Mark Williams - Suicide & Attempted Suicide - Understanding The Cry of Pain-Penguin UK (2002)

SUICIDE AND
ATTEMPTED SUICIDE
MARK WILLIAMS
‘An important book that

deserves a wide readership’
Simon Armson, Chief Executive, -
THE SAMARITANS
S U I C I D E AND ATTEMPTED S U I C I D E
Mark Williams was bom in 1952 and educated at Stockton-on-Tees

Grammar School and St Peter’s College, Oxford. Between 1979 and 1982
he was Lecturer in Applied Psychology at the University o f Newcastle
upon Tyne and from 1983 to 1991 was Research Scientist at the Medical
Research Council’s Applied Psychology Unit in Cambridge. Since 1991
he has been Professor o f Clinical Psychology at the University o f Wales,
Bangor, and, since 1997, Director o f the University’s Institute o f Medical
and Social Care Research. His other books include The Psychological
Treatment o f Depression, Cognitive Psychology and Emotional Disorders (with
F. Watts, C. MacLeod and A. Mathews), and, with Z. Segal and J.
Teasdale, Mindfulness-based Cognitive Therapy for Depression.
Professor Williams is married and has one son and two daughters.
0041486617
Suicide and
Attempted
Suicide
Understanding the Cry o f Pain
MARK W I L L I A M S
/CD
P E N G U IN BOOKS v>
To Phyllis, Rob, Jennie and Annie.
P E N G U IN B O O K S
Published by the Penguin G ro u p

P enguin B ooks Ltd, 80 Strand, L on do n W C 2R o r l , England
Pen gu in Putnam Inc., 375 H ud son Street, N e w Y o rk, N e w Y o rk 10014, U S A
Pen gu in B ooks Australia Ltd, R in gw o o d , V ictoria, Australia
Penguin B ooks C an ad a Ltd, 10 A lcorn A v en u e, T o ro n to , O ntario, C an ad a M 4V 3 B 2
Penguin B ooks India (P) Ltd, 11 C o m m u n ity C en tre, Panchsheel Park, N ew Delhi - n o 017, India
Penguin B ooks ( N Z ) Ltd, C n r R osedale and A irborn e R oads, A lbany, A uckland, N e w Z ealan d
P enguin B ooks (South Africa) (Pty) Ltd, 24 Sturdee A ven u e, R o seban k 2 19 6 , Sou th A frica
Penguin B ooks Ltd, R egistered O ffices: 80 Strand, Lon don W C 2R o r l , England
First published 1997

P revio u sly published under the title Cry o f Pain,
n ew edition published u nd er the present title 2001
C o p yrigh t © J . M . G . W illiam s, 1997, 2001

All rights reserved
T h e m oral righ t o f the au thor has been asserted
Set in 10 /12 .5 pt M on o typ e Dante

T yp e se t b y R o w lan d Phototypesetting Ltd, B ury St Edm unds, Su ffolk
Figu res draw n b y N igel A n drew s
Printed in England b y C lays Ltd, St Ives pic
Except in the U nited States o f A m erica, this b o o k is sold subject

to the condition that it shall not, b y w a y o f trade o r o th erw ise, be lent,
re-sold, hired out, o r o th erw ise circulated w ith o u t the pu b lisher’s
prior con sen t in any form o f binding o r c o v e r other than that in
w h ich it is published and w ith ou t a sim ilar condition including this
condition being im posed on the subsequent purchaser
Contents
List o f Text Figures vi

List o f Tables viii
Preface to the N ew Edition ix
Acknowledgements xi
Introduction xiii
1 A Brief History o f Suicide i

2 Suicide: Facts and Figures 18
3 Psychiatric and Social Factors in Suicide 45
4 Attempted Suicide: Facts and Figures 63
5 The Causes o f Attempted Suicide 81
6 Rational Suicide, Euthanasia and Martyrdom 98
7 Psychodynamics, Biology and Genetics 115
8 The Effect o f the Media 127
9 The C ry o f Pain 136
10 M em ory Traps 155
11 The Prevention o f Suicidal Behaviour 174
12 Therapy for Suicidal Feelings and Behaviour 195
13 Final Thoughts 217
Notes and References 225

Index 243
List of Text Figures
2.1 Number o f suicide deaths by gender, 1911-99, England and

Wales. 21
2.2 Recorded suicide deaths: rate per million population by gender
and age, 1983-95, England and Wales. 22

2-3 Comparison o f suicide rates for men and wom en for different
ethnic groups in the United States, 1950-96. 26
2 .4 Distribution o f recorded suicides by gender, year o f death and
method, England and Wales. 27
2.5 Num ber and percentages o f suicide method by gender, US rate
1996. 30
4.1 Episodes o f attempted suicide referred to the John Radcliffe
Hospital, Oxford, 1976-99. 68
4 .2 Oxford City attempted suicides, 1999, by age and gender. 69
4.3 Methods o f self-poisoning, Oxford: changes between 1976 and
1999- 70
4-4 Rates o f attempted suicide (per 100,000), 1989-93, across
European centres. 71
10.1 Tim e taken to recall specific autobiographical memories. 160
10.2 Stages in retrieval o f specific autobiographical m emory. 163
10.3 Effect o f aborting search for specific m emory. This results in
many repetitions o f intermediate stage, resulting in over
elaborated self-descriptive categories (‘mnemonic interlock’). 163
11.1 Suicide rates in Gotland and the mainland o f Sweden, during
and after systematic education o f Gotland G P s on diagnosis and
treatment o f depressive disorders. 178
List of Text Figures
I I .2 Dum m y data illustrating effect o f hypothetical intervention to

reduce death-rate. 180
i i -3 Graph illustrating effect o f applying five-year m oving average to
data in Figure 11.2 if m oving average is not stopped prior to
intervention. 180
11.4 Graph showing the smoothed means o f actual Gotland data
retaining a five-year m oving average but stopping the window
prior to intervention. 181
12.1 Num ber o f episodes o f self-harm in borderline personality
clients at three phases o f treatment. 209
12.2 Mean number o f episodes o f self-harm in year following D B T
for borderline personality disorder clients. 210
12.3 Mean number o f days in psychiatric hospital in year following
D B T for borderline personality disorder clients. 211
12.4 Mean number o f medically treated episodes o f self-harm in year
following D B T for borderline personality disorder clients. 212
13.1 At important points in the causal pathway to suicidal behaviour
a judgem ent by the individual allows scope for bias in estimates
o f the aversiveness o f stress, its controllability, and how much
social support is available. 221
List of Tables
2.1 Suicide rates for a sample o f countries reporting to the W orld

Health Organization. 24
2.2 Distribution o f ‘other methods', England and Wales. 29
2.3 Deaths from suicide among men (aged 16-64) in different social
classes, England and Wales, 1979-90. 32
3.1 Suicide vulnerability and psychiatric status. 48
3.2 Factors associated with suicide rates among 15-29-year-olds in
Europe, 1960-80. 57
4.1 Proportions o f people endorsing Bancroft & H aw ton’s ‘reasons
for overdose'. 76
5.1 Sample o f items from Marsha Linehan’s Reasons for Living
Inventory. 93
10.1 Responses to cue words given by overdose and matched
control subjects. 161
10.2 Scenarios used to cue memories. 164
10.3 Situations used from the M eans-Ends Problem Solving
Test. 168
10.4 Ratings o f effectiveness o f problem solutions. 169
10.5 Ratings o f specificity o f future images. 171
11.1 Predictors o f suicide. 182
Preface to the New Edition
The first edition o f this book, Cry of Pain, aimed to summarize what is
known about suicide and attempted suicide, and to understand the
biological, social and psychological factors that come together to
increase suicide risk. Since then, there have been some changes in the
statistics, and several new research studies. In the United Kingdom
during the mid-1990s, for example, it looked as if the number o f young
men attempting suicide was increasing so fast that it would catch up or
exceed the number o f wom en. This did not occur, and the situation
looks as if it is settling back into a more familiar pattern whereby
w om en (especially young wom en) are more likely to do themselves
harm that is not fatal, whereas men remain more likely to harm
themselves in ways that are fatal. Other studies have confirmed a
number o f findings that appeared to be emerging then: the close link
between attempted suicide and a history o f child sexual abuse in wom en;
the w ay suicidal behaviour is affected by stories in the media; the
problem o f suicide in rural areas, especially among the farming com
munities; and the greater suicide risk o f people, w ho have been in
psychiatric hospital, during the first few weeks after discharge. Since
then the United Kingdom has also seen the outpouring o f public
emotion that occurred following the death o f a princess, an event that
caused an increase in the number o f suicides and attempted suicides,
and further studies have confirmed that the availability o f the means to
harm oneself remains a large factor in many suicides.
The first edition was written with two sorts o f reader in mind. First,
there are health professionals w ho regularly meet people who are, or
Preface to the New Edition
might become, suicidal. But there are also many general readers who
simply want to know more about what makes someone harm them
selves. It is clear that suicidal feelings and behaviour are much broader
than the world o f professional counselling and therapy. There are few
people who have not, at some point in their lives, been touched by
suicide, either by their own suicidal feelings, or by suicidal urges and
behaviour in family or friends. Central to the argument o f Cry o f Pain
was the idea that suicidal behaviour represents a response to a feeling
o f being defeated combined with a feeling o f not being able to escape
the consequences o f defeat. Like an animal caught in a trap, the struggle
to be free is followed by defeat and hopelessness. Social, psychological
and biological /genetic studies are increasingly confirming this view.
People who are suicidal find themselves seeing evidence o f defeat and
rejection everywhere; periods o f high activity (struggle and protest)
are punctuated by periods o f inactivity and despair. This slide into
hopelessness can occur irrespective o f psychiatric diagnosis, for the
feelings o f inescapability arise as much from the failure to control one's
own internal symptoms (thoughts, feelings and behaviour) as they do
from any failure to deal with 'external' problems in living (family, job
and relationships).
In writing this edition, I am grateful to a number o f colleagues,
especially Andy MacLeod, Leslie Pollock, Paul Gilbert, Caroline
Creasey, Bruce Napier, Simant W estley and Keith Hawton. Thanks also
to Stefan McGrath and Caroline Pretty o f Penguin Books, for help,
advice and encouragement throughout the project. I hope that this
book will continue to be helpful, both to the professionals and to all
those who want to understand more about the enigma o f suicide.
MARK WILLIAMS
February 2001
Acknowledgements
A number o f people have helped with this book at various stages. I am

grateful to Isaac Sakinofsky o f the Clarke Institute o f Psychiatry in
Toronto, and to David Clark o f the Center for Suicide Research and
Prevention, Chicago, for help in obtaining up-to-date statistics from
North America. M y colleague Leslie Pollock has been helpful in keeping
me abreast o f developments within the W H O database. Keith Hawton
kindly made available details o f the Oxford deliberate self-harm data,
and Allan House made available details o f comparable data from Leeds.
Stephen Platt has been generous with advice and help at a number o f
stages, and Breda M cLeavey made available details o f her studies o f the
treatment o f suicidal behaviour in Ireland. Glyn Lewis o f Cardiff gave
helpful comments on the primary prevention chapter, Paul Gilbert
advised on the ‘C ry o f Pain' chapter, Phil Cowen o f Oxford gave helpful
comments on the sections on the psychopharmacology o f suicide and
Neil Cheshire advised on psychodynamic aspects. A number o f col
leagues read other parts in draft: Michaela Swales, Isabel Hargreaves,
Patrick Vesey, Miriam Shieldhouse, Debbie Lovell, Jean Lyon, Caroline
Creasey, Bethan Jones, Mair Edwards and David Nightingale; I am very
grateful for their help.
I am grateful to H M SO for permission to reproduce tables and
figures from Crow n Copyright material (Population Trends, 1992), to
Routledge for permission to use parts o f a previous chapter I had written
on cognitive therapy for suicidal patients (in j. Scott, J. M. G. Williams
and A. T. Beck (eds.), Cognitive Therapy in Clinical Practice, 1989), and to
Carfax Publishing for permission to use parts o f m y paper with Leslie
xi
Acknowledgements
Pollock ('Factors Mediating Suicide B eh avio u r) that appeared in the

Journal o f Mental Health (1993). Brenda Ellis and Sharon Fraser have
helped in the production o f the text, and I am specially grateful to
Caroline Creasey, w ho has borne the brunt o f transforming m y scribbles
into readable drawings and figures, and to Peter Ford, for his excellent
copy-editing and advice on the form o f the final text.
Finally, thanks are owed to close friends and family. JefT Williams-
Jones, Nona Ephraim and Lloyd and Carolynn Evans have shared in
many ups and downs during the writing o f this book, and to them I am
deeply grateful. The project could not have been completed without
the help and support o f m y family. Robert has read parts o f the text
and commented freely and helpfully upon it. Phyllis, Annie and Jennie
have always been prepared to discuss any o f the ideas covered in the
book. All have allowed me to concentrate on writing when they had
every right to expect that I should stop and attend to their needs. For
this love and support I am forever in their debt.
MARK WILLIAMS
April 1996
Introduction
There is much about suicide that presents a real and often tragic puzzle
to be solved by family and friends, by physicians and other professionals
involved with som eone who has committed or attempted suicide. At
the level o f the individual event the questions are most often: ‘W hy did
they do it?' ‘W hy did they not see there was help available?’ At the
larger level, the questions remain whether a biological, sociological or
psychological explanation will provide the best clue to suicidal
behaviour. Each level o f explanation, seen in isolation, has its drawbacks,
so w e need to see how each type o f explanation m ay fit together with
the others to give a more complete picture.
The large growth in interest during recent years in the subject o f
suicide and suicidal behaviour m ay be due, first, to the rapid increase
since the mid-1970s in suicides by young people (particularly young
men). Although the same period has seen a decrease in the numbers o f
older people committing suicide (the traditionally most vulnerable
group), the change in rates has meant that many more suicide victims
have taken their lives ‘at the prime o f life5, challenging society’s view
o f itself as successful.
Second, there has been an increased wish within the World Health
Organization to challenge health care systems to meet targets to
improve the health o f their populations. Suicide rates are seen as a
visible and quantifiable aspect o f the mental health o f a nation.
Third, interest in suicide has been fuelled by a renewed fascination,
especially in the United States, with euthanasia, medically assisted
suicide and rational suicide (self-deliverance). W hen Derek H um phry’s
Introduction
book Final Exit was published in 1991, with its matter-of-fact, detailed
discussion o f the best w ay o f ending life, it sold more than 500,000
copies in its first year. These are issues and debates in which both public
and professionals are interested.
In the main, tw o types o f book are written in response to this
increased interest. The first type is written for the mental health pro
fessionals. The second type is written for suicidal people, their families,
or for suicide ‘survivors’.1 O f all the books which have attempted to be
useful to both the professional and lay communities, Erwin Stengel’s
Suicide and Attempted Suicide2 became a classic. It was written originally
in the early 1960s, and reprinted several times. It gave the main facts
and figures together with sufficient interpretation to help the reader
understand something o f the mind o f those who feel suicidal. The book
kept apart suicide and attempted suicide, but in its revision Stengel
included an extra discussion about the 'appeal function’ o f attempted
suicide, which he believed had been misunderstood, and more on the
separation o f completed from non-fatal suicide attempts, which was
controversial in its day.
There have been many developments in the field since Stengel’s
death. These include changes in the socio-demographic pattern o f
suicide deaths; knowledge about which social factors are most likely to
produce changes in rates o f suicide; a greater awareness o f the strategies
to prevent suicide (and their limits); more focus on the impact o f the
media on suicidal behaviour; and an understanding o f the w ay in which
depression and hopelessness may act as ‘final com m on pathways’ to
suicidal behaviour. The present book aims to deal with these issues in
such a w ay that a lay audience m ay come to know what are the main
facts and theories, but also so that health and mental health professionals
may have a digest o f a subject which will be constantly in the background
in their practice.
For this reason, the book’s subtitle reflects not the puzzle about
suicidal behaviour, but instead summarizes one possible answer. Suicidal
behaviour is most often a ‘cry o f pain’. This perspective is intended to
capture the w ay in which behaviour can have a communication outcome
without communication being the main motive. It is analogous to the
animal caught in a trap, which cries with pain. The cry is brought about
xiv
Introduction
by the pain, but m ay nevertheless communicate distress in a w ay which

