Brit. J. PsychicS.

(i@7ç),125, 355—73

A Hundred Cases of Suicide: Clinical Aspects

By B. BARRACLOUGH, J. BUNCH, B. NELSON and P. SAINSBURY

‘¿It
was the author's object in his paper to establish a fact he believes, of primary importance,—thatthe disposition
to commit self-destruction is, to a great extent, amenable to those principles which regulate our treatment of ordinary
diseases; and that,to a degreemore thanisgenerally supposed, itoriginates in derangementof the brainand
abdominal viscera.'
Forbes Winslow. TheAnatontyof Suicide,1840.
Historically, doctors have not always acknow second. The sampling method excludes residents
ledged that they have an obligationto prevent who diedoutsidetheircoroner's
area,and transients
suicide, partly because they shared the pre who came intothecoroner'sdistrict
and diedthere.
valent idea that most suicides were caused by Interviewprocedure
moral crises, no concern of theirs—and indeed
After each inquest the witnesses who had been
suicide was a criminal matter until quite identifiedas thoselikelyto know most about the
recently; but more, perhaps, because a fatalism suicide were visited without previous warning. We
has characterized their attitudes to its preven found that this procedure made acceptance likely
tion, even where the suicide was clearly suffer and caused least distress. The three interviewers, a
ing from mental illness. Yet two recent American psychiatrist
and two psychiatric
social
workers,
visited
studies have shown more than go per cent of the households of 43,30 and 27 suicides respectively,
suicides to be mentally ill before their death and interviewed the main informant; they used a
(i 7, 8); this finding and the familiar clinical
questionary of 252 categorized and coded items of
observation that suicidal thoughts disappear which z@odealtwith the suicide's previous medical and
psychiatric history, his physical and mental state at
when theillness issuccessfully
treatedprovidea
the time of death, and any treatment he was then
strongcasefora medicalpolicyofprevention. receiving.The itemsreferred toobserved behaviour,
We describehere the psychiatric historiesof objective facts and documented data. Each item of
100 suicidesobtained soon afterthe event from behaviour theinformant was askedabouthad been
their relatives, their doctors, other witnesses and previously defined, and covered those usually in
medical records. Our aims were to see what cluded in a routine examination of mental status.
proportionof suicides had been mentallyill; The items were categorized as present or absent, or
to determinethe diagnoses;to discoverhow rated as severe, mild or not present.
many gave warnings of their intentions, and The interviewer was allowed considerable latitude
find out theirmedical contactsand current as regardsthe questions put to the informant in
order to elicit the information needed to code replies
treatment.Our other objectives,
not dealtwith
to the items, because rapport, and hence validity,
here,were first
toseeifsocial
factors
relatedto ismore readilyobtainedifthe interviewisnot too
theact,and secondto seewhat stressful
events rigidly structured, especially when, as in this enquiry,
precededthesuicide. theinformants wereupsetbecauseofthebereavement.
The intervieweralso recorded the supporting
METHOD evidence for positivelyscored items; and each
Sampling completedcaserecordwas revieweditem by item
One hundred suicides, defined by a coroner at at a joint conferenceof the three interviewers.
inquests, and residents of the County of West Sussex Differencesin interpretation
were then settledin
or of the County Borough of Portsmouthwere terms of the criteria
by which itemswere defined.
studied. They included twosamples: 25of29suicidesDocuments were frequently used to corroborate oral
who died in West Sussex during i966 and 1967 accounts of symptoms, present, past and family
comprised the first sample, and 75 consecutive history and treatment. Psychiatric hospital notes
suicides occurring in both areas during i968, the were seen in 59 cases, general hospital notes in 34,
355
 356 A HUNDRED CASES OF SUICIDE: CLINICAL ASPECTS
MinistryofDefencemedicalrecordsin 7,Executive three psychiatrists agreed; in i8 cases more than one
in all psychiatric diagnosis was made. The percentages of
Councilmedicalcardsand postmortem reports
cases but one; and there were 39 suicide notes. complete agreements were: mentally ill, 95 per cent
The principal informant's relationship to the (N = 93), depression 86 per cent (8o), alcoholism
@ suicide also has a bearing on validity; lived with 87 per cent (is), abnormal personality 75 per cent
thesuicide, i6 visited him daily, 24 visitedtwice (27). We concluded that the procedure provided an
weeklyand 6 visited weeklyorless often.Firstdegree acceptable degree of inter-psychiatrist agreement.
relatives, that is spouse, parent, sib or child were To control the social and some of the clinical
the main informants of 83 suicides, other relatives or data,a populationsamplenot attendingforpsychi
lifelong friends of 13 suicides, and lesser degrees of atric treatment was obtained from the registers of
relationship of 4. Because the main informants could two large general practices, one in Portsmouth and
notgiveallthedetails, otherwitnesses, usually near theotherin West Sussex. The control groupwas
relatives or close friends, were also seen, an average matched with the suicide group for age, sex and
@ of per case, and the evidence of expert witnesses whether ever married, by randomly obtaining quotas
such as pathologists, pharmacists and social workers forthesecategories. One hundredand fifty people
was used where relevant. The general practitioner were interviewed with a questionary which covered
@ was interviewed by visit or by telephone in cases, the same ground as that for the suicides, except the
and medicalornursing staffofpsychiatric hospitalssymptoms section; the details of how this was done to
were interviewed in i i. Exhaustive enquiries were ensure comparability with the suicides are described
therefore made from all relevant quarters to collect elsewhere (f).
valid information.
RESULTS
Reliability The sex, age and marital status distribution
To ensureuniformity thethreeinter and diagnostic classification of the ioo suicides
ofcoding,
viewersjointlyreviewedeach case-record item by will be described first, then the main clinical
item when it was completed. In i i cases the infor findings for the diagnostic groups.
mants were interviewed by two interviewers; one of
the pair interrogated, while both wrote down replies Sex, ageand marital status
and independently coded all the items. Coding The expected values of these characteristics
differences provided a measure of reliability. No for a random sample of i oo people were calcu
disagreements were recordedon 88 per centof the lated by proportion from their distribution
clinical items, and one disagreement on 10 per cent. in the general population of the two counties
Items were also inspected for interviewer bias by
examiningthefrequency ofpositive
responsesobtained
(i6), controlling for sex and age; the expected
by each interviewer foreach item in allthe cases. values were then compared with observed
(Significant differences between them were obtained (Table I). The over-represented groups in the
on less than 5 per cent of the items, that is to say, only suicide sample are men; the older age groups,
tochance.)We therefore especially
asoftenascan be attributed men between 45 and 64, and women
used over 65. Other particularly
had no reasontosupposethattheinterviewers vulnerable male
differing standards. groups are the single, widowed and divorced
and those over 25 years; single men under 25
Diagnosis and married men 25—44are least at risk. For
A panel of three psychiatrists independently re women, beingsingle and over45,and widowed
viewedalltheevidence available on eachsuicide and at any age carry a high risk, while young single
made a diagnosis. A description of the criteria for and young married women are less at risk. The
diagnosingdepressive
illnesswas provided,namelythe sample resembles the nationalsuicidestatistics
presenceofmedicalsymptoms,insomnia, weightloss, of England in these respects and may therefore
anorexia, behavioural signs such as depressed appear
be considered representative of suicides in
ance,reducedactivity andpsychological symptomsof
complaints of depression, listlessness, lack of interest general.
and statements ofguilt, hypochondriasis and various
delusions. It was essential that an unequivocal change
Diagnoses
from the patient's usual state had been reported. Ninety-three of the ‘¿0°suicides were diag
A diagnosis was given to the suicide if two of the nosed mentally ill. The distribution of the
 BY B. BARRACLOUGH, J. BUNCH, B. NELSON AND P. SAINSBURY 357

TABLE I
I)istribution of sex, age and marital status

Depr Miscell mentally ill All Expected

%15—24 %Alcoh
N(64)ession N(15)olism
% N(14)aneous %Not N(7) % N(ioo) % N(ioo)

.. 0 0 0 4 8 I0'2 19
25—44.. .. 5 17 3 25 4 57 125 13 25 15.7 30
45—64.. .. i6 53 9 75 0 0 125 2649 I7@2 32
65+
19Total .... ..4 513 170 0Men 0 30 430 250 10 59 I0@0

