supported the superiority of
the family approach (P < 0.0005).
Benefits in terms of reductions in the clin-
ical and social morbidity of family members
other than the index patient were more dif-
ficult to demonstrate. Although a few fami-
ly members were suffering from significant
levels of morbidity at base line, the overall
levels of distress and burden did not appear
to induce measurable impairments in the
majority of family members. In those family
members where notable morbidity was evi-
dent, family therapy appeared more bene-
ficial. Consistent with the trend toward
slightly increased morbidity for individual
management at 9 months, more family
members in this condition had significant
impairment at this point.
It is important to note again that behav-
ioral family therapy does not aim primarily
to reduce clinical morbidity; rather, the fo-
cus is on achievement of positive functional
changes. The scales used to assess change in
this study were restricted to measuring im-
paired functioning. As a result, positive
functional goals that were achieved in the
lives of our patients and families were less
evident in the rating scale data. Approxi-
Fig. 3. Percentages of families reporting mod-
erate or severe subjective burden attributed to
caring for the index patient.
 Behavioral Family Therapy for Schizophrenia 181
mately half the problem-solving sessions related to similar goal achievement. The
with families concerned issues relating to enhanced quality of life that was achieved
the personal goals of family members other by this approach is only dimly portrayed by
than the index patients, and the majority of our standardized measures of morbidity.
issues that concerned index patients were Nevertheless, it is clear that family-oriented

I!?
.!l1
(5
"0
"0
I/)
"0
C
tOO
'":>
I/)

=
0

.!;
I/)
1ii
0
u

Family management IndIvidual management

Fig. 4. The cost of treatment.

182 I. R. H. Falloon
management was associated with signifi- main source of added costs was the in-
cant, stable benefits in terms of clinical, creased need for crisis services, particularly
social, and family functioning and that hospital care for individual cases.
these benefits were more pronounced than Global indices of clinical, social, and
those achieved through the more usual pa- family morbidity were used to produce a 7-
tient-centered approach. point standardized index of effectiveness
that ranged from 0 = no therapeutic effects
Economic Benefits
in any area to 6 = high therapeutic effec-
tiveness in all areas. The comparative cost
A crucial limitation of many new ap- efficiency of the two management ap-
proaches to health servicves is the cost, proaches was determined by assessing the
which often exceeds that of previous ap- overall cost per unit of therapeutic effec-
proaches, and despite the advantages of im- tiveness. Individual management cases
proved effectiveness, restricts general im- averaged 2.1 units of therapeutic effective-
plementation. In this study, all direct and ness during the 1st year, at a cost of $5167
indirect costs of community management to per unit. Family cases averaged 4.0 units, at
patients, families, health, welfare, and a cost of $2220 per unit. Thus, in broad eco-
community agencies were recorded nomic terms, family management doubled
throughout the 2 years each patient was in the efficiency of the service to achieve the
the study. These costs were tabulated and goals of reduced clinical, social, and family
compared for family and individual man- morbidity. If this cost efficiency can be re-
agement (Cardin, McGill, and Falloon, plicated in everyday clinical practice, family
1985). The results of this cost analysis are management appears to be a highly attrac-
summarized in Figure 4. tive proposition for widespread deploy-
It is evident that the overall costs of the ment in the health services.
family approach were approximately 20%
less than those of individual management
during the 1st year of the study, and this Therapeutic Factors in
despite a hefty differential to account for Family Management
the added time and expense of traveling to The comprehensive nature of the manage-
the homes for family therapy sessions. The ment approaches we used did not allow us

10

8
Q)
Cil 7
u
en
Ol
c: 6
'0.
0
()
5

0 1 1-3 4-6 7-9

Follow-up
I
10-12
Fig. 5. Mean coping ratings during year 1.

Months

o Individual treatment
• Family treatment
 Behavioral Family Therapy for Schizophrenia
to attribute the added benefits associated
with family management to any specific fac-
tor. However, detailed data collected dur-
ing the course of the study enabled us to
consider the probable effects of a number of
potential factors on the outcome (Falloon,
1985).
First, there was little support for the pro-
position that random assignment had pro-
duced a bias that favored family man-
agement. The two conditions were well
matched on prognostic factors. Family
management patients tended to comply
more readily with medication, but skillful
deployment of depot preparations and pro-
grams to improve adherence would seem to
have counteracted the potential clinical
morbidity that may have accrued from in-
adequate pharmacotherapy. Overall there
was a tendency for family cases to ingest
lower amounts of neuroleptic medication
than those receiving individual manage-
ment. The impact of life events that were
not clearly associated with the patient's
condition appeared similar in both condi-
tions. Therapists favored employing family
management in general, but when faced
with specific patients and their families,
showed no differences in their attitudes to-
ward the patient and his assigned form of
management. Conducting family therapy
sessions in the home on a regular basis con-
tributed to the very efficient delivery of this
treatment, but rescheduling of missed ap-
pointments and home sessions when neces-
sary enabled both conditions to receive a
similar amount of scheduled therapy.
Indeed, extra therapy time and effort was
required to prevent clinical and social mor-
References
Anderson, C. M., Hogarty, G. E., and Reiss, D.
J. (1980). Family treatment of adult schizo-
phrenic patients: A psycho educational ap-
proach. Schizophrenic Bulletin, 6, 490-505.
Cardin, V., McGill, C. W., and Falloon, I. R. H.
(1985). An economic analysis: Costs, benefits,
and effectiveness. In Falloon, I. R. H. (ed.),
Family management of schizophrenia: Clinical,
social, family and economic benefits. Balti-
more: Johns Hopkins University Press.
Doane, J. A., Falloon, J. R. H., Goldstein, M. J.
and Mintz, J. (1985). Parental affective style
and the treatment of schizophrenia. Archives
183
bidity in individual management cases.
It was tentatively concluded that the
major source of the increased efficacy of
family management probably resided in the
behavioral family therapy component. Spe-
cific changes resulted from the family edu-
cation (McGill et al., 1983) and coummuni-
cation and problem-solving training
(Doane, Falloon, Goldstein, and Mintz,
1985) that were the core ingredients of
behavioral family therapy. The coping ef-
ficiency of the family unit was specifically
enhanced by the behavioral family therapy.
Figure 5 shows a highly significant linear in-
crease in the effectivieness of the family
unit to resolve the wide range of stressful
situations with which they were faced over
the three trimesters of the intensive phase
of therapy. No such change in coping was
evident for family units in the individual
condition. This increase in coping remained
stable after the intensive phase of behav-
ioral family therapy had been completed.
This improved ability to resolve stress while
achieving functional goals was attributed to
the behavioral family therapy. Further dis-
cussion of the process through which ef-
fectiveness of the family therapy appeared
to have been mediated is provided in Chap.
15.
We have concluded that a family man-
agement approach that combines optimal
pharmacotherapy with comprehensive care
management, family education, and behav-
ioral family therapy is a cost-effective
method of producing sustained reductions
in the clinical, social, and family morbidity
associated with schizophrenia.
of Geneal Psychiatry, 42, 34-42.
Falloon, I. R. H. (ed.) (1985). Family manage-
ment of schizophrenia: Clinical, social, family
and economic benefits. Baltimore: Johns Hop-
kins University Press.
Falloon, I. R. H., Boyd, J. L., and McGill, C. W.
(1984). Family care for schizophrenia: A prob-
lem-solving treatment for mental illness. New
York: Guilford Press.
Falloon, I. R. H., Boyd, J. L., McGill, C. W.,
Razani, J., Moss, H. B., and Gilderman, A. M.
(1982). Family management in the prevention
of exacerbations of schizophrenia: A control-
184 I. R. H. Falloon
led study. New England Journal of Medicine,
306,1437-1440.
Falloon, I. R. H., Boyd, J. L., McGill, C. W.,
Williamson, M., Razani, J., Moss, H. B., Gil-
derman, A. M., and Simpson, G. M. (in press
a). Family versus individual management in the
prevention of morbidity of schizophrenia. I.
Clinical outcome of a two-year longitudinal
study. Archives of General Psychiatry.
Falloon, I. R. H., McGill, C. W., Boyd, J. L.,
and Pederson, J. (in press b). Family versus
individual management in the prevention of
morbidity of schizophrenia: II. Social outcome
of a two-year longitudinal study. Psychological
Medicine.
Falloon, I. R. H., Doane, J. A., and Pederson, J.
(in press c). Family versus individual manage-
ment in the prevention of morbidity of schizo-
phrenia: III. The adjustment of the family unit.
Psychological Medicine.
Fromm-Reichmann, F. (1948). Notes on the de-
velopment of treatment of schizophrenics by
psychoanalytic psychotherapy. Psychiatry, 11,
263-273.
Jackson, D. D., and Weakland, J. H. (1961).
Conjoint family therapy: Some consideration
on theory, technique, and results. Psychiatry,
24 (2) (Supplement), 30-35.
Leff, J., Kuipers, L., Berkowitz, R., Eberlein-
Vries, R., and Sturgeon, D. (1982). A control-
led trial of social intervention in the families of
schizophrenic patients. British Journal of
Psychiatry, 141, 121-134.
Lidz, T., Fleck, S., and Cornelison, A. (1965).
Schizophrenia and the family. New York: In-
ternational Universities Press.
McGill, C. W., Falloon, I. R. H., Boyd, J. L.,
and Wood-Siverio, C. (1983). Family educa-
tional intervention in the treatment of schizo-
phrenia. Hospital and Community Psychiatry,
34, 934-938.
Platt, S., Weyman, A., Hirsch, S., and Hewett,
S. (1980). The social behaviour assessment
schedule (SBAS): Rationale, contents, scoring
and reliability of a new -intervention schedule.
Social Psychiatry, 15, 43-55.
Selvini-Palazzoli, M., Ceechin, A., Prata, G.,
and Boscolo, L. (1978). Paradox and counter-
paradox. New York: Jason Aronson.
Vaughn, C. E., and Leff, J. P. (1976a). Theinflu-
ence of family and social factors on the course
of psychiatric illness: A comparison of schizo-
phrenic and depressed neurotic patients. Brit-
ish Journal of Psychiatry, 129, 125-137.
Vaughn, C. E., and Leff, J. P. (1976b). The
measurement of expressed emotion in families
of psychiatric patients. British Journal of Social
and Clinical Psychology, 15, 157-165.
Weissman, M. M., Prusoff, B. A., Thompson,
W. D., Harding, P. S., Myers, J. K. (1978). So-
cial adjustment by self-report in a community
sample and in psychiatric outpatients. Journal
of Nervous and Mental Disease, 166,317-326.
Wing, J. K., Cooper, J. E., and Sartorius, N.
(1974). The measurement and classification of
psychiatric symptoms. London: Cambridge
University Press.
Wynne, L. C., Ryckoff, I., Day, J., and Hirsch,
S. (1958). Pseudo-mutuality in the family rela-
tions of schizophrenics. Psychiatry, 21,
205-220.
15. Impact of Family Intervention Programs on
Family Communication and the Short-Term
Course of Schizophrenia
M. J. Goldstein and A. M. Strachan
Introduction
Recent interest in the role of the family in
the course of schizophrenia has focused on
aspects of the family environment which in-
fluence the short-term course of the disor-
der once an episode of schizophrenia has
occurred. One research strategy has been to
measure aspects of the family environment
and then follow the natural course of the pa-
tient's psychiatric and social functioning.
This strategy has successfully demonstrated
that relatives' expressed emotion (EE)
about the patient is predictive of relapse in
both British and American samples (Chaps.
2, 8). Further, it has been demonstrated
that this measure of emotional attitudes re-
lates to direct observational measures of
communication with the patient (Chap. 4).
Recent Family Intervention Programs and the Short-term Course
of Schizophrenia
All four studies share a number of common
features. First, all were based on the dia-
thesis-stress model of schizophrenia in
which a biological predisposition interacts
with environmental stressors including both
stressful life events and family pressures.
Secondly, all patients were kept on main-
tenance pharmacotherapy throughout the
aftercare period. The central research que-
stion in these programs was whether the ad-
dition of a family-based therapeutic pro-
gram to regular pharmacotherapy could re-
duce the probability of relapse in the schizo-
phrenic patient.
The first of these studies was carried out
by Goldstein, Rodnick, Evans, May, and
Steinberg at UCLA (1978). The sample
Now, if there are stress-producing as-
spects of the family environment which in-
crease the likelihood of a relapse, then inter-
vention programs designed to modify these
aspects of the family environment should
produce a more favorable course. This has
been the rationale underlying four major
studies, three of which have been com-
pleted (Chaps. 13,14; Goldstein, Rodnick,
Evans, May, and Steinberg, 198) and one
which is close to completion (Chap. 12).
This chapter will provide a brief overview
of the outcome of these four studies and will
then describe in depth the impact of one in-
tervention program on family interactional
variables hypothesized to be important in
indexing family status: affective communi-
cation, problem solving, and coping.
consisted of patients who, for the most part,
were undergoing their first lifetime episode
of schizophrenia, and the period of the con-
trolled trial was the 6 weeks following dis-
charge after a very brief hospitalization. In
a 2x2 design, patients were randomly as-
signed to one of two dose levels of in-
jectable phenothiazine (moderate = 25 mg
or low = 6.25 mg of fluphenazine en-
anthate) and to the presence or absence of a
crisis-oriented family therapy program.
Relapse rates were evaluated at the end of
the 6-week controlled trial and at 6-month
follow-up.
In the Leff et al. (Chap. 13) and Falloon
studies (Chapt. 14), patients were more
chronic, received individually determined
186 M. J . Goldstein and A. M. Strachan
maintenance levels of phenothiazine drugs, Despite the wide variation in samples and
injectible in the former and oral in the lat- the format of treatment, the results of these
ter, and relatives were involved in a longer- three studies were remarkably similar. Fig-
term, family-oriented program which lasted ure 1 presents the relapse rates for these
for a minimum of 9 months. In the Leff studies, contrasting patients treated with
study, the family intervention program was and without a family-based intervention.
focused on the relatives rather than the pa- Although no final report has appeared from
tients, and the relatives were seen in the Anderson et al. study, results sum-
groups. In the Falloon study (Chap. 14), in- marized in this volume (Chap. 12) are very
dividual family therapy sessions did involve similar. Across the three studies, the re-
the patient as well as key relatives and were lapse rates for the family-treated groups are
held in the home. A common feature of the approximately one-fourth of that found in
Leff et al. (Chap. 13) and Falloon studies the groups where the relatives did not par-
(Chap. 14) was that they selected patients ticipate in a focused aftercare program.
primarily from high EE homes so that these The fact that in three of the four studies
interventions were targeted toward the most regular intramuscular neuroleptic medica-
relapse-prone patients and family units. In tion was given suggests that the family pro-
our own study, in contrast, we did not select gram added something above and beyond
on this variable, and took all consecutive insuring medication compliance. Indeed, it
patients who applied for treatment in the certainly appears that something has chan-
local community mental health center ged within the family which has markedly
whose key relatives were living in the area. altered the short-term course of the pa-
tient's disorder.