will affect the behaviour o f other members o f the species. Suicidal
behaviour m ay be overtly communicative in a minority o f cases, but
mainly it is ‘elicited' by the pain o f a situation with which the person
cannot cope - a cry o f pain first, and only then a cry for help.
The distinction thus drawn with earlier ideas about the ‘cry for help'
is quite deliberate. Stengel acknowledged that suicidal behaviour might
have an ‘appeal function’. H ow ever, having raised the concept in the
early editions, Stengel felt he had to try to counteract a fundamental
misunderstanding. He had been taken to mean that suicide attempts
w ere only appeals, only manipulative. He had meant to imply that such
behaviour had an appeal function in the same w ay that physical illness
has an appeal function - i.e. people will change their behaviour to try
and help the person. In this sense, he says, ‘the appeal effect o f a suicidal
act m ay be the greater the less it was intended\3
H ow ever, the scene was set for many years o f a misunderstanding
which still remains. Even now, many suicidal acts are dismissed as
‘mere* cries for help, as if a communication m otive was incompatible
with a serious attempt to end life, and as if the self-damaging act did not
represent a mental health problem which needs to be taken seriously. So
the ‘cry for help’ idea, though originally intended to be a neutral theory
about suicidal behaviour, has outlived its usefulness. First, it has become
limited to non-fatal suicide attempts and thus contributed to a widening
o f the gap between non-fatal suicide attempts and suicide. Second, it is
almost always used pejoratively, or at least to imply that a certain
suicidal act was not so serious but ‘m erely’ a cry for help. Those who
w ork closely with people who feel suicidal and sometimes act on such
feelings know that such behaviour is never ‘m erely’ anything.
Although there are some grounds for maintaining the separation
between completed suicide and attempted suicide,4 it is now widely
acknowledged that differences m ay have been overplayed. There
remain m any differences between these groups, including age and sex
ratio, but not only do the populations overlap, the motivation for both
completed and attempted suicide is complex and crosses the boundary
between them. For example, anger and communication motives can be
found in both, and even relatively low-risk ‘suicidal ideation’ (which
xv
Introduction
often does not lead to suicidal behaviour) m ay be dominated by the

theme o f escape and death rather than communication.
Suicide is usually the most individual o f acts, and I realize how
difficult it is to try and draw conclusions that are general across a
number o f situations. Many readers whose professional or personal
lives have been touched by suicide o r suicidal behaviour will, at many
points, be able to think o f exceptions to much o f what is written here.
Yet if we are to understand and help people in the future, there will
need to be a dialogue between the general rule and the individual
circumstance. This book is offered as a contribution to that dialogue. I
shall start with history.
xvi
1
A Brief History of Suicide
A psychiatrist was talking about suicide risk in her patients. ‘I can

understand w hy some o f m y patients should want to kill themselves. If
they really want to do it, there is nothing I can do/ she said. ‘Even if I
was able to stop them, would it be right to try? Many o f them have
enough insight to know how their illness has ruined their lives/ Contrast
her attitude with that expressed by Professor Gethin Morgan in a U K
Health Advisory Service publication in the same year (1993):
Those at risk o f suicide come to us in our professional capacity to get help.

They have already talked with relatives and friends and we may well be the last
port o f call. They watch us intently for our response. They are usually ambiva
lent about suicide and we have a responsibility to encourage the wish to live
. . . Anyone who has extensive experience in suicide prevention will know that
things can improve in a most unexpected way even in the case o f individuals
facing what may seem enormous adverse odds. It is for us to assume that
change for the better is always possible.1
This represents a disagreement about whether suicide is preventable,

and also a disagreement about whether, even if it could be prevented,
such prevention would be morally justified. If suicide were sometimes
a rational response to an unbearable situation, what business do psy
chiatrists and other mental health professionals have in intervening?
These are not the only conflicts in attitudes to suicide. In fact,
W estern society’s attitude to suicide throughout the ages has been
at best confused, swinging between punitive severity and tolerant
advocacy.2 W alking along the High Street in Bangor some time ago,
1
Suicide and Attempted Suicide
someone wearing a sandwich-board thrust a leaflet into m y hands. It

was concerned with sin, mostly, but included a warning against the
danger o f a person ‘becom ing so discouraged to the extent o f yielding
to the horrible sin o f taking his own life . . . Through suicide the
opportunity to repent is cut o ff and soul and body will be destroyed in
hell/ T o many, this will seem an extreme view from another age, yet
it was only in the early 1960s that attempting to take one’s own life
ceased to be a criminal offence and became the province o f the health
service rather than the courts. By the last decade o f the twentieth
century, several research studies seemed to have shown beyond doubt
that around 90 per cent o f people who commit suicide are suffering
some form o f mental illness. Yet, if so, w hy is there such little evidence
that psychiatric or medical intervention can affect the probability o f
suicidal behaviour? If medicine is not the answer, it would be better if
we were clear about it, and stopped expecting medical practitioners to
w ork miracles on society’s behalf.
Attitudes in classical and medieval times
It is often thought that in Greek and Rom an times there was a tolerance
for suicide. H owever, many philosophers condemned self-destruction.
Pythagoras, for example, compared suicides with soldiers who deserted
their posts. Aristotle said a person should not com m it suicide since such
an act cancelled unilaterally his or her obligations to the state, the
contract between a person and society. On the other hand, Stoic and
Epicurean philosophers believed that suicide could be the right course
o f action in some circumstances, e.g. where there was terminal illness
or unremitting pain. They also allowed that in certain circumstances it
might be an act o f nobility where it was an expression o f political rights
or values.
In the fifth and sixth centuries, attitudes to suicide shifted gradually
away from more permissive Roman philosophical ideals and became
more punitive. St Augustine thought the philosophical support o f sui
cide abhorrent. H owever, the Christian Church has always had to
contend with the difficulty that suicide is nowhere explicitly condemned
2
in the Bible. Christian thinkers and philosophers have always had to

resort to the commandment, T h o u shalt not kill/ Appealing to this
commandment has difficulties o f its own, since most Christian thinkers
have wished to make an exception for people who kill as part o f a just
war. Clearly the commandment, T h o u shalt not kilT, if taken as
absolute in every circumstance, would prevent this sort o f killing too.
If killing could be excused in times o f war, w hy were there no other
circumstances in which the commandment could be set aside? Could
not some o f these circumstances be precisely those envisaged by the
Stoic and Epicurean philosophers?
Since the Bible did not condemn suicide explicitly, Christian thinkers
had to find other arguments. St Thom as Aquinas argued that suicide
was against the natural law. Since God was expressed in natural laws,
suicide was a sin. Adding weight to the Christian thinkers5 rejection o f
suicide was the popular belief in the demonic origin o f self-killing. In
the Middle Ages the view was encouraged that supernatural activity in
the natural world was a relatively frequent occurrence. People who
committed suicide therefore showed evidence that they had been pos
sessed by devils. The diabolical causes o f suicide found many expressions
in literature. One phrase repeated and elaborated throughout the Middle
Ages was that o f Egbert, an eighth-century writer, who blamed ‘self-
murder' on ‘the instigation o f the D evil’.
In England, ‘self-murder’ became thought o f as an offence against
God, against the King and against nature. Those w ho committed suicide
w ere tried posthumously by a coroner’s jury. If they were convicted as
having murdered themselves, all their goods, including all household
items and m oney and debts owed to them, were forfeit to the Crown
or the C row n ’s Agent. The result was that the family o f someone who
committed suicide, especially if it was the head o f the household, would
be reduced to abject poverty.
In addition to these consequences, ‘self-murderers’ were denied
Christian burial. Instead their bodies were buried ‘profanely’ . The
macabre cerem ony surrounding such burials seems to have its origins
in pre-Christian times. But the belief that suicide was a supematurally
evil act encouraged the desecration o f the body o f the person who had
killed themselves. The individual was buried, often at a cross-roads,
3
naked and with a wooden stake through the body. The hole was filled
in, sometimes with the stake showing above the earth so that passers-by
might be reminded o f the awfulness o f the circumstances o f this death.
The clergy did not attend the ceremony. It was carried out by officials
o f the parish, including the church wardens and their assistants.
As part o f the posthumous trial for self-murder, the coroner's court
had to decide whether it was really murder or not. The only mitigating
circumstance was if it was found the person was insane. If someone
killed themselves when mad or mentally incompetent in some w ay,
they were not convicted. Instead o f returning a verdict o f felo de se (a
felon o f himself), the person was deemed to be non compos mentis (not
o f sound mind).
Despite all these popular feelings and the fact that committing suicide
was considered a crime, during the Middle Ages very few juries actually
brought in verdicts o f felo de se. It was not that suicide was rare, but
that juries, often consisting o f local people w ho had sympathy with the
family, were reluctant to see them become paupers. The Crow n had
little control over these local juries, and although from time to time
attempts were made to tighten the enforcement o f the law, this was
rather sporadic. It was not until Tudor times that this aspect o f social
and legal life was more rigorously controlled.
The Tudor revolution in attitudes: government

and folklore combine
During the early sixteenth century, the governm ent tightened up on a

number o f aspects o f English law, enforcing its will where it had not
done so before. The law about suicide was enforced using the Court o f
Star Chamber. This, the King’s Council sitting injudicial session, tried
(and succeeded) in many areas to enforce those laws where there was
a direct financial benefit to the Crow n and government. Since suicides
considered felo de se meant that goods and households were forfeit, the
Court o f Star Chamber had a great deal o f interest in ensuring that the
verdicts were thoroughly carried out and the penalties exacted.
The struggles between local juries and the Court o f Star Chamber
4
continued, however. Local officials would often declare that the suicide
victim had very little possessions. In one extreme case the total value
o f a person's possessions was declared to be precisely the value (to the
last ha’penny) o f the value o f the person's debts. In this way, local
coroners' juries attempted still to protect their fellow-villagers and
townsfolk from the excesses o f central government.
The results o f all the governm ent reforms were reflected in what
appears to be a dramatic increase in suicides, or at least those reported
to the King's Bench. The average number o f suicide inquisitions in 1500
was 61 ( felo de se)y and 1 non compos mentis. By 1600 this number had
risen to 873 felo de se and 7 non compos mentis verdicts.
The Tudor revolution in governm ent was reinforced by the attitude
o f the churchmen, w ho continued to emphasize their belief that self-
murder was an expression o f despair brought about by the devil.
Both Calvinists and non-Calvinists were equally vehement in their
condemnation. For example, George Abbot, one time Archbishop o f
Canterbury, declared that suicide was ‘a sin so grievous that scant any
is more heinous unto the Lord'. He was Calvinist, but the anti-Calvinist
Lancelot Andrewes was similarly explicit: ‘It is worse than beastly to
kill or drown or make aw ay with ourselves; the very swine would not
have run into the sea but that they were carried by the Devil' - a
reference to the story o f the Gadarene swine in the Bible (Mark 5,
Luke 8).
In popular stories and in sermons the idea o f Satan playing on a
man's guilt and luring him to his death by abject despair is prominent.
In morality plays, a point is often reached where the devil casts Man
into a despair or ‘wan-hope'. In John Skelton's Magnyficence, a character
is persuaded by Despair that his sins are so bad that God will not forgive
them. Despair urges him to suicide with the words ‘ryd thy selfe rather
than this lyfe for to lede'. The character Mischief appears, offering the
instruments o f self-murder: ‘Lo here is thy knyfe and a halter, and all
were go ferther, spare not thyself, but boldly thee murder.' The central
character is about to use the knife against him self when Hope suddenly
enters and urges him not to kill h im se lf‘against Nature and Kynde'.3
In such a culture, suicide meant the struggle had been lost, the devil
had won. In Christopher M arlow e’s Dr Faustus, the devil instigates
5
suicide more explicitly perhaps than in many other allegories. The Good
Angel and the Evil Angel try to persuade Faustus to repent on the one
hand and to despair on the other. Faustus cries, ‘My heart’s so harden’d,
I cannot repent.’ W hen Mephistopheles offers Faustus a dagger after he
has yielded to despair, Faustus cries, 'D am n’d art thou, Faustus, damn’d;
despair and die!’
O f course there is little doubt that m any o f the people who tried to
kill themselves or actually succeeded were extremely depressed. The
fact that non compos mentis verdicts were not brought in for them cannot
hide this fact. In many cases the depression was clearly o f psychotic
proportions. A young Puritan, Nehemiah Wallington, imagined him self
'provoked by the D evil’ to suicide and made n attempts. According to
Nehemiah’s own account, Satan showed him self in various manifes
tations: as a crow, as his sister, as a minister and as a disembodied voice.
The disembodied voice immediately suggests auditory hallucinations
o f the psychotic patient. But the pull to life turned out to be stronger
than the pull to death, despite ‘the temptation o f Satan’ .
Then Satan temted me again and I resisted him again. Then he temted me a
third time, and I yielded unto him and pulled out my knife and put it neere my
throat. Then God o f his goodness caused me to consider what would follow if
I should do so. With that I felle out a weaping and I flong away my knife.
As in contemporary accounts o f ‘reasons for living’ v. ‘reasons for dying’,

the person w ho had suicidal thoughts would often stop themselves for
these religious reasons or because o f family. In the early seventeenth
century an account is given o f a wom an whose husband had died and
who was tempted to commit suicide but found her love for her child
saved her from doing so.
But few such considerations made any difference in the case o f
someone psychotically depressed. John Gilpin, an ex-Quaker, reported
that he was possessed by Satan. On one occasion his hand was carried
to take up a knife which lay on the table. His hand was then carried
with it towards his throat and a voice said to him ‘open a hole there,
and I will give you the words o f eternal life’ (a reference to St Jo h n ’s
Gospel: ‘Lord, to whom shall we go; you have the words o f eternal life*
(6:68)). Such involuntary hand movements, in which the limbs appear
6
to be out o f the control o f the person, are not unknown in schizophrenic