....3010012100 71004 100 53 101 53@I0 100

315—24 x' = 8@7;
p < 0.05;d.f. =

.. 0 0 33 I 2 6@8 14
25—44 .. .. 6 i8 I 33 4 57 0 0 II 23 I2@4 26
45—64.. .. 13 38 I 33 I 14 0 0 1532 I5@O32
65+ ....
I2@927Total ..0 150 440 33
IWomen 20 29I 267 2043
....34100399 71003 100 47 100 47.@ 99
315—24 x2 = 9@o'p < 0@o5;d.f.
=
sexes
.. 0 0 14 5 5 17 17
25—44.. .. II 17 4 27 8 57 I 14 2424 2828
45—64 .. .. 29 45 10 67 I 7 I 14 41 41 32 32
65+ ....
2323Total ..4 206 3!0 IBoth 7 50 36I 457 3030
.... 141007 99 100 100 100 100
d.f.Single .649915101 x2= @3.7;
p < o@oi;3

.. ‘¿7 4 25 i8 34@ I2@6 24

Married .. 17 57 5 42 2 29 375 275137.571
Widowed .. 2 7 3 25 I 14 0 0 6 is 2 @83 5
Divorced.. ..II
O@3Total 037 02 2Men 17 057 0I 0 0 2 4 o@I7t
....3010112101 71004 100 53 100 53@I 100
x2= 13@6;
p < o@oi;2 d.f.
(Poisson)Single.. t p = o@oi

0 I 33 13 28 I0@5 22
Married .. II 32 I 33 6 86 I 33 19 40 28@1 59
Widowed .. II 32 I 33 0 0 I 33 13 28 7@83 17
Divorced..
ITotal ..ii I32 30 IWomen
33 014 0I 0 0 2 4 o @67t
....3499399 71003 99 47 100 47., 99 —¿
x2 = 8@5;p < 0@05;2 d.f.
@ (Poisson)Single.. t p<
sexes
13 5 3131 23@I23
Married .. 28 44 6 40 8 57 457 4646 65@666
Widowed .. 13 20 4 27 I 7 I ‘¿4 19 19 io@66 ii
Divorced..
ITotal ..22 I34 22 3Both20 036 0230 0 0 4 4 o 84t
....6410015100 141007 101 100 100 I00@2O 101
x2= 2o@I;P < oooi; 2 d.f.
t p < Ooi (Poisson)
358
A HUNDRED CASES OF SUICIDE: CLINICAL ASPECTS

Depr mentally ill

%SexMale N(64)ession %Alcoh N(I5)olism

%Miscell N(I4)aneous
%Not N(7) %AllN(ioo) %Expec
N(rno)ted

Female.. .... 3447 5312

56Total....6450015100541007
..30 38o 207 750 50457 34353 4753 4743.9 56@I44
500100100100100

p = 0@O4(Binomial) one-tailed

principal diagnoses was, depressive illness 70 per alcoholism account for go per cent of the
cent, alcoholism 55 per cent, schizophrenia 3 per suicides' diagnoses. In the remaining 7 cases the
cent, phobic anxiety state 3 per cent, barbiturate clinical information was insuflicient to make a
dependence and schizo-affective psychosis each diagnosis; but in only one case (M.14) was there
i per cent (Table II). Thus depression and no evidence of illness.
Suicide, it may be inferred, is a rare event for
TABLEII
those with good mental and physical health;
DiagnosesPrincipal
rather is it associated with depressive illness and
Other alcoholism.
diagnosisNDepression
diagnosis
None ..
64ness.. .. .. DIAGNOSTIC GROUPS
Malignant terminal ill
..2Non-malignant .. For convenience the sample has been divided
terrni into four clinical groups : 6z@uncomplicated
..2Barbiturate
nalillness .. depressive; 55 alcoholics; the remainder of the
dependenceIDementia mentally ill, diagnostically heterogeneous, num
and barbi
dependence..I70Alcoholism
turate bering ‘¿4,
which includes the 6 depressives with
other serious illnesses ; and the 7 not diagnosed
None
..4Depression
.. .. mentally ill.
9malignant .. .. The purpose of the description which follows
Depression and non
terminalillness is to show that the clinical evidence was sufficient
..,Non-malignant.. to make a valid diagnosis ; to see whether the
termi suicides' diagnoses are clinically typical or have
..‘5Schizophrenia naliliness.. distinctive features; and to obtain clinical guides
to preventive measures. Most emphasis is given
..33Phobic .. ..
to the suicides diagnosed depression and alco
anxiety None .. .. .. holism, as they are numerically so important.
dependencei3Barbiturate
state Barbiturate
DEPRESSIVES
..IAcute dependence .. ..
A depressive illness uncomplicated by other
..iTotal(schizo-affective) psychosis serious physical or mental disorder was diag
nosed in 64 of the suicides; 30 (@‘@
per cent) men
and 34 (53 per cent) women. Their mean age
..93Total
mentally ill .. .. was 54 years (S.D. x6), for men it was 50 (S.D.
not mentally ill .. .. ..7100 i6) and for women 58 (S.D. 15). Three quarters
were over 45. Their marital status, when com
pared with the distribution in the population at
 BY B. BARRACLOUGH, J. BUNCH, B. NELSON AND P. SAINSBURY 359
risk, showed more single and widowed, fewer Documentary data alone underestimate mor
married and about as many divorced as ex bidity, sojudgements ofpast ill-health were also
pected (Table I). A feature not brought out by made on informants' oral evidence. These
comparisons with the census data is the very indicated that another 12 suicides had probably
high proportion of married suicides who were had previous episodes of depressive illness which
either separated or living unhappily with their had not received psychiatric treatment, making
spouses. a total of 40 (63 per cent) with previous histories
of depression.
Another measure of previous illness is the
Symptoms number of ‘¿spells'of psychiatric treatment. A
The frequency distribution of the suicides' ‘¿spell'
isaperiod ofdocumented psychiatric care
symptoms of depression did not seem atypical ended by dischargeor by lapseofattendancefor
(Table III). To test this clinical impression the more than three months. The range was from
incidence of the 55 leading symptoms was one to twenty spells, the mean 4.4 and the
compared with that found in i@8 consecutive median 3'O. Thus there was a group of suicides
cases of ‘¿endogenous' depression, matched for who were depressedat the time of theirdeath
age and sex, referred to the Chichester District and distinguished by recurrent episodes of the
Psychiatric Service. Their leading symptoms same illness.We were unable to evaluate this
had been rated for severity by a consultant findingby comparison with depressedpatients.
psychiatrist (@, I 2). The rank order of the fre A history of attempted suicide was recorded
quency of the ‘¿5
symptoms was similar, but the only if the hospital notes had been seen, or a
suicides scored consistently higher ratings for witness to the attempt interviewed. Twenty one
all symptoms, especially insomnia. Sympto (33 per cent) of the depressed suicides had
matically, therefore, the depressed suicidesattempted suicide before; more than half during
resembled an unselectedsample of depressives, the year previous to their death, and in the
but they appeared to be more severely course of their final illness. Eight of the 21 had
depressedwhen assessedon symptoms. Differing made two or more attempts at suicide. Another 8
interview procedures may account for the suicides had told an interviewed witness of an
finding. unobserved suicide attempt. If these are also
included 29 (46 per cent) had attempted suicide
before finally succeeding.
Past and family /iistoiy
Only 6 per cent of the unselected Chichester
A history of psychiatric treatment, verified by depressives on the other hand had histories of
consulting the hospital notes, was the index of previous suicide attempts. Thus an important
previousmorbidityused (TableIV). distinction between the two groups is the
Twenty-eight (44 per cent) had a historyof previous tendency of those who kill themselves
psychiatric
treatment,and another5 (8per cent) to think of suicide more and to act on the
a record of a psychiatric consultation only. thought.
Thus halfthe depressives had had a previous
episode serious enough to warrant psychiatric
intervention.Of the 28 previously treated Family histories
suicides 26 had been diagnosed depression, 7 Histories of mental illness could be syste
mania. Reliable oral evidence indicatedthat matically collected only for first degree relatives
a further 4 had had attacks of mania, making ii —¿parents, siblings and children. A history was
@ (i per cent) cases of manic-depressive illness, a scored positively if the hospital notes were seen,
similar proportion (54 per cent) to that found in or an interviewed relative confirmed having
the unselected, disorderedpatients had
affectively a history of treatment; a psychiatric
referred to a psychiatric service in St. Louis consultationonly was not counted. Reports of
(23). So the circular form of the illness is not relatives committing suicide were verified by
apparently over-represented in suicides. their death certificates (Table IV).
 TABLEIII
Symptoms recordedas present duringfour weeks previous to death, by diagnosis
Not
Depression Alcoholism Miscellaneous mentally ill All
Symptoms (N = 64) (N = ,@) (N = 14) (N = 7) (N =
0/ 0/ 0/ 0/ 0
/0 /0 /0 /0 /0