Impact of One Family Intervention Program on Family Factors

and Short-Term Course
The replicated finding that some form of
family intervention lowers the risk for re-
lapse in patients already on maintenance
pharmacotherapy does not indicate what
processes have been modified by the family
program. Most family programs, tested in
these controlled trials, have been com-
prehensive in nature, providing education
for relatives, support for the family system,
as well as specific training in altering intra-
familial behavior. A theoretical model that
emphasizes the affective climate of the fam-
ily as a risk factor for relapse would receive
greater support if it could be shown that the
family programs that reduce relapse do, in
fact, modify this component of family life.
In the report from Leff et al. (Chap. 13),
there is a suggestion that this was the case,
as the majority of participants in the relative
group condition did show a shift in their EE

Family intervention
[]:ID No family intervention

Fig. 1. Relapse rates from three

studies contrasting family interven-
tion program with the absence of
such a program (all schizophrenic
patients on continuous medica-
etal.,
tion).
Six month loliow-up Nine month follow -up Nine month lollow-up
 Impact of Family Intervention Program 187
status from high to low, and where this shift system. Of the 39 families who participated
did not occur relapse was more likely. initially, 3 dropped out, and of these 34
However, as Strachan, Goldstein, and completed both assessments. The total
Miklowitz (Chap. 4) indicated previously, number used in our analyses was reduced to
this merely documents a change in attitudes 33 because a father in one family failed to
expressed to an interviewer during a repeat show up for the posttest assessment. The
Camberwell Family Interview, but does not data will be described below. Further de-
indicate that family behaviors have chang- tails of the analyses have been reported
ed. In Los Angeles, we have attempted this elsewhere (Doane, Falloon, Goldstein, and
type of evaluation with data from the Fal- Mintz, 1985; Doane, Goldstein, Miklowitz,
loon et al. study summarized in this volume. and Falloon, in press).
The reader is referred to Chap. 14 (Falloon) First, we examined whether there was a
for a full description of that study. To sum- greater decline in the negative affective
marize , it involved a comparison ofindivid- style codes of criticism and intrusion for
ual and family therapy for schizophrenic pa- parents in family therapy than parents of
tients maintained on phenothiazine med- individually treated patients. To test this,
ication. the total number of these types of parental
Before each family entered their respec- statements were summed over the two
tive psychosocial treatment program, they problem discussions, and analysis of co-
participated in a pretherapy family assess- variance was carried out in which the pre-
ment task. This was the direct interaction therapy scores served as the covariate for
task described by Strachan et al. (Chap. 4), the posttherapy scores. These results are
in which two problems were discussed with presented in Figure 2 where we see that the
an experimenter present. After the first 3 incidence of negative affective style mes-
months of treatment, this procedure was re- sages was markedly lower in the family-
peated. The problems discussed were de- treated group [F = 9.41 (1,30) P < 0.005].
veloped in a similar fashion as for the pre- A breakdown of the data for criticisms and
therapy assessment. The problems used intrusions revealed highly significant reduc-
were those identified as idiosyncratically tions for both types of negative affective
relevant to each family at the time each as- communications for relatives in the family
sessment took place . as contrasted with the individual treatment
The interactions were audio-recorded, groups [F for criticisms = 5.49 (1,30) P <
and verbatim transcripts of these discus- 0.05; Ffor intrusions = 8.1 (1,30) P < 0.01].
sions were coded blindly by raters using the This sample consisted of an equal num-
Doane Affective Style (AS) (Doane, Fal- ber of single- and dual-parent families. We
loon, Goldstein, and Mintz, 1985) coding then tested whether the contrast between
20

18

16
14
o
OIl

~ 14
.,
s::
.0 10
~
"8 16
u
~ 6
4
Fig. 2. Total number of critical and intrusive
statements at 3 months post-test, by therapy 2
group (adjusted means).
Family Individual
F = 8.052 . dl 1/29. therapy therapy
P< 0.008
For treatment group
188 M. J . Goldstein and A. M. Strachan
24

20
18
rn ~ Pre-therapy assessment
0
.;::; 16 _ 3 Month post-the rapy assessment
«l
.r::
<Il
.D 14
"0
<Il
"0
0
12
(.)
(f)
« 10
'0
Q; 8
.D
E
;:)
c:
6
(ij
(5 4
I-

single parent dual pare nt single parent dual parent

families families famil ies families
Individual therapy Family therapy
Fig. 3. Pre-post changes in AS by family composition and therapy group.

the parents in the family or individual treat-
ment groups was valid for both types of fam-
ilies. Figure 3 gives pre- and post-AS data
for both types of family structures by treat-
ment condition. We see that there is a com-
parable reduction in the sum of criticisms
and intrusions for both types of family
structures in the family treatment condi-
tion. Interestingly, while there was little or
no change in the level of negative AS state-
ments for single parents in the individual
treatment condition, there are suggestions
of a worsening affective climate in the dual-
parent families when only the patient is
treated. This was quite intriguing, but un-
expected, as we had anticipated that the
single parents, because of their limited
marital support system, might deteriorate
more without family treatment.
The pattern of group means suggests the
Falloon family management approach does
indeed reduce the average level of negative
affective communication in parents of
schizophrenics. However, an average does
not tell us how many families reduced this
behavior in each treatment condition nor
does it establish a precise association be-
tween the reduction of negative affective
communication in a parent, or parents, and
the likelihood of relapse in the patient off-
spring. To deal with these two questions,
the 33 families were divided into three
groups based on the patterns of pre- and
post-AS scores. Since the British work on
EE treats criticism and overinvolvement as
independent dimensions in defining EE sta-
tus, we considered the rate of change in crit-
icism separate from the rate of change in in-
trusiveness. Families were thus classified
into three categories: (1) dual increase, in
which both criticism and intrusiveness in-
creased from pre- to posttest, (2) dual de-
crease, in which both declined during the
same period, and (3) a mixed group in
which one attribute increased while the
other decreased. Figure 4 presents the fre-
quency of each type of family change pat-
tern by treatment group, and it also pre-
sents the number of relapse cases that oc-
curred in each of these pattern-based cate-
gory groups. First, we see that the dual in-
crease pattern is found in half of the individ-
ual therapy families but only rarely appears
in the family therapy group. Conversely,
 Impact of Family Intervention Program 189
Family therapy Individual therapy
(n = 17) (n = 16)
10

(/)
Q)
6 o No relapse
(/) • Relapse
C1:l
0 4

0
01 MX 00 01 MX 00

Change in criticism/intrusiveness
Fig. 4. Patterns of change in AS in individual and family therapy. DI = dual increase; MX = mixed;
DD = dual decrease.

the dual decrease pattern is prominent in
the family-treated cases, but less frequent
in the individual therapy condition. The fre-
quency of the mixed pattern does not vary
by treatment groups. Thus, there is a clear
trend for reduction in one or both forms of
the negative AS behaviors in the family tre-
atment condition.
All but one of the relapses occurred in the
individual therapy condition, and all but
one occurred in the dual increase or mixed
groups. The probability of relapse for indi-
vidually treated patients in or of these two
groups is 67% compared with 25% for the
dual decrease group. These data show that
88% of families in family therapy have a re-
duction in one or both dimensions of AS in
contrast to 50% of those in the individual
therapy condition. Increases in AS are as-
sociated with relapse in that condition as
well. What is puzzling is that there are some
cases in the family therapy condition who
show the dual increase pattern (n = 2) or
the mixed pattern (n = 6) who did not re-
lapse. Of course, this reassessment of the
families occurred after only 3 months of
family therapy, and it is possible that with
sustained therapeutic contacts over the sub-
sequent 6 months, changes in the affective
climate of these families did occur which
prevented relapse.
Data supporting this hypothesis have re-
cently been reported by Doane et a1. (1985)
which involved the ratings of coping effec-
tiveness alluded to earlier. Doane et a1.
classified families at the 3-month assess-
ment as relatively benign or negative in AS,
on the basis of parental behavior in the in-
teraction task. Her classification was based
on whether the parent(s) reduced the num-
ber of harshly critical remarks (personal cri-
ticisms, guilt-inducing remarks, or critical
intrusions) or reduced the number of in-
trusions below a previously determined cut-
ting score of six. Within the family therapy
condition, negative AS status was associat-
ed with lower coping than the benign AS
families at the 3-month point. However,
after 3 more months of family therapy, the
differences between the two AS groups had
disappeared and both, in fact, showed sig-
nificantly higher levels of coping than either
AS group in the individual therapy condi-
tion, who had in fact deteriorated in coping.
Thus, while there are some families who
had not reduced their negative AS by 3
months, they did show improvement with
sustained family therapy, and this improve-
ment in coping is associated with a signifi-
cantly lower rate of patient relapse than was
found in the individual therapy condition.
In the latter condition, there seemed to be
little change in family coping effectiveness
following the 3-months assessment.
190 M. J. Goldstein and A. M. Strachan
Discussion
The various studies summarized in this
chapter indicate that the affective climate of
the family with a schizophrenic young adult
offspring relates to the short-term course of
the disorder. Further, when family-based
interventions are associated with a longer
period of community tenure for the patient,
parallel reductions in this negative affective
climate are noted in direct observations of
family behavior. The optimum results were
found when both components of the AS
system, criticism and intrusiveness, were
reduced in frequency from pre- to a 3-
month posttest assessment, although be-
nefits were observed when only one com-
ponent was attenuated.
Whenever one type of family behavior is
reduced in frequency, it naturally raises the
question of what characterizes the new pat-
tern of family transactions. A key objective
of the Falloon et al. study was to train family
members in a defined set of strategies for
dealing with problems inside and outside of
the family. A coding system was developed
by Doane et al. (1985) for recording in-
stances in the direct family interaction tasks
where family members utilized these re-
commended problem-solving strategies
(PSS). When pre- and 3-month direct in-
teraction tasks data were contrasted in the
number of PSS behaviors expressed, there
was a highly significant difference between
treatment conditions at 3 months favoring
the family-treated group (adjusted mean
FT = 24.59; adjusted mean IT = 10.39, F =
6.60, P < 0.02). Also, when the pattern of
change in AS presented in Figure 4 was
examined for PSS change levels, sharp dif-
ferences were found between the dual de-
crease groups in family therapy and dual in-
crease group in individual therapy in the 3-
month data (mean PSS for dual decrease FT
= +20.0; mean dual increase IT = -0.20).
This means that in the former group there
were 20 more instances of PSS behavior
than was evident at pretest, while in the lat-
ter there was virtually no change from base
line. The remaining AS change groups sho-
wed slight positive changes in scores that
were similar in the two treatment condi-
tions.
These data suggest that a task-centered
form of problem solving takes the place of
the negative interaction patterns of criti-
cism, blaming, and intrusiveness into the
inner state of the patient. When a large re-
duction in the negative AS behaviors (in the
form of redcutions in both criticism and
intrusiveness) has occurred, a major shift in
this more affectively neutral, task-centered
behavior is observed; the reverse is noted
when heightened levels of both forms of AS
are evident. These data suggest a marked
reciprocity between uncontrolled negative
affective expression focused upon the pa-
tient and more controlled, cognitive form of
behavior focused upon common family
problems. Since we have only two points in
the Falloon study where direct interaction
data were obtained, we cannot speak to the
sequential linkage between these two di-
mensions of family behavior. Does family-
based intervention first inhibit negative af-
fective expression, thereby permitting the
acquisition of problem-solving skills? Or,
does the early acquisition of problem-solving
skills simply replace negative affective ex-
pression without any specific instructions to
limit or inhibit these behaviors? Further
studies with repeated observation of the
family therapy process are needed to tease
apart these issues - issues that are critical to
understand if were are to plan more effective
programs for optimizing family change.
An interesting finding was that family
coping responses to external life events im-
proved over time in the family therapy con-
dition, particularly in the families with
negative AS profiles. Unfortunately, the
measure of coping was very global, and thus
it is not known whether the improvement in
coping represents changes in behavioral,
cognitive, or emotional responses. Hatfield
(1981) studied coping in the families of the
mentally ill and found that coping varied
on a number of dimensions: emotional
mastery, cognitive skill in understanding
the illness and learning management strate-
gies, and the extent to which relatives main-
tained outlets for their own personal needs.
An important avenue of research would be
to understand the development of family
member's abilities to cope over time and
the interaction with their style of emotional