states and certain forms o f brain damage.
During the sixteenth century there was little change in the overall
attitude to the type o f ‘mitigating circumstances' that might be allowed
by the coroners' courts. One might suppose that, because melancholy
was seen, as today, as the final com m on pathway to suicide, this would
constitute enough grounds for a non compos mentis verdict. However,
the religious thinking o f the time, combined with the popular belief
that melancholy was a sign that the devil had taken over a person’s
soul, meant that society found it difficult to shake itself free from the
conviction that suicide was the outcome o f diabolical possession and
not caused by madness. The result was that less than 5 per cent o f the
men and w om en who committed suicide between 1485 and 1660 were
judged to be non compos mentis.
Nevertheless, in the sixteenth century, as in other periods o f history,
there was a lack o f uniformity about attitudes. The elements o f more
tolerant attitudes, later to become more general beliefs, can be found
even where the Zeitgeist was less tolerant. For example, in Thom as
M ore’s Utopia people w ho are ill with incurable diseases were imagined
to be able to kill themselves with the permission o f their priests. This
was considered, at least in pagan terms, to be a ‘good and wise act’ ,
since the death o f the person would put an end to torture rather than
to enjoyment. In those terms, then, it was considered a pious and
holy act. It is not surprising that the intellectual elite o f the country,
influenced by the Renaissance emphasis on classical literature, should
revisit many aspects o f such Epicureanism.
The seventeenth century: gradually changing attitudes
During the seventeenth century an increasing range o f motives began

to be imputed in the case o f suicidal thoughts, feelings and behaviour.
W ith it came an increased secular view o f suicide as being caused by
economic circumstances or psychological state o f mind. After the mid
1600s coroners appeared much more ready to accept an alternative to
thereto de se verdict. H ow ever, characteristically o f a period o f transition,
7
no one could be certain that, should a suicide occur, there would not
be a felo de se verdict followed by the requirement to forfeit the person’s
goods. This is well illustrated in Samuel Pepys’s Diaries.
A businessman called Anthony Joyce, kinsman to Pepys, tried to
commit suicide by throwing him self into a pond. Pepys takes up the
story in his diary entry o f 21 January 1667:
Comes news from Kate Joyce that, if I would see her husband alive, I must
come presently. So I to him, and find his breath rattled in the throat; and they
did lay pigeons to his feet, and all despair o f him. It seems, on Thursday last,
he went, sober and quiet, to Islington, and behind one o f the inns, the White
Lion, did fling himself into a pond: was spied by a poor woman, and got out
by some people, and set on his head and got to life: and so his wife and friends
sent for.
Joyce explained to Pepys that he had been led by the devil and that he
had ‘forgot to serve God as he ought’. H owever, it is interesting to note
that Pepys him self believed the real reason for his suicidal behaviour
was that his business had failed after losses sustained in the Great Fire.
Here we see evidence o f the increased secularization o f attitudes. But
side by side with it another theme emerges.
Pepys goes on to give a first-hand account o f the panic a suicide
instilled in the bereaved family: panic that their goods would be forfeit
(even though technically Joyce survived the suicide attempt a few days).
Pepys even agrees to hide some o f the fam ily’s goods to prevent their
being seized and some o f the panic transfers to him. Finally, he uses his
contacts in high places to avert the crisis:
The friends that were there, being now in fear that the goods and estate would
be seized on, though he lived all this while, because o f endeavouring to drown
himself, my cosen did endeavour to remove what she could o f the plate out o f
the house, and desired me to take my flagons; which I did, but in great fear all
the way o f being seized; though there were no reason for it, he not being dead.
So, with Sir D. Gauden, to Guild Hall, to advise with the Towne-Clerke about
the practice o f the City and nation in this case; and he thinks it cannot be found
selfe-murder; but if it be, it will fall, all the estate, to the King. So I to my cosen's
again; where I no sooner come but find that her husband was departed. So at
8
their entreaty, I presently to White Hall, and there find Sir W. Coventry: and
he carried me to the King, the Duke o f York being with him, and there told my
story which I had told him: and the King, without more ado, granted that, if it
was found, the estate should be to the widow and children. I presently to each
Secretary’s office, and there left caveats, and so away back to my cosen’s. When
I come thither, I find her all in sorrow, but she and the rest mightily pleased
with my doing this for them; and which, indeed, was a very great courtesy, for
people are looking out for the estate.4
Shifts in opinion about suicide during the seventeenth century might

not have been so absolute had they not been supported by shifts in the
arguments from the intellectual elite o f the country similar to those
witnessed in M ore's writing a century before. Once again, the continued
interest in classical philosophy, with its undertone o f tolerance and
understanding, was responsible. The momentum o f interest in the
humanism o f the Renaissance, founded on a reverence for classical
literature, philosophy and history, was maintained, involving an increas
ing influence by classical customs. This, in turn, revived interest in Stoic
view s which were m ore tolerant o f suicide, and indeed recommended
suicide as a right act in certain circumstances.
Yet the ambivalence to suicide continued. Even when a w ork was
written that is celebrated as one o f the greatest turning-points in the
thinking about suicide in the seventeenth century, Biathanatos by John
Donne, its author dared not publish for fear o f where his own arguments
had taken him. This w ork explored the ‘Paradox or Thesis, that Self
Homicide is not so naturally Sin as it m ay never be otherwise’. The
importance o f this w ork is that Donne relied on theological analysis,
rather than returning to classical literature. In this w ay he was able to
undermine the arguments the Church had used for centuries. His
arguments included the fact that, first, self-killing is nowhere forbidden
in the Bible and that, second, many who have committed suicide in the
history o f the Church have been excused or seen as martyrs.
Third, the reference to the Sixth Commandment, T h o u shalt not
kill’, which St Augustine used, ignored the fact that on many occasions
killing is found to be legitimate, e.g. in war, or execution o f criminals.
Donne criticizes Augustine’s arguments that biblical and early Christian
9
suicides were permissible and poured scorn on the view that these acts
were carried out by the secret command o f God. This, Donne says, is
mere supposition and cannot be considered a rational argument. Indeed
Donne goes further at one point to argue that one could see the death
o f Jesus o f Nazareth as suicide.
Donne's treatise constituted an extremely rare set o f arguments
since most defendants o f suicide o f the time relied on classical literature.
Even Thom as More, who had written Utopia envisaging that suicide
would be permissible for humanists in that society, had not envisaged
that suicide would be permissible for Christians.
By contrast, the Stoic view was that death ‘unlocked the chains o f
suffering' and no law could forbid it. As Montaigne put it, ‘as I offend
not the laws made against thieves when I cut my own purse, so am I
nothing tied unto laws made against murderers, if I deprive m yself o f
m y own life'.5 Far from imagining that suicide was against the natural
law, writers relying on Stoic arguments argued that no one was obliged
to live when his or her death might promote some greater good. In
fact, they thought that suicide in certain circumstances was consonant
with the natural law.
Yet John Donne was so worried about how Biathanatos would be
received that he refused to permit it to be published in his own lifetime.
In presenting the manuscript to Sir Robert Ker, he commented that it
was a book ‘written by Jack Donne and not by Dr Donne'. It was finally
published in 1647. It appears that Donne him self shied away from his
own conclusions - not an uncommon occurrence. It was one thing to
use arguments for the justification o f suicide to undermine the state’s
wish to punish severely a person who took their own life, another to
use arguments in such a w ay as to give the impression that suicide was
an option implying that life could be ended lightly as an acceptable way
o f dealing with life’s difficulties. W hy such reluctance to publish? Donne
was a poet, but also, as Dean o f St Paul’s, a pastor. As a pastor he may
have had to deal with people bereaved by suicide, and m ay well have
felt it anathema to take a punitive line. But the complexity o f what is
said as part o f a private pastoral relationship does not transfer well into
public pronouncement or public policy.
10
Madness as illness
By the end o f the 1700s, the felo de se verdicts were as rare as the non
compos mentis verdicts had been in the early Tudor period. The prevailing
belief had become that suicide was either a product o f rational choice
or a medical calamity.
Once again a case-by-case leniency exercised by individual courts
confirmed the change. The coroner's ju ry became increasingly reluctant
to enforce penalties for self-murder. T w o trends drove the change:
increasing resistance to a law which seemed too draconian, and increas
ing willingness to see suicide as the product o f an unbalanced mind.
Assisting the first trend against property forfeit was the increased belief,
expressed by John Locke, that no governm ent should interfere with the
cult o f private property. Governm ent should not ‘take to themselves
the whole or any part o f the subjects property without their own
consent'. Such beliefs had been confirmed and emphasized by the
English Revolution o f 1688.
Assisting the second trend towards mental illness as an explanation
was the rejection o f religious enthusiasm, whether Puritan or Catholic.
Although it is likely that popular belief in supernatural forces continued
(and still continues), the intellectual elite gradually, during the Enlight
enment, came to reject almost absolutely the belief in supernatural
intervention in the natural world. This gave w ay, in the case o f suicide,
to a medical interpretation which greatly influenced the middle-ranking
men w ho served on coroners' juries. N ow , evidence o f melancholy,
previously evidence that the person had given in to the devil, was found
sufficient proof that a person’s balance o f mind was disturbed.
The suicide verdicts reflect the trends. Less than 7 per cent o f the
suicides reported to the King’s Bench were declared to have been non
compos mentis in the middle o f the seventeenth century. By the 1690s,
around 30 per cent o f suicides were thought to be non compos mentis,
and this proportion steadily increased over the next few decades into
the eighteenth century. Forty per cent o f verdicts were non compos
mentis in 1710, and by the last third o f the eighteenth century this had
become the usual verdict.
11
Indeed, in Norwich the figures between 1670 and 1799 show the
percentage non compos mentis at the early part o f that period as being
quite low, below 30 per cent; but by 1770-1800100 per cent o f the verdicts
were non compos mentis. The non compos mentis verdict itself implicitly
rejected the folklore and religious interpretations o f suicide as a super
natural intervention by the devil. Instead it substituted a medical exp
lanation which declared explicitly that the suicide was excusable.
The increased tolerance for self-murder can be seen in various writers
o f the time. William Ram esey in The Gentlemen's Companion (1672)
suggested that those who killed themselves, because they were fre
quently the victims o f mental illness,
should rather be objects o f our greatest pity than condemnation as murderers,

damned creatures and the like. For, ’tis possible even for G od’s elect having
their Judgements and Reasons depraved by madness, deep melancholy, or
somehow otherwise affected by Diseases o f some sorts, to be their own
executioners. Wherefore let's be slow to censure in such cases.
One consequence o f this change was that no more were non compos
mentis verdicts used simply as a device for protecting some and not
others. During the medieval and Tudor periods a person’s social stand
ing, his personality, his relationship with his neighbours, and his sur
vivors’ claims on the sympathy o f the local community all played a part
in determining which verdict was brought in by the coroners’ courts
(as in the Pepys example). Families o f the rich who committed suicide
were able to bribe the coroner or the jury, or perhaps more com m only
could afford to bring in an expert medical opinion to certify that
the person had undoubtedly been unbalanced at the time o f death.
Gradually, as coroners’ juries became more reluctant to find the suicide
guilty o ffelo de se, and more likely to bring a non compos mentis verdict,
these differences and influences became less and less significant.
Despite these changes, the law o f self-murder remained on the
statute books. It was deemed necessary to discourage people from
suicide (it was thought suicide would cause a disintegration o f society)
and offered a means o f condemning men and wom en who had com m it
ted suicide to escape punishment after being found guilty o f an offence.
In such cases, the rituals associated with self-murder were used as a
12
w ay o f stigmatizing the dead for some offence committed previously.

In 1790, The Times reported a story o f a wom an who poisoned 16 inmates
at the Epworth W orkhouse for w hom she was responsible. Four o f
them died, and she afterwards overdosed herself fatally with the arsenic
she had used. The coroner s ju ry was shocked by this crime and not
only brought in a verdict o f felo de se but also ordered her body to be
buried in the public highway with two stakes driven through it.
But this use o f the felo de se verdict for extreme cases o f the punish
ment o f known criminals emphasizes how much the culture had
changed and how difficult and unusual it was becoming to bring in
a similar verdict for suicides w ho had not committed such crimes.
Increasingly the suicide itself was not judged to be the crime. The
coroners’ juries judged the rest o f the person’s life, not the act o f suicide.
The situation in other countries
In the early days o f the colonies in America, most colonies adopted the
traditional English punishments though not all enforced the laws o f
forfeiture. Pennsylvania and Delaware explicitly abolished forfeiture in
1701. After the Revolution, Maryland and N ew Jersey decriminalized
suicide as part o f their constitutions. Thom as Jefferson attempted to
abolish forfeiture and argued against punishing suicide. He also pointed
out that juries disapproved o f such severity and it was therefore sensible
to get rid o f the severe rules altogether.
The enforcement o f laws on suicide in Germany, Spain and Italy is
an unknown quantity. Few records remain. H ow ever, there is some
evidence from France and parts o f Switzerland. French law consisted
o f a collection o f regional customs and codes, and different punishments
were stipulated in different regions o f the country. In some places
penalties were worse than those imposed in England, in others they
were mild. The m ajor difference, however, was that throughout France
the laws against self-murder were applied very infrequently. In 1670
they were standardized in a law decreeing that every convicted suicide
should forfeit his goods and be drawn on a hurdle to a profane grave.
But interestingly, after this standardization, the customs and laws were
13
enforced even less often. By the time it took place, the growth in Enlight
enment philosophies had begun to undermine the case for such punitive
handling o f suicide. In 1770 in France, the Rites o f Desecration were abol
ished. Suicide was decriminalized completely 21 years later in 1791.
Switzerland, or at least the Swiss city states, Geneva and Zurich,
were more rigorous in attempts to enforce regulations against self-
murder. Their Calvinist regimes attempted to impose strict religious
rules and customs. H ow ever, while suicides were often convicted, they
seldom gave rise to severe punishment. As in France, by the time the
authorities tried to reform the punishment o f suicide, opposing trends
had already m oved towards the secularization o f suicide. The last
example in Switzerland o f dragging the corpse o f a suicide victim
through the streets occurred in 1732, and although the old law remained
in force, as it did in France, until the early 1790s, the city officials in
Geneva, for example, declared in 1735 that all suicides w ere insane and
were to be spared punishment.
Ambivalence within the Church
Although the Church had roundly condemned self-murder for most o f

its history, it did not actively participate in the rituals o f desecration. It
did not stand in the w ay either. Many o f these rituals m ay have pre-dated
Christianity, and it was sufficient that the Church did not object to
them. It did not say prayers for the dead and did not permit a Christian
burial in consecrated ground for suicides. O f course, if somebody was
found to have committed suicide because the balance o f their mind was
disturbed, there was nothing to stop Christian ministers from burying in
consecrated ground. H ow ever, even into the sixteenth and seventeenth
centuries many conservative clergymen felt uneasy about such burial
even for someone declared ‘insane\ In the 1662 Book o f Com m on
Prayer, a rubric was added (not there in the 1549 and 1552 versions),
saying that a clergyman could not bury in consecrated ground any ‘who
have laid violent hands upon themselves'. W hoever drafted the rubric
failed to insert the word ‘feloniously’ . The rubric as it stood might
exclude all people who killed themselves.
14
In the light o f such confusion, clergymen looked for a compromise