Looked miserable .. .. .. 89 40 43 0 69
Insomnia .. .. .. .. 86 8o 50 29 76
Taking hypnotics .. .. .. 70 73 43 29 64
Weight change .. .. .. 69 67 57 57 66
Looked anxious .. .. .. 67 53 43 29 6o
Complained of sadness .. .. 6@ 6o 14 14 53
Weight loss .. .. .. .. 61 47 43 43 55
Difficulty in working .. .. 61 33 21 0 47
Reduction in work .. .. .. 53 53 43 0 48
Less interest .. .. .. .. 53 40 29 14 45
Pessimistic or hopeless about future 52 47 21 0 43
Anorexia .. .. .. .. 50 6o 21 0 44
Less social activities .. .. 47 27 43 0 40
Less energy .. .. .. .. 47 33 29 0 39
Slower movements .. .. .. 45 40 43 14 42
Reproached self .. .. .. 44 40 7 14 36
Difficulty in concentration .. .. 42 33 2I 0 35
Weeping .. .. .. .. 42 6o 14 14 39
Restless .. .. .. .. 41 40 14 0 34
Diurnalmood variation .. .. 38 20 43 0 30
Hypochondriacal .. .. .. 36 53 7 14 35
Indecisive .. .. .. .. 33 20 7 0 25
Thought self a burden .. .. 33 20 29 14 29
Slower speech .. .. .. 31 40 57 0 34
Useless or worthless.. .. .. 31 33 14 0 27
Thought let people down .. .. 28 27 7 0 23
Complained of anxiety .. .. 28 40 43 I4 31
Trembling/shaking .. .. .. 22 53 14 0 24
Failing memory .. .. .. 7 6o 14 0 22
Seriousunfoundedworries.. .. 17 33 C) 0 II
Less care with appearance .. .. 17 o 21 54 20
Said thought symptoms due to cancer i6 33 7 i6
Irritability
.. .. .. .. II 20 14 14 13
Ruminations .. .. .. II 13 7 0 10
Panicattacks .. .. .. 7 7 0 9
Delusions .. .. .. .. 9 13 29 0 12
Specific phobias .. .. .. g 0 14 0 8
Ideas of poverty .. .. .. 7 7 o 8
Weight gain .. .. .. .. 8 20 14 14
Thought illness a punishment .. 8 13 () () 7
Hypochondriacal delusions.. .. 8 27 0 C)
Sexual symptoms .. .. .. 8 7 0 14 7
Somatic anxiety symptoms.. .. 8 7 7 7
Changed smoking habits .. .. 6 o 0 14 5
Ideas of reference .. .. .. 6 0 29 0 8
Dirty .. .. .. .. 6 20 0 14
Heavydrinking .. .. .. 8 100 7 14 22
Physical violence .. .. .. 5 7 7 14 6
Threatened violence .. .. 5 27 0 i4 8
Depersonalization .. .. .. 5 0 0 0 3
Barbiturate dependence .. .. 2 7 36 0 7
Hallucinatiotis .. .. .. 0 0 21 C) 3
Confused .. .. .. .. 0 13 0 0 2
Ideas of influence .. .. .. o 0 14 0 2
Flatteningofaffect .. .. .. o o 2! 0 2
Lackofvolition .. .. .. o 21 0 3
Thought disorder .. .. .. 0 0 14 0 2
Abnormal personality .. .. 25 47 21 14 27
 BY B. BARRACLOUGH, J. BUNCH, B. NELSON AND P. SAINSBURY 36!
TABLEIV
Past histo,y ofrnental illness and attempted suicide,family histor, ofmental illness and suicide, by diagnosis
NotDepression
ill AU
(N = 64) (N = 15) (N = 14) (N = 7) (N = ioo)
0/ 0/ 0/ 0/ 0/
/0Alcoholism
/04453645441402!0 /0Miscellaneous /0mentally /0

Past history of psychiatric treatment 46

Past history of depressive illness
treated by psychiatrist . . .. 35
Past history of mania treated by
psychiatrist .. .. .. II 0 0 0 7
Past history of affective disorder
treated by psychiatrist .. .. 41 53 2! 0 37
Past history of alcoholism treated by
psychiatrist .. .. .. 0 53 0 0 8
Number of episodes of psychiatric
treatment excluding present epi
sode:
I .. .. .. .. 13 13
2 .. .. .. .. 6 13 21 0 9
3+ .. .. .. .. 25 2714
244036042100 2914 013
Probable previous episodes of de
pression not treated by psychiatrist 48
Probable previous episodes of other
mental illness not treated by
psychiatrist .. .. .. 26
Previous history of attempted suicide 33 477
3°‘3 70 ‘¿433
Number of previous attempts:
20
2 .. .. .. 53 337 054
531370II0548623275414200004 017
Unverifiable self report of suicide
attempt .. .. .. .. ‘¿3
No evidence of previous mental dis
order or suicide attempt . . .. 23
History of affective disorder treated
by psychiatrist in parent, sib or
child .. .. .. .. 20
History of suicide in parent, sib or
child .. .. .. .. 6
No history of mental illness or
attempted suicide or family history
of mental illness or suicide .. i6 0 7@ 7'
Twenty per cent of the suicides diagnosed thirds (Table V). The illness present at the
depression had first-degree relatives with histo time of death could not therefore be regarded
ries of affective disorder; frequently more than as chronic in the majority of suicides.
one family member had been ill. A similar The duration of ‘¿present
illness' of the de
proportion of Chichester referrals had family pressed patients referred to the Chichester
histories of mental illness. Six per cent of cases psychiatric service was briefer than that of the
had a history of suicide in first-degree relatives, suicides—72 per cent had been ill for less than
which is the same proportion as found in the six months at referral compared with 55 per cent
matched controlgroup ofwellpeople. of suicides. Thus, as Copes (6) inferred, suicide
may tend to occur later in an episode of de
Duration of illness pressive illness.
The duration of the spell of illness which ____
Summarized, the main clinical findings about
preceded the suicide was less than six months in the 64 suicides with depression were that
half of the cases and less than a year in two their symptoms and their family history of
 362 A HUNDRED CASES OF SUICIDE: CLINICAL ASPECTS
TABLEV
(a) Age atfirst
diagnosisNotDepression
treatmentby a psychiatrist, by

illAll(N Alcoholism Miscellaneous mentally

= 64) (N = 15) (N = ‘¿4) (N = 7) = ico)
0/ 0/ 0/ 0/ 0/
/0Age, /0 /0 /0 /0(N

years:
200rless
0425—30 .. .. .. 5 7 0
.. .. .. .. 5 13 29 0
3'—40 . .
1545—50 .. .. .. 52 13 36 o9
.. .. .. .. 12 27 0 14
55—60..
565—70 .. .. .. 8 o o o13
0271+ .. .. .. .. 2 7 0
03Not .. .. .. .. 5 0 0
previously treated by a
8648N.K.
psychiatrist .. .. .. 50 33 36
.. .. .. .. 0I101
2 0 0
100100(b) 100 101

diagnosisTime, Timebetweenfirst treatmentbya psychiatristanddeath,by

years:<I
05i—5 .. .. .. .. 5 13 0
0126—so.. .. .. .. 9 27 14
01011—20
.. .. .. .. II 7 14
0132landover
.. .. .. .. 9 13 36
14IINot .. .. .. 14 7 0
previously treated by a
8648N.K.
psychiatrist .. .. .. 50 33 36
.. .. .. .. 0I100
2 0 0
00500(c) 100 I 00 I

diagnosisTime, lime betweenestimatedonset ofpresent p.@@chiatric

illness and death, by
months:<3
0273—6.. .. .. .. 33 20 21
0197—12
.. .. .. .. 22 27 7
.. .. .. .. 57 0 14 0
13—24 .. .. .. .. i6 13 0 0 12
25andover .. .. .. II 7 57 0 i6
NA.
12N.K. .. .. .. .. o 33 0 10013
.. .. .. .. 0I101
2 0 0
500500* IOO@ 99
Duration refers to depressive illness episodes, not alcoholism.