expression.
Much of this chapter has emphasized the
impact of a combined program of phar-
maco- and family therapy on relapse rate
over a 9-month to I-year period after dis-
charge for an episode of schizophrenia. It is
obviously an important criterion for a suc-
cessful maintenance program, but relapse
prevention is a minimal criterion for a treat-
ment program. Can a family intervention
program extend beyond this minimal ob-
jective to facilitate improved social role
functioning? The data here are far from
complete. Findings from the Falloon et al.
study do suggest that significantly greater
improvements in social role functioning
were observed in the family than in the indi-
vidually treated group. Also, the recent re-
port by Doane et al. (1985) indicates that a
reduction in negative AS from pre- to 3-
month assessment was associated with high-
er levels of social functioning than when this
was not the case. Also, Anderson et al.
(this volume) report improvements in social
role functioning when families were in the
family therapy condition. Thus, it appears
that modifications of the affective climate
of the family achieves more than relapse
prevention and is associated with greater
social improvement as well. Despite these
encouraging findings, it is still not clear
what level of social improvement is possible
when such a multimodal program is con-
tinued beyond the 9-month point of active
treatment used in the Leff et al. and Falloon
et al. programs. Certainly, the optimistic
results to date warrant continued efforts to
test the limits of efficacy of such programs.
Along these lines, it would be very help-
ful to gather data concerning the long-term
impact of these family-focused programs
after they officially cease. Most of the con-
trolled trials described in this paper have
offered drug and family interventions for
roughly half of their sample for a fixed time
period. At the end of the time period, these
References
Doane, 1. A., Falloon, 1. R. H., Goldstein, M.
1., and Mintz, 1. (1985). Parental affective style
and the treatment of schizophrenia: Predicting
course of illness and social functioning. Ar-
chives of General Psychiatry, 42, 34-42.
Impact of Family Intervention Program 191
interventions typically cease or patients are
transferred to some other clinical unit or
service. What happens to those patients and
their families at this point? Do they start to
relapse rather rapidly or do they continue to
hold the gains observed during the "offi-
cial" treatment period? Evidence concern-
ing maintenance or deterioration in both
the patient and intrafamilial relationships in
general would be useful. If there is little re-
tention of the gains observed during the
controlled trials, it does not imply that these
drug and family programs are without
value. Instead, they could suggest that the
treatment of schizophrenia requires a type
of longer-term commitment than most re-
search studies can typically guarantee.
One final issue, because the emphasis of
this chapter has been upon the family of the
schizophrenic and family-oriented treat-
ment programs, we have not emphasized
issues concerning the long-term pharmaco-
therapy of the patient. We should never for-
get that all of the successful family pro-
grams are built upon a foundation of regular
and continuous pharmacotherapy. How-
ever, the successful addition of family-
based intervention programs to regular
pharmacotherapy raises the question of
whether a potent social therapy permits
alteration in the strategy of drug treatment
in terms of dose level and pattern of treat-
ment. Given the evidence that a com-
prehensive family intervention program
produces some degree of stabilization of the
stress level in the social environment, we
can now consider whether this stabilization
permits the use of lower dose levels of phe-
nothiazine medication or even more radical
strategies, such as intermittent or targeted
use of medication. An ironic side effect of
an effective family program may be that it
permits greater latitude for experi-
mentation in the complementary area of
psychopharmacology.
Doane, 1. A., Goldstein, M. 1., Miklowitz, D. 1.,
and Falloon, 1. R. H. (in press). The impact of
individual and family treatment on the affective
climate of families of schizophrenics. British
10urnal of Psychiatry.
192 M. J. Goldstein and A. M. Strachan
Goldstein, M. J., Rodnick, E. H., Evans, J. R.,
May, P. R. A., and Steinberg, M. R. (1978).
Drug and family therapy in the aftercare of
acute schizophrenic. Archives of General
Psychiatry, 35,1169-1177.
Hatfield, A. B. (1981). Coping effectiveness in
families of the mentally ill: An exploratory
study. Journal of Psychiatric Treatment and
Evaluation, 3. 11-19.
Vaughn, C. E., and Leff, J. P. (1976). The influ-
ence of family and social factors on the course
of psychiatric illness: A comparison of schizo-
phrenic and depressed neurotic patients. Brit-
ish Journal of Psychiatry, 129, 125 -137.
16. The Psychosocial Program of Treatment
of Schizophrenic Patients in the
Crakow Psychiatric Clinic
B. Barbaro, A. Cechnicki, A. Szymusik, K. Zawadzka, and J. Zadecki
Community Psychiatry in Poland
Polish psychiatry is still under the influence
of the nineteenth century model, which em-
phasized large psychiatric hospitals. The
image of these hospitals corresponds to that
of the large state hospitals as previously de-
scribed by Goffman (1961).
The ideology of "hospital centralization"
is a direct descendent of the medical model
of the organization of health care. Recent
changes within the hospitals, as well as the
development of community-based forms of
treatment, have occurred in only a few
General Data on the Crakow Clinic
The psychiatric clinic is situated in the cen-
ter of Crakow among other university clin-
ics. At present the inpatient department
houses about 120 patients, while in the en-
larged outpatient department about 1000
patients are treated.
A Glance at the Sociology of the Institution
In 1951, the psychiatric clinic gained auton-
omy and a separate site. Previously, it had
been part of the psychiatric-neurological
clinic. In the last 3 decades, the clinic de-
veloped from a closed, uniform organiza-
tion system toward an open and differen-
tiated one (Szymusik, Cechnicki and,
Zadecki, 1982). This development, which
paralleled the sociopolitical transformation
in Poland as well as that in psychiatry world-
wide, went through three phases.
Phase 1: "Asylum" (1951-1955). The
clinic operated during this phase to segre-
gate patients from society and to give shel-
university centers. They have not eliminat-
ed the key position of hospitals in the health
care system. The lack of modern legal re-
gulations, lack of social organizations ready
to aid the movement seeking to provide
help for those suffering from psychiatric
diseases, financial difficulties, difficulties in
the coordination of the existing social serv-
ices, and the lack of suitably trained "social
workers," etc. have all hindered the de-
velopment of community psychiatry in
Poland.
The wards of the clinic (for young people,
children, geriatric, neurotic, and two psych-
otic patients) take patients regardless of
their place of residence, although most of
them come from the town of Crakow.
ter "to the individuality and personality of
the patient and physician". It corresponded
to a more "familiar" style of relationship
among the medical personnel (in spite of
the hierarchical structure of the hospital)
and a great limitation of the patients' free-
dom. The theoretical foundations of treat-
ment were influenced by E. Kretschmer's
and E. Bleuler's classical German psychia-
try. Pavlos's influence was felt and to some
slight extent, that of classical psychoanaly-
SIS.
Phase 2: "Renovation of the Hospital and
the Organization of the Clinical Based on
194 B. Barbaro et al.
Therapeutic Groups" (1955-1975). During
this phase, the change within the clinic fol-
lowed the English emphasis on therapeutic
communities publicized in the mid-1950s.
This shift in theoretical model was a crucial
change. Most of the changes were part of a
dynamic process which is still in progress
today. During this period, the emphases of
American dynamic psychiatry on the one
hand and of the existential-phenomenologi-
cal movement on the other were assimilated
into the hospital system.
Phase 3: "Branching of the Outpatient
System" (1975-1981). This phase corres-
ponded to the "separation" of therapeutic
Group-Oriented Community Treatment of Schizophrenia
One of five therapeutic groups in the clinic
is devoted to the treatment of schizophrenic
patients. Paranoidal states, schizoaffective
psychoses, borderline states, and deep dis-
turbances of personality with noncharac-
teristic psychotic phenomena are also in-
cluded within this group. Patients are usual-
ly between 18 and 35 years old and live in
Crakow. There are 20 persons in the
groups: physicians, occupational thera-
pists, psychologists, nurses, a social wor-
ker, and a sociologist.
Pragmatic Plan
In our social model of therapy, we are con-
cerned with concentrating on the process of
communication and on social training for
the relationship between the patient and
those around him or her. Thus, we oppose
the tendency noticed in those who suffer
from schizophrenia to increase the isolation
system.
Professional psychotherapy constitutes
only a small segment of the treatment. An
integrated approach is possible only
through the linkage of various types of ther-
apy: linking inpatient with outpatient ther-
apy and 24-h care with part-time care and
partial hospitalization.
It is necessary to "introduce" those
people closest to the patient, particularly
the family, to the process of therapy and to
treat the various subsystems, e.g., the body
subsystem, the information subsystem, or
groups, which undertook their work in the
newly organized treatment settings (day
wards, hostels, therapeutic camps, differ-
entiated ambulatory treatment of the pa-
tient and his family, etc.). Although the in-
stitutions founded on the therapeutic com-
munity notion and group psychotherapy
developed parallel in time to the English ex-
periments in the 1950s, the newer forms of
partial hospitalization and of therapeutic
lodgings appeared later. During phase 3,
the psychiatric hospital became a part and
no longer the center of a complex system of
psychiatric care.
the particularly human-value subsystem,
with similar attention.
Organization Model
The psychotic ward of less than 20 beds
forms the basis of the hospital. The day
hospital with the therapeutic family group
was opened 3 years ago when the other
parts of the system were already completed.
The day hospital is situated on one floor of
a multifamily housing complex in the center
oftown. The club offormer patients and the
section of the Polish Tourist Association
run by them are also situated there. The day
center is also the meeting place for the
patients who have left after having been in
ambulatory care for several years. A hostel
rented by the clinic from private owners
forms another part of the system. The inpa-
tient ward is prepared to receive, when ne-
cessary, an outpatient or an inmate of the
hostel. Therapy is also carried out in a shel-
ter in the mountains where therapeutic
camps are organized every year, as well as
in the patient's home.
Inpatient Ward
Therapy of acute psychotic states is carried
out in the inpatient ward, where an attempt
is made to reduce the length of the patient's
stay to a minimum. The basic problem dur-
ing inpatient treatment involves maintain-
ing a balance between reality and the world

of the patient's imagination and delusions.
One major goal is to rebuild a sense of social
and individual identity. An emotional rela-
tionship with the individual therapist is a
basic conditon in this process. Individual
care is undertaken not only by phsyicians
but also by experienced nurses and psycho-
logists, with the physician serving as a con-
sultant in the matter of drugs.
The purpose of all the therapeutic, phar-
macological, individual, and group meth-
ods is to rebuild emotional communication
with the people around the patient before
the psychotic phenomena disappear. Non-
theless, there is a progressive shift from in-
dividual to group contacts. In addition to
the daily, large group meetings of the thera-
peutic community, there are a wide variety
of group experiences which reflect the so-
ciocentric model of therapy (sports, group
psychotherapy three times a week, excur-
sions, occupational therapy organized by
the leaders of the ward). The purposes of
the therapy in the day ward are clear: we try
to introduce the patient into the "common
area." Several almost ritualistic activities
are dealt with during the therapeutic com-
munity meetings, such as the election oflea-
ders, the establishing of the hours of duty at
the door of the ward, reports on the week-
end spent outside the hospital, planning of a
common afternoon, introduction of new
patients and saying goodbye to those leav-
ing, which tend to reinforce a constant
structure of common situations. This type
of communication bears fruit in increased
activity and enhanced initiative among the
patients. This emphasis on self-efficacy is
critical from the first moment the patient is
received into the inpatient setting. There is
frequently an inherent conflict between
these socially oriented therapy programs
and the expectations of the schizophrenic
patients, probably related to the main sym-
ptoms of the disease, such as suspicious-
ness, social withdrawal, and anhedonia.
This should be understood as one aspect of
the complex process of establishing com-
munication between patient and staff on the
unit.
The stay of the patient in the ward aver-
ages about 3 weeks. From among 140 pa-
tients treated during the year, half transfer
subsequently to the day ward, if they fulfill
The Psychosocial Program of Treatment 195
the following criteria: (a) the acute psycho-
tic symptoms have disappeared, (b) the sui-
cide risk has disappeared, (c) the optimum
therapeutic dosage of drugs has been
reached, (d) basic group abilities have been
achieved, (e) basic practical abilities have
been realized, e.g., going to town inde-
pendently, and (f) a positive motivation to
continue psychiatric treatment has been
achieved.
These criteria require a short commen-
tary. The attenuation of acute psychotic
symtoms is relatively easy to evaluate, yet
persisting hallucinations or delusions do not
constitute a contraindication for moving the
patient to the day ward. The level of anxiety
and secondary adaptation to the symtoms
determines what course of action is subse-
quently taken. The necessity of evaluating
the risk of suicide needs no explanation.
Reaching the optimal dosage of drugs in-
volves evaluation of the individual's bio-
logical reaction to the neuroleptics. We
have noticed that despite the appearance of
additional symptoms connected with taking
neuroleptics the dosage should not be ra-
pidly reduced for a patient who after trans-
fer to the day ward encounters a more com-
plicated social situation. Perhaps it is ana-
logous to a family with a high index of ex-
pressed emotion (Brown, Birley, and
Wing, 1972; Kuipers, 1979; Vaughn and
Leff, 1976). The achievement of basic group
abilities is connected with a readiness to
receive feedback from the group. Finally,
the practical abilities are limited to going to
town on one's own and basic self-care.
The motivation for psychiatric treatment
is usually feigned when presentation is the
result of pressure from the patient's family
or other contact persons and the patient is
sometimes hospitalized involuntarily.
Ambulatory treatment is the result of at-
tachment to the institution or the therapist
who established emotional contact with the
patient. When transferring the patient to
the day ward, we expect at least a trace of
authentic motiviation. This motivation is
not always associated with full insight into
the disease. Often the patients lose interest
in the psychotic contents of their dreams,
but on the other hand, in the group they fre-
quently become aware of severe personality
problems.
196 B. Barbaro et al.
Day Hospital the plan of the day. The style of work in the
group is similar to the interactive groups
As mentioned previously, greater impor- described by Kayser et al. (1973), where
tance is attached to the patient's own mo- stress is placed on the present situation of
tivation to participate in therapeutic activi- the patient and a minimum of attention is
ties. Patients transferred from the inpatient devoted to the conflict from the past. In a
ward of the clinic constitute 75% of the day group of this kind it is easier to correct pres-
hospital population, and the remaining ent behavior with the support of the group.
25% come from various psychiatric wards Psychodrama is central for demonstrating
or outpatient care. The average length of actual conflicts before the group. Group
stay is about 7 weeks. The day hospital ther- psychotherapy forms an institutional
apy area is shared together with the family framework for communicating, and
group, the club of former patients, and the through socialization it lessens the feeling
outpatient clinic. Thus, the day hospital of isolation and loneliness of patients suf-
performs its part of the rehabilitation pro- fering from schizophrenia.
cess (Howells, 1968) and is directed mainly Another type of group is oriented to crea-
at patients who, having gone through psy- tive activity, in which spontaneity and try-
chosis, are left with feelings of emptiness, ing ou of new parts are encouraged. It takes
insufficiency, ruptured or feeble bonds with place during paratheater, choreotherapy
their previous peers, a readiness for with- and art sessions. Self-help is required and
drawal, which has been increased by the the patients themselves prepare breakfasts
disease and a stay in the hospital, and a state and lunches from supplies provided, per-
of postpsychotic depression while searching mitting the practice of skills, the lack of
for their place in society. The whole sur- which is often described as the "life aprag-
rounding reality of the ward has as its aim to matism" of those suffering from schizo-
create a "corrective family atmosphere," phrenia. Finally, many therapeutic activ-
which allows the patient's social abilities to ities are oriented toward creating a thera-
develop. The family-oriented therapy pro- peutic milieu with an emphasis upon mutual
gram is possible only with close cooperation coexistence and cooperation such that a so-
between the families and the ward staff. cial field is created for individual decisions
Through the direct association with the and personal responsibility. This is found
primary group, the family, and the thera- everywhere within the therapeutic com-
peutic group on the ward, it is possible to munity: the day hospital, the classes held
work through the current conflicts and for the former patients club, the section of
mutual involvements in the primary group. the Polish Tourist Association, and at the
One of the sessions is called "My family therapeutic camp (which is the climax of
today and tomorrow" and is conducted by each year's work at the day hospital and
the family therapists, who hold similar ses- where most of the patients who had been
sions for the patient's families. treated during the year meet).
The day hospital offers a 40-h program of Those parts of the system not described
therapy 5 days a week and once a month a here in more detail emphasize the patients'
weekend excursion into the mountains with own numerous activities, while the part
the Polish Tourist Association. The patients played by the psychiatric institution is mini-
are gradually prepared to take up work or malized.
study. The group cooperates closely with
the Trade Advisory Center, trying to
Family Therapy Unit
achieve a return to work in full-time jobs for
this group of young patients 18-35 years As mentioned above, this unit occupies the
old. Only a small percentage of patients re- same area as the day hospital and exerts a
turn to institutions offering protected work decisive influence through its close co-
(sheltered workshops). operation with the day hospital. Our under-
The day program of the ward is filled with standing of schizophrenia is expressed in a
various group activities. The daily 90-min lecture: "Values, aims, and aspirations and
psychotherapy group has a central place in the crisis of identity experienced by those