which they hoped would meet the private need without making a
public statement. They allowed burial to be in churchyards, but buried
such people on the north side o f the church, alongside executed felons,
excommunicates and unbaptized infants. This compromise had one
unfortunate consequence. The north side o f the church gradually came
to be considered by local people as polluted, with the result that by the
late eighteenth century, despite some churchyards running out o f space
to bury bodies, clergymen could not easily persuade families to bury
relatives there. Families would rather dig up the bones o f ancestors to
make space for burial on the south side o f the church. The Church was
often relatively powerless to argue against popular superstition. But
having been partly responsible for fostering such superstitions, it per
haps deserved to become victim to them.
The end of the crime of self-murder
Parliament did not abolish the religious penalties for self-murder until
1823, and did not abolish the secular punishments until 1870. W hy it
took so long to change these laws remains something o f a mystery. The
main reason seems to have been the relation between the law on suicide
and other criminal laws.
The problem o f deterrence was the main issue. It is somewhat
bizarre now to recall that divines such as John W esley called for victims
o f suicide to be gibbeted and for their corpses to be left to rot. In 1823,
however, Parliament acted to repeal the custom o f profane burial. Sir
Jam es Macintosh in the House o f Com m ons declared that punishing
suicide was ‘an act o f brutal folly’. Although the Com m ons defeated
Macintosh’s resolution, as soon as it had failed in its passage through
the House T. B. Lennard introduced a separate Bill to abolish the ritual
penalties for felo de se and substitute the milder sanction o f night-time
burial. He called the Rites o f Desecration ‘an odious and disgusting
cerem ony’ . Such practices were useless as regards the dead and ‘only
tortured the living*. The Bill passed with only one amendment to
make it clear that this change in the law did not mean that clergymen
15
were required to perform the usual rites over the bodies o f suicides.
Instead the services they provided were left to their own devising.
Some conservative clergymen resisted pressure to read the burial
rites, and there is evidence that in the dying years o f the nineteenth
century the controversy am ong the Anglican clergy raged as fiercely
as ever.
The provisions for the forfeiture o f property for self-murder were
repealed in 1870 as part o f an omnibus Bill that covered felonies o f every
kind. Felo de se is mentioned only once in the Bill itself, and suicide was
not discussed in the debates on the Bill, presumably because forfeiture
for suicide was such a rare event and no longer a current topic.
One interesting development which arguably allowed the law on
suicide to change was the gradual emergence o f a common-law crime
o f attempted suicide. In earlier centuries, suicide had been against the
law, but it had not been a crime to try and kill oneself. During the 1700s,
people who ‘attempted suicide' were sent to madhouses, gaols or
workhouses to prevent them killing themselves. H owever, during the
nineteenth century there is evidence that, in London at least, people
increasingly began to be arrested for ‘attempting suicide’ . By 1850 this
crime was recognized by the courts, and by the later part o f the century
the courts all over England were arresting and occasionally trying
people for attempted suicide.
In part this trend arose from the increase in police forces throughout
the country. Previously there would have been very few officials who
could have enforced such a law had it existed. But the rise in prosecutions
for attempted suicide was also an answer to those who complained that
the law o f suicide punished the innocent (the family) and spared the guilty
(the one who committed suicide). After about 1850, the law could be used
as a measure to punish (and perhaps rehabilitate) these would-be suicides.
The criminal status o f suicidal behaviour did not end for the United
Kingdom until 1961, when Parliament repealed the com m on-law felony
o f self-murder. Coroners and physicians subsequently no longer needed
to attest that someone who killed him or herself had done so when ‘the
balance o f the mind was disturbed’ .
16
Concluding remarks
The history o f attitudes to suicide could be interpreted as a journey

from the darkness o f ignorance and belief in the supernatural to the
light o f a m odem and tolerant era. It is immensely more complex,
however. While these attitudes have varied throughout the ages, and
every age has seen a range o f value systems brought to bear for a variety
o f reasons, we have also seen that some in our own day would argue
that suicide is a giving in to sinful despair brought about by the devil,
as witnessed by the leaflet alluded to at the start o f this chapter.
Ambivalent attitudes to suicide and towards people who harm them
selves may partly reflect our failure to understand how someone is
capable o f taking this ultimate step. Part o f us may admire their apparent
control over when their life should end. Another part abhors it, for fear
w e might take the same route (rather like stepping back from the
platform edge for fear we may be seized by the impulse to throw
ourselves under a train).
W hy does the diabolical metaphor survive? One obvious reason is
that a proportion o f society still believes Satan exists and influences
human affairs. Additionally, the ‘devil’ vocabulary m ay persist because,
for those who have been seriously depressed, it can feel as though they
have been taken over by something outside themselves. The devil
metaphor matches and gives shape to this inner feeling, even for those
who would normally reject supernatural explanations o f experience. In
fact, one o f the advantages o f the metaphor is that it locates the
source o f the problem ‘outside’ the individual. In this respect, modem
psychotherapies that emphasize the collaboration between therapist
and client against the ‘common enem y’ o f depression use a secularized
version o f the devil myth to achieve their ends. It is ironic that such
forms o f psychotherapy have shown the most promise in helping people
to deal with suicidal thoughts and behaviour.
17
2
Suicide: Facts and Figures
Suicide is the most individual o f acts, yet it happens too often for us to
ignore the possibility that there m ay be trends within the data when
these individual acts are examined together. Some questions can only
be answered by taking account o f entire populations. Is the number o f
suicides on the increase or decrease - and in which groups o f people?
H ow is the figure related to gender, and to social class? W hat means do
most people use? Are any trends limited to a single country - do suicidal
trends respect national, cultural or ethnic boundaries? H ow is suicide
related to socio-economic conditions? And what about unemployment
and other such sources o f stress - can we say definitively whether they
affect the suicide rate, and if not, w hy not? This chapter aims to address
each question as essential background to later chapters which ask for
explanations.
Durkheim defines suicide as ‘the termination o f an individual's life
resulting directly from a positive or negative act o f the victim him self
which he knows will produce this fatal result'.1 But are such definitions
useful in the real world o f the inquest? Interpreting suicide statistics is
often said to be difficult because they are inherently unreliable ‘Suicide’
is a legal definition, so official statistics inevitably underestimate the
number o f suicides. The authorities have to be certain beyond reason
able doubt that death was caused by self-inflicted injury and that the
deceased intended to kill him or herself One result, for example, is that
passive methods o f dying are less likely to be called suicide than active
methods. Thus drowning is called suicide in 54 per cent o f cases whereas
hanging is called suicide in 98 per cent. Another result o f definition
18
problems is that a verdict o f suicide m ay be less likely when a young

person is involved. A further limitation o f statistics is that m any coun
tries do not report suicide rates. The W orld Health Organization collates
these data, but not all member states return statistics reporting suicide
rates. Even for those that do, w e cannot be sure that the reported rates
reflect the use o f the same criteria to bring in a verdict o f suicide. Surely,
then, the data, even where available, are almost uninterpretable.
Such a conclusion would be too pessimistic, however. W here there
is insufficient evidence to justify the conclusion that a sudden death
was suicide, the alternative is most likely to be an open verdict (‘undeter
mined’). W hen these verdicts are examined, the sex, age and social-class
profile o f the deceased resembles closely the profile for people known
to have committed suicide. Som e researchers have therefore included
such open verdicts in their statistics when trying to estimate the total
number o f people w ho kill themselves. H owever, research has shown
that such trends are affected very little whether the open verdicts are
included or excluded.2 Further evidence for the reliability o f suicide
figures from country to country is that statistics largely agree on which
sub-populations are most at risk for suicide.3
The scope of the problem - UK statistics
In males within the United Kingdom, suicide is the second most

com m on cause o f death in 15-34-year-olds (after motor-vehicle acci
dents).4 For example, in 1990, am ong 15-24-year-olds, 14 per cent o f all
deaths (1 in 6 male deaths and 1 in 8 female deaths) were by suicide.
Between the ages o f 25-34, 17 per cent o f all deaths am ong males were
by suicide. Overall, suicide and undetermined deaths (referring mostly
to open verdicts which parallel suicide statistics very closely) account
for 8.5 per cent (males) and 3.8 per cent (females) o f years o f life lost
before the age o f 65. These figures are similar in most Western countries.
Given these statistics, it is not surprising that governments should be
turning attention to trying to reduce the suicide rate. Their concern is
not only the untold suffering the statistics represent, but also the huge
economic cost o f such loss o f life. It has been calculated that each
19
suicide o f a young person in the United States (aged 15-24 years)

represents $432,000 o f lost economic productivity.5 In total, therefore,
suicides in this age group alone represent 276,000 years o f life lost, and
$2.26 billion loss o f economic productivity. Over 645,000 years o f
productive life are lost each year in the United States as a result o f
deaths by suicide. Figures for Europe are likely to be at least three times
these amounts.
Figure 2.1 shows the number o f suicides in England and W ales since
1911 for men and w om en separately. The data, taken from the Office o f
Population Census and Surveys,6 has been smoothed by averaging each
four-year period. In these data one can see the effect o f the First and
Second W orld Wars (where there was a reduction in suicides) and the
rise in suicides in the Depression between the wars (which peaked, as
in the United States, in 1932). After 1945, the suicide rate rose steadily to
a peak for both men and wom en in the late 1950s and early 1960s
(though the peak was slightly later for wom en, in 1963).
Since that time a number o f quite dramatic changes have taken
place. First, the suicide rate in wom en has fallen continuously. Second,
the suicide rate for men fell until the early 1970s, then started to rise.
This is the first time in the twentieth century that suicide rates for males
and females m oved in opposite directions.
In Figure 2.2, w e see the England and W ales data for five-year age
groups for men and wom en, making use o f the three-year average
death rates. For wom en, the suicide rates for all age groups, which
started to fall after 1953, continue to fall apart from the youngest (15-24
years), which were very low then and have remained low since. Older
wom en are now less likely to commit suicide. The result o f the conver
gence towards low er rates in older wom en is that for w om en over 45
years there is little evidence o f an age-related increase in risk.
Figure 2.2 confirms that since the early 1980s suicide rates for men
aged 45 years and over have fallen. H owever, the rate for men under
45 years has risen considerably, and for the 25-44 age group it now
exceeds the rate for all but the oldest men (those over 85). There has
been a particularly dramatic rise (75 per cent) in suicide rates among
males aged 15-2.4 years since 1982. Although the rate in this age group
is still low er than that o f the older age group, in general the younger
20
o
o
CD
O
O
r<S
IT\
O
O
O
O
O
of
ITN
O
O
O
of
o
o
LTJ
sipeap jo J9quinj\i
o
o
O
M
o
O
in
O
21
Figure 2,1 Number of suicide deaths by gender, 1911-99, England and Wales.
Source : Office of National Statistics.
40
Male
35
Rate per 100,000 population
30
25
20
15
10
35-44 —+ 75-84
45-54 — 85 and over
1983 1985 1987 1989 1991 1993 1995
Year
Rate per 100,000 population
Year
Figure 2.2 Recorded suicide deaths: rate per million population by

gender and age, 1983-95, England and Wales.
Source: Office of National Statistics
22
and older adult rates for suicides have been converging over the past 30
or 40 years. Currently over 40 per cent o f all suicides occur in males
between the ages o f 15-44 years, and this differs from previous times
when the rates were much higher for older people. The situation by
the end o f the 1990s was that for men between 25 and 75 there was little
difference in age-related suicide risk.
W hy this steep increase in young male suicides? One clue lies in the
changing pattern o f marriage and divorce. Divorced men have higher
rates o f suicide, so if the divorce rate rises, the suicide rate will rise
with it.
International suicide rates
Table 2.1 shows the suicide rates for several different countries reporting
to the W orld Health Organization. The rates for many different coun
tries follow the patterns seen in the United Kingdom in many respects.
In every country but China, the suicide rate is low er for wom en than
for men. These statistics do not show the trends over time, which differ
between men and wom en. In many European countries, for example,
the suicide rate in wom en has remained flat or decreased while the rate
for men has increased, most markedly in Ireland, Spain, Netherlands,
N orw ay, Belgium, France, Luxem bourg, Denmark, Finland and North
ern Ireland. A notable exception to international trends is in West
Germany, where there has been a substantial decline in both male and
female suicide.
Are the trends within the suicide statistics from North America
similar to European trends in general and U K trends in particular?7
During the past century, the rate in the United States has averaged a
rate rather similar to that in the United Kingdom: 12.5/100,000. A high
rate o f 17.4 was observed during the Depression in 1932 and a low point
o f 9.8 in 1957. This rose again steadily to a peak o f 13.3 in 1977, then
dipped to a rate o f 12.7 in 1987. The US data since 1950 are shown in
Figure 2.3. These rates are also very similar to those o f Canada, where
the 1995 rate was 13.5 per 100,000 (21.5 per 100,000 in men and 5.4 per
100,000 in women).
23
Table 2.1 Suicide rates for a sample of countries reporting to the

World Health Organization (most recent year data available in
brackets)
All Ages
Male Female Male and

Female average
Southern Greece (1996) 5-7 1.2 345

Europe Italy (1993) 12.7 4 -0 9-3
Portugal (1996) 10.3 3.1 6.7
Spain (1995) 12.5 3-7 8.1
Western Austria (1997) 30.0 10.0 20.0

European Belgium (1992) 26.7 I I .0 18.85
France (1995) 30.4 10.8 20.6
Germany (1997) 22.1 8.1 15.1
The Netherlands (1995) 13.1 6.5 9.8
Switzerland (1994) 30.9 12.2 21.55
Scandinavia Denmark (1996) 24.3 9.8 17.05

Finland (1996) 38.7 10.7 24.7
Norway (1995) 19.1 6.2 12.65
Sweden (1996) 20.0 8.5 14.25
Old World United Kingdom (1997) 11.0 3.2 7I

Ireland (1995) 17-9 4.6 11.25
Asia Japan (1994) 17-3 7-2 12.2

Singapore (1994) 13.1 8.5 10.7
Hong Kong (1987) 11.4 91 10.3
China (1994) 24.1 27.8 25.8
Korea (1994) 12.6 55 8.8
Thailand (1994) 54 2.3 3.8
New World Australia (1995) 19.0 51 12.05

Canada (1995) 21.5 5-4 1345
N ew Zealand (1994) 23.6 5.8 14.7
U SA (1996) 193 4-4 11.85
24
There have been changes in age distribution in the United States

over this period. Similar to the United Kingdom, over the 1970s and
1980s the percentage o f deaths by suicide by persons 15-24 years has
increased, whereas the percentage o f total suicide deaths by people over
45 years has decreased. In 1970 the median age o f people who died by
suicide was 47.2; by 1980 it was 39.9. This pattern is also evident in
Canada, where the 1992 rate for 20-29-year-old men (29 per 100,000)
exceeds that for older men (25, 26, 20 and 28 per 100,000 for the age
groups 40-49, 50-59, 60-69 and 70-79 years respectively).
Methods of suicide
Figure 2.4 shows the distribution o f recorded suicides by sex, year o f

death and method for England and W ales between the late 1940s and
1990. In the United Kingdom, the methods by which people take their
own lives reflect the availability o f different means from those used in
the United States, with many few er firearms being used. In 1970, the
leading cause o f suicide death was poisoning (31 per cent o f males and
65 per cent o f females), hanging (23 per cent o f males, 9 per cent o f
females). As can be seen in Figure 2.4, how ever, there were several
changes between 1970 and 1990. In 1990, 35 per cent o f male suicides
were due to ‘gases and other vapours’ (the vast m ajority o f these were
vehicle exhaust deaths). Thirty-one per cent o f men committed suicide
by hanging, and only 14 per cent died by poison. In wom en, there was
a decrease between 1970 and 1990 in the numbers using poisoning (from
65 per cent to 44 per cent) and domestic gas (reduced from 9 per cent
to nil as the non-toxic natural gas replaced coal gas). Instead, there was
a switch to hanging (a rise from 9 per cent to 23 per cent) and vehicle
exhaust (a rise from 1 per cent to 13 per cent). These trends suggest that
while it used to be true that w om en used less lethal methods than men
in their suicides, this difference is reducing.
More recent statistics show that during the 1990s there were further
changes. The most significant was a drop in car exhaust deaths, falling
to 23 per cent o f male deaths in the mid 1990s, and to 16 per cent in
1998. H ow ever, the proportion o f men w ho committed suicide by
25
26
Figure 2.3 Comparison of suicide rates for men and women for different ethnic groups in the United States,
1950-96.
Source: The National Center for Health Statistics, Vital Statistics of the United States
Figure 2.4 Distribution of recorded suicides by gender, year of death and method, England and Wales.
0
}00
00
ir\
00
C1O
00
0
00
00
r\
m rC
1
r\
o
7'
00
VO
Vin
O
cu , ////\
<
1
Jo
VO
wmk .................... .............

