Psychiatry Research 220 (2014) 1019–1027

Contents lists available at ScienceDirect

Psychiatry Research
journal homepage: www.elsevier.com/locate/psychres

Intelligence as a predictor of outcome in short- and

long-term psychotherapy
Paul Knekt a,b,n, Taru Saari a, Olavi Lindfors a
a
National Institute for Health and Welfare, Helsinki, Finland
b
Biomedicum Helsinki, Helsinki, Finland

art ic l e i nf o a b s t r a c t

Article history: Intelligence has been suggested as a suitability factor for short-term therapy whereas its possible effect
Received 24 June 2013 on short-term versus long-term therapy still is unknown. The aim of this study was to compare the
Received in revised form prediction of intelligence on the level of psychiatric symptoms and psychosocial functioning in
2 February 2014
psychotherapies of different lengths. A total of 251 outpatients from the Helsinki Psychotherapy Study,
Accepted 5 October 2014
aged 20–46 years, and suffering from mood or anxiety disorders were allocated to two long-term and
Available online 16 October 2014
two short-term therapies. Intelligence was assessed at baseline with the Wechsler Adult Intelligence
Keywords: Scale (WAIS-R). Psychiatric symptoms and psychosocial functioning were assessed 5–10 times during a
Depression 5-year follow-up using two primary symptom measures (HDRS and HARS) and one primary measure of
Intelligence
psychosocial functioning (GAF). Short-term therapy was more effective than long-term therapy during
Long-term
the first year of follow-up. During the second to fourth follow-up year no differences between short- and
Prediction
Psychotherapy long-term therapies or the intelligence groups were found. At the fifth follow-up year, however, long-
Short-term term psychotherapy showed a statistically significantly larger change in all three primary measures
compared to short-term therapy among those with higher intelligence. No differences between therapy
groups were noted in those with lower intelligence. People with higher intelligence may benefit more
from long-term than from short-term psychotherapy. These findings should be confirmed.
& 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction
It has been suggested that short- and long-term psychotherapies
apparently demand different attributes from the patient (Laaksonen
et al., 2013a). In short-term psychotherapy the aims of the therapy
are relatively clearly defined and the therapist has an active role in
facilitating fast development of the time-limited treatment process.
In contrast, in open-ended psychoanalysis and long-term psycho-
dynamic psychotherapy there is less focus on predefined goals and
therapist guidance and more weight is given to improvements in
self-awareness and insight (Gunderson and Gabbard, 1999) and to
implicit learning processes (Wong and Haywood, 2012). These
require that the patient, besides having the ability to commit to
the long treatment, is also able to form a fruitful alliance with the
therapist, has tolerance for anxiety, and has the motivation and
capacity to explore his/her internal world actively and thoroughly
(Gabbard, 2004). Likewise, patient's good cognitive capacities and
ego functions, such as an adequate level of reflective ability (Fonagy,
n
Corresponding to: National Institute for Health and Welfare, P.O. Box 30 00271
Helsinki, Finland, paul.knekt@thl.fi, Tel.: þ 358 29 524 8774.
E-mail address: paul.knekt@thl.fi (P. Knekt).
2010; Watzke et al., 2010), the ability for abstract thinking and
mastery of interaction with the environment (Allen et al., 1986;
Gabbard, 2004), all suggest greater potential for a more sustained
recovery in long-term explorative psychotherapy.
As far as these authors are aware, intelligence as a predictor of
psychotherapy outcome has been studied exclusively in short-
term therapies and after the 1980s covering only cognitive-
behavioral therapies (Haaga et al., 1991; Luborsky et al., 1996;
Fournier et al., 2009; Rizvi et al., 2009; D’Alcante et al., 2012 ) and
non-directive supportive counseling (Doubleday et al., 2002). The
results of these studies are contradictory. Intelligence did not
predict the outcome of cognitive-behavioral therapies in out-
patients suffering from depressive or anxiety disorders (Haaga
et al., 1991; Doubleday et al., 2002). However, higher intelligence
predicted more reduced anxiety in older adult patients treated
with non-directive counseling, thus suggesting its potentially
greater demands on abstract thinking than what is needed in
cognitive-behavioral therapy (Doubleday et al., 2002). In one
study, lower intelligence predicted poor response after cognitive
therapy in patients with depression (Fournier et al., 2009). Also, in
patients with obsessive–compulsive disorder, especially higher
verbal intelligence predicted a greater reduction of symptoms in
cognitive-behavioral therapy (D’Alcante et al., 2012). However,