suffering from schizophrenia," in which we
describe more fully the initiation and de-
velopment of disturbances within the family
system and in the complex pathological
communication within the family. We em-
phasize that the individual and social reac-
tions to the illness do not exist separately
from each other. An understanding of the
implications of the illness must take into ac-
count both the person and the milieu in
which he/she lives.
Clinical Experience
During the first 2 weeks of the patient's stay
in the inpatient hospital, the therapist initi-
ates a home visit without the patient's pres-
ence. This is done after all the family mem-
bers have agreed. On the part of the thera-
pists these visits are participant observa-
tions. The aim of the visit is to produce a
dynamic diagnosis of the family which de-
fines, at least, the complexity of communi-
cation among the members of the family
group. One therapist from the family treat-
ment group and one from the day hospital
or the inpatient hospital take part in each
such visit. The goal is to experience the real
atmosphere in the family as perceived
through the medium of the therapists' emo-
tions, which are later analyzed. The scheme
of the diagnosis follows, that published by
Howells (1968) and is subsequently trans-
mitted to the therapists conducting the fam-
ily therapy. In addition to the family thera-
py described above, work with a multifami-
ly group with patients has recently been
undertaken. Both the patients and their
families have earlier experiences from
groups in which they have participated that
have a decisive influence on the group
dynamics.
We have noticed advantages of such a
form of therapy in addition to the greater
economy compared with each family having
its own therapist. Thus, meetings with indi-
vidual family units most often have a diag-
nostic character, giving one more glance at
the family as a whole from another per-
spective, or allow intervention when the pa-
tient is already in ambulatory care and the
family is in a crisis. Similarly, therapists oc-
casionally come into contact with the family
of a former patient who has had a recur-
The Psychosocial Program of Treatment 197
rence and requires "home care." With
every contact, the family's ability to carry a
further burden and the harmfulness of the
patient's increasing isolation in his own
home are evaluated.
Live-in Community Hostel
The hostel is a transitional form with a de-
finite social program. The therapy in the
hostel is directed at the development of the
person through real life experiences. The
institution's role consists in introducing into
the group a "hostel adviser" from the thera-
peutic group, who lives in and shares every-
day life with the patients. The general aims
are to develop, in a small group, important
emotional bonds and to enhance the abili-
ties needed to deal with life so that a more
independent life is feasible. With a periodic
separation from the family there is a good
chance of regulating the disturbed relations
which have existed in the family group,
e.g., an excessive overinvolvement with the
patient. Patients are assigned to a hostel
straight from the day hospital or while in
ambulatory care. Two therapists from the
hospital visit the hostel once a week at the
hostel community meeting and analyze the
course of the previous week. The close co-
operation of the hostel adviser, the visiting
therapists, oriented to the development of
the group as a whole, the individual thera-
pist, who is most often one of the members
of the day hospital, and the family therapy
group makes it possible to implement a
coherent treatment strategy based on
mutual agreement. It often happens that
the families of the inhabitant remain in the
multifamily group with the patients while
the patients reside in the hostel. It allows
still another opportunity for the whole
family to assimilate new experiences.
Hostel Adviser
The hostel adviser of the therapeutic com-
munity hostel, who lives with the psychotic
patients for about a year, has a difficult
task. Since the hostel came into being in
1976, eight persons have undertaken this
role. They need to overlook professional
limitations and to avoid any conflict of
loyalty between the group of patients and
198 B. Barbaro et al.
the therapeutic group. The feeling of be-
longing to the group of inhabitants must be
emphasized. The expression of aggressive-
ness is usually a critical event and is usually
followed by good communication between
the adviser and the members of the group.
The adviser functions in the hostel as a
model to be imitated. His close proximity
exposes him to intensive feelings, transfer-
red and countertransferred. The hostel ad-
viser requires intensive help and an analysis
of his own emotions with the help of the the-
rapists visiting the hostel.
Such prolonged living together with schiz-
ophrenic patients is an invaluable method
of developing the abilities needed to coexit
with patients in close proximity. The specif-
ic range of problems and difficulties associ-
ated with such an experience would require
a separate chapter. As in a lens, the practi-
cal problems of therapy are brought into
focus. When seen at such close quarters the
psychopathological symptoms do not look
like symptoms, and they need nontradition-
al solutions.
Principle of the Small Group
The system of therapy described above is
based on small groups of 10-12 persons.
The hostel has only seven residents. The
developemnt of these groups enriches the
therapy by providing realistic life experi-
ences. It is done mostly outside the psychia-
tric hospital, on the basis of emotional
bonds that can be sustained when the dis-
tance between the members of the group is
small.
Problem of Distance
The autism described in schizophrenia re-
quires putting the patient in contact with his
milieu (therapist) but also the milieu (thera-
pist) must be put in touch with the patient.
The various systems of therapy can be eval-
uated from the point of view of this
distance. The distance is almost automatic-
ally connected with differently organized
structures. Thus, the psychopathological
symptom is a dynamic function of the entire
complex social structure in which it is mani-
fested and observed, depending on the
point of observation.
Principle of Continuity in the Process of
Therapy
The continuity of the treatment depends on
an administrative structure which encour-
ages entry into transitional forms of treat-
ment after the patient leaves the hospital,
e.g., day hospital, hostel, and, depending
on the individuals' needs and abilities, into
varied ambulatory care programs such as
the club of former patients, individual ther-
apy, family therapy, therapeutic camp, and
the Polish Tourist Association section.
Continuity of care requires at the very least
a number of years of the various therapies.
This continuity is guaranteed by partici-
pation in the same therapeutic group over
the years. The group fulfills, for the pa-
tients, various therapeutic functions by pro-
viding a chance for manifold relationships
to develop.
Principles of Work in the Group
The complex tasks are orchestrated by the
interplay between the main therapists and
the auxiliary therapists. Although each pa-
tient always has his own individual thera-
pist, it is important that he/she can tum to
different members of the group for assist-
ance depending on his therapeutic situa-
tion.
The functioning of the groups is based on
the principle of rotation. Once a year a
meeting of the group is held, which lasts a
few days, in which a summary of the work
accomplished to date and plans for the com-
ing years are presented. The rotation of
people into particular subgroups takes into
account both the patients' and the thera-
pists' needs, the need for reciprocal school-
ing, and deadlines established for thera-
peutic or research tasks in the groups.
With this multiplicity of therapeutic
tasks, constant ongoing supervision is nec-
essary. This is carried out in small thera-
peutic groups; for example, one of the ther-
apists from the family therapy section is a
supervisor of the inpatient hospital. The so-
called Balint groups (Balint and Balint,
1962) undertake other problems of patients
and their therapists. In Crakow there are
five such groups and the majority of the
therapists participate in them.

The Most Difficult Task
The most difficult task consists of treating
the schizophrenic patient without creating a
deep dependance on the therapist or the
therapeutic system. It happens sometimes
that the patient passes from a deep depend-
ence on the therapist and family to a type of
Final Remarks
Psychosocial models have begun to play an
increasingly important part in the treatment
of schizophrenia (Stierlin, Wynne, and Wir-
sching, 1983). The model described here pro-
vides more intensive treatment and re-
habilitation than is possible in traditional
ambulatory treatment. Through its many
links, such as the day hospital or the hostel,
high expectations for change are rein-
forced.
The greatest dilemma confronting the cli-
nician preparing such a social treatment
program is whether the patient can profit
from a less intensive ambulatory treatment,
such as traditional individual psychothera-
py and supportive neuroleptic treatment.
There are not patient attributes which ena-
References
Anthony, W. A., Cohen, M. R., and Ray, V.
(1978). The measurement of rehabilitation out-
come. Schizophrenia Bulletin, 4, (3).
Balint, M., and Balint, E. (1962). Psychothera-
peutische Techniken in der Medizin. Stuttgart:
Klett.
Brown, G. W., Birley, J. L. T., and Wing, 1. K.
(1972). The influence of family life on schizo-
phrenic disorders: A replication. British Jour-
nal of Psychiatry, 121,241-258.
Cechnicki, A., Zawadzka, K., Rostworowska,
M., Zadecki, J., and Barbaro, B. (1984). Die
sozialpsychiatrische Tagesklinik fUr schizo-
phreniereagierende Patienten [The Socio-
psychiatric day clinic for patients with schizo-
phrenic reactions]. Paper presented at the
XVth Symposium of the DAP, Munich.
Goffman, E. (1961). Asylums. Essays on the
social situation of mental patients and other
inmates. Chicago: Aldine.
Howells, J. G. (1968). Theory and practice of
family. Edinburgh: Oliver and Boyd.
Katschnig, H. (1985). Methods of measuring so-
cial adjustment. In T. Helgason (Ed.), Metho-
The Psychosocial Program of Treatment 199
pseudoindependence which is manifested
by breaking all of the bonds established
during therapy. The work of motivating the
patient for the therapy should be based on
the expectation that a phase of dependence
will be followed by a phase of cooperation
based on mutually established conditions.
ble onte to predict which patient will profit
from a rehabilitation program (Klar, Fran-
ces, and Clarkin, 1982). It is a positive de-
velopment that various systems of therapy
take into account the patient's individual
needs. This permits therapists to compare
their experiences and the results of research
(Anthony, Cohen, and Ray, 1978; Katsch-
nig, 1985). The fairly homogeneous popula-
tion of our patients in terms of age, the pre-
cisely determined diagnostic criteria, and
the specific description of various forms of
therapy used in our system permit further
longitudinal research to evaluate both the
total model and the suitability of the various
components for any given patient.
dology in evaluation of psychiatric treatment.
Cambridge: Cambridge University Press.
Kayser, and Mavers, W. (1973). Gruppenarbeit
in der Psychiatrie [Group work in psychiatry].
Stuttgart: Thieme.
Klar, H., Frances, A., and Clarkin, 1. (1982).
Selection criteria for partial hospitalization.
Hospital and Community Psychiatry, 33,11.
Kuipers, L. (1979). Expressed emotion. A
review. British Journal of Social and Clinical
Psychology, 18,237-243.
Stierlin, H., Wynne, L. C, and Wirching, M.
(Eds.) (1983). Psychosocial intervention in
Schizophrenia. An international view. Heidel-
berg: Springer.
Szymusik, A., Cechnicki, A., and Zadecki, J.
(1982). Die Entwicklung der Idee der dynami-
schen Psychiatrie in der Psychiatrischen Klinik
der Medizinischen Akademie in Krakow
[Development of the idea of Dynamic Psychia-
try in the Psychiatry Department of the Aca-
demy of Medicine in Crakow]. Paper presented
at the XIVth Symposium of the DAP, Munich.
200 B. Barbaro et al.
Vaughn, C. E., and Leff, J. P. (1976). The influ- phrenic and depressed neurotic patients. Brit-
ence of family and social factors on the course ish Journal of Psychiatry, 129, 125-137.
of psychiatric illness. A comparison of schizo-
Epilogue
17. Coping and Competence as Protective Factors
in the Vulnerability-Stress Model of Schizophrenia
R. P. Liberman
Introduction
Schizophrenia is not to be understood as
psychopathology alone, or as a set of bio-
logical impairments or deficits; rather, the
nature of schizophrenia is to be found in the
interfaces among environmental, behav-
ioral, and biological factors. Some of these
factors endure over time and change only
slowly in response to changes in the en-
vironment. Biobehavorial factors operate
on the person as vulnerability to episodes of
schizophrenic symptoms when they endure
through periods of remission as well as re-
lapse. Abnormalities in information pro-
cessing and attention as well as autonomic
reactivity have been implicated as enduring
vulnerability factors across illness episodes
as well as in children at risk for schizophre-
nia (Nuechterlein and Dawson, 1984). In-
adequate learning opportunities during
one's developmental years can also lead to
subsequent, enduring deficits in social and
occupational capabilities that may increase
a person's susceptibility to schizophrenic
symptoms. For example, certain schizo-
typal traits have been posited as vulnerabil-
ity factors in schizophrenia (Chapman,
Edell, and Chapman, 1980). Furthermore,
studies have demonstrated that premorbid
social and occupational deficiencies can
predict subsequent symptomatic status and
rehospitalization of schizophrenic patients
(Strauss and Carpenter, 1981). Enduring
biobehavioral vulnerabilities interact with
transient, time-limited stressful life events,
longer-term environmental potentiators of
risk to symptoms, and protective factors
that operate at the environmental and per-
sonallevels.
Progress has been made in recent years in
identifying factors that can protect a vul-
nerable person from stress-related, bio-
medical determinants of schizophrenic re-
lapse. One major source of protection
emerges from studies of the familyenviron-
ment where supportive, tolerant, and
understanding relatives who exhibit realis-
tic expectations for social performance
from a schizophrenic member can reduce
relapse to 15% in 9 months following dis-
charge from the hospital (Vaughn et aI.,
1982). On the other hand, ambient levels of
stress and tension within the family can
negatively affect patients who suffer from
schizophrenia. When individuals with the
specific psychobiological vulnerability to
schizophrenia experience stress and tension
from family transactions, relapse or exac-
erbation of symptoms is made more likely,
with relapse rates exceeding 50% in the 9
months after hospitalization. Family pat-
terns of stress are interactional as schizo-
phrenic persons also produce stress on their
relatives by virtue of their symptoms and
social impairments. This interactional pat-
tern leads to emotional overinvolvement
and excessive criticism from relatives in
about two-thirds of Anlgo-American fami-
lies (Vaughn et aI., 1984). Thus, family
interaction processes, both unfavorable
and supportive, go in both directions and
can affect patients and their relatives in re-
verberating cycles.
Brief cross-sectional views of course and
outcome seen at anyone point in time in our
assessments and follow-ups of patients with
schizophrenia can be interpreted from
a vulnerability-stress-coping-competence
model. These are really not "outcomes" but
rather momentary "snapshots" of how the
family environment, the social network,
 8 ~ Personal vulnerability factors
3~ f3
('1) I-' r-------------~i i
o .
;:;J> Reduced Autonomic
Schizotypal
g..g Dopaminergic available hyperreactivity
personality :;tl
N-= dysfunctions processing to aversive
o :::t_ traits .."
'1:1 '" capacity stimuli
0" O.
.... (")
('1) (")
~
('1)
~-g
f6 Intermediate states Outcomes 3
-g ::l
'"
Personal protectors
=
e.. ~--------~i ~i----------~
Processing Social
q> capacity functioning
Coping overload
Antipsychotic
and
medication
I self-efficacy