s O
rXVv^-rv;--•- ;
I 1 - 00
in
in
F* 'T^g"xr-J in
'm m
o
V
00
a
03
0 £
t00
00
in
00
1
m
00
Source : The Office of National Statistics
0
00
1
00
r\
1n
£
o
c V'
<u 'm m m 00
'O
in
V
T
O
------------- ---------- ro
VO
, mu,’ v . ; -!— . >.t , ’ K • ■» ; j " 7 -
O
^ ,/v -^ . " ^ 5 x -* . •*
a____________ _ __________ 00
in
7*^7
o
______________________I 0T0
a
o o o o o o o o
o
<s
00 V
O (S o 00 o (S
uopEjndod u o i j j i u i j^d aiE'y
27
hanging rose from 31 per cent in 1990 to 52 per cent in 1998. In addition,
the increase in hanging continued for wom en too, with 31 per cent
using this method in 1998. (Note that these proportions are higher
than estimates based on combining suicides with open verdicts, where
proportions for hanging o f 31 and 15 per cent for men and wom en
respectively were reported for 1996 to 1998. This difference is due to
the fact that o f all methods o f death, hanging is the most likely to result
in a verdict o f suicide, whereas other methods, such as self-poisoning
or drowning often end in an open verdict. The data for 1998 given here
are for suicides alone, since the data in Figure 2.4 are based on verdicts
o f suicides only.)
Table 2.2 shows the distribution o f methods included in the ‘other’
category. As can be seen the largest proportion o f other methods is
drowning. The table also shows that there has been a recent decrease
in the use o f firearms for both men and wom en. H owever, jum ping
from a height or in front o f m oving vehicles remains a relatively high
category o f ‘other methods’ . Over the last 45 years there has been a
reduction in the proportion o f drowning for both men and wom en.
The apparent increase during the mid 1900s for men m ay have been
temporary, since latest figures (for 1998) show a reduced proportion o f
12 per cent.
When methods o f suicide are examined for the US population, there
are clear differences from the United Kingdom. Males still choose more
violent means than females, but use many m ore firearms (64 per cent).
As in the United Kingdom, females are more likely to choose non-violent
means, e.g. drug overdose (38 per cent). (See Figure 2.5.)
Factors affecting suicide rate
Gender
In almost all countries, the rates for men exceed those for wom en by a
factor o f between 2:1 (Denmark, Netherlands and Sweden) and 5:1
(Finland).8 This sex ratio has existed from the earliest days. Historical
data show that men have committed suicide at least twice as often as
wom en in almost all samples o f data that survive. For example, the sex
28
Table 2,2 Distribution of'other methods', England and Wales (from

Charlton, Kelly, Dunnell, Evans, Jenkins and Wallis, Trends in Suicide
Deaths in England and Wales', Population Trends, 1992): S. Kelly and
J. Bunting, 'Trends in Suicide in England and Wales, 1982-1990',
Population Trends, 1998. Source : The Office of National Statistics
Method 1948-50 1988-90 1994 -
Men
Drowning 35 14 20
Firearms 21 27 16
Cutting 20 10 8
Jumping 9 16 22
Other 15 33 34
Women
Drowning 60 33 33
Firearms 3 5 1
Cutting 10 9 4
Jumping 15 23 19
Other 13 30 43
ratio in the King’s Bench data for 1485-17149 was 1.8 men for every 1
wom an; for Norwich in the eighteenth century it was 2.3:1; and for
greater London it was 2:1. Completed suicide seems always to have
been an act more associated with men than women.
As we have seen in the data for England and Wales (Figure 2.1), the
pattern during most o f the twentieth century, in which wom en and
m en’s suicide rates followed each other fairly closely (albeit at different
overall levels), has now changed. The rates for wom en fell during the
1970s and 1980s, while those for men rose. The net effect was to reduce
the overall figures for wom en over this period, but to leave the rates
for men (taking all ages together) rising steadily.
29
Suicide method
Q
3
u
toJO
C3 p
u
C rP
*-* C
O OJ
T3 C/)
+-»
<L) <v
C/3
C/3 C/3
30
X
g
c03
'So
W)
£
6
it
C/)
03
o
o
o
o
O
o
<s
CO
a
c
ctJ
v
<
u
4->
uu
<u
'bJO
Figure 2.5 Number and percentages of suicide method by gender, US rate 1996.
Source: The National Center for Health Statistics, Vital Statistics of the United States
Ethnic group
The suicide rate for different ethnic groupings is different: whites in the
United States are approximately twice as likely to commit suicide as
non-whites (see Figure 2.3). This difference is particularly pronounced
for the older age group. In the United States, native American young
people have a very high rate o f suicide, the highest rate being in those
tribes undergoing the greatest and fastest cultural assimilation. In the
United Kingdom, there have been reports that the suicide rate among
Asian w om en is almost double that for other wom en o f the same age,
and although research has not yet established the reason, difficulties in
resolving cultural clashes is a clear possibility.
Sexual orientation
A 1989 report from the US Department o f Health and Human Services,
the Secretary's Task Force Report on Youth Suicide, suggested that gay
youth were tw o or three times more likely to complete suicide than
other young people.10 It suggested that lesbian and gay young people
m ay comprise up to 30 per cent o f youth suicides annually. However,
this part o f the report was not backed by any evidence. Elsewhere it
admitted that this was a neglected area, with virtually no research on
youth suicide having taken sexual orientation into account. The two
studies that have been published (both US surveys) suggest much more
conservative figures: that between 2.5 and 5 per cent o f suicides are gay,
and that these figures are not significantly different from the proportions
one would expect by chance. One study11 surveyed 283 adolescent and
adult suicides, finding that 5 per cent were gay. Another study, using
psychological autopsy techniques, studied 120 consecutive suicides
under the age o f 20 years, and 147 community age-, sex- and ethnic-
matched controls living in the Greater N ew York City area.12 Three
(3.5 per cent) o f the 95 teenage suicides in this sample (2.5 per cent o f
the total sample o f 120) were reported to have had a homosexual
experience. Both studies concluded that psychiatric disturbance and/or
substance abuse were critical predisposing factors regardless o f sexual
orientation.13
31
Religious affiliation
Rates vary according to religious affiliations. A number o f studies have
found that Protestants are more likely to commit suicide than Catholics
or Jew s. The Catholic countries o f Spain, Italy, Portugal, Greece, Ireland
and Poland have fairly low suicide rates. Yet there are other Catholic
parts o f Europe, such as France, where the suicide rate is relatively
high, though the increased secularization o f France, and reduced church-
going, may be responsible. In one study in N ew York, rates for Prot
estants per 100,000 were 31.4, those for Catholics 10.9 and those for Jew s
15.5. H owever, many studies have also found that churchgoing in
general is associated with low er suicide risk, presumably because o f the
social network and sense o f meaning in times o f suffering it can som e
times provide.
Social class
Table 2.3 shows the proportional mortality rates calculated for different
social classes in England and Wales. Social class I is defined as pro
fessional; II as intermediate (lower professional and executive); III as
skilled (manual and non-manual); IV as partly skilled; V as unskilled.
The proportional mortality ratio (PM R) is a ratio which enables the
impact o f a disease upon an exposed population to be examined.
Table 2.3 shows that the risk o f suicide is greater in social class I and
social class V, with the highest risk o f all carried by those with no
occupation at the time o f the census.
Table 2 3 Deaths from suicide among men (aged 16-64) in different

social classes, England and Wales, 1979-90
Social Class PMR Observed deaths
I 116 1 319
II 99 5 202
IIIN 102 3 083
HIM 87 9 004
IV 100 5 649
V hi 3 582
‘Unoccupied’ 126 3885
32
The pattern whereby people o f higher socio-economic status show

as high a suicide rate as those o f low status may be a feature o f the
United Kingdom not found elsewhere. Studies in the United States
and Canada have found a strong association between differences in
socio-economic status between communities (e.g. in proportion o f
substandard housing) and the suicide rate in that community. The
poorer the housing, the higher the rate. The exception to this trend is
the poor areas occupied by a high proportion o f African-Americans,
w ho have a lower suicide rate, most usually attributed to the strong
social bonds within and between families.
Unemployment
Economic conditions have been studied many times, with some conflict
ing results. Studies on unemployment in suicide have found that men
who are unemployed and seeking w ork are at two or three times greater
risk o f suicide death than the average. Suicide is thought by many investi
gators to be one o f the possible consequences o f unemployment, and yet,
in the United Kingdom at least, there seems little relationship between
change in unemployment levels and the suicide rate over tim e.14
National unemployment rates in the United Kingdom increased
substantially at the end o f the 1970s, stayed high until the mid 1980s,
started to decline until the late 1980s and then began to rise again. In
the 1981-90 period male suicide rates were at their highest at the time
when unemployment rates were lowest. H owever, the relationship is
a complicated one. Most unemployment is short term (under six
months) and unemployment totals are net figures taken at a fixed point
in time. This means that unemployment rates never give the whole
picture o f the distress caused by unemployment. Stephen Platt has
pointed out that the impact o f unemployment depends on whether an
entire firm has been made unemployed in a community (in which case
there may be some increase in social support - an ameliorating factor
in suicide risks). By contrast, if an individual is made unemployed,
independent o f other people becoming unemployed, then this may be
more o f a stress factor.
In fact, other studies have found that the United Kingdom is an
exception in not finding a connection between unemployment and
33
suicide. Boor (in 1980)15 compared the unemployment rates with the
suicide rates in the following eight countries: United States, Canada,
Sweden, France, Great Britain, Germ any, Italy and Japan. The suicide
and unemployment rate correlated positively in all but two countries,
Italy and Great Britain. Boor also found that the relationship between
unemployment and suicide held for both men and wom en. The rate o f
unemployment among suicide victims is 50 per cent across a number
o f studies.
Rural areas
Geographical region also plays a part in different suicide rates across
countries. As in the United Kingdom, where rural areas have the highest
rates, so in the United States the mountain regions have the highest
overall rate, Nevada ranking highest at 17.6 per 100,000. In a study
with m y colleagues Leslie Pollock, John Hollis and Patrick Vesey, we
examined the suicide rates in Powys, Wales, one o f the least densely
populated areas o f Europe. W e found a suicide rate greatly in excess o f
the U K average (around 20 per 100,000). Looking closely at the data for
a three-year period, w e found that farmers or farm labourers accounted
for 22 per cent o f this figure, though they only represented 11 per cent
o f the population. W e also examined whether it was the lack o f access
to primary care services in rural areas that determined the high rates,
but could find no evidence that these individuals had visited their
doctor any less frequently than happens in more urban areas. Rather, it
appeared to be the greater availability o f lethal weapons that accounted
for the high rates, with firearms used more com m only than the national
average. A study by Keith Hawton and colleagues reinforces this con
clusion.16 Between 1981 and 1993 there were 719 suicides or open verdicts
o f farmers in England and Wales. Firearms were used in 40 per cent o f
cases (compared with the national average o f 4.3 per cent), and hanging
in 30 per cent (compared with the national average o f 23 per cent).
Hawton pointed out how the easy availability o f such lethal means was
a critical factor in the high suicide rate within this community.
34
Imprisonment
The number o f prison suicides a year has risen considerably since the
early 1960s. In 1961 there were 15 suicides in the United Kingdom, and
this figure had increased to 67 per year by 1997. Most o f these are in the
15-44 age group o f men. This increase represents 4 per cent o f the
increase in the young male suicide rate over this period. The increase
has given cause for concern, especially since most o f the suicides o f the
under-45 age group are male prisoners on remand, not yet found guilty
o f any crime. D ooley examined 300 prison deaths between 1972 and
198717 and analysed the reasons for each suicide. Forty per cent o f
suicides could be attributable to the prison environment (overcrowding,
etc.). Fifteen per cent were attributable to outside pressures, 12 per cent
to guilt feelings for the offence and 22 per cent to already diagnosed
mental illness.
Medical illness
There are several medical conditions that increase the risk o f suicide.
For example, the risk o f suicide among those with epilepsy is four times
that o f normal controls, and for temporal-lobe epilepsy it is 25 times
greater than controls. Cancer has been associated with increased risk o f
suicide, the risk highest immediately following diagnosis, and in those
receiving chemotherapy. (Those with a severe and morbid fear o f
cancer are also at increased risk.) People with peptic ulceration have a
high risk, probably because o f the prevalence o f alcoholism as a cause
o f ulceration. With Huntington’s chorea there is a sixfold increase in
the risk o f suicide compared with the general population, both in those
suffering from the disease and in family members who may also be at
risk but as yet have no overt symptoms. People undergoing renal
dialysis have been said to have a greater incidence o f suicide compared
with the general population, but there are not enough reliable studies
to estimate the risk exactly. There is an increased risk in people who
have spinal-cord injuries (a fivefold increase compared with the general
population), and in those who have multiple sclerosis (a threefold
increase for men and a twofold increase for women). The risk o f suicide
in those who have Aids was some 36 times greater than the general
35
population in the 1980s, but this figure appeared to reduce during the
1980s to a sevenfold risk.18
In many cases the co-occurrence o f depression with physical illness
combines to increase the suicide risk. Medical illness can precipitate
severe depression and also produce an organic mental disorder leading
to perceptual, cognitive and m ood changes. In other cases, the suicide
appears to be a rational act where the person decides that the prospect o f
suffering and loss o f dignity is intolerable. H owever, research evidence
suggests that suicide in the physically ill very rarely occurs in the absence
o f psychiatric disorder.
It is important, therefore, for the physician and other health pro
fessionals not to ignore associated psychiatric problems that co-occur
with physical illness. For example, depression associated with terminal
cancer is often ignored or thought to be ‘normal’ . It therefore goes
untreated despite the fact that psychological treatment is often very
helpful in alleviating the depression. Depression in such patients goes
beyond sadness and includes a sense o f worthlessness and failure which
is not justified on the basis o f the person’s experience.19
Lifespan issues
In all countries, suicide is extremely rare in children under 12 years, but

it becomes steadily more common after puberty, and the risk then
increases with age. For older people to be more at risk o f suicide appears
to be the predominant pattern across all countries and cultures, though
there is some variation in whether the increase is linear, or has peaks
and dips across the lifespan. The highest rates are found for elderly men
(over 75 years) in almost all countries, but in many (especially for
Scandinavian countries) the peak for wom en comes earlier. Is suicide
the same phenomenon in young adults as in older adults?
One factor explaining lifespan differences would be the increase in
substance abuse which has occurred to a much greater extent in younger
members o f the population and in young males in particular. Further
more, there is some evidence that the most vulnerable young men are
those from working-class backgrounds w ho have fewest employment
36
opportunities. These are just the groups that have suffered the greatest
rise in suicide rate over the past 20 or 30 years.
Different rates o f medical illness across the lifespan might also contrib
ute to differences between different ages. Illness is associated with 50
per cent o f adult completed suicides, contributing more prominently
for older adults. Medical illness in older adults contributes to suicidal
risk in a number o f possible ways. Depression m ay be worsened by the
fact that the physical illness puts an increased burden on the fewer
social supports that exist. There are also cognitive impairments often
associated with ageing in general and with physical illness in the aged
in particular. Finally, the use o f certain medications to alleviate the
physical illness sometimes lowers m ood and impairs judgement.
Do different psychiatric disorders increase the risk o f suicide at differ
ent stages in life? W e shall see in the next chapter that the individuals
most at risk are those suffering from depression and schizophrenia.
Right across the lifespan, if these conditions co-occur with particular
personality traits such as impulsivity and aggression, and with alcohol
and substance abuse, the risk is increased. H owever, there are changes
over the lifespan in the particular combinations o f these disorders.
Adolescents who commit suicide tend to show a combination o f
depression and conduct disorders, though the depression is often only
recognized after the event, as research by Alan Apter and Israel Orbach
shows.20
It is possible that it is psychiatric disorder per se which is the important
issue. Different psychiatric disorders m ay appear to raise the suicide
risk at different points in the lifespan simply because they are more
prevalent at that point. Some have argued that it is changes in the action
o f the major neurotransmitters over the lifespan which are implicated.
For example, brain serotonin metabolism is believed to change during
the adult lifespan. This remains speculative.
Alcohol and substance abuse