http://dx.doi.org/10.1016/j.psychres.2014.10.011
0165-1781/& 2014 Elsevier Ireland Ltd. All rights reserved.
1020 P. Knekt et al. / Psychiatry Research 220 (2014) 1019–1027
among women with posttraumatic stress disorder, treated with
cognitive processing therapy, no prediction of intelligence on
outcome in symptoms was found (Rizvi et al., 2009).
Davanloo (1978) developed a comprehensive set of essential
suitability selection criteria for short-term psychodynamic psy-
chotherapy considering different aspects of ego functions, including
intelligence. Greater level of intellectual resources has also been
suggested beneficial for long-term psychodynamic psychotherapy
and psychoanalysis (Bacharach and Leaff, 1978; American
Psychiatric Association, (APA) (1985)). Since it is reasonable to
hypothesize that individuals with higher intelligence may be able
to benefit more in the long run from long-term psychotherapy and
since this hypothesis has not been studied previously, the present
study investigates the prediction of intelligence on changes in
psychiatric symptoms and psychosocial functioning among patients
receiving short- versus long-term psychotherapy during a 5-year
follow-up from the start of treatment. Since long-term therapy is
less effective during the first year of follow-up and more effective at
the end of follow-up (Knekt et al., 2013), the comparisons are
carried out separately for short and long follow-up.
2. Methods
2.1. Patients
This study is part of the Helsinki Psychotherapy Study, in which 506 eligible
outpatients (459 for short- and long-term psychotherapy and 47 for psycho-
analysis) were recruited from psychiatric services between June 1994 and June
2000 and followed for 5 years (Knekt and Lindfors, 2004). The protocol was
approved by the Helsinki University Central Hospital's ethics council. Written
informed consent was obtained from each patient. The patients and settings,
therapies and therapists, and assessment methods and statistical methods have
been presented in more detail elsewhere (Knekt and Lindfors, 2004; Knekt et al.,
2008), and are summarized briefly here.
Patients considered eligible were 20–45 years of age and had a long-standing
( 41 year) disorder causing dysfunction in work ability. They had to meet the
Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (American
Psychiatric Association, 1994) criteria for depressive or anxiety disorder. Patients
were excluded from the study if they were suffering from psychotic disorder or
severe personality disorder, bipolar I disorder, adjustment disorder, substance
abuse, organic brain disease or other severe organic disease, or mental retardation.
Individuals treated with psychotherapy within the previous 2 years, psychiatric
health employees, and persons known to the research team members were also
excluded.
Of the eligible patients, 139 refused to participate, and 326 of the remaining
367 patients were randomly assigned to solution-focused therapy (SFT), short-term
psychodynamic psychotherapy (SPP) and long-term psychodynamic psychotherapy
(LPP). In addition, 41 patients were selected for psychoanalysis (PA). The present
study was conducted on a randomly selected subpopulation of 251 patients who
had been assessed using the abbreviated Wechsler Adult Intelligence Scale-Revised
(WAIS-R) (Wechsler, 1981). Of these patients, 73 were in the SFT group, 71 in the
SPP group, 66 in the LPP group, and 41 in the PA group.
2.2. Therapies and therapists
SFT is a brief, resource-oriented, goal-focused therapeutic approach, which
helps clients change by constructing solutions (Johnson and Miller, 1994). The
orientation was based on an approach developed by De Shazer et al. (1986). The
frequency of sessions in SFT was flexible: usually one session every 2 or 3 weeks, up
to a maximum of 12 sessions, over a period of no more than 8 months.
SPP is a brief, focal, transference-based therapeutic approach which helps
patients by exploring and working through specific intrapsychic and interpersonal
conflicts. The orientation was based on approaches described by Malan (1976) and
Sifneos (1978). SPP was scheduled for 20 treatment sessions, with one session
per week.
LPP is an open-ended, intensive, transference-based therapeutic approach
which helps patients by exploring and working through a broad area of intrap-
sychic and interpersonal conflicts. Therapy includes both expressive and supportive
elements, depending on the patient's needs. The orientation followed the clinical
principles of LPP (Gabbard, 2004). The frequency of sessions in LPP was 2–3 times a
week during a period of approximately 3 years.
PA is an open-ended, highly intensive, transference-based psychodynamic
therapeutic approach which helps patients by analyzing and working through a
broad area of intrapsychic and interpersonal conflicts. The therapeutic setting and
technique are characterized by facilitating maximum development of transference
by the use of a couch and free association for exploring unconscious conflicts,
developmental deficits, and distortions of intrapsychic structures (Greenson, 1985).
The frequency of sessions in PA was four times a week during a period of
approximately 5 years.
Altogether 60 therapists participated in the study. A total of five therapists gave
solution-focused therapy, 11 short-term psychodynamic psychotherapy, 29 long-
term psychodynamic psychotherapy, and 25 psychoanalysis. The therapists had
practiced for at least 2 years after training in the specific form of psychotherapy.
The mean number of years of experience in the therapy provided was 9 in SFT and
SPP, 18 in LPP, and 15 in PA.
The therapists giving psychodynamic psychotherapy or psychoanalysis had
received standard training in psychoanalytically oriented psychotherapy (Knekt
et al., 2008). During their training the therapists received a minimum of 3–6 years
of analytical (psychoanalysis or psychotherapy) training, and those giving short-
term therapy received 1–2 years of additional short-term focal psychodynamic
psychotherapy training. Also therapists giving SFT had been trained to use the
method and received a qualification for the method from a local institute. None of
the therapists had received any training in psychodynamic psychotherapy and
vice versa.
In psychodynamic psychotherapies, the treatment was provided in accordance
with clinical practice using “manual like general guidelines”, i.e., the therapists
could modify their interventions according to patients' needs (Knekt et al., 2008).
In contrast, the solution-focused therapy was given according to a manual, and
adherence control was organized. Since the effectiveness of the two short-term
therapies did not differ in this data (Lindfors et al., 2012;Knekt et al., 2013;
Laaksonen et al., 2013b) they were combined to one group.
2.3. Assessment methods
Intelligence was assessed at baseline using an abbreviated version of the test
Wechsler Adult Intelligence Scale-Revised (WAIS-R) (Wechsler, 1981), covering the
full scale intelligence quotient (IQ), the verbal IQ, and the non-verbal performance
IQ. The tests Digit Span, Arithmetic, Comprehension, and Similarities were used for
the verbal IQ and the tests Picture Completion, Picture Arrangement, Block Design,
and Digit Symbol for the performance IQ. In this study, all three quotients were
used as predictors and were for the purposes of the analyses divided into two
categories by the median: low and high.
Descriptive and potential confounding factors were assessed at baseline using
questionnaires and interviews. Psychiatric diagnoses (American Psychiatric
Association, 1994) were assessed using a semi-structured interview (Knekt and
Lindfors, 2004). Psychiatric history (age at the onset of primary psychiatric
disorder, duration of primary psychiatric disorder, and separation experiences at
childhood), previous psychiatric treatment (i.e., psychotherapy, psychotropic med-
ication, or psychiatric hospitalization), and socioeconomic factors (age, sex, marital
status, education, occupation, and current employment status) were assessed using
questionnaires. Personality functioning was assessed by the scales Level of
Personality Organization (LPO) (Kernberg, 1996; Valkonen et al., 2012), Quality of
Object Relations (QORS) (Azim et al., 1991), Immature Defense Style as per the
Defense Style Questionnaire (DSQ) (Andrews et al., 1989), Affiliation Toward Self as
per the Structural Analysis of Social Behavior questionnaire (SASB) (Benjamin,
1996), Interpersonal Problems (IIP) (Horowitz et al., 2000), and the Suitability for
Psychotherapy Scale (SPS) (Laaksonen et al., 2012) and social functioning by the
scales Social Adjustment (SAS-SR) (Weissman and Bothwell, 1976) and Sense of
Coherence (SOC) (Antonovsky, 1993).
The outcome measures of this study were psychiatric symptoms and psycho-
social functioning assessed during a 5-year follow-up from start of therapy. The
primary measures were based on interviews conducted by experienced clinical
raters at baseline and four times thereafter (at 7, 12, 36, and 60 months from
baseline) and the secondary measures were based on self-report questionnaires
filled in at baseline and nine times thereafter (at 3, 7, 9, 12, 18, 24, 36, 48 and 60
months from baseline) (Knekt and Lindfors, 2004).
The three primary measures were the interviewer-assessed Global Assessment
of Functioning scale (GAF) (American Psychiatric Association, 1994), the 17-item
Hamilton Depression Rating Scale (HDRS) (Hamilton, 1960), and the 14-item
Hamilton Anxiety Rating Scale (HARS) (Hamilton, 1959). The secondary measures
were the self-report 90-item Symptom Check List, Global Severity Index (SCL-90-
GSI) (Derogatis et al., 1973), 13-item Symptom Check List, Depression Scale (SCL-
90-DEP) (Derogatis et al., 1973), 21-item Beck Depression Inventory (BDI) (Beck
et al., 1961), and 10-item Symptom Check List, Anxiety Scale (SCL-90-Anx)
(Derogatis et al., 1973).
The assessment of compliance was based on information regarding waiting
time from randomization to the initiation of treatment, the completeness of the
treatment (i.e., withdrawal after randomization and discontinuation of treatment),
and the use of auxiliary treatment (i.e., additional psychotherapy, psychotropic
medication use, and hospitalization) at baseline and the nine measurement points
during the 5-year follow-up (Knekt et al., 2011). Auxiliary treatment was assessed
by questionnaires, interviews, and based on nationwide health registers.
 P. Knekt et al. / Psychiatry Research 220 (2014) 1019–1027 1021