psychiatric treatment resembled those found with alcoholics identified in a comprehensive

in a sample of living depressives, except that a survey of alcoholism in Cambridgeshire* (is).
history of attempted suicide was found eight There were i 2 men and 3 women, the same
times more frequently. ratio (4 : I) as was found in Cambridgeshire.
So the sex of the alcoholic does not, apparently,
ALCOHOLISM
* The survey aimed to identify all alcoholics aged over

@ The group of 15 suicides diagnosed alcoholism 15 the County in the three-year period 1961-I4,
will be described and compared on some features and assesstheir main characteristics.
 BY B. BARRACLOUGH, J. BUNCH, B. NELSON AND P. SAINSBURY 363
predispose to suicide. The mean age of all the cases a neurologist had been consulted, the
@ alcoholic suicides was 51 . 3 (S.D. ‘¿3
.o); complications were so severe. Alcoholic pen
for the men it was 50 . 9 (S.D. g .i@)and for the pheral neuritis had been diagnosed in one case,
@ women 53 years (S.D. 22 2). Cambridgeshire hepatic cirrhosis in another, gastritis in four, and
@ male alcoholics' mean age was 8 years one case was in liver failure at the time of death.
(S.D. ‘¿4.3); they were, therefore, significantly In half of the cases work performance had been
younger than the male alcoholic suicides (t = impaired and jobs lost, in a third drunkenness
@ 2 36; p < 0 oi) ; the older alcoholic is the had led to arrests and prosecution. Clearly they
greater suicide risk. The longer alcoholism were a seriously addicted group. When the
persists the more likely it is to cause the adverse alcoholics were classffied into an adaptation of
personal, social and health changes which may Jellinek's types (2 i), there were 5 chronic and
increase risk of suicide. 5 symptomatic drinkers; 3 uncontrolled or
@ Three of the i alcoholic suicides were compulsive drinkers ; and 2 periOdiCal or bout
divorced, seventeen times the expected number drinkers. Many depressive symptoms were
@ ( o , 8) calculated from the census. As only one recorded, and their relative incidence was
of the 85 non-alcoholic suicides was divorced, similar to those of the uncomplicated depressive
@ compared with an expected figure of o 66, suicides. So extensive were the symptoms of
it seems that the high suicide rate of the di depression that 9 of the i@ were also diagnosed
vorced, consistently found in suicide statistics depressive illness (Table III).
may largely be accounted for by divorced Without a sample of living alcoholics to
alcoholics. compare the suicides with, conclusions about
Compared with the Cambridgeshire alco their distinguishing characteristics must be
holics, the alcoholic suicides again had a high conjectural. The most obvious clinical features
incidence of divorce and widowhood. Twenty were the combination of severe alcohol addic
per cent were divorced, in contrast to 6 per cent tion and depression occurring in a recently
of the Cambridgeshire alcoholics, and 27 per disturbed domestic and social setting.
cent were widowed as compared 5 per cent.
Thus loss ofspouse may predispose the alcoholic Duration of drinking
to suicide. Such comparisons cannot show the All had lengthy histories of heavy drinking,
separations, or convey the unhappy relations of the mean duration being 25 years, and the
those who are married and still living with their range from io to @oyears. Such histories suggest
spouses, which seems to be such a special feature an early start; the median age at which drinking
of alcoholic suicides. began was 20, all but 3 beginning before thirty
(TableVI).
Symptoms and diagnosis
The record of drinking; the opinion of rela Past andfamily history
tives and friends that it was excessive and thus One half had previous histories of psychiatric
beyond the norms of the suicide's cultural treatment, usually in-patient care to control
group; the effects on health and the social drinking on the one hand or to treat an
consequencesof excessivedrinkingprovided the affective disorder on the other. Multiple
factsfor the diagnosticclassification. admissions for recurrent severe bouts of de
The more conspicuous clinical features were pression were recorded for a quarter; a similar
these: finding to that for the depressive suicides
In 8o per cent healthor socialfunctioning (TableIV).No casesofmania were reported.
was seriously impaired by drinking, and one Previous suicide attempts were verified in half,
half had physical illness due to excessive drink and ifthesuicide's self-reports
ofpastattempts
ing.The centralnervoussymptom had been are included then nearly two-thirds had
affected at some time in nearly every case; 4 had attempted suicide before; one-third of them
had delirium tremens, 5 grand ma! fits, and in twice or more often. Of the Cambridgeshire
3 memory was permanently impaired. In three alcoholics40 per cent had histories of previous
 364 A HUNDRED CASES OF SUICIDE: CLINICAL ASPECTS
TABLEVI To summarize the main points about the
Leading features of alcoholic suicides fifteen suicides diagnosed alcoholism : Alcohol
(N= 15) addiction had led to physical damage, especially
0/
/0 to the central nervous system, and caused serious
Heavy drinking, causing physical or mental social and domestic consequences. Depressive
illness, or social disability .. .. .. 8o illness was common, its onset recent and in its
Heavy drinking, not causing such disability 20
clinical features similar to the depressed suicides
Informant says suicide drank too much .. I00
Family says suicide drank too much . . .. and to the unselected living depressives. Com
93
Suicide said he drank too much .. .. 40 pared with Cambridgeshire alcoholics, the
Frequency of drinking: suicides were older and more likely to be
Daily . . .. .. .. .. .. 67 divorced or widowed, and histories of attempted
Two-three times weekly .. .. ..
@ Bout drinking . . .. .. .. .. suicide were present six times more often.
Illness due to drink .. .. .. ..
@ Trouble at work .. .. .. .. MISCELLANEOUSDIAGNOSES
Arrested because of drunkenness .. .. 33 Depressive illness and alcoholism accounted
Heavy thinking at time ofsuicide .. ..
47 for 79 of the 93 suicides diagnosed mentally ill;
Recent change in drinking habits:
Increase .. .. .. .. .. 40 the remaining fourteen with a miscellany of
D@ase .. .. .. . ... 53 diagnoses will now be discussed. They were:
No change . . .. .. .. .. 7 (I) Advanced physical disease in
Help to stop drinking .. .. .. 40 association with depressive illness 4
Duration of heavy drinking (in years):
0-9.. .. .. .. ..
7 (2) Schizophrenia 3
!0—I9 .. .. .. .. .. 20 (3) Phobic anxiety state 3
20-29 .. .. .. .. .. 40 (4) Barbiturate dependence (in
30—39 . . . . . . . . .. 27 association with depressive illness) 3
40—49 .. .. .. .. ..
Age of onset of heavy drinking:
7 (5) Acute schizo-affective psychosis i
10—19 . . . . . . . . .. 33 The physical diseases associated with a depressive
20—29 . . . . . . . . .. 47 illness were : recurrent and extensive pulmonary
30—39 . . . . . . . . .. 7 infarction; multiple sclerosis with quadriplegia
40—49 . . . . . . . . .. 7 and a rodent ulcer; carcinoma of the stomach
50—59
60-69
.
..
. .
..
. . .
......
. . ..
; with
peritoneal
metastases
; and
carcinoma
of
the prostate with peritoneal metastases. The
psychiatric treatment, but only io per cent had aetiological relation between mood disorder
histories of attempts at suicide. Thus evidence of and organic disease is impossible to disentangle;
previous suicidal actions is an important distin the interesting but complex interaction between
guishing feature of the alcoholic suicide, as it is these two major factors is well illustrated with
with the depressive suicide. these two brief histories.
None of the alcoholics had a family history of
I. A 66-year-old retired railway engine driver lived with
treatment for alcoholism, though the relatives of
his wife. Throughout his adult life ups and downs of mood
3 were said by informants to be heavy drinkers; periodicallyaffectedhis behaviour; he was at times phobic
the families of 4 (27per cent)cases,however, about underground travel, and outspokenly overconcerned
had had treatment for an affective disorder. about his health. A minor railway accident just before his
The past and family histories of mood disorder retirement caused him to take a year offwork, the condition
being diagnosed neurasthenia by his family doctor.
in the alcoholic suicides resemble those of the Shortly before his death he suffered recurrent pulmonary
depressives, and this observation may indicate emboli, undiagnosed ante-mortem, with haemoptyses
a fundamental similarity between the two and dyspnoea. Then, quite suddenly, he developed a
groups. Excessive drinking by depressivesto severe agitated depressive illness when he said he thought
he might have lung cancer. A fortnight later he shot
control unpleasant affect has been suggested himself. There was no family history of mental illness.
as one of the explanations for alcoholism (23).
2. A 93-year-old widower had lived alone in a seaside
It may well be that the alcoholic especially at hotel for 15 years since his wife's death. His life had been
risk for suicide falls into that category. a good one; a happy marriage, a family of successful
 BY B. BARRACLOUGH, J. BUNCH, B. NELSON AND P. SAINSBURY 365
children, commercial prosperity in middle life, a circle of instances, the intention to die did not seem high:
friends who liked and respected him. He had successfully
the act was clearly impulsive and their deaths
coped with a series of cthes—migration to the East when
young, near bankruptcy in late middle age, and finally may well have been accidents.