Alcohol and substance abuse has been on the increase since the 1970s.
One indication o f the increase is the number o f people each year who
die from alcohol- or drug-related deaths (other than suicide). The data
suggest that, since 1968, there has been a sixfold increase in drug-related
37
deaths am ong men aged 15-24, and a fivefold increase for males aged
25-44. There are similar trends am ong wom en, but the rates are
generally much lower. The data for alcohol show a similar pattern,
with men aged 25-44 showing the largest increase in rates (more than
fourfold between 1968 and 1990) followed by men aged 15-24. Once
again, rates for younger w om en rose as well, but the levels were very
low compared with those for men.
Alcohol and substance abuse represent major risk factors for suicide
right across the lifespan. The prevalence o f alcohol and drug abuse
increases steadily from 15 until about 45, then declines. The number o f
years somebody who completes suicide has typically been abusing
alcohol is between 20 and 25, possibly because, as alcoholism progresses,
it destroys those factors known to protect against suicide. First, it
destroys social supports because o f the alcoholism-related guilt and
anger. Second, it destroys intellectual function through brain damage
and brings about reduced health and increased incidence o f medical
complications. Third, chronic alcoholism reduces personal control and
increases helplessness. Since, however, 60 to 70 per cent o f patients with
a diagnosis o f alcohol problems have additional psychiatric diagnoses, it
is possible that the substance abuse exacerbates the course o f any
psychiatric illness. George M urphy’s studies show that alcohol and drug
abuse are most lethal when they occur alongside an affective disorder
(such as depression).21
H ow ever, alcohol and substance abuse raises the risk o f suicide as
soon as the abuse starts. Substance abuse has been diagnosed in over
one third o f young people committing suicide.22
The rise in young male suicide
In the 1990s the United States found itself with one o f the highest suicide
rates for young men in the world, even exceeding Japan and Sweden.
The question o f w hy there had been such an increase in youth suicide
became very pressing indeed. It appeared to parallel the increase in risk
factors associated with suicide in young people: depression, conduct
disorders and substance misuse. In addition there are population effects
38
such as that the proportion o f youth in society is high, implying

increased competition for opportunities ( jobs and education). This is
consistent with the finding that there was also a high youth suicide rate
at the beginning o f the twentieth century when young people also
comprised a higher proportion o f the population. Other factors blamed
have been the increased divorce rate, increased geographical mobility
(with its consequent loss o f important attachments figures), changes in
family structure and decreased religious affiliation. H owever, some o f
the evidence for these remains patchy. The rise in the number o f
younger males committing suicide in the last few years has been almost
universal.
There are two classes o f explanation. The first type sees the rise as
part o f an unexplained change in rates within a group o f people bom
within a few years o f each other, a cohort effect. The second type lays
the blame at the door o f prevailing social conditions, employment
opportunities, drug and alcohol abuse, disempowerment. That is, it
sees the change in rates as a period effect. Clearly there can be interactions
between cohort and period effects, as where a certain cohort is exposed
to increase in availability o f some substance (e.g. illicit drugs). There is
evidence o f higher rates o f depression, bipolar m ood disorders and
substance abuse for the generation bom just after the w ar (the 'baby
boom 5 generation), which may explain some o f the changes in the
suicide rate.
A cohort effect?
Studies by Murphy and W etzel for the United States, by Solomon and
Hellon in Alberta, Canada, and by Goldney and Katsikitis for Australia
followed through the suicide rates for people bom within a few years
o f each other - each ‘cohort's’ suicide rate could then be compared
with cohorts bom at other times.23 Each study found that if a cohort
had a high suicide rate at a young age, then the increased rate continued
throughout the life o f that cohort. Is this also true o f England and
Wales?
Cohort data published by the Office o f Population Census and
Surveys in 199224 confirm the existence o f a cohort effect for both men
and wom en for the suicide rate in England and Wales. For men, more
39
recent cohorts have higher age-for-age mortality than earlier ones. That
is people bom between 1962 and 1971 have a higher risk o f suicide than
those bom between 1952 and 1961, who in turn have a higher risk o f
suicide than those bom between 1942 and 1951. The rates for the more
recent cohorts may not yet have peaked, in which case we may expect
a continued rise in suicide rates by younger men for some years to
come (as the high-risk cohorts m ove through the different age ranges).
Recent data show that the cohort bom around 1971 has a still higher
rate.
For wom en over 45 years, successive cohorts show lower age-for-age
suicide risk, people bom in the 1910s, 1920s, 1930s and 1940s showing
gradually reducing rates. If this trend continues, we m ay expect that
rates for wom en will remain low.
But how could a cohort effect come about? Some studies have
pointed to the size o f a birth cohort. Where there is a large number o f
babies bom during the same period, there is increased competition for
resources (as occurred at the turn o f the twentieth century). This
results in higher unemployment and decreased access to educational
opportunities for that cohort.
A period effect?
It is also possible that the recent trend for increases in young male
suicides does not represent a cohort effect (i.e. people bom recently
m ay not have a greater disposition to suicide), but might rather be a
period effect (something in current society is pushing up the rates).
Taking the data for the 1936 cohort for wom en, when they were 25-29,
their curve was rising sharply at this time, but it turned out to result
from the 1960s period effect. On that occasion, the same period effect
was evident for older age groups too.
Although the recent increase in young male suicides appears not to
be reflected in older males, it may nevertheless be a period effect
combining with a cohort effect. W hatever current stress factors there
may be, they may be differentially affecting young males. This is
increasingly likely the larger the size o f the birth cohort. Throughout
their lives, more people in the larger cohort will be chasing whatever
resources are available. The effect o f the cohort size will depend upon
40
the amount o f resources. A large birth cohort will not always have this
effect, if, by chance, there are sufficient resources to sustain members
o f the cohort throughout their lives.
H owever, if the number o f jobs in a country falls during a recession,
then the effects are likely to fall disproportionately on those in the
population who are members o f the large cohort, whatever age the
cohort has reached. Suddenly too many people o f about the same age
will be chasing too few resources.25 This is consistent with the sugges
tion that the reduced status o f men caused by changes in employment
prospects and family cohesion, together with reduced anticipation o f
long-term roles as husband and parent, has produced a rise o f alcohol
and drug abuse in this subpopulation that contributes to the increased
suicide rate. Is there evidence to support this particular causal pathway?
Decreased family cohesion

Charlton et al estimate that about half the increase in the number o f
men committing suicide is likely to be caused by the increase in the
number o f young men who are single or divorced .26 They estimate this
by comparing the suicide rate for men aged 15-44 in 1972-4 (11.5 per
100.000) with the rate prevailing by the end o f the 1980s (18.6 per
100.000). Using the marital status distribution in suicide rates, they
calculated what the suicide rate in the late 1980s would have been if
there had been no change in the proportions o f men in different marital
groups. They calculate that it would have been 14.9 per 100,000; hence
their conclusion that about half the increase in rates o f suicide in young
men may be attributed to the smaller proportion who are married. This
means we must seek additional reasons for the increase. A major piece
o f evidence about exactly which young people are becoming most
vulnerable emerges from data o f Norman Kreitman and his colleagues
at Edinburgh, who found that the increase in young male suicides was
explained by an increase in the more working-class populations.27
These are the populations which have been most exposed to reduced
employment opportunities and increased misuse o f alcohol and other
drugs.
41
Changes in rates of depression

Part o f the rise in young male suicide m ay also reflect changes in
age-related depression over the twentieth century. Data from the Cross
National Collaborative Group28 show that people bom more recently
(since 1955) have the highest rates o f depression across several countries.
People are becoming depressed at younger ages. Further analyses o f
these data suggest this change is especially true o f males. The cause o f
the change is unknown, but researchers into depression have come to
similar conclusions as those investigating suicidal behaviour: that social
and demographic changes such as the changing structure o f families,
social mobility and increased isolation, disengagement from the com
munity and limited access to resources combine to explain the pattern
o f change. Both changes in rates o f depression, and changes in rates o f
suicide, have occurred in parallel with important economic changes
and changes in family structure over the past 20 or 30 years.
Older adults' reduction in suicide: period or cohort effect?

Finally, could the trend towards low er rates in older people since the
early 1960s be attributed to a period effect: has society found a w ay to
protect older people from suicide risk during this period? Such a claim
is open to considerable doubt. The current rates for older people o f
both sexes are no low er than those for younger people: they are simply
low er than the very high rates seen earlier in the twentieth century for
older people. So the currently ‘reduced’ rates are largely attributable to
the fact that the cohorts bom since 1910 have had more ‘average’ suicide
rates throughout their lives (despite the cohort peaks) compared to the
very high rates o f those bom in the period 1870-1900 (i.e. those who
contribute to the older age data in the suicide figures o f 1946 onwards).
It is not that we have done something to reduce the suicide rates for
old people, but that we are seeing low-risk cohorts reaching old age.
The very high rates seen among some very old people (e.g. over 85s)
that remain despite the fall for other old people, are simply due to the
fact that they belong to the high-risk cohort bom early in the twentieth
century (1905 to 1915).
42
Concluding remarks
T o return to the questions posed at the beginning o f this chapter and

summarize what the statistics tell us: is the number o f suicides on the
increase or decrease - and in which groups o f people? It seems they are
on the increase in most countries, but in men rather than wom en. This
is particularly true o f younger men, among whom suicide rates are
rising. By contrast, the suicide rate is coming down in older people, a
trend likely to be explained by the fact that a low-risk (or average-risk)
cohort is reaching old age.
H ow is the trend related to gender, and to social class? The data o f
interest here show that the rise in young males is almost wholly caused
by a rise among working-class men, those most vulnerable to economic
hardship in a recession, and most likely to abuse alcohol and drugs.
Many studies show a bias, such that low er socio-economic groups are
more vulnerable. Other studies (UK) show a U-shaped curve, with
social class I more likely to commit suicide than middle-class people,
with an increased risk in working-class people. The increase in young
male suicides, how ever, is due to working-class men committing
suicide more frequently, apparently a combination o f cohort and period
effects.
W hat methods do most people use? W e have seen that men use
more lethal methods than wom en, though the difference is reducing.
Do suicidal trends respect national, cultural or ethnic boundaries? While
the methods used from country to country m ay vary (firearms being
much more common in the United States, for example), the trends
across virtually all W estern nations are remarkably similar. Within
each country, however, there are ethnic differences to be taken into
account. African-Americans are much less likely to commit suicide
than Anglo-Americans, but native Americans have a higher rate than
either.
Finally, do unemployment and other such stress factors affect the
suicide rate. Studies across a number o f countries show the United
Kingdom to be an exception to the general rule that changes in
unemployment rate are associated with changes in suicide rate. Even
43
in the United Kingdom, however, those who commit suicide are very
much more likely to have been unemployed at the time than a person
matched for age.
44
3
Psychiatric and Social. Factors
in Suicide
‘Suicide has been so closely associated with insanity only by arbitrarily

restricting the meaning o f the w ords/ So wrote Emile Durkheim in
rejecting a psychiatric interpretation in his study o f social causes o f
suicide in 1897.1 Some hundred years later Susan Blumenthal, C hief o f
the Behavioral Medicine Program at the US National Institute o f Mental
Health, said: ‘Over 90 percent o f patients committing suicide have a
psychiatric disorder/ These authors appear to contradict each other
completely about the significance o f psychiatric disorders in suicide. Is
the difference explained by the different eras in which they wrote? Do
w e have more data now that show Durkheim ’s analysis to have been
wrong? That is possible, though it is worth bearing in mind that many
prior to Durkheim concluded that ‘insanity’ was the most probable
cause. At the end o f the seventeenth century, to say someone committed
suicide because they were ‘insane’ was one o f the few ways o f being
non-censorious, o f showing gentleness to the m em ory o f the victim
and care for the family.
Durkheim found an alternative w ay o f being non-censorious. Suicide
should be seen as a feature o f society, not o f the individual. He accepted
that ‘insanity’ was sometimes involved as a cause o f suicide, but rejected
the argument that understanding psychiatric disturbance was the key
to understanding suicide. Instead, he pointed to a number o f social facts
about the incidence o f suicide and its correlation with social integration
and disintegration. Most famously, he pointed to anomie - the state
associated with detachment from society - as a central feature o f many
suicides.
45
Many studies looking for a link between suicide and psychiatric

disturbance have been done since Durkheim wrote his treatise, and I
shall examine them and conclude that Blumenthal's proposition is
correct, though this does not entail that Durkheim was wrong. Even if
it proved to be that psychiatric disturbance precedes suicide in over 90
per cent o f cases, its causal status would remain unclear. Mental illness
may precede suicide because some circumstances drive people both to
psychiatric breakdown and to suicide.
Evidence for psychiatric illness as a cause of suicide
Researchers have come to the conclusion that suicide is closely associ

ated with psychiatric illness by using two methods. In the first, the
researcher follows up people who have been psychiatrically disturbed
in the past to see what proportion eventually die by suicide. Since the
risk o f suicide in the population as a whole is between 1 and 2 per cent,
it can then be calculated how much greater risk this particular group
carries. The second method is known as psychological autopsy. This
involves careful interviewing o f relatives and friends following a suicide,
and asking questions about the m ood and behaviour o f the deceased in
the period leading up to the death. The results from this method have
been used to support the conclusion that the m ajority o f suicides were
suffering psychiatric problems beforehand.
Research by Brian Barraclough in the United Kingdom has shown
that 70 per cent o f suicides would have been diagnosed as suffering
from major depression, 15 per cent from alcoholism and 4 per cent from
schizophrenia or schizoaffective disorder (though other studies have
found that a much greater proportion o f suicide victims, up to 30 per
cent, have suffered from schizophrenia in the past).2 The total pro
portion o f his sample o f a hundred who would have attracted a psychi
atric diagnosis was 93 per cent. He also showed that those who complete
suicide are ten times more likely to have attempted suicide than other
depressed patients.
Taking these two methods o f assessment, Table 3.1 represents the
current best available evidence on the likelihood o f completed suicide
46
Psychiatric and Social Factors in Suicide
in groups having different psychiatric diagnoses. The results agree with