After assignment to the treatment groups, 17 patients (four SFT patients, two
SPP patients, 10 LPP patients, and one PA patient) refused to participate. Of the
patients starting the assigned therapy, 32 (seven assigned to SFT, seven to SPP, nine
to LPP, and nine to PA) discontinued the treatment prematurely. Altogether 89 of
the 227 patients (30 in SFT, 33 in SPP, 16 in LPP, and 10 in PA) received auxiliary
treatment at the end of follow-up.
The mean drop-out rate over the measurement occasions during the follow-up
was 16% in the SFT, 14% in the SPP, 11% in the LPP, and 1% in the PA group. The major
reason for the drop-out from measurement was refusal on the ground that the
study occasion was felt to be mentally stressful or the patient was disappointed
with the treatment.
2.4. Statistical methods
A cohort study design with repeated measurements was used. The main analyses
were based on the ‘intention-to-treat’ (ITT) design which included all the patients
who had been randomized. Complementary ‘as-treated’ (AT) analyses were also
performed (Knekt et al., 2008). The primary analyses were based on the assumption
of ignorable dropouts from the outcome measures (Harkanen et al., 2005). In
secondary analyses, missing values of one outcome measure were replaced by
multiple imputations using existing information on the other outcome measures
(Rubin, 1987). The imputation was based on Markov chain Monte Carlo methods. The
variables in the imputation model were assumed to follow a multinormal distribu-
tion. The ITT and AT models were carried out based on both the original data and the
imputed data. As there were no notable differences between the ITT and AT models
or between the original ITT and the imputed models (data not shown), the results
based on the ITT model were presented in the results.
Linear mixed models (Verbeke and Molenberghs, 1997) with random effects
for patient and time (Harkanen et al., 2005) were used in the statistical analysis.
The dependent variables were the outcome measures (HDRS, HARS, GAF, SCL-90-
GSI, BDI, SCL-90-DEP, and SCL-90-ANX). The basic independent variables of the
ITT model included the IQ measured at baseline, therapy group (short-term (SFT
or SPP) versus long-term (LPP or PA)), and the time of measurement during
follow-up, their first- and second-order interactions, and a correction term
(including the interaction of the difference between theoretical and realized date
of measurement, time, and intelligence). The following confounding factors,
measured at baseline, were included in the full model: age, sex, marital status,
education, age at the onset of primary psychiatric disorder, duration of primary
psychiatric disorder, previous depressive episodes, separation experiences at
childhood, psychotropic medication, and the scores of LPO, QORS, DSQ, SASB,
IIP, SAS, SOC, all satisfying the criteria for confounding (Rothman and Greenland,
1998). The AT model was created by adding to the ITT model the waiting time
from randomization to the initiation of treatment, the completeness of study
treatment, and auxiliary treatment during follow-up mainly as time-dependent
covariates.
The independent variable of main interest was the interaction term between
the IQ, therapy group, and time. The statistical significance of this term was tested
with the Wald test. The significance of the change in the outcome variable from
the baseline point to the different measurement points for each therapy group
(short-term and long-term) and category (low and high) of intelligence consid-
ered was also tested using the Wald test. The differences in outcome between the
therapy by categories of IQ at the different measurement points were model-
adjusted (Lee, 1981). The delta method was used for the calculation of their
confidence intervals (Migon and Gamerman, 1999). All statistical analyses were
performed with SAS software, version 9.2 (SAS Institute Inc., 2008).

Table 1
Baseline characteristics by full scale IQ tertile.

Characteristic All Full scale IQ P-value for

heterogeneity
Low (N ¼ 87) Medium (N¼ 84) High (N ¼ 80)
(range¼ 71–107) (range ¼108–114) (range¼ 115–137)

Demographic variables
Women (%) 71.7 78.2 75.0 61.3 0.04
Age (years), mean (S.D.) 32.0 (6.7) 31.1 (5.3) 31.4 (7.5) 33.5 (7.1) 0.04
Living alone (%) 40.2 41.4 47.6 31.3 0.10
University degree (%) 27.1 14.9 26.2 41.3 o0.001
Low-level employee (%) 28.3 42.5 23.8 17.5 o0.001
Currently employed or studying (%) 83.9 83.5 81.0 87.5 0.52

Diagnoses
Mood disorder (%) 52.6 46.0 51.2 61.3 0.14
Anxiety disorder (%) 19.1 23.0 15.5 18.8 0.46
Co-morbid mood and anxiety disorder (%) 28.3 31.0 33.3 20.0 0.13
Personality disorder (%) 23.1 27.6 22.6 18.8 0.40

Symptoms
Symptom Check List Global Severity Index (SCL-90-GSI), 1.27 (0.5) 1.31 (0.5) 1.32 (0.5) 1.18 (0.5) 0.16
mean (S.D.)
Beck Depression Inventory (BDI), mean (S.D.) 17.9 (7.7) 18.1 (8.3) 17.9 (7.1) 17.6 (7.8) 0.91
Hamilton Depression Rating scale (HDRS), mean (S.D.) 15.4 (4.6) 15.0 (4.8) 15.2 (4.3) 16.0 (4.8) 0.31
Hamilton Anxiety Rating scale (HARS), mean (S.D.) 15.1 (5.3) 15.7 (5.3) 14.4 (5.1) 15.2 (5.5) 0.28
Symptom Check List Anxiety scale (SCL-90-Anx), mean (S.D.) 1.24 (0.7) 1.2 (0.6) 1.3 (0.7) 1.2 (0.7) 0.29
Symptom Check List Depression scale (SCL-90-Dep), mean (S.D.) 2.0 (0.7) 2.0 (0.8) 2.1 (0.7) 2.0 (0.7) 0.36
Global Assessment of Functioning scale (GAF), mean (S.D.) 55.5 (7.2) 56.3 (6.9) 55.0 (5.9) 55.0 (8.7) 0.37

Psychiatric history
Previous psychotherapy (%) 23.7 24.7 21.4 25.0 0.84
Psychotropic medication (%) 21.4 24.1 24.4 15.2 0.27
Psychiatric hospitalization (%) 1.2 0 0 4.0 0.04
Separation experiences at childhood (%) 44.6 49.4 45.2 38.8 0.38
Proportion of patients whose age at the onset of primary 62.9 61.9 65.1 61.5 0.88
psychiatric disorder under 22 years (%)
Duration of primary psychiatric disorder more than 5 years (%) 32.9 39.5 27.7 31.3 0.25

Personality and social functioning

Suitability for Psychotherapy Scale (SPS), (proportion of poor 17.9 17.2 17.9 18.8 0.97
score values, %)
Level of Personality Organization scale (LPO), mean (S.D.) 4.16 (0.68) 4.23 (0.70) 4.18 (0.67) 4.07 (0.68) 0.32
Quality of Object Relations Scale (QORS), mean (S.D.) 5.07 (0.63) 5.04 (0.55) 5.00 (0.70) 5.17 (0.62) 0.21
Immature defense style score (DSQ), mean (S.D.) 3.90 (0.74) 3.89 (0.81) 3.84 (0.71) 3.97 (0.70) 0.53
Affiliation toward self score (SASB), mean (S.D.) 4.89 (62.2) 16.6 (60.4) 0.22 (60.2) -2.43 (65.1) 0.11
Interpersonal problems score (IIP), mean (S.D.) 87.3 (31.0) 86.8 (31.9) 90.6 (30.6) 84.4 (30.3) 0.43
Social Adjustment Scale (SAS), mean (S.D.) 2.17 (0.38) 2.15 (0.40) 2.21 (0.37) 2.16 (0.38) 0.57
Sense of Coherence scale (SOC), mean (S.D.) 113 (20.4) 119 (20.6) 109 (19.4) 111 (20.1) 0.003
1022 P. Knekt et al. / Psychiatry Research 220 (2014) 1019–1027