widowhood, lonelinessand financial dependency. Shortly These six, the phobics and the addicts share
before his suicide by asphyxiation an agitated depressive many common features ; persistent experiences
illness suddenly began. At post mortem a large, spreading of unpleasant anxious and depressed mood, the
gastric carcinoma, undiagnosed in life, was found. A
brother had a severe mental illnessin middle life. use of drugs to control this, and the social and
interpersonal consequences of drug abuse. In
Schizophrenia was diagnosed three times. these respects the resemblance to the alcoholics
All three cases were young, single people, is close, indicating possible shared features, pre
under psychiatric care whose ample clinical disposing to suicide, and perhaps also to the
records substantiated the diagnosis. Two were depressives, some ofwhom also used barbiturates
long-term in-patients, one in a rehabilitation excessively.
unit and the other in a back ward. The third The last of the fourteen, a young single man,
was a day patient, ill some six months with her being treated by his G.P. with protriptyline 30
second attack, having recovered from the first mgm. daily, was diagnosed an acutepsycliosis, but
two years before. had symptoms characteristic of both a schizo
No mood disturbances were recorded nor phrenic and a depressive psychosis of great
previous attempts at suicide, nor talk of it. severity. After being placed under a compulsory
Each of them used a highly effective method: admission order he slipped away as the
hanging in a little-used public lavatory; drown ambulance door opened at the mental hospital
ing, with suitcases of stones tied to the wrists; admission bay, ran to the nearest railway track
a massive aspirin overdose while living alone. and was crushed by the next train.
The deaths were unexpected and apparently
inexplicable : the assessment of suicide risk in
schizophrenia may be difficult. SUICIDES NOT DIAGNOSED AS MENTALLY lu.
The three cases of phobic anxi4y slate, all Seven cases were not diagnosed as mentally
youngish married people, had anxiety symptoms ill, but in only three instances were the three
of such degree that their domestic and social psychiatrists unanimous. Thus there is less
lives had been severely affected ; all required confidence about this classification than about
daytime sedation, and all had seen psychiatrists. the others. The following case histories give
The two women were dependent on barbitu some indication of ill health, but insufficient to
rates, and the man, although taking tran make a diagnosis. An opinion can be formed
quillizers, drank heavily. All three killed them therefore of the criteria used in diagnosing
selves after unpleasant experiences affecting mental illness.
self.regard; in two it was the loss of keenly
anticipated jobs, and in the third it was the
loss of a boy friend. One had attempted suicide CaseP.52
A retired nurse, single, aged 77, lived reclusively in a
before, and another had taken a number of rented bedsitting room. No relatives or Mends could be
accidental overdoses. found after her death, but her landlady had taken good
The threesuicides by barbiturateaddicts,all care of her. Physical and mental health had always been
middle-aged married women, were associated good until ten days before her death when continuous
with depressive mood disorders in two and a abdominal pain began, for which the general practitioner
prescribed conservative treatment The pain continued,
long-standing depressive personality disorder she could not eat and lost weight. She hinted that she
resembling the ‘¿Munchausen'syndrome in the thought she had cancer. Arrangements were being made
third. All three were living in difficult personal for her transfer to a nursing home so that her landlady
circumstances, and used barbiturates to alleviate could have a fortnight's holiday, but she asphyxiated her
self with a plastic bag before they were completed. At
unpleasant subjective experiences. None had post-mortem an abscess in the pouch of Douglas arising
attemptedsuicidebefore,but one had several from a ruptured diverticulitis coli and consequent pelvic
times taken an accidental overdose. In two peritonitis was discovered.
 366 A HUNDRED CASES OF SUICIDE: CLINICAL ASPECTS
Gas. M.i@ before death. Eight months previously she had been
A 21-year-oldmarriednurse,an Asianimmigrantfrom widowed. Her mental health had always been g@,od,but
Tanzania, lived with her husband, her four-month-old coronary disease, which caused angina and a myocardial
baby and another male relative and his wife.Physical and infarction eighteen months before death limited her
mental health had always been good. The marriage was activities. Bilateral labyrinthectomy for Méniêre's
disease
said to be happy; but there was some worry and an affected her balance. She left notes which indicated a
happiness at work. She drowned herself impulsively and realistic appraisal of her future health and a dislike of
without warning after a briefdomestic quarrel. living without her husband. She had spoken of suicide
frequently.
CaseM.@z6
A se-year-old successful married nurse lived with his
wife and daUghter. His past mental health had always PERSONALITY
been good. The family reported no symptoms or stresses, The three psychiatrists who were provided
whereas evidence from his workmates suggested definite
with definitions based on Schneider (is),
changes: weight loss, tenseness, irritability, not working
well, not as well groomed, and hints ofmarital difficulties classified 27 of the suicides as having an ‘¿abnor
and money troubles. His death by poisoning with an mal personality' (Table III) ; they obtained a
prescribed barbiturates followed a sudden financial strain. fair measure of agreement, 20 of the 27 cases
CaseM.4o being classified unanimously. However, the
A 71-year-Oldsingle former private soldier and labourer reliability of an alternative classification accord
had lived in a mental hospital for 23 years. The diagnosis
ing to personality types was too low for use.
on admission was an epileptic psychosis and alcoholism.
A sib had been admitted to a mental hospital with epilepsy. Personality factors may play a key part in
Two years before his death he slashed his wrists in a suicide. They determine not only some of the
suicide attempt At the time of his death he was mentally social situations associated with it, but also a
and physically well, and transfer to a local authority person's responses to being ill and his reactions to
welfare home was being arranged. On the day ofhis death
his ward was moved and integrated with a female ward. adverse circumstances occasioned by chance.
He went out, got drunk (blood alcohol i 15 mgm.%), and The association of illness and certain abnor
jumped, or fell, from a pier and quickly drowned. There malities and traits of abnormal personality may
was doubt in the interviewer's mind as to the verdict's therefore affect the suicide risk. The subject is
correctness. complex and will be discussed in future papers.
CaseM.49
A 72-year-old childless retired civil engineer had lived
WARNINGS
with his wife in a prosperous seaside retirement suburb for
sixteen years. His physical and mental health had been The proportion of suicides who can be re
good until glaucoma made eye surgery necessary a year garded as having given warnings depends on
before his death. To his doctor and opthalxnologist be
made remarks about future blindness, and they reassured what evidence the investigator and the in
him that his sight was not in serious danger; nevertheless formant, both with enlightened hindsight, are
he still seemed preoccupied with that thought. No other prepared to accept as a warning. Unequivocal
symptom or stress was recorded. He gassed himself statements of intention to kill shade through to
withoutwarning. ambiguous hints which may only have signifi
CaseM.58 cance after the event. If only unquestionable
A 29-year-Old married Anglo-Indian, a skilled trades..threats are counted, such as that ofthe man who
man, lived with his wife and two young children. Six years
before his death he had a head injury resulting in 48 hours said when threatened with a court appearance,
@ unconsciousness. A personality change followed; he ‘¿I
will kill myself rather than appear', then
became awkward and irritable, and made threats of per cent gave such warnings (Table VII). The
suicide at various times. Although generally cheerful,
number of alcoholics recordedas havingmade
he suffered from sudden depressive moods with associated
violent behaviour, whicheventually ledhiswifetotake an overt suicide threat was significantly higher
out a separation order. The day she did this he gassed than the number of other suicides who did so
@ himself while apparently depressed. Scarring of the cortex (x2 = p <0.05). The reasonmay be
was evident at post mortem. His mother had been treated that by reason of personality disorder, or
fordepression afterhisbrotherhad been drowned in
circumstances suspicious of suicide.
because of the effects of alcohol, alcoholics are
less inhibited about frightening people or mani
CaseM.59
A childless,widowed, medical practitionerof 65 lived pulating them. In one third of the suicides the
alone in a flat into which she had moved two months threat was uttered more than one month before
 BY B. BARRACLOUGH, J. BUNCH, B. NELSON AND P. SAINSBURY 367
TABLE VII
Threats and warnings, b,
diagnosisNotDepressionAlcoholismMiscellaneousmentally

illAll(N
ioo)0/
= 64)(N = 15)(N = 14)(N = 7)(N =
/0Threats /00/ /00/ /00/ /00/

suicide—timebetween,threat
of
death:7days
and
.. .. ..