Barraclough’s findings that the three diagnoses carrying the highest risk
are depression, alcoholism and schizophrenia.
Depression as a normal m ood state is a com m on experience. In minor
depressive states, the person ruminates on negative themes. He or she
feels resentful, irritable or angry much o f the time, feeling sorry for
themselves and constantly needing reassurance from those around
them. Often they have a variety o f physical complaints that do not
seem to be caused by any physical illness.
As depression deepens, more symptoms become apparent. These
include further emotional changes (feelings o f extreme sadness and
hopelessness); changes in the content or process o f thinking (low self
esteem, guilt, m em ory and concentration difficulties); changes in
behaviour and motivation (feeling agitated or slowed down, reduced
interest in social or recreational activities); and bodily changes (sleep,
eating and sexual problems, loss o f energy). If the depression is intense
enough to include five or more o f these symptoms for more than a
two-week period, it is called ‘m ajor’ or sometimes ‘clinical’ depression.
T w elve per cent o f men and 20 per cent o f wom en will experience an
episode o f such m ajor depression at some time in their lives. At any
one time, around 5 per cent o f the population is suffering depression o f
this severity. Twenty-five per cent o f these episodes o f depression last
less than a month; a further 50 per cent recover in less than three
months. H ow ever, the depression can develop into a longer-term
problem, with around a quarter being seriously depressed one year
after symptom onset and one-fifth remaining depressed two years later.
The lifetime risk o f suicide in major depression severe enough to require
hospitalization is between 10 and 15 per cent, with the greatest risk
associated with those whose episodes o f depression last longer. The
lifetime risk for people who have depression and have not been hospi
talized has been estimated to be around 3 per cent.3
Alcoholism: Alcohol intake is not evenly spread through a population.
In the United States, 50 per cent o f alcohol is consumed by 10 per cent
o f the population. Studies in the 1980s found that 13 per cent o f the US
population suffer alcohol dependency at some point in their lives, with
the ratio o f male to female being between 2:1 and 5:1. Such dependency
47
Table 3, 1 Suicide vulnerability and psychiatric status
Diagnosis Long-term risks Notes
Major depressions 10-15% 1. These rates apply only to those

patients whose depression has
been severe enough to require
hospitalisation (see text)
2. No difference between
endogenous and non-endogenous
sub-types
3. Psychological autopsy
post-suicide reveals 70% suffered
from depression
4. Insomnia, self-neglect and
impaired memory are key
predictive symptoms
5. Hopelessness, loss o f pleasure
and mood cycling in index
episode predict suicide
6. The presence o f severe panic
attacks, psychic anxiety, anxious
ruminations and agitation has also
been found to be associated with
increased suicide risk in
depression4
Bipolar affective 10-30% Suicide risk is associated with

disorder depressed rather than manic
symptomatology, although at
times it is related to apparent
improvement
Schizophrenia Around 10% 1. First admission schizophrenia

annual rate o f 500-700/100,000
2. Psychological autopsy
post-suicide reveals 30% suffered
from schizophrenia
3. Mean age o f suicide = 33 years
(younger than normal suicide)
48
Diagnosis Long-term risks Notes
4. No correlation between suicidal

thoughts o f behaviour and
hallucinations
5. History o f depressive features
(including past history o f E C T or
antidepressant treatment)
increases vulnerability to suicide5
6. Male patients most vulnerable
to suicide after 4.8 years o f
disorder; female patients after 9.8
years. Period after discharge most
vulnerable
Alcoholism 187/100,000 Alcoholism present in 15-25% o f

suicides. Mean length o f excessive
drinking prior to suicide = 20
years
Neuroses 2-5%
Personality disorder 130/100,000
can show itself in one o f three ways: by regular excessive intake o f

alcohol on a daily basis; by regular excessive intake limited to weekends;
and by periods o f abstinence interspersed with extremely heavy drinking
binges lasting weeks or months. The lifetime risk o f suicide in alcoholics
is similar to that o f depressed patients, 15 per cent, with male alcoholics
six times more likely than wom en alcoholics to commit suicide. The
mean age for suicide in alcoholic patients is 47, following an average
20-year drinking history.
Schizophrenia, which up to 1 per cent o f the population suffer at any
one time, involves a range o f symptoms. These include delusions, a
disorder in the content o f thought (e.g. the belief that one’s thoughts
are being broadcast from one’s head, or being controlled by a dead
person), ideas o f reference (e.g. that events or people have a special and
unusual significance, such as the change in a traffic signal, or the
expression on the face o f a television newscaster meaning one has been
49
chosen for a special mission). A second set o f symptoms includes

disorder in the form o f thought. These m ay include the loosening o f
associations, in which ideas flow from one to the other without any
apparent connection, sometimes to such an extent that speech is incom
prehensible. A third set o f symptoms involves disorders o f perception,
especially auditory hallucinations, in which (most commonly) the
person hears voices speaking to him or her, and perhaps commenting
on his/her behaviour. Sometimes, and very dangerously, the voices
command the person to carry out certain acts, including suicide. Fourth,
there is often a change in mood, with normal expression o f emotion
being disturbed in one o f two ways. Either the m ood is ‘flat’, with little
sign o f emotion, the voice and face remaining expressionless. Or the
person seems to have moods inappropriate to the situation, laughing
or smiling at events or descriptions that would norm ally evoke sadness
or pity. Finally, and unsurprisingly in the light o f these other symptoms,
the person m ay experience a disturbance in their sense o f ‘who they
are’. There may be a disruption o f the will, with a person finding it
almost impossible to initiate any activity at all. In consequence, there is
often major disturbance in a person's relationships with other people.
The person withdraws, and becomes detached, though m ay sometimes
cling to or get too close to others.
Schizophrenia may take many forms, so that some have suggested
the term itself is meaningless, or we should talk o f ‘the schizophrenias’.
Others have maintained that as long as we realize the complexity o f the
disturbances, and that giving something a name does not really explain
it, then continued use o f the term is permissible. Several studies point
to the fact that up to 15 per cent o f patients with a diagnosis o f
schizophrenia (especially males) will end their lives by suicide.
Other psychiatric disorders found to have a higher than average risk
o f suicide include anxiety disorders in general, and panic disorder in
particular, but to date there have been too few studies to establish the
exact risk in these groups.6
Personality disorder (especially borderline and antisocial personality
disorder) carries an increased risk o f suicide. ‘Borderline’ patients were
originally so called because it was believed they lived at the edge
between psychosis and neurosis. They have affective instability, a his
50
tory o f self-damaging acts and damaging relationships, chronic feelings

o f emptiness and boredom, intolerance ofbeing alone, and brief dissocia
tive episodes, often associated with flashbacks o f sexual abuse. Recent
evidence suggests that between 4 and 10 per cent will eventually kill
themselves. In antisocial personality disorder, although 46 per cent have
some form o f suicidal behaviour, the risk o f completed suicide appears
to be around 5 per cent.
Gender and age effects not explained by

psychiatric illness
Despite these apparently strong relationships between psychiatric diag

nosis and suicide, there are several reasons to look for causes other than
psychiatric illness. One such is the gender difference. W om en are more
likely to become clinically depressed, but men are more vulnerable to
suicide (see Chapter 2). Neither can psychiatric illness explain the way
such gender differences change with age. Whereas female vulnerability
tends to rise linearly but very slightly with increasing age, the male
vulnerability rises markedly from ages 15-24 to 25-54. It falls again
between 55 and 84, then rises again for the over 85s.
This rise in the very old is most probably due to the fact that people
bom early in the twentieth century carried a slightly higher suicide risk
with them throughout their lives (see previous chapter's discussion o f
cohort effects in older adults). It is added to by the increased risk
following bereavement. Men over 65 years who are married have a
suicide rate o f only 13 per 100,000, whereas those who are widowed
have a suicide rate o f 51 per 100,000. Men o f all ages, however, are
vulnerable to loss o f their wife by death. Young men whose wives die
have the highest rates o f all. W om en w idowed young are at an increased
risk, but this is not so marked as the effect o f bereavement on men.
The vast majority o f older people, even those who are depressed,
who have been bereaved, w ho are suffering from a medical illness, or
even those who are terminally ill, do not end their lives by suicide,
however. This raises the question o f what are the important risk factors
in the older age group. The following have been identified: a history o f
51
poor adaptation to life’s stress, vulnerability to loss and disruptions, loss

o f mastery and control, cognitive impairment caused by organic mental
disorder. Once again, w e need to look further than simply the presence
or absence o f psychiatric problems to understand the data.
Durkheim's view of psychiatric causation
Before Durkheim the connection between suicide and psychiatric dis

turbance had been made many times. Some arguments had been
relatively weak. Some had suggested that, even where no other symp
toms o f mental illness were present, the fact o f suicide itself justified a
psychiatric diagnosis. Suicide had been thought by some to be a disease
in itself, sui generis, a specific form o f insanity. In such a ‘m onom ania’
(a sick person whose mentality is perfectly healthy in all respects but
one), suicide was argued to be the sort o f behaviour ‘not to be found
in sane persons’ . Emile Durkheim easily showed how this was a circular
argument. T o establish that suicide is caused by mental illness, we need
to be able to say more than that suicide itself constitutes such a diagnosis.
Instead, Durkheim considered three types o f psychiatric suicide.
The first was maniacal suicide (caused by hallucinations or delirious
perceptions), in which the person acts to escape an imaginary danger,
or obey a mysterious order from on high. Today, such experiences
would be associated with a diagnosis o f schizophrenia. The second
type was melancholy suicide, associated with extreme depression and
exaggerated sadness. Such a state ‘causes the person no longer to realize
the bonds which connect him with people and things about him ’. The
third type was obsessive suicide, in which suicide was thought to be
caused by no other motive, but only by a fixed idea o f death. The
person was tormented by the idea, since they were also aware that
there was no reason to kill themselves. Trying to resist the idea seemed
hopeless, and the person, it was thought, sometimes simply gave
up the struggle. Additionally, Durkheim pointed to the possibility o f
impulsive or automatic suicide, as unmotivated as obsessive suicide, but
differing in that the idea comes suddenly with full force ‘not preceded
by any intellectual antecedent’, little apparent warning for themselves
52
or others. The person sees a knife, or walks near a cliff-edge, and the
idea o f suicide comes suddenly into mind. If the patient survives, they
m ay say afterwards that they were not aware o f or cannot remember
any reason for their action.
In summarizing the psychiatric suicides, Durkheim notes that all
such suicides are either ‘devoid o f any m otive' (obsessive and impulsive
suicides) or are determined by ‘purely imaginary m otives’ (maniacal
and melancholy suicides). W here psychiatric disturbance itself explains
the motivation, it does so by indicating how the illness acts to produce
the behaviour either in the absence o f m otive or through the w ay it
distorts the w ay the person views their world. Durkheim points out
that ‘many voluntary deaths fall into neither category; the majority
have motives, and motives not unfounded in reality. Not every suicide
can therefore be considered insane, without doing violence to language/
The alternative - the importance of social facts
Durkheim ’s main argument was that social facts have to be taken

into account as realities external to the individual. Social institutions
(families, churches, non-religious groups) were extra-personal forces,
definite realities whose influence needed to be subject to scientific
analysis, an argument that gave most impetus to the evolving science
which would be called ‘sociology’ . (Durkheim was not translated into
English until 1951.) The incidence o f suicide was one such element - a
reality explicable only by looking for links to other social facts about
society. Each society was thought to have a ‘collective inclination to
suicide’ . This inclination was found in the suicide rate within the
society or sub-group, and would not change while the character o f
the society or sub-group did not change. Thus a certain number o f
suicides was to be expected in every society, and it would be problems
in the structure o f society that led to an increase in suicide rates. It
followed that the more strongly any individual was integrated into a
social group (e.g. close-knit families or religious groups), the less was
the likelihood o f suicide. Any changes in society that caused greater
disintegration would increase the suicide rate.
53
The three types o f suicide delineated by Durkheim reflected the

three categories o f breakdown that might occur in the relationship
between an individual and the society. In egoistic suicide, a person
comes to have no concern for the community, and no interest in
being involved with it. This category includes people with physical
or mental illness, together with those who suffer deprivation and
bereavement. The result was a reduction in society’s control, and
weakened immunity against society’s natural collective inclination
towards suicide.
By contrast, in altruistic suicide, society has too strict a hold, and a
person has too little individualism. Self-destruction is motivated by
altruism, inspiring respect and admiration am ong other members o f
the group. The suicide bom ber would be counted am ong those who
commit altruistic suicide, as would religious ‘m artyrs’ - Japanese
samurai warriors who committed hara-kiri rather than fall into the
hands o f their enemies, and kamikaze pilots o f the Second W orld War.
More controversially, those people who kill themselves because they
are old or terminally ill, and do not wish to be a burden to family,
friends or society, m ay be said to have committed altruistic suicide.
However, it is often difficult to determine how much their sense o f
being a burden, or that ‘everyone would be better o ff if I were not
here’, arises from their depression rather than from a clearly thought
out and tested set o f reasons (see Chapter 6).
In anomic suicide, society has failed in its regulation and integration
o f its members. Changes in family structure, reduced employment
opportunities, declining religious beliefs and practices, changes in mari
tal codes - all were manifestations o f anomie, resulting in disturbances
o f collective organization. The result was a reduction in individuals’
immunity against suicidal tendencies. In strict societies and subcultures,
suicide would remain low because such integration o f the individual
with his or her social group would remain (e.g. suicide has traditionally
been low in Catholic countries, but has increased as these countries
have gradually become more secularized).
54
Evidence to support the social theory
Much o f the data on suicide within and between populations is broadly

consistent with the notion o f anomie and societal disintegration as a
major factor in explaining differences in proneness to suicide. A study
o f localities in the Bristol area in the United Kingdom found the
occurrence o f non-fatal suicidal behaviour was highly correlated with
deprivation (assessed by a number o f factors, such as the proportion o f
people that own houses or cars, how many households are over
crowded, levels o f unemployment). Furthermore, the list o f factors that
increase the risk o f suicide in those who have made a previous attempt
makes the importance o f social factors obvious:
• older age;
• gravity o f previous attempt;
• living alone;
• psychosis;
• left: a suicide note;
• unemployed or retired;
• from a broken home;
• men more than wom en;
• multiple attempts;
• not married;
• poor physical health;
• lethal method used before; and
• infrequent use o f health agencies.
The picture that emerges is one o f poor circumstances with few

resources to sustain the individual. Most significantly, there is an obvious
lack o f other people to support the person. This is consistent with
Durkheim ’s hypotheses.
In many countries, suicide is more common among those members
o f society most affected by economic downturns: semi-skilled and
unskilled manual workers. It is more common in rural communities,
where there is a special vulnerability to economic downturn (because
55
o f the possible loss o f farms that generations o f a family might have

built up), combined with a mechanization o f farming that has seen
decreased employment opportunities for many and increased social
isolation o f those few left.
Unemployment, whether in a rural or urban setting, appears to
make suicide more likely. Even after controlling for social class, men
aged 15-64 who are unemployed have a standard mortality rate greatly
in excess o f that o f men in work. H owever, the relation between
unemployment and suicide is complicated by two factors: first, the
possibility that some m ay be vulnerable both to becoming unemployed
and to suicide;7 second, the fact that, in some cases o f sudden unem ploy
ment in whole communities following closure o f factories, people may
(at least for a while) feel more integrated with their community as they
fight the 'com mon enem y'. H ow ever, such fellow-feeling m ay not
offset the effects o f economic hardship for long. Within individual
families, increased depression in the form er wage-eam er reduces energy
levels. This can then be the cause o f family friction as one partner
accuses the other o f ‘not trying’ to find employment. It is not uncommon
to find single mothers talking frankly o f how their unemployed (and
now estranged) partner was only a drain on their fam ily’s resources,
especially if he had been spending m oney on alcohol or gambling.
Studies o f how socio-demographic factors contribute to changing
rates o f suicide over time also confirm the main thrust o f Durkheim ’s
theory o f anomie. Suicide rates in eighteen countries across Europe
between i960 and 19808 showed that increases were associated with:
(a) reduction in the population aged 15 years and under, taken to be an
indicator o f the extent to which people were not living in family groups;
(b) an increase in the percentage o f the population aged 65 years and
over, i.e. the age group with the relatively highest rate in European
countries; and (c) an increase in w om en’s tertiary education, taken to
reflect changes in the family structure. Changes in suicide rate over the
same time interval among 15-29-year-olds found some factors which
overlapped with those o f the earlier report, and some new factors (see
Table 3.2).
The correlation between suicide and the divorce rate emerges from
several studies. What is more questionable is whether this correlation
56
Table 3.2 Factors associated with suicide rates among