3. Results improvement of functioning in short-term therapy during the first

Descriptive variables of the 251 patients were presented in the
whole sample and in tertiles of the full scale IQ in Table 1. The
means of the full scale IQ, the verbal IQ, and the performance IQ
were 110.4, 109.8, and 109.9, respectively (Table 2). The standard
score of different sub-tests in different psychotherapy groups
ranged from 10.2 to 13.3. No significant differences were found
with regard to intelligence in different psychotherapy groups.
The patients experienced and maintained a statistically signifi-
cant symptom reduction for all six measures, HDRS, HARS, GSI, BDI,
DEP, and ANX, and a significant improvement in psychosocial
functioning (GAF) at the measurement points during the 5-year
follow-up in comparison with the baseline point in all four
categories of the combined therapy (short- and long-term) and
intelligence (values under the median and above the median)
groups considered (Table 3). The difference in the changes between
short- and long-term therapies was not statistically significant
between individuals in the higher and lower intelligence categories
over the total 5-year follow-up (P for interaction¼ 0.11–0.99). At
single measurement points, however, differences emerged.
During the first year of follow-up, short-term psychotherapy was
more effective than long-term psychotherapy in the reduction of
depressive symptoms with a statistically significantly larger reduc-
tion at either the 7- or 12-month measurement point for all three
depressive symptom measures, HDRS, BDI and ANX, in both full
scale IQ groups (Table 3). No differences were found with regard to
changes in anxiety symptoms. During the next 3 years of the
follow-up no statistically significant differences were found
between long- and short-term therapies in either IQ group.
At the 5-year follow-up point, however, individuals with IQ
above the median showed a statistically significant change in the
three primary measures GAF (Mean difference (M):  6.11, 95%
confidence interval (CI):  10.8,  1.40), HDRS (M: 2.37, CI: 0.20,
4.54) and HARS (M: 2.95 CI: 0.79, 5.12) in the long-term therapy
group in comparison with the short-term therapy group. The
corresponding non-significant differences in individuals with IQ
under the median were 3.04, 1.34 and 1.69, respectively.
Study of verbal IQ (Table 4) separately showed only
minor differences with the full scale IQ (Table 3). In divergence,
however, statistically significantly more symptom reduction and
year of follow-up was seen only in patients with higher perfor-
mance IQ (Table 5).
4. Discussion
As far as these authors know, the present study is the first to
compare the prediction of intelligence on the reduction of psy-
chiatric symptoms in short-term and long-term psychotherapy.
We found that, although intelligence did not predict differences
between long- and short-term therapies during the first years of
follow-up, individuals with higher intelligence were likely to
benefit more from long-term psychotherapy in the long run. For
people with lower intelligence, long-term therapy was no more
beneficial than short-term therapy at any follow-up point. This
finding may be explained by the suggestion that individuals with
higher intelligence have more potential to obtain better results
from long-term therapy due to the fact that the explorative
orientation of long-term psychodynamic psychotherapy and psy-
choanalysis demands good capacity for abstract and metaphoric
thinking and reflective functioning (Gabbard, 2004; Valbak, 2004).
Accordingly, persons with higher intellectual capacity have more
potential for increasing their self-understanding and for gaining
from the relational corrective experiences by a combination of
thorough explicit cognitive and implicit emotional learning (Wong
and Haywood, 2012). Our study thus supports the importance of
taking the patient's level of intelligence into account when
choosing whether to refer the patient to short-term therapy or
to some other treatment.
The finding was more evident for the interviewer-rated mea-
sures on symptoms (HDRS and HARS) and global functioning
(GAF) than for the questionnaire measures BDI and SCL-90. It is
possible that patients with higher intelligence might be able to
verbally present their symptoms and functional status in more
detail and complexity, thus increasing the interviewer's ability to
detect more problems in symptomatic and functional status,
potentially biasing the evaluation. However, the fact that the
self-reported findings were in line with interview assessment
does not support this hypothesis.

Table 2
Full scale, verbal, and performance IQ and standard scores of part tests by psychotherapy group.

Intelligence test All Therapy group P-value for

heterogeneity
Solution- focused Short-term psycho- Long-term psycho- Psycho- analysis
(N ¼ 73) dynamic (N¼ 71) dynamic (N ¼ 66) (N ¼41)

WAIS-R full scale IQ, mean (S.D.) 110.4 (10.1) 110.0 (11.5) 111.2 (9.4) 108.6 (9.7) 112.5 (8.7) 0.21
WAIS-R verbal IQ, mean (S.D.) 109.8 (10.6) 109.5 (12.4) 110.2 (10.0) 107.9 (9.4) 112.9 (9.3) 0.12
WAIS-R performance IQ, 109.9 (10.8) 109.7 (10.8) 111.0 (10.8) 108.5 (11.8) 110.5 (9.4) 0.59
mean (S.D.)
Digit span, standard score, 10.9 (3.1) 10.9 (2.9) 10.4 (2.1) 11.4 (2.9) 0.33
mean (S.D.)
Arithmetic, standard score, 11.2 (3.0) 11.6 (2.4) 11.2 (2.9) 12.0 (2.3) 0.48
mean (S.D.)
Comprehension, standard score 11.8 (2.1) 12.0 (1.8) 11.6 (1.7) 12.4 (1.6) 0.11
mean, (S.D.)
Similarities, standard score mean 12.1 (2.7) 12.5 (2.3) 12.4 (2.3) 13.3 (2.1) 0.10
(S.D.)
Picture Completion, standard score, 10.5 (2.1) 10.9 (2.6) 10.2 (2.7) 10.7 (1.9) 0.38
mean (S.D.)
Picture Arrangement, standard 10.6 (3.1) 11.4 (2.6) 10.5 (2.8) 11.2 (2.7) 0.23
score, mean (S.D.)
Block Design, standard score, mean 10.9 (2.6) 11.6 (2.8) 11.7 (3.0) 12.0 (2.6) 0.16
(S.D.)
Digit Symbol, standard score, mean 11.8 (1.9) 11.6 (2.4) 11.9 (2.3) 12.0 (3.0) 0.85
(S.D.)
Table 3
Mean valuesa of psychiatric symptoms and psychosocial functions in short-term (S) and long-term therapy (L) groups and mean value differencesb (95% confidence intervals) at 7-, 12-, 24-, 36-, 48- and 60-month follow-up
according to the low and high values of patients' full scale IQ.