8—3odays
931—365 .... .. ..II 827 207 714 013
..II13141412@o6o292834No days ....

threat recorded
..70.407'716610010010099lOGRecent ....

talk about death,dying

orsuicide
53572955No
.. .... ..58.

recent talk about death..4247437!45100100ZOOZOO100

death, so before the thought of suicide becomes It should not be thought that a clear threat is
deed, there is a period of gestation when inter a more serious predictor of suicide than an
vention may be possible. oblique hint ; our findings show the reverse
If a less stringent definition of warning is may be true. Both need serious attention when
taken, and statements such as ‘¿You'll
have my spoken by a depressive or by an alcoholic,
gratuity—I won't be here', or ‘¿I'll
do something especially if he has made a previous suicide
stupid before that happens', are accepted as attempt, and particularly ifhis present domestic
clear indications of suicidal intent, then over a life is in disarray.
@ half per cent) gave warnings. The suicides in
each diagnostic group did not vary in the extent
to which they made these less obvious warnings. N4EDICAL CONTACT
The proportion ofsuicides recorded as having The period between the suicide's last medical
given warnings must be an underestimate, for consultation and death was elicited by inter
only some of their contacts were interviewed, viewing the family doctors and psychiatrists,
although these were necessarily the ones who and in most cases was verified by examining the
were in a position to act—relatives, close friends medical record (Table VIII). Two thirds had
and doctors. Suicide is, therefore, not always a visited their family doctor in the month before
surprise. Warnings have a preventive value in death, 40 per cent in the week before. Only
labelling the mentally ill who are likely to 25 per cent of the control sample had seen a
attempt suicide. More speculatively, they may doctor in the previous month, and 7 per cent
also be useful in considering the function of had done so in the previous week. Thus the
such frightening talk for the potential suicide. suicides had recently seen a doctor, and there
Without entering into detailed discussion, one fore had medical contact to a far greater degree
may see that the ‘¿cryfor help' formulation is than the normal population, indicating that
clearly only a part of the story. Talk indicating they either felt some need of help or perceived
thoughts ofsuicide are sometimes a statement of themselves as sick. Seventeen had not seen their
fact ; or coercive ; or may imply a mental conifict family doctor for more than one year, but three
between the wish to die and the wish to live; of those were under psychiatric care and one
or may indeed be a real supplication. was a general practitioner who treated herself.
368 A HUNDRED CASES OF SUICIDE: CLINICAL ASPECTS
TABi.x VIII

Period
diagnosisNotDepressionAlcoholism
betweenmost recent medical contact and death, by

illAllControls(N Miscellaneous mentally

= 64) = ,@) .(N = 14) (N = 7) = ioo) = 150).
0/ 0/ 0/ 0/ 0/ 0/
/0(N /0 /0(a) /0 /0(N /0(N

doctorO—7days Family
..
434078—@odays ..346o 36
..
14191731.-.9Odays ..1913 29
.. .. 7 14
28366+days..
91—365
days .. ..i6 1313 7 0 1414 1023
..17 29 14
1000—7days 997 100 101 9917 10025

P.@ychiatrist
0II8—3OdayS .. ..9(b) 27 7
..
‘¿473i—godays ..80 7
039I—@65days..
.. ..37 0
.. 032234 14

Total under psychiatric

care
1424Not .. ..20 28
under psychiatric
care .. . .78 71 86
1000—7days 10067 101 99 100

(c)Mostrecent
centact
ofeither76
.. .. 43 43
8—@odays .. .. 20 7 36 29 2!
31—9Odays .. .. ‘¿7 7 7 ‘¿4 ‘¿4
91—365days.. .. II 7 7 14 10
366+days .. ..42 9 0 7 0 7
99So 101 lOG 10048 lOG

A quarter of the sample were seeing a psychi tentions to be recognized or for their obtaining
atrist, and half of them had visited him in the treatment for the illnesses they have been shown
week before death. Most were severely ill; the to be suffering from—or, on the other hand, for
three cases of schizophrenia, nearly all the obtaining the means for peaceful self destruction.
manic-depressives and the more difficult cases of
alcoholism were those receiving psychiatric MEDICAL Ti@s@mn@NT
treatment. Of those not under psychiatric care, We asked the subjects' doctors what treatment
two had recently been discharged and one was they had been given and also looked at their
in the process ofcompulsory admission. Recency medical notes (Table IX).*
of contact varied between diagnostic groups; Eighty per cent were prescribed psychotropic
8o per cent of the alcoholics, for example, had drugs, which implies that the suicides were
seen either a doctor or a psychiatrist in the recognized by their doctors as being psycho
week before death, compared with 40 per cent logically disturbed. Hypnotics were prescribed
of depressives. most often, and were given to the depressives
The salient findings, then, were that the and alcoholics, and frequently for many years.
majority of suicides had recently seen their All but two suicides were receiving their
general practitioner and many were receiving
S The counselling and supportive psychotherapy given
psychiatric treatment, consequently there was will not be assessed here ; nor the provision of social care
no lack of opportunities for their suicidal in by statutory and voluntary bodies.
 BY B. BARRACLOUGH, 3. BUNCH, B. NELSON AND P. SAINSBURY 369
TABI2 IX
Prescribed medical treatment at time of death by diagnosis

Miscellaneous ill
(N = 64) (N = 15) (N = 14) (N = 7) (N = ioo)
0/ 0/ 0/ 0/ 0/
/0Prescribed /0NotDepression
/0Alcoholism /0mentally /0All

psychotropic drugs
Barbiturates .. .. .. ..
53 53 57 43 53
Antidepressants .. .. ..
30 7 7 —¿
21
Phenothiazines
2!Calcium .. .... ..8i 2087 20100 3643 —¿82

(abstem)..—7——IMinor
carbimide
tranquilizers ..
ECT.. .. .. .. ..
5 —¿
7 —¿
3
Haloperidol
IHypnotics .. .. .... ..23 220 —¿21 —¿â€” —¿21