Q
* Tonic Schizophrenic
= Prodromal
autonomic psychotic
symptoms
~ Environmental Protectors hyperarousal symptoms
fa
8.5- Family Supportive
OQ problem psychosocial
g. solving interventions Deficient
('1) Occupational
< processing
functioning
:::t_
'" of social stimuli
2:
('1)

'"
[ T Enviro~menital potentiators and ~tressl"'io_r_s__________--,
a
('1)
0- Critical or Over-
....
o emotionally stimulating Stressful
overinvolved social life events
§ family climate environment
;jl
('1)

8.o Feedback loop

Premorbid or remission period Prodromal period - - - - - - - - - -....-- Episode
!f.

and the individual's psychobiological vul-
nerability and personal coping are interact-
ing to determine current levels of symptoms
and social functioning. Figure 1 depicts a
conceptual framework for organizing a vul-
nerability-stress-coping-competence model
Case Example
To elucidate the vulnerability-stress-cop-
ing-competence model of schizophrenia,
we might think of a farm boy, Michael, who
at the age of 18 is recruited into military
service. He has spent all of this life in a rural
and serene pastoral county, without very
much contact with people. He went to a
small school but avoided contact with other
children. Most of his time was spent on the
slopes of his native mountains, tending
sheep and cows. He has poorly developed
social skills with peers. His parents were
also socially withdrawn and did not provide
opportunities to learn expressiveness or
intimacy. There was, at the genetic level, a
grandfather who spent many years in the
state pychiatric hospital, presumably for a
schizophrenic disorder.
At the age of 18, Michael entered military
training. He lived in a barracks full of other
young men, most of whom came from lar-
ger cities. They talked openly, fast, and
loud. He was not used to this active and
stimulating social intercourse and he expe-
rienced difficulty in processing the incoming
information. He did not know how to make
jokes with his fellow recruits. He did not
know how to brag about his sexual exploits
because he did not have any. After several
stressful weeks of trying vainly to cope with
the rigorous military discipline and peer
pressure, he experienced confusion in his
thinking and autonomic symptoms of anxie-
ty. These prodromal symptoms of psychosis
were quickly followed by ideas of reference
and then delusions that thoughts were being
put into his head by his officers. He showed
lapses in attending his regular duty, was
awake at night, and was brought to the mili-
tary hospital with florid hallucinations and
delusions where the psychiatrist pronounc-
ed him as suffering from a schizophrenic
disorder.
Michael then went into a hospital ward
Copying and Competence 203
of schizophrenia. At present, we probably
know more about the influences at the envi-
ronmental and behavioral levels affecting
schizophrenia than we know about the bio-
logical factors.
which was supportive, nurturing, warm,
and friendly. The staff were helpful to him
and seemed to understand his needs for pri-
vacy and quiet. They did not expect much
from him. He did not have to polish his
shoes or make his bed in the morning. He
received antipsychotic medication and
slowly but surely his symptoms subsided.
He began to tentatively socialize, and his
level of function returned to its premorbid
level.
At this point, he was discharged from the
military hospital and returned to his family
home. Unfortunately, the protective fac-
tors provided by the military hospital did
not get communicated to Michael's after-
care system. The psychiatrist in the milita-
ry did not talk to the psychiatrist respons-
ible for the rural clinic near Michael's home
nor to Michael's parents. As a result,
Michael returned home to family members
who were baffled and confused about what
had happened to him. He looked normal.
There were no outward signs of illness. He
was taking some pills, which his parents felt
made him sleepy and were a sign of weak-
ness. His parents told him that he did not
really need to have those pills as long as he
ate his ample farm meals and drank a lot of
good milk. His parents urged him to again
work on the farm. Their expectations for his
work capacity were as high as before, if not
higher, because they were frustrated and
annoyed with him for having failed to com-
plete his military service.
As the performance demands on the farm
and pressures within the family mounted,
Michael experienced the return of weird
feelings. He felt unreal and became obses-
sed with concerns about bodily decay and
dying. The long-acting antipsychotic med-
ication was slowly metabolized and ex-
creted, and Michael again developed the
symptoms of schizophrenia.
204 R. P. Liberman
Reference to Figure 1 will enable the rea-
der to chart the evolution and course of
Michael's disorder and to identify points of
departure for therapeutic intervention. The
prototypical patient, Michael, had both
genetic and schizotypal personal vulnerabi-
lity to schizophrenia. As he was exposed to
the stressors of leaving home and entering
the military, his personal and coping skills
and environmental support were inade-
quate to protect him from the emergence of
psychotic symptoms. Prodromal symptoms
were associated with autonomic hyper-
arousal and cognitive overload and were
precipitated by an overstimulating environ-
ment in the military.
The supportive framework of the military
hospital combined with neuroleptic med-
ication temporarily reconstituted his cogni-
Schizophrenia: A Multidimensional Disorder
Counteracting unidimensional thinking
about schizophrenia is difficult in contem-
porary psychiatry because of the strong
trends toward biological or psychopatho-
logical determinism. Psychopathology is
reified and viewed as having permanency
and pervasiveness well beyond its actual
parameters. The ascendant neurosciences
are uncovering more of the secrets of the
brain, but we too quickly accept early data
on PET scans, neurotransmitters and re-
ceptors as the holy grail. We do not need to
engage in an ideological struggle over nur-
ture vs nature in schizophrenia, but we do
need to resist premature and simplistic bio-
logical theories of mental disorders.
The chapters of this book begin the ardu-
ous journey of searching for the complex in-
teractions among the biological, environ-
mental, and behavioral determination of
schizophrenia. Data are becoming available
to go beyond the general outlines of a vul-
nerability-stress model of schizophrenia to
an empirical framework that possesses pre-
dictive Validity. Jobs, work, participation in
social clubs, social skills training, familyed-
ucation and survival skills training, and trai-
ning in problem solving have been identi-
fied in quantitative and objective studies as
having the power to reduce impairments in
social functioning and the frequency of re-
tive capacities and social functioning; how-
ever, returning home to family members in
a tense and unsupportive home environ-
ment led quickly to relapse when the pro-
tective benefits of medication were with-
drawn.
Several entry points for therapeutic inter-
vention present themselves, including fam-
ily education and management, mainte-
nance neuroleptic drug therapy, social skills
training, and less demanding vocational
and peer settings. The diagram in Figure 1 is
meant to highlight the reverberating cycles
and interactions among factors influential
in the course of schizophrenia. Protective,
potentiating, vulnerability, and stress fac-
tors interact to determine the symptomatic
and functional status of a person at anyone
time.
lapses.
The research reported in this book also
raises the possibility that temporary re-
lapses may be worth the price to pay to
achieve greater gains in social functioning.
In other words, patients and their families,
as well as responsible clinicians, might ac-
cept and tolerate a relapse if that relapse is
in the service of the individual's confronting
and eventually mastering the stress requir-
ed to compensate for premorbid deficits in
social functioning. It may be necessary to
encounter new life situations, risking exac-
erbations and symptoms while learning
skills to improve social and occupational
functioning. In fact, data supporting this
view were presented in Chap. 1 by Moller
and his colleagues from the Max Planck
Institute of Psychiatry. They found that
patients with frequent episodic relapses
also had the best occupational adjustment
at the time of long-term follow-ups. Simi-
larly, in Chap. 12 on psychoeducatoional
family therapy by Anderson, increases in
relapses from 4% to 24 % were noted for pa-
tients from the 1st to the 2nd year of psycho-
social treatment, presumably because more
patients were holding jobs and involved in
school and other kinds of normalizing work
and educational activities. Thus, we might
view relapses as being temporary but neces-

sary costs for improved social functioning.
In turn, strengthened social skills and fam-
ily coping may reduce the vulnerability of
the individual at the behavioral level, en-
abling that person to be more resistant to
subseqent stressful events and family ten-
sion.
As highlighted by this book, it is impor-
tant to extend our studies both longitudinal-
ly over time and intensively across relevant
biological, behavioral, and environmental
variables. No longer is it sufficient to design
studies that begin with initial assessments at
point A and then recapitulate with follow-
ups at 1, 2, 3, or 30 years later. We have
gained knowledge from these follow-up or
retrospective studies that assess patients at
long intervals; however, we will learn much
more from intensive follow-through stu-
dies, such as the one reported by Falloon
in Chap. 14, which continuously and re-
peatedly rate biopsychosocial variables that
influence the course of illness. This type of
intensive research design is exemplified by
a current study being carried out at the
UCLA Clinical Research Center for the
Study of Schizophrenia by Nuechterlein
and his colleagues (1984) who are studying
patients within 2 years of their first episode
of schizophrenia. The patients are followed
for at least 2 years while they participate in
a clinic where they are seen every 2 weeks as
they come in for their medication evalua-
tions. Ratings of psychopathology and life
events are made every 2 weeks, and meas-
ures are taken of their autonomic psycho-
physiology and cognitive attention capabili-
ties every 3 months. Every year ratings are
made of their home emotional climate
through the CFI.
Preliminary results from this study indi-
cate that 3 months after discharge from the
hospital, there is a strong trend for family
"expressed emotion" (EE) to predict both
symptoms and poor occupational status.
From direct behavioral observations made
of ecological variables during the index hos-
pitalization, those young schizophrenics
who show greater social interaction (i.e.,
talk with and listen more to other patients
and staff and are in closer proximity to
other patients and staff) have a much better
rate of social functioning at the 3-month fol-
low-up point. At the autonomic level, high
Copying and Competence 205
arousal and low reactivity on electrodermal
responses during the initial hospital period
appear to predict higher symptomatology at
the 3-month follow-up. Thus, through more
frequent assessments in intensive follow-
through studies, we may improve predic-
tion of factors that influence the course of
schizophrenic disorders. Measures also
need to be drawn from psychobiological
factors that may be markers or indicators of
the underlying vulnerability of schizophre-
nia; that is, factors that do not change when
symptoms change, but that remain fairly
constant during pre morbid , symptomatic,
and remitted periods.
A frequent criticism of research on psy-
chosocial stressors in schizophrenia is its
lack of specificity to schizophrenia. Critics
who take this view appear not to realize that
nonspecificity of many protectors, stres-
sors, and potentiators is integral to the vul-
nerability-stress model of illness and
health. It is the nature of the interaction
among nonspecific stressors, potentiators,
and protective factors with specific psycho-
biological vulnerability that predict onset,
course, and outcome of schizophrenia. Psy-
chobiological vulnerability, perhaps in-
herited or influenced by congenital prenatal
or postnatal trauma, infection, or immuno-
logical events, determines specificity of ill-
ness and its symptomatic impairments.
However, nonspecific factors may be even
more powerful in determining severity and
course of illness. Moreover, factors that are
nonspecific for diagnostic categories may
possess specificity for dimensions of course
and outcome. The data reported on family
factors in this volume are pointing the way
toward more specific and dimensional pre-
dictors of outcome in schizophrenia. For
example, high expressed emotion may spe-
cifically predict return of positive schizo-
phrenic symptoms, but not predict social
outcome. On the other hand, indices of di-
rect family interaction may better predict
patients' social adjustment. We must in-
vestigate variables that not only predict
relapse, but also other specific dimensions
of outcome. We will undoubtedly discover
that some of our assessment instruments
are better than others at predicting the vari-
ous domains of outcome of importance in
schizophrenia.
206 R. P. Liberman
Advances in Assessment
Improvements in assessment methods have
been highlighted in chapters of this volume
and represent the most critical advances in
furthering our understanding of the vul-
nerability-stress model of schizophrenia.
Improvements in diagnostic and assessment
instruments have made generality and re-
plicability of research on schizophrenia pos-
sible for the first time. Studies now can be
compared across countries and cultures be-
cause we share common criteria for diagno-
ses of schizophrenia, and even more impor-
tantly, common instruments for eliciting
the symptoms of schizophrenia. The Pre-
sent State Examination (PSE), the SADS,
the SCID, and the DIS are now being used
in many parts of the world (Andreason,
1986). Results from studies in Europe are
now comparable with those being done in
centers in the United States. On both sides
of the Atlantic, investigators are using an
expanded approach to the assessment of
psychopathology and functional behavioral
states that assumes that manifestations of
schizophrenia are heterogeneous and di-
verse. We no longer have to limit our focus
on characteristic symptoms of schizophre-
nia as long as we take care to gather diag-
nostic information very specific and reliable
ways. We also have begun to better specify
measures of social process and outcome
variables. As illustrated by several chapters
in this book, methods for measuring protec-
tive and potentiating variables within fami-
lies are being developed in Germany, the
United States, and the United Kingdom;
for example, the KPI (see Chapter 7), ex-
pressed emotion, and Affective Style (AS,
see Chapters 4 and 15) are bringing the role
of the family system into clearer focus.
Instruments for measuring social adjust-
ment, social skills, and social competence
are coming into use that will refine our ap-
preciation of course and outcome at the so-
cial and occupational levels and of treat-
ment effects as well (Liberman, 1982). A
very promising approach to measuring and
predicting social and behavioral status
comes from time-sampling techniques of
direct observation of patients in natural, cli-
nical environments (Licht, 1984; Paul and
Lentz, 1977). Methods of behavioral assess-
ment have permitted differentiation of di-
agnostic groups, accurate predicaton of
short- and long term outcome, and evalua-
tion of the effects of different psychosocial
treatment settings (McGuire and Polsky,
1982; Paul and Lentz, 1977; Paul, 1986;
Alevizos, DeRisi, Liberman, Eckman, and
Callahan, 1978).
Improvements in assessment help us to
more specifically and efficiently identify
areas for intervention. We now can see as-
sessment, not as an isolated exercise in pa-
tient description, but as intertwined with
treatment. The value of an assessment in-
strument is enhanced if it enables a clinician
to pick out problems and goals that can
serve as targets for treatment. The instru-
ments described in this book for measuring
family factors related to relapse - the CPI,
KPI, and AS codes - are excellent ex-
amples of how assessment technology can
provide clinicians road maps for directing
treatment.
The research reported in this book on the
role of expressed emotion in depression
also reminds us that depression is a very
common problem for people suffering from
schizophrenia, particularly in the post-hos-
pitalization phase when the patient has re-
constituted from first-rank symptoms. We
can look forward to assessment instruments
that may help us to identify phase-specific
vulnerabilities to certain social stressors.
We may discover, for instance, that individ-
uals with schizophrenia who are in a de-
pressed phase of their illness may be more
sensitive to relapse induced by critical com-
ments made by workmates or family mem-
bers. Identification of phase-specific vul-
nerabilities can help us to refine and tailor
our treatment programs to better fit pa-
tients' needs. One very significant area
which has benefited from advances in as-
sessment is the conceptualization of "re-
lapse."