Measure Therapy Low full scale IQ (71–110) High full scale IQ (111–137) P-valuec

0 7 12 24 36 48 60 0 7 12 24 36 48 60

GSI S 1.27 0.81 0.83 0.86 0.83 0.73 0.64 1.24 0.93 0.83 0.84 0.80 0.74 0.70 0.99
L 1.39 1.04 1.00 0.90 0.84 0.72 0.62 1.28 0.99 0.95 0.78 0.72 0.57 0.55
S–L  0.18n  0.12  0.01 0.03 0.05 0.07  0.09  0.15 0.03 0.04 0.12 0.10
( 0.35, 0.03) (  0.30, 0.06) (  0.22, 0.20) (  0.19, 0.24) (  0.14, 0.25) (  0.12, 0.26) ( 0.26, 0.07) (  0.31, 0.02) (  0.16, 0.23) (  0.17, 0.24) (  0.06, 0.30) (  0.07, 0.28)
BDI S 17.3 9.32 9.38 9.04 8.80 7.83 8.15 17.0 10.0 9.22 8.79 9.36 7.96 8.61 0.50
L 20.1 13.7 12.7 10.2 9.99 7.06 6.27 17.6 12.0 11.3 8.02 7.41 4.92 5.49
S–L  2.99n  2.09 0.24 0.01 1.86 2.99  2.92n  2.93n  0.22 1.17 2.18 2.02
( 5.89,  0.09) (  5.14, 0.96) (  3.17, 3.64) (  3.68, 3.69) (  1.52, 5.24) (  0.26, 6.24) ( 5.62,  0.22) (  5.76, 0.10) (  3.36, 2.92) (  2.28, 4.61) (  0.93, 5.29) (  0.97, 5.02)
ANX S 1.24 0.89 0.80 0.81 0.81 0.68 0.62 1.27 0.98 0.90 0.87 0.77 0.72 0.66 0.99
L 1.17 0.89 0.90 0.76 0.76 0.66 0.52 1.30 1.01 0.96 0.79 0.67 0.50 0.50
S–L  0.17  0.12 0.03 0.02 0.00 0.08  0.08  0.11 0.04 0.06 0.16 0.12
( 0.39,0.05) (  0.34,0.11) (  0.23,0.29) (  0.24,0.28) (  0.22,0.23) (  0.15,0.32) ( 0.29,0.13) (  0.32,0.11) (  0.20,0.28) (  0.19,0.30) (  0.05,0.37) (  0.10,0.33)

P. Knekt et al. / Psychiatry Research 220 (2014) 1019–1027

HDRS S 14.4 9.66 10.6 9.88 8.33 15.9 10.7 10.6 10.6 10.3 0.49
L 15.4 11.7 12.2 10.0 7.99 15.8 12.5 12.4 9.21 7.51
S–L  1.53  1.00 0.34 0.79  0.97  2.19n 0.93 2.37n
( 3.40, 0.34) (  3.10, 1.09) (  2.17, 2.84) (  1.55, 3.13) ( 2.72, 0.79) (  4.16,  0.22) (  1.50, 3.37) (0.20, 4.54)
HARS S 14.5 9.12 9.31 9.13 7.61 15.2 11.4 10.9 9.89 10.1 0.54
L 15.7 11.4 11.0 9.03 7.51 15.2 11.7 11.9 8.70 7.03
S–L  1.59  1.04 0.72 0.73  0.41  1.07 0.97 2.95n
( 3.36, 0.19) (  3.00, 0.93) (  1.56, 2.99) (  1.61, 3.06) ( 2.08, 1.25) (  2.91, 0.78) (  1.22, 3.16) (0.79, 5.12)
DEP S 1.96 1.26 1.30 1.26 1.29 1.12 0.99 1.97 1.40 1.19 1.24 1.14 1.08 1.06 0.73
L 2.11 1.65 1.62 1.39 1.38 1.13 1.02 2.05 1.60 1.55 1.13 1.17 0.77 0.82
S–L  0.34n  0.26  0.07  0.04 0.04 0.02  0.23  0.41n 0.05  0.12 0.23 0.17
( 0.63,  0.05) (  0.57, 0.04) (  0.43, 0.29) (  0.40, 0.32) (  0.30, 0.37) (  0.30, 0.34) ( 0.50, 0.05) (  0.69,  0.12) (  0.29, 0.38) (  0.46, 0.22) (  0.08, 0.54) (  0.13, 0.47)
GAF S 56.6 67.5 68.0 69.0 74.1 55.0 63.4 63.4 66.6 68.4 0.11
L 55.9 62.8 62.0 66.0 69.9 54.6 64.4 62.0 66.0 69.9
S–L 3.46 4.70n 1.74 3.04 0.15 2.69  1.00  6.11n
( 0.21, 7.11) (0.63, 8.78) (  3.13, 6.60) (  2.02, 8.10) ( 3.29, 3.60) (  1.15, 6.53) (  5.73, 3.73) (  10.8,  1.40)

a
The full ITT model.
b
The full ITT model further including the baseline level of the outcome measure considered.
c
P-value for interaction between the full scale IQ measure and the therapy group throughout the follow-up.
n
P-value for difference between S and L o 0.05.

1023
 1024
Table 4
Mean valuesa of psychiatric symptoms and psychosocial functioning short-term (S) and long-term therapy (L) and mean value differencesb (95% confidence intervals) at 7-, 12-, 24-, 36-, 48- and 60-month follow-up according to the
low and high values of patients' verbal IQ.

Measure Therapy Low verbal IQ (71–110) High verbal IQ (111–137) P-valuec

0 7 12 24 36 48 60 0 7 12 24 36 48 60

GSI S 1.27 0.87 0.83 0.90 0.87 0.76 0.67 1.24 0.89 0.82 0.81 0.78 0.73 0.68 0.92
L 1.32 1.02 1.01 0.92 0.80 0.71 0.59 1.29 1.01 0.94 0.77 0.77 0.59 0.58
S–L  0.13  0.16  0.01 0.08 0.07 0.10  0.13  0.12 0.04  0.01 0.12 0.08
( 0.30, 0.04) (  0.34, 0.02) (  0.22, 0.20) (  0.13, 0.29) ( 0.12, 0.26) (  0.08, 0.29) (  0.29, 0.03) (  0.29, 0.04) (  0.16, 0.23) (  0.21, 0.19) ( 0.06, 0.30) (  0.10, 0.25)
BDI S 17.4 9.87 9.79 9.53 8.88 8.28 9.13 17.0 9.66 8.84 8.38 9.39 7.68 7.77 0.30
L 19.3 13.3 12.5 9.88 9.24 6.27 5.22 18.3 12.4 11.5 8.29 8.29 5.78 6.52
S–L  2.90n  2.28 0.23 0.15 2.40 4.35n  2.84n  2.77  0.12 1.01 1.68 0.84
( 5.78,  0.03) (  5.81, 0.07) (  3.14, 3.59) (  3.53, 3.82) ( 0.92, 5.72) (1.20,7.51) (  5.56,  0.12) (  5.61, 0.07) (  3.27, 3.04) (  2.44, 4.45) ( 1.45, 4.80) (  2.19, 3.86)
ANX S 1.22 0.79 0.75 0.85 0.81 0.71 0.60 1.30 0.94 0.93 0.84 0.79 0.71 0.68 0.92
L 1.15 0.90 0.92 0.80 0.72 0.65 0.50 1.31 1.01 0.94 0.76 0.72 0.51 0.53
S–L  0.14  0.20 0.02 0.04 0.03 0.06  0.09  0.04 0.07 0.04 0.16 0.14
( 0.36, 0.09) (  0.43, 0.03) (  0.23, 0.28) (  0.22, 0.30) ( 0.19, 0.25) (  0.16, 0.29) (  0.30, 0.12) (  0.25, 0.18) (  0.18, 0.31) (  0.20, 0.29) ( 0.05, 0.37) (  0.08, 0.36)