.. .. ....7073432964

supplies of barbiturates from general practi
tioners. A detailed description of the hypnotics
prescribed and their relation to the impulsive
suicide has been published (3).
One third of the depressives had been pre
scribed antidepressants, which poses the question
of why the drug did not improve the depression
and prevent the suicide ? A close look at the
prescribing showed the following: Of the i@
depressives given antidepressants, 4 had a mono
amine oxidase inhibitor (MAOI) and 15 one of
the tricycics. The four cases prescribed MAOI
had severe symptoms ; the drug is not found to
be effective (is, 20) with this type of depression.
Five of the I5 given a tricycic had been taking
the drug continuously for more than six months
without any improvement in their symptoms;
treatment had therefore failed. Four had been
prescribed ECT, the suicide occurring at the
beginning of treatment; in four others the
relatives were sure the drug was not being
taken; in one the drug was started the day
before death ; and in only one case had a
tricycic been accompanied by some improve
ment before death. The average dose was also
well below that recommended. Thus in 9 cases
the drug had failed, in 5 it had not had a fair
trial, and in 4 MAOI had been given when the
indication was for a tricyclic.
earlier episodes of depression had been success
fully treated by ECT, but it was only used in
four instances in the spell that preceded the
suicide. Both doctors and patients apparently
dislike repeating it too often.
Antidepressant drugs had not been prescribed
to any of the cases ofdepression with alcoholism
or with physical disease; it is not known whether
this was because the depression had not been
diagnosed or because it was believed that
depression complicated by other conditions does
not respond to drugs.
Prophylaxis with lithium carbonate, un
monitored, was being given to only one case
of manic-depressive psychosis; it had not yet
come into wide use. Phenothiazines were used
to treat the three cases of schizophrenia, but
i8 other patients were also taking phenothia
zines. It is possible that the drugs may have been
harmful in some instances (i).
To summarize : a wide range of psychotropic
drugs were prescribed, showing that the doctors
had diagnosed a psychiatric disorder in most
cases. One third of the depressives were given
antidepressants, mostly in doses below those
recommended. Barbiturates, phenothiazines and
minor tranquilizers were probably over
prescribed.
DISCUSSION
The most important of the findings is the
high proportion of suicides who had been
diagnosed as mentally ill. Could this finding be
In many cases an artefact introduced through bias on the part
of coroners, informants, interviewers or the
committee of psychiatrists who made the
diagnoses?
It could be argued that coroners might only
bring in a verdict of suicide where there was
370 A HUNDRED CASES OF SUICIDE: CLINICAL ASPECTS
evidence of present or past mental illness, error. Each item of behaviour was predefined
and allocate other suicides with clean bills of to avoid leading questions and subjective
health to the Open Verdict and Accidental judgements. The interviewer also had to
Death categories, thereby inflating the propor record the supporting evidence for positively
tion of mentally ill suicides. This is most un scored items, and the completed case record
likely, because coroners, nearly all of whom are was reviewed item by item at a conference of
lawyers, are required to use a legal definition of the three interviewers. Differences in inter
suicide which relies almost exclusively on preting the replies and the criteria by which
evidence ofintent (i i). Further, false accidental items were defined were then settled. Docu
or open verdicts are appealed against by Life mentary evidence for symptoms and such items
Insurance companies, who stand to lose by as present, past and family history and treat
wrong decisions. And even if the coroner has ment was often available. And the level of
evidence of a disturbed mental state he may not agreement in the joint interviews was high and
construe it as psychiatrists do. no evidence of a systematic bias was found.
The converse may also be argued, that The panel of psychiatrists who reviewed the
accidental deaths with evidence of mental ill data and then, without discussion, decided
health will falsely be called suicides, again whether the suicide was mentally ill might
inflating the percentage of mentally ill. The have tended to make diagnosis too readily in
constraints imposed by the legal definition of cases of suicide, but in 93 per cent of cases they
suicide, and the coroners' bias against a suicide were in agreement that a mental illness was
verdict because of its social stigma, make such present. Their standards for making a diagnosis
decisions improbable, yet in three cases this are illustrated in the case histories of those
may have happened (see case M.4o). But if classified as not mentally ill, and can be re
those three are removed from the series the garded as on the conservative side (see p. 365).
proportion diagnosed as mentally ill does not Using the common definition of depressive
change. This bias is not therefore a source of illness, the inter-psychiatrist agreement was
overestimation ofmental illness in our sample. 86 per cent ; similarly when they were asked to
Relatives might inflate the proportion of confine a diagnosis of alcoholism to cases in
mentally ill by attempting to ‘¿explain'
to them which it was clear from their histories that
selves the reason for the suicide. Again this alcoholism had had severe physical or social
seems improbable ; they give their statements to consequences, there was also an 86 per cent
the coroner's officer before the inquest, when consensus.
they are overwrought and less likely to dissi Independent evidence of the validity of their
mulate, even supposing they had the medical diagnosis was supported by the documentary
knowledge to do so. In fact a number of evidence, which is not subject to bias of the
relatives firmly stated the suicide's state of mind kind we have been considering. Half the
was lucid, to the extent that the coroner's suicides had documented reports of a previous
sympathetic phrase ‘¿suicide
while the balance mental illness, and if attempted suicide is
of his mind was disturbed' angered them. And included the proportion rises to two thirds.
our impression during the interviews was that Moreover, their family histories, course and
relatives underestimated symptoms and hence symptoms were precisely those found for an
illness, partly because of difficulty in recalling unselected sample of depressives. In addition
the details after a month, and partly because to the psychiatrists' considered decisions based
they could only report behaviour; they were on the history and symptoms obtained from
denied the subjective experiences the psychiatrist relatives, there were supporting data derived
relies on when making a diagnosis. from their previous medical records and current
Interviewer bias when questioning relatives treatment,for three quarters had attended a
and rating their replies might also have led to doctor in the previous month and over 8o per
an overestimate of mental illness, although the cent were taking a psychotropic drug of some
interview was designed to overcome such an kind.
 BY B. BARRACLOUGH, J. BUNCH, B. NELSON AND P. SAINSBURY 37!
Finally, it should be noted that our results pressives differed from their comparison group.
resemble those obtained in St. Louis (i7) and From which it can be inferred, obviously
in Seattle (8). The proportions of mentally ill enough, that suicides are drawn from those
suicides found in those studies were g@per cent depressives who have previously had suicidal
and ioo per cent; their diagnoses were also thoughts and acted upon them. Such thoughts
similar; depression and alcoholism comprising may be specific to the illness and have the form
the great majority—85 per cent in St. Louis of delusions or obsessions, or they may be a
and 66 per cent in Seattle. Our study, a more response to the experience of illness determined
extensive replication of theirs, has produced a by personality. In either case the ultimate
similar result, making due allowance for explanation for such thinking may lie as much
differences ofecology and diagnostic standpoint. in the culture as in the illness, because attitudes
In summary, therefore, a very high proportion to suicide which may come to the fore when
of suicides have been found to be mentally ill; illness supervenes, or form the content of
the detailed data support the validity of the symptoms, are acquired during social matura
diagnoses, and possible sources of bias do not tion.
account for the findings. The high proportion of suicides found to
have an abnormal personality does suggest that
depressives and alcoholics may be predisposed
to suicide on that account. The suicidogenic
Virtually all the suicides studied were there
effects of an abnormal personality may be those
fore mentally ill; mood disorders predominating.
which make it difficult to sustain steady rela
Most of them were in contact with the medical
tionships; those conducing to a morbid do.-
services, received treatment with psychotropic
pendency so that the individual is predisposed
drugs, and warnings about suicide were
to depression should he become separated or
commonly given. These, starkly summarized,
bereaved; or those which generate the stressful
are the findings relating to the stated aims of the
social situations commonly found associated
study. What implications then do the findings
with suicide.
have for the causes ofsuicide and its prevention?
Social and interpersonal factors are clearly of
great importance in suicide, and it seems likely
Causes that social stresses, particularly those affecting
Mental illness is an essential component of domestic life increase the suicide risk of de
suicide; our findings, and those ofRobins and of pressives and alcoholics. These social stresses
Dorpat suggest that in Western society suicide may be caused by the illness, or by abnormal
in the healthy person is a rare event. Yet it is behaviour, or they may be the result of chance
obvious from follow-up studies that although and perhaps then play a part in causing the
depression and alcoholism, the disorders most illness itself. The subject is complex and will be
commonly reported, are associated with high discussed in future papers.
suicide rates, only a minority, about i@ per cent
over a lifetime die in that way (to). Other factors Prevention
must therefore be involved. Does the study What useful conclusions may be drawn about
indicate what they might be? preventive measures ? The very high proportion
Conceivably they might arise from some of suicides with mental illnesses of a kind
special characteristic of the illness, the symptom usually regarded as treatable suggests that
pattern and length of illness for example, or suicide could be decreased by the effective use
from the type of personality and its response to of modern treatments. But from a practical
illness, or from the suicide's social setting, view point such a confident appraisal might be
especially isolation, or the stresses he encounters. misplaced. The findings that a high proportionof
The incidence and pattern of symptoms suicides had lengthy histories of treatment for
appeared typical, and it was only in the history recurrent or persisting illness, and that 25 per
of more previous suicide attempts that de cent were having psychiatric treatment at the
 372 A HUNDRED CASES OF SUICIDE: CLINICAL ASPECTS
time oftheir death, suggest that they may either ally disordered suicide, because the findings of
have varieties ofiliness not amenable to existing our survey gave some clear indications that
methods of treatment and management, or preventive measures are feasible. But that is not
that psychiatric treatment and medical services to say that other approaches have no place,
are not effectively deployed. Our fIndings and other studies suggest that extending psychi
suggest that the truth may be between the two, attic services to the community (22) may reduce
that some may not receive sufficiently high the incidence of suicide in patients, and that
standards of diagnosis and care, while others the systematic after-care of patients suffering
are unresponsive to available methods. Recent from conditions known to increase the risk of
developments in the treatment and prevention suicide (I 8) may also be effective.
of recurrent mood disorders which seem to
‘¿present an especially difficult treatment pro