Operationalization of Relapse
Relapse in schizophrenia tends to be de-
fined too broadly and loosely by most inves-
tigators (Falloon, 1984). If we are to com-
pare the results from studies and build on
each other's work in a truly scientific way,
then we need to know how each research
group or team is actually defining and ope-
rationalizing relapse. Quantified scales for
measuring changes in psychopathology
have been developed at the UCLA Clinical
Research Center for the Study of Schizo-
phrenia which have proven useful in relia-
bly assessing relapse as increases in symp-
toms exceed certain threshold criteria. The
methods derive operational criteria from
changes in severity of psychotic symptoms
and signs as measured by standardized in-
terview using the Psychiatric Assessment
Scale (Krawiecka, Goldberg, and Vaughn,
1977) and the Brief Psychiatric Rating Scale
(Overall and Gorham, 1962).
Psychiatric Assessment Scale (PAS)
Criteria
Ratings on this scale are based on a brief 10-
to 20-min interview utilizing questions that
elicit psychotic symptoms from the PSE.
The PAS contains eight items which are
rated on a 0-4 scale of increasing severity,
and each scale point is operationally defin-
ed. For the purpose of defining relapse,
however, only the items assessing schizo-
phrenic symptoms were employed, e.g.,
hallucinations, delusions, incoherence.
The type I and II criteria presented in
Table 1 were developed after extensive dis-
cussion and prior application of various
scoring strategies to a sample of patients
Copying and Competence 207
who it was agreed would fall into "relapse"
and "no relapse" categories on clinical
grounds. The criteria were then applied to
28 schizophrenic patients who were partici-
pating in a controlled clinical trial of psy-
chosocial and drug therapy. The patients'
ratings on the three psychotic scales at dis-
charge were used as the base line from
which changes were assessed. Follow-up as-
sessments were conducted at 1, 3, 6, 9, 12,
18, and 24 months. Monthly telephone calls
or visits were made by the project's social
worker to ascertain whether a deterioration
in functioning or increase in symptomato-
logy might be occurring. Any suspected in-
dications of a change in symptoms or func-
tioning triggered an immediate PAS evalu-
ation by a psychiatrist or psychologist. Of
the 28 patients in the study, 24 could be
rated by these criteria. The remaining four
patients had such high levels of persisting
symptoms at discharge that even rating a re-
lapse as an exacerbation of existing symp-
toms could not be made. Interrater reliabi-
lities obtained by two sets of interviewers in
the United Kingdom and the United States,
utilizing the PAS, are shown in Table 2, and
are acceptably high.
Brief Psychiatric Rating Scale (BPRS)
Criteria
Operational criteria for two categories of
relapse were developed for an expanded
version of the BPRS that included three
new scales suitable for rating decompensa-
tion of schizophrenic outpatients: bizarre

Table 1. Relapse Criteria Based on the PAS Psychotic Symptom Scale of Delusions, Hallucinations,
and Incoherence
Type I:
If a change from discharge rating occurs on only one of the three scales, a 2-point increase is designated
as a relapse, providing that a maximum severity score of 4 on that scale occurs. Thus, score increases
from 0 to 2 or from 1 to 3 would not be considered a relapse, but an increase from 2 to 4 would be a re-
lapse.
Type II:
A total increase of 3 points on one or more of the three scales is designated a relapse with the caveat
that single-point changes from 0 to 1 are not counted. (Zero represents symptoms absent, while 1 indi-
cates symptoms not clearly pathological. )
208 R. P. Liberman
Table 2. Interrater Reliabilities of the PAS Psychotic Symptom Scales
Kendal's coefficient of
concordance W for five
psychiatrists' ratings of 10
videotaped interviews
Delusions .83
Hallucinations .78
Incoherence .64
behavior, self-neglect, and suicidality. Be-
cause the BPRS was designed primarily for
use with an inpatient psychiatric population
in psychopharmacology outcome studies, a
brief interview and anchor points applica-
ble to an outpatient setting were developed
for its use with outpatients participating in
longitudinal studies of course and outcome.
The two types of relapse coded were: (1)
psychotic relapse which is based on the core
BPRS psychotic symptom scales of. hal-
lucinations, unusual thought content (de-
lusions), and conceptual disorganization
and (2) other types of relapse, which signal
gross impairments in the patient's func-
tioning and thinking but which are not as
clearly related to schizophrenic psychotic
processes, that is, depression, suicidality,
self-neglect, bizarre behavior, hostility.
The BPRS items that were used in these cri-
teria for relapse are presented in Table 3
with their behavioral anchors and interrater
reliability coefficients. BPRS items are
rated on a 1- 7 scale of increasing severity.
A rating of 6 (severe) or 7 (extremely
severe) on the critical item signifies a re-
lapse.
In future studies of schizophrenia, it may
be advisable to avoid using qualitative
terms describing "states" of remission or re-
lapse. Many schizophrenic patients never
Research Instruments Have Clinical Valoe
There is the widespread, but misguided,
view that assessment instruments, such as
the Camberwell Family Interview (CFI)
yield rich lodes of information but are only
useful in research. Because they take so
much time to learn and use reliably, it is felt
that they cannot be readily disseminated to
Pearson Rs for 2 raters
co-interviewing 51 patients
.94
.99
.85
recover from an episode sufficiently to be
considered in remission. About 10% of
first-admission patients and 22% of longer-
term patients have such high persisting sym-
ptoms, even when they are discharged from
the hospital, that there is no way that one
could subsequently consider a rating a re-
lapse. These patients are near the top ofthe
psychopathology ceiling at all times. Thus,
instead of using "all or none" qualitative
concepts of symptomatic status, it would be
better to adopt methods that permit meas-
urement of psychopathology on a con-
tinuum. This is best done, as described in
Chap. 14, by frequent and repeated meas-
ures of the patient's key symptoms. This
method of monitoring target symptoms
would also be useful for assessing needs for
changes in antipsychotic medication and for
determining the lowest dose consistent with
sustained minimization of symptoms. Such
an approach has been used for decades in
medicine where repeated measures of lab-
oratory tests of organ function direct the re-
sponsible clinician to treatment decisions.
There is no reason why objective and stand-
ardized ratings of patient's psychopatholo-
gy cannot enable mental health profes-
sionals to titrate social therapies as well as
drug treatment.
clinicians in the field. I strongly diagree
with that view. We denigrate the interests,
attitudes, and motivation of clinicians to
suggest that the CFI is too cumbersome and
difficult for them to learn. We are now get-
ting to the point where our assessment in-
struments are providing predictive and
 Copying and Competence 209

Table 3. BPRS Scales and Behavioral Anchors Used to Define Various Types of Relapse. Interrater
Reliability is Expressed as Median Intraclass Correlational Coefficients.

Psychotic Relapse

Scale Rating Definition Reliability

item coefficient*
Unusual 6 Severe Full delusion( s) present with much preoccupation OR
thought many areas of functioning are disrupted by delusional
content thinking
.93
7 Extremely Full delusion( s) present with almost total preoccupation
severe OR most areas of functioning are disrupted by delusional
thinking
6 Severe Severeal times a day OR many areas of functioning are
disrupted by hallucinations
Hallucin- .97
ations 7 Extremely Persistent throughout the day OR most areas of
severe functoning are disrupted by hallucinations
Conceptual 6 Severe Speech is incomprehensible due to severe impairments
disorgan- most of the time
ization .73
7 Extremely Speech is incomprehensible throughout interview
severe
Other types of relapse
6 Severe Deeply depressed most of the time OR many areas if
functioning are disrupted by depression
Depression .90
7 Extremely Deeply depressed constantly OR most areas of
severe functioning are disrupted by depressive thinking
6 Severe Wants to kill self; searches for appropriate means and time
OR suicide attempt that is a potentially serious threat to
life with patient knowledge of possible rescue
Suicidality .97
7 Extremely Specific suicidal plan and intent (e.g., "As soon as _ _ _ ,
severe I will kill myself by doing X") OR suicide attempt
characterized by plan that the patient thought was lethal
or an attempt in a secluded environment
6 Severe Hygiene and eating potentially life-threatening, e.g., eats
and/or bathes only when prompted
Self- .78
neglect 7 Extremely Hygiene and eating life-threatening; does not eat or
severe engage in hygiene
6 Severe Unusual petty crimes, e.g., directing traffic, public nudity,
contacting authorities about imaginary crimes
Bizarre .84
behavior 7 Extremely Unusual serious crimes, e.g., setting fires, asocial theft,
severe kidnapping committed in bizarre fashion or for bizarre
reasons
6 Severe Has assaulted others but with no harm likely, e.g., slapped
or pushed others OR destroyed property, e.g., knocked
overfumiture, broken windows
Hostility .89
7 Extremely Has attacked others with definite possibility of harming
severe them or with actual harm, e.g., assault with hammer or
weapon
* Median ICC among the seven Developmental Processes BPRS raters.
210 R. P. Liberman
prognostic infonnation, at scientifically ac-
ceptable levels, that would be highly bene-
ficial to clinicians for guiding treatment.
Instruments have been carefully honed to
effectively elicit and rate high-quality infor-
mation on problems and goals that could be
chosen for treatment, the indications for
certain types of treatment, and most im-
portantly, assessments that reflect the pro-
gress that our patients are making over
time.
At some future time, we will have the so-
cial psychiatric equivalent of a thenno-
meter or a panel of blood chemistry tests to
aid our diagnosis and monitoring efforts.
We may ask the "family laboratory" for
KPI or EE "readings" on our patient and
family. These assessment instruments also
help to engage patients and families in
treatment, through their ability to produce
catharsis, to promote motivation, and to
make the family and patient more aware of
the problems they face. Most people who
have administered CFIs to families speak of
the powerful impact the interview has on
the establishment of rapport and the feeling
by the family members that they are being
understood for the first time. Thus, moti-
vating and preparing families to participate
in treatment is another strong justification
to incorporate these instruments into every-
day clinical practice.
However, the most important of the fac-
tors that will detennine whether any of our
assessment instruments will be adopted into
psychiatric and psychological practice are
their profitability and their "exportability."
Profitability is a major consideration for the
widespread professional acceptance of as-
sessment instruments. Neuropsychology is
developing rapidly as a clinical subspecialty
Intervention Design for the Future
Specifying the relevant domains of func-
tional outcome for therapy, based upon as-
sessments of patient and family deficits, is
vital for setting goals in our treatment ef-
forts. Planning for feasible and realistic out-
comes, consonant with the desires of pa-
tient and relatives, will enable us to apply a
minimum-maximum approach to treat-
ment. A minimum-maximum strategy of
because neuropsychological test batteries
are in great demand and insurance com-
panies will pay for their cost. When patho-
logists began doing laboratory studies in
hospitals, fees for the tests could be charged
that netted income to doctor and hospital;
they provided diagnostic and prognostic in-
fonnation that guided the treatment; and
they could be analyzed and used to assess
the progress patients were making. Labora-
tory tests were instituted, not as part of the
practice of the nurse or general practition-
er, but by new specialities of pathology,
clinical chemistry, and radiology. Radio-
logists do not ask whether they should re-
serve CAT scans for research only because
they are so complex, so expensive, and so
difficult to teach people to use. No, they use
them because they have scientific utility and
they provide useful infonnation.
"Exportability" of an assessment proce-
dure is an example of the transfer of knowl-
edge. Many factors have been identified
that determine how, when, and why clinical
research practices get adopted by other
practitioners (Libennan, 1983; Libennan et
aI., 1986). One of the most important of
these factors is having an "internal champi-
on" , a person with positive informal or for-
mal status within an institution who is wil-
ling to fight for that innovation, to "raise
the flag" and install the CFI, the KPI, or the
PSE. Once mental health professionals ex-
pand the clinical use of objective and scien-
tifically validated diagnostic and prognostic
tools, as radiologists and pathologists have
already done, we will see hospitals compet-
ing with each other to hire a KPI technician
or a CFI interviewer to provide assessment
services.
intervention has worked very well in the
area of sex therapy (Annon, 1974). In
behaviorally oriented therapy for sexual
dysfunctions, a certain proportion of cou-
ples only need a small amount of informa-
tion to correct their misbeliefs; for ex-
ample, a single session with some reassur-
ance and infonnation can produce improve-
ment in 10% of the couples. Depending

upon their assets and defiats, more severely
disordered couples require more extensive
counseling, adivice, and reassurance while
still others need more systematic behavioral
interventions such as training in sensate
focus. A residual number of couples need
discrete behavioral training plus couple
therapy that focuses on areas of their mar-
riage beyond the sexual sphere, such as
communication.
We can make our family interventions
more efficient by offering only so much as is
desired and needed. Some families will do
very well with just a few sessions of educa-
tional counseling or even a single an am-
Modular Interventions
Cost constraints in the delivery of mental
health service will move our interventions
beyond minimum-maximum strategies to
modular approaches that specify the ele-
ments of treatment that are best for certain
patient needs. We know, for example, that
neuroleptics seem best for the positive
symptoms of schizophrenia. Workers in
psychiatric rehabilitation also are develo-
ping methods for remediating specific social
deficits through social skills training (Liber-
man, Massel, Mosk, and Wong, 1985b).
One example of a modular approach to
training social problem-solving skills to
chronic mental patients comes from work at
the UCLA Clinical Research Center for
Schizophrenia (Foy, Wallace, and Liber-
man, 1983). Modules have been developed
for training a variety of social and indepen-
dent living skills, including personal hy-
giene, conversational skills, use of public
agencies, job finding, leisure and recrea-
tion, medication management, and money
management. The modules are constructed
to teach a patient specific functional skills,
to train the patient to solve problems that
may be encountered while attempting to
employ these new skills, and for practicing
the skills in vivo.
Because most chronic mental patients
suffer from a spectrum of deficits in social
and independent living skills, a comprehen-
sive yet modular program for remediating
these deficits has been designed and operat-
ed at a 500-bed Veterans Administration
Copying and Competence 211
nestic CFI, which often yields an emotion-
ally satisfying catharsis. Many more fami-
lies will have to go through a survival skills
workshop or an extended educational pro-
gram, perhaps sponsored by self-help or ad-
vocacy groups. It is possible that well-de-
signed videocassettes could help to fill the
need for family education, some of which
could be carried out in the privacy of the
home (Backer, Liberman, and Kuehnel,
1985). Families with less coping skills will
have to go through intensive problem-solv-
ing training and communication skills train-
ing. Some families will require a combination
of family, individual, and group-therapies.
Medical Center in Los Angeles (Liberman
and Foy, 1983). The program focuses on
goal setting and problem solving. Modules
are provided for training these skills that a
patient lacks. A module is composed of a
category of skills falling within a major life
domain and is constructed so that problem-
solving and interpersonal skills are taught
simultaneously.
Each module is divided into separate skill
areas with each area having specific behav-
iors that are taught to achieve personal ef-
fectiveness and competence. For example,
the module that teaches "medication self-
management" contains skill areas on (a)
learning about the benefits and side effects
of medications, (b) acquiring the skills of
medication self-administration, and (c)
coping with the side effects of medication.
Patients proceed through each module in
sequence, starting with an "introduction"
that aims to highlight the values and advant-
ages of the module to motivate subsequent
participation. After acquiring requisite
knowledge and skills in the "training"
phase, patients learn how to gather the tan-
gible and social resources required to put
the skills to use. For example, to negotiate
medication side effects or dosage issues
with the prescribing physician, a patient
must be able to use a telephone to make an
appointment to see the doctor and find
transport to take him to the doctor. After
solving such "resource management prob-
lems," patients anticipate "outcome prob-
 N
I-'
N