P. Knekt et al. / Psychiatry Research 220 (2014) 1019–1027

HDRS S 14.6 10.4 11.0 9.72 8.84 15.8 11.3 10.1 10.7 10.1 0.76
L 15.2 11.9 12.6 9.93 8.28 12.3 12.0 9.27 7.24
S–L  0.88  0.95 0.51 1.34  1.48  2.38n 0.71 2.09
( 2.75, 0.99) (  3.05, 1.16) (  1.98, 3.01) (  0.95, 3.63) (  3.28, 0.33) (  4.38,  0.39) (  1.68, 3.10) (  0.14, 4.33)
HARS S 14.6 10.0 9.92 9.20 8.50 15.2 10.6 10.3 9.85 9.58 0.92
L 15.7 11.9 11.4 8.99 7.72 15.2 11.2 11.5 8.79 6.81
S–L  0.96  0.68 1.09 1.69  0.89  1.43 0.58 2.38n
( 2.74, 0.83) (  2.67, 1.31) (  1.19, 3.38) (  0.62, 3.99) (  2.60, 0.82) (  3.31, 0.44) (  1.60, 2.75) (0.13, 4.63)
DEP S 1.92 1.31 1.27 1.31 1.29 1.10 1.04 2.01 1.36 1.21 1.22 1.14 1.12 1.02 0.67
L 2.07 1.56 1.59 1.36 1.30 1.09 0.91 2.08 1.63 1.56 1.16 1.26 0.82 0.93
S–L  0.24  0.31  0.03 0.02 0.03 0.15  0.03n  0.38n 0.02  0.17 0.25 0.06
( 0.53, 0.05) (  0.61, 0.00) (  0.38, 0.33) (  0.34, 0.37) ( 0.30, 0.36) (  0.17, 0.45) (  0.58,  0.03) (  0.67,  0.10) (  0.31, 0.35) (  0.51, 0.16) ( 0.06, 0.56) (  0.24, 0.36)
GAF S 56.6 66.0 66.3 69.1 72.6 54.8 64.7 65.1 66.5 69.0 0.42
L 55.0 62.3 61.3 65.1 69.7 55.5 65.0 62.6 69.5 75.7
S–L 2.06 3.19 2.22 1.22 1.33 4.04n  1.22  5.30n
( 1.60, 5.72) (  0.93, 7.32) (  2.61, 7.05) (  3.77, 6.20) (  2.19, 4.85) (0.14, 7.95) (  5.86, 3.42) (  10.2,  0.44)

a
The full ITT model.
b
The full ITT model further including the baseline level of the outcome measure considered.
c
P-value for interaction between the verbal IQ measure and the therapy group throughout the follow-up.
n
P-value for difference between S and L o 0.05.
Table 5
Mean valuesa of psychiatric symptoms and psychosocial functioning in short-term (S) and long-term therapy (L) and mean value differencesb (95% confidence intervals) at 7, 12, 24, 36, 48 and 60 month follow-up according to the
lower and higher values of patients' performance IQ.

Measure Therapy Low performance IQ (71–110) High performance IQ (111–137) P-valuec

0 7 12 24 36 48 60 0 7 12 24 36 48 60

GSI S 1.27 1.07 0.87 0.92 0.85 0.74 0.72 1.23 0.86 0.79 0.78 0.78 0.71 0.63 0.32
L 1.29 0.99 0.93 0.79 0.85 0.70 0.59 1.32 1.04 1.02 0.84 0.72 0.59 0.58
S–L  0.09  0.06 0.13  0.01 0.03 0.13  0.17n  0.22n  0.09 0.06 0.12 0.06
(  0.26, 0.08) (  0.23, 0.16) (  0.08, 0.34) ( 0.22, 0.21) (  0.16, 0.29) (  0.07, 0.32) ( 0.34,  0.01) (  0.39,  0.06) (  0.29, 0.10) (  0.14, 0.25) ( 0.05, 0.30) (  0.12, 0.24)
BDI S 16.8 10.4 9.98 9.87 9.81 8.07 9.45 17.5 9.09 8.60 8.03 8.55 7.72 7.61 0.12
L 19.1 11.8 11.7 8.41 10.5 6.61 6.01 18.5 13.7 12.5 9.70 7.15 5.35 5.73
S–L  0.32  0.62 2.38 0.26 2.15 4.03n  5.29n  4.58n  2.25 0.89 1.88 1.26
(  3.16, 2.51) (  3.59, 2.34) (  0.99, 5.75) ( 3.42, 3.95) (  1.23, 5.54) (0.77, 7.29) ( 7.95,  2.62) (  7.39,  1.78) (  5.33, 0.83) (  2.48, 4.25) ( 1.24, 5.00) (  1.75, 4.26)
ANX S 1.29 0.86 0.91 0.95 0.86 0.75 0.73 1.23 0.86 0.79 0.75 0.73 0.65 0.57 0.60
L 1.22 0.94 0.87 0.74 0.79 0.61 0.54 1.25 0.97 0.98 0.80 0.64 0.53 0.47
S–L  0.11  0.00 0.17 0.01 (  0.25, 0.27) 0.08 0.14  0.13  0.22n  0.08 0.06 0.09 0.08
(  0.33, 0.11) (  0.23, 0.22) (  0.09, 0.43) (  0.14, 0.30) (  0.10, 0.37) ( 0.34, 0.08) (  0.43,  0.00) (  0.32, 0.16) (  0.18, 0.30) ( 0.11, 0.29) (  0.14, 0.29)

P. Knekt et al. / Psychiatry Research 220 (2014) 1019–1027

HDRS S 14.5 10.5 10.3 10.4 9.51 15.8 11.2 10.8 10.2 9.42 0.34
L 14.8 9.14 11.3 9.86 7.32 16.5 13.6 13.4 9.48 8.11
S–L  0.88  0.95 0.51 1.34  1.48  2.38n 0.71 2.09
(  2.75, 0.99) (  3.05, 1.16) ( 1.98, 3.01) (  0.95, 3.63) ( 3.28, 0.33) (  4.38,  0.39) (  1.68, 3.10) (  0.14, 4.33)
HARS S 14.5 9.68 9.24 9.17 8.87 15.2 10.9 10.9 9.90 9.21 0.78
L 15.1 10.5 10.9 9.16 7.21 15.8 12.6 12.1 8.77 7.26
S–L  0.34  1.21 0.38 1.97  1.49  0.99 1.19 2.21n
(  2.11, 1.43) (  3.17, 0.76) ( 1.93, 2.68) (  0.37, 4.31) ( 3.15, 0.17) (  2.84, 0.86) (  0.98, 3.36) (0.07, 4.34)
DEP S 1.95 1.37 1.33 1.32 1.31 1.12 1.13 1.98 1.30 1.15 1.19 1.11 1.06 0.94 0.16
L 2.04 1.58 1.49 1.22 1.36 1.06 0.94 2.11 1.61 1.66 1.28 1.19 0.84 0.89
S–L  0.20  0.15 0.12  0.05 0.05 0.20  0.35n  0.54n  0.12  0.12 0.19 0.02
(  0.49, 0.09) (  0.45, 0.14) (  0.24, 0.45) ( 0.41, 0.31) (  0.28, 0.39) (  0.13, 0.52) ( 0.62,  0.07) (  0.82,  0.26) (  0.45, 0.20) (  0.45, 0.21) ( 0.12, 0.50) (  0.28, 0.32)
GAF S 66.6 67.7 67.9 71.0 64.1 63.7 67.4 70.5 0.97
L 64.8 63.6 66.6 73.1 62.4 60.0 68.2 72.6
S–L 1.19 3.40 0.62  2.79 2.21 4.23n  0.37  1.82
(  2.45, 4.82) (  0,63, 7.43) ( 4.25, 5.49) (  7.93, 2.36) ( 1.19, 5.62) (0.43, 8.02) (  4.96, 4.22) (-6.52, 2.87)

a
The full ITT model.
b
The full ITT model further including the baseline level of the outcome measure considered.
c
P-value for interaction between the performance IQ measure and the therapy group throughout the follow-up.
n
P-value for difference between S and L o 0.05.