blern may have an impact on the incidence of SUMMARY

suicide. Lithium prophylaxis may be such an One hundred suicides were investigated
advance (7). retrospectively by interviewing surviving rein
Over 40 per cent of the sample had previous tives. Nine-three per cent were diagnosed
episodes of affective disorder, and a stringent mentally ill, 85 per cent suffering from de
anal@ of the chronology of their relapses pression or alcoholism. Eighty per cent were
suggests that lithium might have prevented seeing a doctor and 8o per cent were prescribed
relapse, and hence suicide, in at least half of psychotropic. drugs. Over a half had given
these 40 per cent (2). Prophylaxis of attacks of warnings of suicidal thinking. Some suicides
depression with tricyclics may similarly prove may be preventable with modern psychiatric
to be of value in suicide prevention (14). treatment, but our investigation showed that
. But there is nevertheless, likely to be a propor these methods were not always being effectively
ton inaccessible to existing methods, and in deployed.
whom other methods designed to cope with
chronic disability will be required. In some
Acx@owiinoaisr@crs
cases, however, there were indications that the
The followinggave valuable help: H.M. Coronersfor
standard methods of treating depression could Portsmouth (P. D. childs) and West Sussex (G. L. F.
have been deployed more effectively. First, only Bridgxnan and F. F. Haddock) ; the Clerk to the West
‘¿9
of the 64 suicides with a depressive illness, Sussex Executive Council (J. R. Knighton) ; the @J1erk
to
and only I of the 19 alcoholics with depression, the Portsmouth Executive Council (A. R. Goddard);
were receiving antidepressant drugs; none of Mr. A. Corkindale, General Register Office; Drs. N.
Capstick, J. C. Grad de Alarcon, B. E. Heine, A. Levey,
the 6 cases with depression associated with other 3. D. Morrissey,D. Pallis,N. H. Rathod; Mrs.J. Coffins;
conditions were prescribed them. Second, the Mrs. V. J. Kewdll; the relatives of the suicides, who
antidepressants were not effective because the unselfishly agreed to be interviewed; and the family
type and dose were not appropriate. Third, doctors who answered enquiries.
EUF had been neglected in those with whom it
had previously beeen successful. Fourth, barbi
turates, the commonest cause of suicidal death, I. DE ALAROON, R. & C@'uu@zv, W. M. P. (@6g) Severe
were liberally prescribed by general practi depressive mood changes following slow release
tioners, and often repeated without seeing the intra-muscular fluphenazine injection. BTitÃœh
MedicalJournal, iii, 564—7.
patient, who either telephoned or left a note. 2. BARRAcLOUGH, B. M. (1972) Suicide prevention, re
Although it is unlikely that a determined suicide current affective disorder and lithium. Britith
would be deterredby denying him drugs,there Journal ofP@ychiatry,121, 391-2.
seem to be some who take them impulsively. 3. —¿ NELSON,B., Brmcii, J. & S@n@en@uav, P. (@97@)
Their deaths might be avoided if drugs as Suicideand barbiturate prescribing. Journal of
the Royal Collegeof GeneralPractitioners,21, 645—53
toxic as barbiturates were not ready to hand (s). 4. Bur4cis, J., BARRACLOUGH,B. M., NELSON, B. &
The emphasis has been on the clinical aspects SAINSBURY, P. (i97i) Suicide following bereave
and the medical management ofthe psychiatric ment of parents. SocialPsychiatty,6, 193—9.
 @-@r@1

BY B. BARRACLOUGH, J. BUNCH, B. NELSON AND P. SAINSBURY 373

5. c@LAYTON, P. J. & Pfl@S,F. N. (@6@)Aff@th@ 14. MINDRAM, R. H. S., Howi@im, C. & S,ssp,iaan, M.
disorder. IV. Mania. Comprehensive
Psychiafr,,6, (‘973)An evaluation of continuation therapy
313—22. with tricyclic anti-depressants in depressive
6. Ooi'ns,J.B., FREEMAN-BROWNE,
D. L. & ROBIN,
A. A. illness. Psyc1iolo@ Medicine,3, 5-57.
(‘97') Danger periods for suicide in patients 15. Mess, M. L. & Bannsi'oan-DAviu, E. (zg67) A Siwvej
undertreatment. PsychologicalMedicine,i, @oo—@o@j. ofAlcoiwlism in an English Count,. Cambridge.
7. Coppasi, A., NOGUERA, R., B@nay, J., Buiues, B. H., i6. Rs@oism@@ut
Gaseniw@(i967) Sample Census ig66.
Sw@i, M. S., Hi@az,E. H., GARDNER,
R. & County Reports for Hampshire and West Sussex.
MAGG5, R. (i97i) Prophylactic lithium in London.
affective disorders. Lancet, ii, 275-9. I 7. ROBINS, E., Muiuxv, G. E., WILICINSON, R. H.,
8. DORPAT,T. & Rwx@, H. S. (,g6o) A study of suicide G@swan,S. & K@vxs,J. (i@@@) Some clinical
intheSeattlearea.Con@prehensiePsychiafry,i,@-59. considerations in the prevention of suicide based
9. Ga@D,J. and SAINSBURY,P. (1966) Evaluating the on a study of 134 successful suicides. American
Community Psychiatric Service in Chichester: Journal ofPublic Health, 49, 888-98.
Results. The Milbanic Memorial Fund Quarterly i8. S@inmuay,P. (,968) Suicideand depression,in Recent
XLIV 246-278. Developmentsin AffectiveDisorders(ed. Coppen, A. &
@o.GUZE,S. B. & RoBins, E. (1970) Suicide and primary Walk, A.). London.
affective disorder. British Journal of Psychiatry.
59. SCHNEIDER,K. 0. (1950). PsychopathicPersonalities.
117, 437—8.
ii.Jervls on Coroners(i@@7)9th edition (ed. Purchase & Vienna.
Woflaston). London. 20. Saapssasw, M., L.@sa, M. & Ronsno,rr, R. (i968)
12. Kaas'ne@n, N., SAin8BURY, P.,MoIuussEY, J.,TowEas, Clincal Psychopharmacology.
London.
@ J. & Sciuvr.nan,J. (‘96')The reliability of 2!. E. & Roi'ss, M. (1970) ClinicalPsjthiat,y.
psychiatric assessment. Journal of Mental Science, London.
107, 887—908. 22. WAUC, D. (‘967)Suicide and community care.
13. MEDICAL RESP.ARcss Councn. (1965) Clinical trial of British Journal of Psychiatry,113, 1381-91.
the treatment of depressive illness. British Medical 23. WnsoKim,G., CLAYTON,
P. J. & REICH,T. (‘969)
Journal, 1, 88i-6. Manic DepressiveIllness. St. Louis, U.S.A.

A synopsis of this paper was published in the January 1974 Journal.

B. M. Barraclough, F.R.A.C.P., M.R.C.PSyCh.,

J. Bunch, M.A.,

B. Nelson, Dip. Soc.Ad.,

P. Sainsbury, M.D., F.R.C.P@th.,
M.R.C. Clinical Psychiatry Unit, Graylingwell Hospital, Chic/jester,Sussex

(Received 17 July 5973)

A Hundred Cases of Suicide: Clinical Aspects
B. BARRACLOUGH, J. BUNCH, B. NELSON and P. SAINSBURY
BJP 1974, 125:355-373.
Access the most recent version at DOI: 10.1192/bjp.125.4.355

References This article cites 0 articles, 0 of which you can access for free at:
http://bjp.rcpsych.org/content/125/587/355#BIBL
Reprints/ To obtain reprints or permission to reproduce material from this paper, please write
permissions to permissions@rcpsych.ac.uk

You can respond /letters/submit/bjprcpsych;125/587/355

to this article at
Downloaded http://bjp.rcpsych.org/ on April 24, 2016
from Published by The Royal College of Psychiatrists

To subscribe to The British Journal of Psychiatry go to:

http://bjp.rcpsych.org/site/subscriptions/