(:i
:0
t::
cr'
(1)

Structured 3
Florid symptoms activities Intensive Il>
of schizophrenia psychopharma- ::I
graded
cotherapy
expectations

Specialized
behavior therapy
Sx's remit? token economy Increase dose
eod
social learning

Depot drugs
and
med education
Low dose drugs
0<
mtermittant
supportive, drugs
goal-oriented
psychosocial Ax 0'
drug-free trial
Vocrehab

Social skills
tramlng
Psychosocial rehab
and community support
program
(social prostheses)

Consider
drug-free
trial

Fig. 2. Decision tree for clinical management of schizophrenic disorders: drug and psychosocial
treatment strategies

lems" which might interfere with the suc-
cessful implementation of the skills in the
natural environment. Thus, patients learn
how to deal with disappointing events, such
as the lateness or unpredicted absence of
Drug-Psychosocial Treatment Interactions
A modular approach to the treatment of
schizophrenia is diagramed in Figure 2. A
decision-tree model is used to delineate the
choices available to the practitioner in im-
plementing treatment that is modular and
flexibly responsive to the specific needs and
attributes of the patient, including the on-
going effectiveness oftreatment. This treat-
ment model, then, guides the clinician to in-
crementally provide additional modules of
intervention conditional upon the cumula-
tive impact of previous modules, and course
of illness.
The flow chart depicted in Figure 2 also
integrates drug therapy with psychosocial
therapy, thereby making the interactional
nature of treatment congruent with the
multifaceted determinants of course and
outcome of schizophrenia. For example,
patients with acute onset of schizophrenic
symptoms, good pre morbid adjustments,
and low EE families could be considered for
drug-free psychosocial treatment if special-
ly structured and richly programed inpa-
tient milieus were available (Carpenter,
McGlashin, and Strauss, 1977; Paul and
Lentz, 1977). In one comparison study of
intensive behavior therapy with custodial
state hospital treatment, only 10.7% of the
former patients required continuation neu-
roleptic therapy compared with 100% of
the latter patients (Paul and Lentz, 1977).
Similarly, a supportive family-style residen-
tial treatment program, where only 8% of
the patients received neuroleptic medica-
tion, was clinically as effective as standard
hospital treatment with regular doses of
medication. Furthermore, the patients
treated in the special-group home milieu
achieved more independent community
living and superior vocational outcomes
(Mosher and Keith, 1979).
Unless most or all of the prerequisites for
a drug-free trial are present, a patient with
florid symptoms of schizophrenia should be
Copying and Competence 213
the physician at the appointment. Patients
can enroll in one or more of these modules,
depending upon the extensiveness of their
deficits and the nature of the goals estab-
lished for their treatment.
given the benefit of optimal doses of neuro-
leptic drug in the context of a supportive
milieu - whether inpatient, day hospital,
home care, or residential care. Psychosocial
treatment during the period of acute symp-
tomatology should be aimed at calming the
patient, buffering and reducing levels of so-
cial and physical stimulation and stress, and
assisting the patient in integrating and
understanding the symptoms as part of an
illness process. The most effective psycho-
social treatment, in any clinical locale, con-
tains elements of practicality, concrete
problem solving of everyday challenges,
low-key socialization and recreation, en-
gagement in tasks that are attainable, and
specific goal orientation (Liberman, Fal-
loon, and Wallace, 1984). It is important
that a positive therapeutic alliance - begun
during acute episodes and continued inde-
finitely into the aftercare period - occupy a
central position in the overall strategy for
treating the schizophrenic patient, no mat-
ter how much drug or psychosocial treat-
ment contributes to the overall plan. This
relationship can be with the prescribing psy-
chiatrist or with a paraprofessional case
manager.
If the patient's symptoms respond to this
environmentally assisted drug regimen,
more specific psychosocial interventions
can be considered depending upon those
stressors in the environment and deficits in
personal functioning that appear to play
specific roles in relapse and community
maladjustment. Thus, family intervention
can be offered when high EE, negative af-
fective climate. and communication de-
viance are present in the family relation-
ships. Similarly, vocational rehabilitation
and social skills training can be provided
when deficits in these areas are interfering
with a patient's community adjustment and
quality of life. Family communication, em-
ployment, and social skills may be consider-
214 R. P. Liberman
ed protective factors - at the personal and
environmental levels - against vulnerabili-
ty to stress-linked relapse.
While space precludes a full annotation
of the decision tree for clinical management
of schizophrenic disorders shown in Figure
2, one further point needs to be made. Psy-
chosocial treatment, as well as drug thera-
py, should be long-term. It is clear from a
number of studies that the benefits accruing
from psychosocial intervention do not be-
come apparent before 12 months and are
even greater at 2 years following initiation
(Liberman et aI., 1984). It is likely that in-
definite, if not lifelong, psychosocial sup-
port, guidance, and training may be optimal
for most chronic schizophrenics. Since neu-
roleptic drugs are most effective in main-
taining symptomatic improvement when
continued indefinitely, it should not be sur-
prising that psychosocial rehabilitation ef-
forts are similarly optimized by continuing
care. The module in the flow chart labeled
"specialized behavior therapy, token econ-
omy, and social learning" is purposely
meant to be part of a closed loop. For
refractory patients who resist most efforts
at deinstitutionalization, this arduous form
of intervention may have to be sustained for
2 or more years. Such long-term commit-
ment to a consistent and predictable
psychosocial treatment environment has
been shown to yield highly beneficial out-
comes (Paul and Lentz, 1977; Banzett,
Liberman, Moore, and Marshall, 1984).
The value of a modular and individualiz-
ed approach to the combined drug and
psychosocial treatment of schizophrenia
References
Alevizos, P. N., DeRisi, W. J., Liberman, R. P.,
Eckman, T., and Callahan, E. (1978). The de-
velopment and application of a method of
direct observation for program evaluation.
Journal of Applied Behavior Analysis, 11:
243-257.
Andreason, N. (1986). Schizophrenia update. In
A. Frances (Ed.) Psychiatry 1986. Washington,
D.C.: American Psychiatric Press.
Annon, J. S. (1974). The behavioral treatment of
sexual problems, Honolulu: Enabling Systems,
Inc.
stems from the overarching importance of
individual differences in schizophrenia, dif-
ferences which are grounded in the wide
diversity of biological vulnerability, stres-
sors, and protective and potentiating fac-
tors that together determine course and
outcome of schizophrenic disorders. In-
deed, individual differences are found in all
medical disorders, and medical therapeut-
ics has been guided by an appreciation of
and adherence to the special needs and re-
sponses of the individual patient.
From yet another point of view, a mod-
ular approach to treatment makes sense. If
methods designed and tested in clinical re-
search settings are to be "exported" to serv-
ice settings and practitioners, the methods
must be adaptable to local conditions and
constraints. Each treatment and rehabilita-
tion system of care for chronically mentally
ill individuals has its own strengths and
weaknesses, its unique set of resources, and
its political, social, and economic concerns.
A comprehensive treatment program that
works in an urban setting may need con-
siderable "retooling" before it can be used
in a rural setting. Thus, by dismantling com-
prehensive programs into discrete elements
or modules, each treatment setting can
organize its own system of care that fits its
location, organization, and patient popula-
tion. Modularization of treatment for the
chronically mentally ill has already begun,
including a resource book containing ele-
ments of effective service for this important
group of patients (Liberman, Kuehnel,
Phipps, and Cardin, 1985a).
Backer, T., Liberman, R. P., and Kuehnel, T. G.
(1985). Living on the edge: An educational
videotape for helping families cope with chro-
nic schizophrenia. Available from Camarillo-
UCLA Research Center, c/o Training Co-
ordinator, Box A, Camarillo, CA 93011.
Banzett, L. K., Liberman, R. P. Moore, J. W.,
and Marshall, B. D. (1984). Long-term follow-
up of the effects of behavior therapy. Hospital
and Community Psychiatry, 35: 277-279.
Carpenter, W. T., McGlashin, T. H., and
Strauss, J. S. (1977). The treatment of acute

schizophrenia without drugs. American Jour-
nal of Psychiatry, 134: 14-20.
Chapman, L. J., Edell, W. S., and Chapman, J.
P. (1980). Physical, anhedonia perceptual
aberration, and psychosis processes. Schizo-
phrenia Bulletin, 6: 639-653.
Falloon, I. R. H. (1984). Relapse: A reappraisal
of assessment of outcome in schizophrenia.
Schizophrenia Bulletin, 10: 293-294.
Foy, D. W., Wallace, C. J., and Liberman, R. P.
(1983). Advances in social skills training for
chronic mental patients. In R. D. Craig and R.
J. McMahon (Eds.) Advances in Clinical.
Behavior Therapy. New York: Brunner/
Maze!.
Krawiecka, M., Goldberg, D., and Vaughn, M.
(1977). A standardized psychiatric assessment
scale for rating chronic psychotic patients. Acta
Psychiatric Scandanavica, 55: 299-308.
Liberman, R. P. (1982). Assessment of social
skills. Schizophrenia Bulletin, 8: 62-84.
Liberman, R. P. (1983). Sociopolitics of behav-
ioral programs in institutions and community
agencies. Analysis and Interventions in De-
velopmental Disabilities, 3: 131-259.
Liberman, R. P., Backer, T., and Kuehnel, T. G.
(1986). Dissemination and adoption of psycho-
social treatment innovations. Journal of Con-
sulting and Clinical Psychology, in press.
Liberman, R. P., Falloon, I. R. H., and Wallace,
C. J. (1984). Drug-psychosocial interactions in
the treatment of schizophrenia. In M. Mirabi
(Ed.) The chronically mentally ill: research and
services. New York: SP Medical and Scientific
Books.
Liberman, R. P. and Foy, D. W. (1983). Psychia-
tric rehabilitation for chronic mental patients.
Psychiatric Annals, 13: 539-545.
Liberman, R. P., Kuehnel, T. G., Phipps, C. C.,
Copying and Competence 215
and Cardin, V. (1985a). Resource book for
psychiatric rehabilitation. Available from
Camarillo/UCLA Research Center, c/o Train-
ing Coordinator, Box A, Camarillo, CA 93011.
Liberman, R. P., Massel, H. K., Mosk, M. D.,
and Wong, S. E. (1985b). Social skills training
for chronic mental patients. Hospital and Com-
munity Psychiatry, 36: 396-403.
Licht, M. H. (1984). Assessment of client func-
tioning in residential settings. In M. Mirabi
(Ed.) The Chronically mentally ill: research
and services. New York: SP Medical and
Scientific Books.
McGuire, M. T., and Polsky, R. H. (1982). An
ethological analysis of behavioral change in
hospitalized psychiatric patients. In S. A. Cor-
son and E. Corson (Eds.) Ethology and non-
verbal communication in mental health. New
York: Pergamon Press.
Mosher, L. and Keith, S. (1979). Research on the
psychosocial treatment of schizophrenia. Ame-
rican Journal of Psychiatry, 136: 623-631.
Nuechterlein, K. H. and Dawson, M. E. (1984).
A heuristic vulnerability-stress model of schiz-
ophrenic episodes. Schizophrenia Bulletin,
1984,10: 158-203.
Overall, J. E., and Gorham, D. R. (1962). The
Brief Pyschiatric Rating Scale. Psychological
Reports, 10: 799-812.
Paul, G. L. (1986). Assessment in residential
treatment settings. Champaign: Research
Press.
Paul, G. L., and Lentz, R. (1977). Psychosocial
treatment of the chronic mental patient. Cam-
bridge: Harvard University Press.
Strauss, J. S. and Carpenter, W. T., Jr. (1981).
Schizophrenia. New York: Plenum Medical
Books.
Subject Index