1025
1026 P. Knekt et al. / Psychiatry Research 220 (2014) 1019–1027
The faster early symptom reduction in short-term therapy
did not depend on the intelligence level, implying that factors
other than intelligence may be more important to successful
short-term psychotherapy. This effect did not emerge, however,
with regard to anxiety symptoms, which is in accordance with
the general results on effectiveness of the HPS, which showed
that short-term psychotherapy was equally or more beneficial
than long-term psychotherapy during the first year of follow-up,
especially in the reduction of depressive symptoms (Knekt et al.,
2008).
Previous research on the prediction of intelligence on therapy
outcome is limited. Apart from a dozen early studies of relatively
poor methodology, reviewed by Luborsky et al. (1996), which
indicated intellectual functioning to be a moderate predictor of
treatment success in diverse, mostly short-term therapies, there
has been very limited interest during the last decades. A few
recent studies on mostly short-term cognitive therapies have
given controversial results on the prediction of intelligence on
treatment outcome (Haaga et al., 1991; Doubleday et al., 2002;
Fournier et al., 2009; D’Alcante et al., 2012). No previous studies
are available on psychodynamic therapies or on comparisons
between short- and long-term psychotherapy. The importance of
assessing patients' pre-treatment level of intelligence has been
highlighted by a study indicating that the ability to benefit from
mental health services is better among patients with higher IQ
(4 110) (Slade et al., 2006), a level similar to what appeared to
predict a different outcome in long-term versus short-term psy-
chotherapy in this study.
The results on the prediction of IQ were largely similar when
the predictor variable was the full scale, the verbal, or the
performance quotient. However, for patients with higher pre-
treatment performance IQ, the benefits of short-term therapy
were more evident during the first year of follow-up than among
those with higher full scale IQ and verbal IQ, perhaps indicating
their greater behavioral propensities, beyond verbal capacity, to
adopt the brief therapy orientation.
The present prospective study has several strengths: Firstly, we
used a validated intelligence assessment method and several
validated outcome measures, which were based on both inter-
views and questionnaires. Secondly, the relatively large sample
size, the long follow-up period, and the frequent assessments of
the outcome measures during follow-up gave a reliable basis for
the analyses (Knekt et al., 2008). Finally, the small dropout, the
wide selection of potential confounding factors considered, and
the inclusion of auxiliary treatment in the models minimized the
risk of biased results.
There are, however, also issues that may complicate the
interpretation of the results. Firstly, despite the relatively large
sample the two short-term therapies and the long-term therapy
and psychoanalysis groups were combined to achieve more
strength in the statistical analyses. For this reason, possible
differences within these groups have been neglected in this study
and the emphasis was placed on the length of therapy. Secondly,
residual confounding cannot be fully excluded despite the com-
prehensive adjustment for potential confounding factors. Thirdly,
due to the multiple comparisons performed, the possibility of a
chance finding cannot be excluded. Fourthly, despite the similarity
of the results of the ITT and the AT analyses, lack of compliance
with the study treatment may potentially cause bias (Knekt et al.,
2011). Fifthly, although no notable differences were found
between the results derived from the original and imputed data,
it cannot be excluded that still other than ignorable dropouts may
bias the results (Harkanen et al., 2005). Finally, although the WAIS-
R is considered a reliable measure giving an adequate estimate of
the patients' intelligence for the purposes of this study, it does not
give a complete picture of intelligence.
In conclusion, individuals suffering from anxiety or depressive
disorder with a higher intelligence benefitted more from long-
term than from short-term psychotherapy. This finding suggests
that the intelligence quotient could be a useful tool in identifying
patients who would benefit more from long-term than short-term
psychotherapy. Using this and other suitability criteria would help
create clinically meaningful recommendations for treatment pro-
viders. This finding should, however, be confirmed in large-scale
studies before firm conclusions can be made.
Acknowledgment
The study was financially supported by the Academy of Finland
(Grant no. 138876).
References
Allen, J.G., Coyne, L., David, E., 1986. Relation of intelligence to ego functioning in an
adult psychiatric population. Journal of Personality Assessment 50, 212–221.
American Psychiatric Association (APA), 1985. The psychodynamic psychotherapies.
Criteria for their prescription and concurrent review, APA, Manual of Psychiatric
Peer Review, 3rd ed. American Psychiatric Association, Washington, DC, pp.
93–96.
American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental
Disorders. American Psychiatric Association, Washington, DC.
Andrews, G., Pollock, C., Stewart, G., 1989. The determination of defense style by
questionnaire. Archives of General Psychiatry 46, 455–460.
Antonovsky, A., 1993. The structure and properties of the sense of coherence scale.
Social Science and Medicine 36, 725–733.
Azim, H.F., Piper, W.E., Segal, P.M., Nixon, G.W., Duncan, S.C., 1991. The quality of
objects relations scale. Bulletin of the Meninger Clinic 55, 323–343.
Bacharach, H.M., Leaff, L.A., 1978. “Analyzability”: a systematic review of the clinical
and quantitative literature. Journal of the American Psychoanalytic Association
26, 681–920.
Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., Erbaugh, J., 1961. An inventory for
measuring depression. Archives of General Psychiatry 4, 561–571.
Benjamin, L.S., 1996. A clinician-friendly version of the interpersonal circumplex:
structural analysis of social behavior (SASB). Journal of Personality Assessment
66, 248–266.
D’Alcante, C.C., Diniz, J.B., Fossaluza, V., Batistuzzo, M.C., Lopes, A.C., Shavitt, R.G.,
et al., 2012. Neuropsychological predictors of response to randomized treat-
ment in obsessive-compulsive disorder. Progress in Neuropsychopharmacology
& Biological Psychiatry 39, 310–317.
Davanloo, H., 1978. Evaluation and criteria for selection of patients for short-term
dynamic psychotherapy. Psychotherapy and Psychosomatics 29, 307–308.
De Shazer, S., Berg, I.K., Lipchik, E., Nunnally, E., Molnar, A., Gingerich, W., et al.,
1986. Brief therapy: focused solution development. Family Process 24, 207–221.
Derogatis, L.R., Lipman, R.S., Covi, L., 1973. SCL-90: an outpatient psychiatric rating
scale – preliminary report. Psychopharmacology Bulletin 9, 13–28.
Doubleday, E.K., King, P., Papageorgiou, C., 2002. Relationship between fluid
intelligence and ability to benefit from cognitive-behavioural therapy in older
adults: a preliminary investigation. British Journal of Clinical Psychology 41,
423–428.
Fonagy, P., 2010. Psychotherapy research: do we know what works for whom?
British Journal of Psychiatry 197, 83–85.
Fournier, J.C., DeRubeis, R.J., Shelton, R.C., Hollon, S.D., Amsterdam, J.D., Gallop, R.,
2009. Prediction of response to medication and cognitive therapy in the
treatment of moderate to severe depression. Journal of Consulting and Clinical
Psychology 77, 775–787.
Gabbard, G.O., 2004. Long-Term Psychodynamic Psychotherapy: A Basic Text.
American Psychiatric Publishing, Inc., Washington, DC.
Greenson, R.R., 1985. The Technique and Practice of psychoanalysis. The Hogarth
Press and The Institute of Psycho-Analysis, London.
Gunderson, J.G., Gabbard, G.O., 1999. Making the case for psychoanalytic therapies
in the current psychiatric environment. Journal of the American Psychoanalytic
Association 47, 679–704.
Haaga, D.A., DeRubeis, R.J., Stewart, B.L., Beck, A.T., 1991. Relationship of intelligence
with cognitive therapy outcome. Behaviour Research Therapy 29, 277–281.
Hamilton, M., 1959. The assessment of anxiety states by rating. British Journal of
Medical Psychology 32, 50–55.
Hamilton, M., 1960. A rating scale for depression. Journal of Neurology, Neurosur-
gery & Psychiatry 23, 56–61.
Harkanen, T., Knekt, P., Virtala, E., Lindfors, O., 2005. A case study in comparing
therapies involving informative drop-out, non-ignorable non-compliance and
repeated measurements. Statistics in Medicine 24, 3773–3787.
Horowitz, L.M., Alden, L.E., Wiggins, J.S., Pincus, A.L., 2000. Inventory of the
Interpersonal Problems Manual. The Psychological Corporation, San Antonio.
 P. Knekt et al. / Psychiatry Research 220 (2014) 1019–1027
Johnson, L.D., Miller, S.D., 1994. Modification of depression risk factors: a solution-
focused approach. Psychotherapy: Theory, Research, Practice, Training 31,
244–253.
Kernberg, O.F., 1996. A psychoanalytic theory of personality disorders. In: Clarkin,
aJ.F., Lenzenweger, M.F. (Eds.), Major Theories of Personality Disorders. The
Guilford Press, New York, pp. 106–140.
Knekt, P., Lindfors, O., 2004. A randomized trial of the effect of four forms of
psychotherapy on depressive and anxiety disorders. Design, methods, and
results on the effectiveness of short-term psychodynamic psychodynamic
psychotherapy and solution-focused therapy during a one-year follow-up.
Available from 〈http://www.thl.fi/hps〉 (accessed 11.08.11.).
Knekt, P., Lindfors, O., Harkanen, T., Valikoski, M., Virtala, E., Laaksonen, M.A., et al.,
2008. Randomized trial on the effectiveness of long-and short-term psycho-
dynamic psychotherapy and solution-focused therapy on psychiatric symptoms
during a 3-year follow-up. Psychological Medicine 38, 689–703.
Knekt, P., Lindfors, O., Renlund, C., Sares-Jaske, L., Laaksonen, M.A., Virtala, E., 2011.
Use of auxiliary psychiatric treatment during a 5-year follow-up among
patients receiving short- or long-term psychotherapy. Journal of Affective
Disorders 135, 221–230.
Knekt, P., Lindfors, O., Sares-Jäske, L., Virtala, E, Härkänen, T., 2013. Randomized trial
on the effectiveness of long- and short-term psychotherapy on psychiatrics
symptoms and working ability during a 5-year follow-up. Nordic Journal of
Psychiatry 67, 59–68.
Laaksonen, M.A., Knekt, P., Lindfors, O., 2013a. Psychological predictors of the recovery
from mood or anxiety disorder in short-term and long-term psychotherapy
during a 3-year follow-up. Psychiatry Research 208, 162–173.
Laaksonen, M.A., Knek,t, P., Sares-Jäske, L., Lindfors, O., 2013b. Psychological
predictors on the outcome of short-term psychodynamic psychotherapy and
solution-focused therapy in the treatment of mood and anxiety disorder.
European Psychiatry 28, 117–124.
Laaksonen, M.A., Lindfors, O., Knekt, P., Aalberg, V., 2012. Suitability for Psychother-
apy Scale (SPS) and its reliability, validity, and prediction. British Journal of
Clinical Psychology 51, 351–375.
Lee, J., 1981. Covariance adjustment of rates based on the multiple logistic
regression model. Journal of Chronic Diseases 34, 415–426.
Lindfors, O., Knekt, P., Virtala, E., Laaksonen, M.A., The Helsinki Psychotherapy Study
Group, 2012. The effectiveness of solution-focused therapy and short- and long-
term psychodynamic psychotherapy on self-concept during a 3-year follow-up.
The Journal of Nervous and Mental Disease 200, 946–953.
1027
Luborsky, L., Crits-Christoph, P., Mintz, J., Auerbach, A., 1996. Who will benefit from
psychotherapy? Predicting therapeutic outcomes. Basic Books, New York.
Malan, D.H., 1976. The Frontier of Brief Psychotherapy: An Example of the
Convergence of Research and Clinical Practice. Plenum Medical Book, New York.
Migon, H.S., Gamerman, D., 1999. Statistical Inference: An Integrated Approach.
Arnold, London.
Rizvi, S.L., Vogt, D.S., Resick, P.A., 2009. Cognitive and affective predictors of
treatment outcome in cognitive processing therapy and prolonged exposure
for posttraumatic stress disorder. Behaviour Research and Therapy 47, 737–743.
Rothman, K.J., Greenland, S., 1998. Modern Epidemiology, second ed. Lippincott-
Raven, Philadelphia.
Rubin, D.B., 1987. Multiple Imputation for Nonresponse in Surveys. John Wiley,
New York.
SAS Institute Inc., 2008. SAS Online Doc. SAS/STAT User's Guide, Version 9.2. SAS
Institute Inc., Cary, NC.
Sifneos, P.E., 1978. Short-term anxiety provoking psychotherapy. In: Davanloo, H. (Ed.),
Short-Term Dynamic Psychotherapy. Spectrum, New York, NY, pp. 35–42.
Slade, M., Leese, M., Gillard, M., Kuipers, E., Thornicroft, G., 2006. Pre-morbid IQ and
response to routine outcome assessment. Psychological Medicine 36, 1183–1191.
Valbak, K., 2004. Suitability for psychoanalytic psychotherapy: a review. Acta
Psychiatrica Scandinavica 109, 164–178.
Valkonen, H., Lindfors, O., Laaksonen, M.A., Knekt, P., 2012. Association between the
Rorschach Ego Impairment Index (EII-2) and the Level of Personality Organiza-
tion. Psychiatry Research 200, 849–856.
Watzke, B., Ruddel, H., Jurgensen, R., Koch, U., Kriston, L., Grothgar, B., et al., 2010.
Effectiveness of systematic treatment selection for psychodynamic and
cognitive-behavioural therapy: randomised controlled trial in routine mental
healthcare. British Journal of Psychiatry 197, 96–105.
Wechsler, D., 1981. Manual for the Wechsler Adult Intelligence Scale – Revised.
Psychological Corporation, New York.
Weissman, M.M., Bothwell, S., 1976. Assessment of social adjustment by patient
self-report. Archives of General Psychiatry 33, 1111–1115.
Verbeke, G., Molenberghs, G., 1997. Linear mixed models in practice: an SAS-
oriented approach. Springer, New York.
Wong, P.S., Haywood, D.M., 2012. Foundations of psychodynamic therapy: implicit
emotional learning. In: Levy, R.A., Ablon, J.S., Kächele, H. (Eds.), Psychodynamic
Psychotherapy Research: Evidence-based Practice and Practice-based Evidence.
Humana Press, New York, pp. 269–282.