Acceptance 47, 79, 80, 82, 90 Camberwell Family Interview (CFI) 25. 26,
Acute phase 4 32. 38. 39.44. 51, 52. 59. 60. 61. 63. 73. 74.
- psychotics 111. 112 76.79,80.83.87,94.98,104. 110. 113. 158.
Adherence to medication 119 159,160,166.174,189.207,208,210.211,
Adopted children 3 212
Affection 47. 99 - Anglo-American 38. 39.41. 42. 43
Affective climate 193 Mexican-American 45
- communication 187. 190 - spanish 39
- relations 37 -. interrater reliability 63
- Style (AS) 52.53.55.57. 189. 190, 191. -. interviewing techniques 98
192.208 -, retest compliance 65
Aftercare X. 4 Catharsis 212
Aggression 102 Chicano 37
Alliance 124 Chronic patients 83
AMDP 82 Chronicity 2
Amotivation 151 Closeness 37
Anger 43.45.47.99.104. 113. 149 Cognitive overload 206
Anhedonia 197 Communication 99, 118, 119. 149. 174, 185.
Antiparkinsonian medication 150 187
Antipathy 43 - deficits 119
Antipsychotic medication 124. 125.210 - Deviance (CD) 104. 214
Anxiety 82. 135. 149 - skills 135
Apathy 105.151 - skills training 119. 125. 175
Approval 56 -. amorphous forms 112
Arousal 29 Community hostel 199
- hypothesis 165 - tenure 192
Assessment 119. 208 Competence 203, 205
Autonomic hyperarousal 206 Compliance 125. 126. 150. 156
- reactivity 203 Conceptual disorganization 210
Autonomy 31, 100 Conflict XI. 1. 118. 147
A voiders 136 Contact time 29
Conversational skills 212
Beck Depression Inventory (BDI) 87, 89 Convulsive treatment 2
Behavioral analysis 119. 174. 175 Coping XL 61. 82, 83,99, 101. 105. 119, 136.
- family therapy 173. 185 147.148.150.152.157.160, 185,187.191.
Belladonna 1 192, 203, 205, 206, 207, 212
Benefits 125 - skills 118
Benign criticism 53, 54 - styles 31. 46. 136. 139
Biobehavioral factors 203 Costs 184
- vulnerabilities 203 Cost-effectiveness 184. 185
Bizarre behavior 210 Crisis 149. 175
Blame 114. 135. 192 - intervention 113
Borderline 4 Critical 56
Brief Psychiatric Rating Scale (BPRS) 57. 179. - comments 39,43.44.45. 88. 90. 98. 161
209, 210 - comments. content analysis 44
Burnt-out 105 - remarks 32. 51. 52. 63
218 Subject Index
Criticism X. 26. 27. 45. 46. 51. 52. 53. 55, 57. - dual high 54
60. 79. 80, 81. 82, 83, 85, 89, 90, 99, 104, 123. - dual low 54
158. 165, 166. 189. 190. 192. 203 - high 27, 28, 29, 30. 31. 32, 33, 36, 41, 43,
- (CRIT) 65 45,46,47, 53, 55, 56, 57,59,60, 61, 63, 64,
65, 66, 67, 73, 74, 77, 85, 86, 88, 89, 90, 93,
94, 98, 99, 101, 102, 103, 104, 105, 113, 118,
Daily management 36
123, 147, 152, 155, 157, 158, 160, 161, 165,
DAS 90,93
188
Decompensation 209
- high, mothers 54
Deficit symptoms 112, 113
- low 27, 28, 29, 30, 31, 32, 33, 41, 43, 44, 45,
Deinstitutionalization 147.215
47, 53, 55, 56, 57, 60, 61. 64, 66, 67. 73. 74.
Delusions 32, 210
77, 85, 86, 88. 90. 93, 94, 98. 99. 100. 101.
Depression X. 32, 87. 89. 100. 208. 210
103, 104. 105. 118. 157, 158, 160
- endogenous 3
- modifications 95
Developing nations 35
- research 97
Diathesis stress model 123, 126, 187
Differential prognosis 35 - criticism 97
Disagreements 90 -. cross-cultural valididty 40. 41, 42
Discharge 14 -. determinants 103
-, development 33
-. psychopathological 16
Disengagement 151 -. ethnicity 42
Distress 43 -. fluctuation over time 103
Divorce 47 -. interactional correlates 51
Dominance 25 -. patient characteristics 42
Double-bind 104 -, psychophysiology 102, 118
-, response styles 31
Drugs antipsychotic 3
-, sociocultural context 47
-, antipsychotic X
- neuroleptic 4 -. socioeconomic status 41,42
DSM-III 109. 110. 111. 112 -. stability of 60
- Emotion, family 36
Dual-parent families 189. 190
- emotion. level of contact 25
Durations of looking 86
Dyadic Adjustment Scale (DAS) 89
Face-to-face contact 28.60. 74. 76. 155. 156.
158. 160. 163. 166
Economic benefits 184 - interaction 85. 86
Education 175 Familial commitment 47
- program 156, 158, 159, 160, 161, 166 - involvement 36
Educational program 131,212 Families mixed EE 54
- programs 109, 110, 112, 113, 114, 120 Familism 37
Electroconvulsive therapy 2 Family assessment 4
Electrodermal activity 29, 30 - assessment task 189
Emotinal overinvolvement 98 - bonds 45,47, 151
Emotional atmosphere 98 - burden 114.182
- Family Atmosphere (EFA) 79,80,81, - communication 119. 173
82.83 - Compliance Index (FCI) 72, 73
- overinvolvement (EOI) X. 25, 26, 27, 39. 53. - conflict 123
54. 55, 56, 57, 60. 63. 65. 68. 81. 83. 86. 98, - education IX, 117, 118, 119, 121, 126, 127,
99, 104, 123, 203 128, 129, 131, 133. 135, 136, 137, 138, 139,
- response 97 185, 206, 212
Empathy 100 - education program 120, 121, 123, 124, 134
Employment status 14 - education, live format 121
Ergotherapy 1 - education, mass media distribution 121
Ethnicity 38 - education, videotape format 121
Events 29 - etiology 123
Expectations 36, 123, 126, 135 - factors 208
Expressed emotion (EE) X, 5, 12,25,26.27, - functioning 30
28, 30, 32, 33, 36, 37, 38, 39,40, 41, 43, 51, - groups 119
52, 53, 57, 59, 61, 63, 65, 67, 73, 76, 77, 79, - interaction XI, 52, 80, 97, 98, 173, 203
85, 86, 89, 94, 97, 98, 102, 104, 105, 118, 119, - interaction, standardized situation 52
130, 147, 148, 149, 152, 155, 158, 159, 163, - intervention programs 113, 117
165, 166, 187, 190, 197, 207, 208, 211 - intervention. acute psychotics 113
 Subject Index 219
- involvement 119 Interactions 30, 54, 85, 94
- life XL 97, 98, 104, 147 - direct 53, 98
- management 179, 180, 182, 185, 190 - pattern 90. 93. 94
- ombudsman 134 - style 93
- problem solving 175 - task 53, 57, 189
- programs XI Interactional behavior 52, 55
- stress 123 Intermittent medication, intermittent 193
- support groups 138 Intermorbid adjustment 33
- system 188 International Pilot Study of Schizophrenia
- tension 207 (IPSS) 13. 109
- therapy X, 76, 77, 83,117.119.130.148. Interpersonal boundaries 151
152 - skills 212
- therapy indirect 61 Interrater reliability 63
- therapy psychoeducational X. 60. 61 Intervention research 95
- transactions X Interviewing techniques 98
- treatment. models of X Intolerance 100
Flashbacks 111 Intrusion 189. 190
Flexibility 101. 126. 127 Intrusiveness 31. 46, 53. 55. 56, 57, 98.
Frankfurter Beschwerdebogen (FBF) 82 102. 192
Frustration 43,47.99,100. 102. 113. 133 Irritability 100
Functional analysis 174 Isolation IX. 157

Generational boundaries 151 Job finding 212

Gittelman-Klein scale 17,18.19 Justifications 90
Goal setting 212
Gobal Assessment Scale (GAS) 13. 14. 15. Kategoriensystem fUr Partnerschaftliche Inter-
18.20 aktion (KPI) 87. 208. 211
Goldstein scale 17. 18 Kin network 37
Group therapy 3 Knowledge interview 157. 158
Guilt 53. 123. 126. 133. 135. 138. 149
Leisure 212
Life events 30. 77. 114. 174. 192. 203. 207
Habituation 29 - goals 174
Hallucinations 32. 210 Lithium 3. 111
Haloperidol 13 Locus of control 45
Helplessness 99
Hemodialysis 3 Maintenance medication 29
Home environment 36 Mania 111
Hope 123. 131. 137. 139 Manics young 111
Hopelessness 133 Marijuana 111. 114
Hospital discharge 56 Marital adjustment 88
Hospitalization 2.4. 13. 14,27,31. 57, 81. - discord 151
110,137,147. 150 - interaction 88
Hostility X. 25, 26. 27. 39. 65. 79. 80. 88, 98. - satisfaction 86. 89. 90. 94
104.158.210 - therapy 161. 165
Humor 56 Medication 28.76.77. 97. 105. 125. 126, 129,
Hygiene, personal 100.212 131. 133, 134. 136. 137, 138, 148
- compliance 28, 76
leo 13 - depot 2
Inactivity 46. 151 - dosage 2
Independent living skills 212 - evaluations 119
Indifference 79. 80, 81. 82. 83, 105 - intermittent 193
Indignation 43. 45 - neuroleptic 11
Individual management 179, 180, 185 - self-management 212
Industrialized countries 35 - targeted 193
Information processing 203 -. benefits 125
Inpatient Multidimensional Psychiatric Scale Megavitamins 150
(IMPS) 13 Mental health education 105. 119
Insulin 2 - health-professionals IX
Integration, psychosocial 4 - illness 37
220 Subject Index
Mexican-Americans 36, 37, 38, 39, 41, 42, 43, Premorbid adjustment 12, 19. 31. 57. 99
45,47 - social adjustment 180
- familism 47 Premorbid personalities, poor 103
Minimum-maximum strategies 211, 212 Present State Examanization (PSE) 27, 38, 57,
Modeling 176 60,61,63,87,110,111,158,159,160,208,
Modular interventions 212 209,211
Money management 212 Preventive psychiatry 28
Moral treatment 1, 3 Pride 57
Morale 123 Primary caretaker X
Morbidity 178, 180, 182 Problem-Solving XI, 105, 125, 173, 174, 175.
- clinical 182, 185 180, 183, 185, 187, 192, 206, 212, 214
- family 184, 185 - deficits 119
- social 182, 185 - skills 100, 103, 118, 119,212
Munster Family Interview (Munsteraner - strategies (PSS) 192
Familienbogen, MFB) 79, 80, 81, 83 - training 119,175,176
Prodromal symptoms 2, 205, 206
Prognosis 12, 14, 124, 150
Negative feelings 176
- symptoms 3, 105, 112 Prognostic criteria 2
- scale 19
Neuroanatomy 124
Propranolol 3
Neuroleptic medication 2, 3, 4, 13, 29, 51, 60,
Protective factors 28, 36, 203
81,83,110,112,155,159,163,188,206,212,
Psychiatric Assessment Scale (PAS) 209
214
Psychoeducational family therapies 147
- depot 13, 69
- model 148
- medication continuous 69 - programs 111, 120, 122
- therapy 2 Psychopathology, at discharge 14
Neuroses 3 -,severity of 99
Nonresponders 11 Psychophysiological studies 29
Nonverbal behavior 80 Psychosis 1, 2. 3, 12, 13
- acute 109,110,113,114
Occupational disintegration 14, 20 - affective 15. 19,20,22, 112
Ombudsman 149, 150 -. atypical 12
Opium 1 - paranoid 13, 15
Outcome 14 -, psychodynamics of 3
-, poor 12 -, reactive 12
-, predictors of 12, 16 - schizoaffective 12, 13. 15, 19,20,22
Overinvolvement 43,45,77, 158, 161, 162, -, schizophrenic 19,22
165, 166, 190 Psychosocial treatments 60
Overprotectiveness 79, 80, 81, 82, 165 Psychotherapy 3, 4, 150
Overstimulation 83, 102, 150 - long-term 4
- supportive 13

Parental conflict 103 Quinine 1

Patients, chronic 13, 114
Personality change 14, 16, 19 Recovery period 126. 131, 137, 138
Pessimism 133, 138 - process 119, 123, 125. 132, 133. 136, 139
Pharmacotherapy 2, 185, 193 -, social X
- maintenance IX, X Recreation 212
-, depot 2 Rehabilitation IX, 3
-, long-term 2 - psychosocial 3
Phenothiazine 4 Rehospitalization 83
Phillips scale 12, 17, 18, 19 Rejection 47, 79, 80, 82
Placebo Treatment 2 Relapse X, XI, 2, 11, 12, 25, 27, 30. 32. 33,
Positive feelings 176 36, 52, 56, 59, 63. 67, 68, 70. 72. 73, 74, 77.
- remarks 26, 27. 98. 104 83.97,98,105,118.123,147,150,152,155.
- symptoms 212 159, 163, 164. 165. 187. 191.203.206.207,
Pre-illness relationship 99 208, 209
Predictor 14, 18,36, 67 - criteria of 64. 81, 84
- of relapse 98 - definitions 81
Predominance 80 - operationalization 209
 Subject Index 221
-- predictors 12, 14, 18, 36, 67, 81, 98 - improvement 193
prevention 32, 193 - interaction 207
-- process of 85 - Interview Schedule 20
-- questionnaire 63 - learning 173, 215
- rate 2, 28, 29, 193 - management 20
- rate, in depression 32 - network 119,132
-- rates 28, 32, 69, 85, 86, 139 - reinforcement 176
-, risk for X, 29, 3 L 83 - role functioning 193
Relationship heterosexual 16 - skills 148,152,207,208
Relative-patient contact 28 - skills training 60, 151, 152, 206, 212, 214
Remission 210 - status 83
Residual symptomatology 57, 99 - stress 33
Resignation 79, 80 - support 100, 138
Role performance 152 - Support 5
- therapies 4
Sadness 43, 45, 149 - withdrawal 102, 112, 197
Schizophrenia 123 Sociocentric model 197
-courseof 11,51 Sociocultural processes 36
-, daily management 36 Socioeconomic status 41
- diagnosis of 121. 122, 131 Sociotherapy 13
-, dopamine hypothesis 124 Sorrow 43
- ethnopsychiatric models 44, 45 Stephens scale 12
-, etiology 123, 124 Stigma 45
-, family factors 36 Straus-Carpenter scale 19
-, genetic factors 123 Strauss-Carpenter scale 12, 17, 18
-, International Pilot Study (IPSS) 11 Stress XL 29, 30, 33, 36, 52, n 102, 103, 105,
-, misconceptions 121, 122 1 HL 123, In 133, 135, 137, 139, 147, 148,
-, misinformation 135 n
1 175, 185, 193, 203, 205, 206, 214, 215
-, neutrotransmitter theory 124 - level 137
-nuclear 112,114 management XL 129
-, outcome 11 Suicidality 210
- paranoid 3 Suicide 13, 15
-, sociocultural processes 36 Support 36, 37, 56
- spectrum disorders 32 - groups 119
-, symptoms 122 Supportive statements 56
Schizophrenics young 59, 60, 74, 76 Survival skills 120, 138
-, chronic 59 - skills workshop 119,121, 123, 126, 150,
Schizophrenogenic response styles 104 151, 206, 212
Schneiderian first-rank symptoms 110 Sympathy 43
Self-control 101 Systems theory 174
Self-disclosures 90
Self-neglect 210 Tardive dyskinesia 2
Self-statements negative 57 Target symptoms 179, 210
- positive 57 Targets of indifference 83
Sequential analyses 57 Tension 100,203
Shame 45 Therapeutic camps 196
Side effects 2,69,71, 105, 113,125,131,134, Therapy, behavioral 1
135,138,150, 163,212 -, pschodynamic 1
Silence 56 -, psychoanalytic 3
Single parents 190 Time budget 158, 160
Single-parent families 189 Token economy 215
Sleep 44 Tolerance 31, 99, 100
SociaL activities 180 Training 166
- adjustment 11, 12, 13, 15, 18, 20, 180, 208 Tranquilizer 125
- competence 208 Tranqulizer 124
- contact 164 Treatment, effectiveness 3
- control 101 -, inpatient 4
- distance 3 L 100 -, intermittent 2
- functioning 182, 205, 206, 207 -, long-term 2, 11
- impairment 105, 182 -, long-term effects 4
222 Subject Index
Treatment, low dose 2 Vaillant scale 12
-, maintenance 2
-, neuroleptic 11
-, outpatient 4
-, psychosocial 60
Twins 3
Understimulation 83
Unemployment 14,46, 103, 105
Vocational rehabilitation 214
Vulnerability 30, 36, 61, 83, 102, 103, 118,
148, 149, 205, 206, 207, 215
Vulnerability-stress model 203, 207, 208
Warmth 26, 27, 43, 52, 79, 80, 82, 90, 98, 104,
158
- remarks 51
Work 14, 180