15-29-year-olds in Europe, 1960-80
% unemployed (+)
% o f population under 15 years (—)
% wom en employed (+)
divorce rate (+)
homicide rate (+)
change in alcohol use (+)
change in church affiliation (+)
+ = increase in factor associated with increased suicide rate

— = decrease in factor associated with increased suicide rate
will continue as divorce becomes more common and society develops

more ways o f dealing with and normalizing single-parent families. For
example, in the Netherlands the suicide rates among divorced people
are tending to stabilize or decline.
Social facts and the individual
Durkheim has been criticized from various quarters. First, some have
pointed out that he needed a better and more operational definition o f
social integration. Later sociologists have come up with clearer defi
nitions o f terms, e.g. o f ‘status' (a category o f people with clearly defined
roles). Thus a person may have a variety o f statuses and roles: within
society (male, white), within the family (father, husband), within w ork
(teacher, counsellor). In this w ay it is possible to calculate the frequency
with which any person's combination o f statuses conforms to the
combinations most com m on in the society in which they live. In similar
vein, sociologists have much better definitions o f ‘social support' than
those proposed by Durkheim.9
Second, he based his conclusions on studies o f a W estern society,
and there is evidence that in other societies risk factors are different.
For example, although the association between divorce rates and suicide
holds for W estern nations (the correlation for the United States has
57
been found to be 0.78 for the 48 continental states), it does not hold for
Taiwan. The correlation between the divorce rate and suicide rate
among the sixteen counties o f Taiw an is an insignificant 0.05.
Third, Durkheim too readily separated melancholic from other
depressive suicides. He allowed that many normal persons w ho kill
themselves m ay also be depressed and dejected, but maintained that
they did not fall thereby into the category o f melancholic mental
illness. The difference, he suggested, is that in melancholy the person's
depression was unrelated to their external circumstances, whereas in
those he wished to call ‘norm al’ the state o f depression and the act o f
suicide had an objective cause. In this respect, Durkheim ’s analysis was
wrong. There is no clear-cut difference between different types o f
depression, based on some having been preceded by negative life events
and others coming ‘out o f the blue’ .
Whereas it used to be thought that some depressions were ‘reactive’ ,
caused by life circumstances, and others were ‘endogenous’, caused by
biological factors (‘endogenous’ means ‘originating from the inside’),
research has shown that negative events and circumstances precede all
types o f depression equally often, and all types o f depression have some
biologically driven features. Even in depressions which follow directly
after m ajor loss or disappointment, there is evidence to show that, as
the depression deepens, so certain neurochemical pathways in the brain
undergo a change, affecting a person’s eating and sleeping patterns,
energy levels and capacity to enjoy previously enjoyed hobbies, interests
and social contacts. Such changes sometimes self-correct in time, and
where they do not, antidepressant medication or types o f psychotherapy
that provide a structure within which the person can regain control
over their moods have been found to help.
Nevertheless, the central thrust o f Durkheim ’s argument remains
true: we need to look for factors other than mental illness to explain
w hy risks o f suicide differ between individuals and between societies.
The importance o f social support cannot be overestimated in moderat
ing the impact o f other stresses. Not only is it important at times o f
stress, but people who have social support tend to adhere to the
treatment suggested, and it is therefore no surprise that they respond
better. With social support, people are more likely to take the opportun
58
ities given to begin to solve the problems in their life, with or without
the help o f mental-health professionals.
H ow ever, understanding the social facts needs to be combined with
understanding the individual circumstances in causing particular suicidal
acts. If this were not so, suicidal behaviour would be much more
com m on than it is.
The causal significance of psychiatric illness and

social facts
It certainly seems that suicide is closely associated with psychiatric

disturbance. W hat, then, are we to make o f Durkheim ’s scepticism
about psychiatric illness, and how can w e make sense o f the equally
compelling data linking suicide to factors external to the individual?
Although psychiatric disturbance precedes suicide in many cases, it
remains unclear whether its presence explains very much. Certainly we
can readily see that, although up to 15 per cent o f those who have been
hospitalized for depression commit suicide, the majority do not. We
therefore have to understand what is different about those who do
commit suicide. Second, even within the 15 per cent who do commit
suicide, w hy did they choose that particular moment, when the disorder
from which they suffered had been with them a long time? Finally,
there remains a proportion o f people w ho commit suicide either as a
catastrophic response to a negative event (usually a loss, or perceived
loss), or as a matter o f honour, or because they feel on rational grounds
it is time to die.
These cases are linked in the feelings o f uncontrollability they
involve. Such feelings arising from inside a person are particularly
damaging, yet are often ignored. Why? Perhaps because psychiatrists
and psychologists often take account only o f negative life events (such
as bereavement, loss o f job, marital breakdown) in explaining the onset
o f a psychiatric illness. Once the psychiatric illness has been diagnosed,
however, the search for causes often stops. The illness has been
explained. But this does not explain how the disturbance is maintained,
w hy it sometimes lasts a long rather than a short time or gets worse
59
rather than better. Occasionally there are further negative external

events that explain such prolongation. But we need also to take account
o f the negative effects o f experiencing unpleasant psychiatric symptoms
themselves: e.g. hearing voices that cannot be switched off, in schizo
phrenia; feeling constantly tired but unable to sleep, in depression;
feeling at the mercy o f craving, in alcoholism. Such uncontrollable
stresses that arise from within are as likely to produce a state o f
helplessness as uncontrollable stresses that arise from outside. It is these
that explain w hy psychiatric illness so often precedes suicide, and why
most people who become psychiatrically ill do not commit suicide.
I suggest that the presence o f mental illness does add significantly to
the explanation o f suicide in many cases, but, when it does, does so to
the extent that the symptoms o f the condition engender hopelessness.
T o state the hypothesis more strongly: psychiatric illnesses carry an
increased risk for suicide only to the extent that the person feels they
cannot escape their symptoms. For example, it m ay not be the frequency
o f ‘bizarre' symptoms that predicts suicidality in schizophrenia, but
how the person feels about having such symptoms: the extent to which
the person feels entrapped by such symptoms, as research by Max
Birchwood and his colleagues has indicated.10 The entrapment hypoth
esis also predicts that it will be the longest and most persistent depression
that is associated with suicide, and that other predictors o f suicide will
all be elements that decrease a person's sense o f control over external
and internal events.
The evidence is quite compelling. In depression, it is the people who
have been depressed longest, and who are therefore most likely to feel
hopeless about recovering a normal state o f mind, who carry most risk
o f suicide. Furthermore, retrospective analysis o f suicidal behaviour in
people who have been depressed in-patients, and o f suicide in those
who have previously attempted suicide, has found that the greatest risk
was for those who showed severest depressed mood, alcohol problems,
long-term use o f sleeping pills and long-term physical illness.11 Finally,
whereas attributing suicidal behaviour to the diagnosis o f depression
does not explain w hy the period immediately after discharge is the most
vulnerable time, especially for men, such increased risk is more easily
explained by the exacerbation in a sense o f hopelessness brought about
60
by the change in circumstances. Jouko Lonngvist reviews evidence

showing that level o f hopelessness accounts for the variation in intensity
o f suicidal feelings even when depression level is taken into account
statistically.12
Similarly in schizophrenia, the suicide risk is particularly strong for
those who feel hopeless, have suicidal ideation, fear mental disinte
gration, have made previous suicidal attempts, do not adhere to treat
ment and experience many relapses in symptoms. Indeed, it is a history
o f depressive features (past use o f E C T or antidepressants in treatment)
which is an important predictor, not the extent o f the schizophrenic
symptoms (e.g. voices). This supports the view that it is not the voices
themselves which increase the risk o f suicide, but how the person feels
about them. The external stresses with which schizophrenic patients
have to cope are well known: decreased job prospects, family disinte
gration, poor network o f relationships. Entrapment theory would sug
gest that such events have most impact upon those who can compare
current reality with what might have been.13 Studies have found that
most schizophrenic suicides are young males, functioning at a high level
prior to the onset o f the schizophrenia but now unemployed. Those
who previously had the most promising careers are at greatest risk; the
contrast is for them the most stark. As in depression, any event that
increases a sense o f helplessness can increase the suicide risk, and
change in circumstances is particularly likely to do so. This may include
changing wards when an in-patient, or discharge from hospital. Each
finding points away from the particular psychiatric diagnosis o f schizo
phrenia per se as the element that increases suicide risk and towards the
increased helplessness and hopelessness such a disorder shares with
other serious mental problems.
The link between suicide and helplessness in the face o f feelings o f
entrapment in alcoholics is striking. Once again, discharge is a vulnerable
time, often made worse by the fact that family contacts have broken
down. Such people often have vulnerable or fragile personalities, and
social problems that may have precipitated their drinking but are then
exacerbated by the drinking itself. Connected to this is the fact that
alcoholic individuals who commit suicide are very likely to be either
unmarried or divorced. Life events (especially loss events) are most
61
closely connected to suicide in alcoholics, with 50 per cent having

suffered loss o f an important relationship in the year preceding the
suicide (the figure for depressed patients is 20 per cent). Indeed, one
third o f alcoholic suicides have suffered such a loss within six weeks o f
the suicide.
These patterns in the suicide data both confirm the significance o f
psychiatric illnesses and show how they combine with a psychosocial
perspective in explaining suicide risk. Psychiatric diagnoses, especially
depression, schizophrenia and alcoholism, act as markers o f increased
suicide risk. But suicide is not unique to any particular diagnosis. It
arises from a secondary aspect o f mental illness. In such mental states,
the person experiences symptoms that can give rise to an inner turmoil
that seems inescapable and uncontrollable. It is the combination o f
uncontrollable stress factors arising from sources external and internal to
the individual that increases hopelessness. The most potent preventative
factor when such uncontrollable stress threatens to overwhelm the
individual is the availability o f social support from friends and family.
W hen the suicidal feelings are too strong, however, even support o f
the highest quality will be ignored by the suicidal individual; the
hopelessness they experience includes extreme pessimism about
whether anyone can help them.
62
4
Attempted Suicide: Facts and
Figures
‘I have to admit that I am a failed suicide. It is a dismal confession to

make, since nothing, really, would seem easier than to take your own
life/ So wrote A 1 Alvarez in his book Savage God.1 Alvarez knew first
hand what it was like to experience suicidal despair, and be driven to
take the final step. The crisis took place ten years before, on Christmas
Eve. After a ‘final, terrible quarrel' with his wife, she left to stay
elsewhere for the night. Alvarez went upstairs to the bathroom and
swallowed 45 sleeping pills. Impulsive? In some senses, yes, but the act
had been prepared for some time. His account illustrates, among other
things, the premeditation o f suicidal behaviour (he had been hoarding
sleeping pills), the precipitation o f the act by some ‘final straw' and the
unwillingness o f the police to intervene in the late 1950s.
I had been collecting the things for months obsessionally, like Green Stamps,
from doctors on both sides o f the Atlantic . . . hoarding them in preparation for
the time I knew was coming. When it finally arrived, a box was waiting stuffed
with pills o f all colors . . . I gobbled the lot.
My wife got back at noon, took one look and called the ambulance. When
they got me to the hospital I was, the report says, ‘deeply unconscious, slightly
cyanosed, vomit in mouth, pulse rapid, poor volume’ .
I was still unconscious the next day and most o f the day after th at. . . During
the afternoon o f the third day, December 28, I came to . . . In a fog I saw my
wife smiling hesitandy, and in tears. It was all very vague. I slept . . .
At some point the police came, since in those days suicide was still a criminal
offense. They sat heavily but rather sympathetically by my bed and asked me
63
questions they clearly didn’t want me to answer. When I tried to explain, they
shushed me quietly. ‘It was an accident, wasn’t it, sir?’ Dimly, I agreed. They
went away.
Alvarez’s experience raises many o f the questions that need to be

answered in relation to attempted suicide. First, what proportion o f
people who think about harming themselves actually go on to do it,
and what proportion o f those who do, go on at some time in the future
to kill themselves? Second, if ending one’s life is not the only motive
for such self-harm, what are the other motives? Third, how can one
distinguish a failed suicide attempt (where the full intention was to end
life) from self-harm intended to change some aspect o f oneself or others?
Definitions
T o answer these questions, we need a clear definition o f what we mean

by 'attempted suicide’ . Is all and any self-harm to be deemed an attempt
at suicide, even self-cutting, or burning with cigarette ends? And what
o f the elderly wom an who takes four sleeping pills, and telephones the
physician in great distress. In this case the physician would be well
advised not to ignore the behaviour, even though it represents no risk
to life. Such a person may be used to taking half a sleeping pill to help
her get to sleep at nights, and therefore believes four such tablets (eight
times her normal dose) will kill her. Turning a blind eye may be to miss
someone who is extremely suicidal.
On the other hand, a young person who takes a hundred paracetamol
m ay be at considerable risk o f dying without medical intervention as a
result o f liver failure (especially if they have taken alcohol with the
paracetamol), yet may at no point have thought o f killing him or herself.
Indeed, the availability o f large quantities o f such lethal substances over
the counter in the past gave the completely false impression o f safety.
Such a person can hardly be said to have been attempting suicide,
though their risk o f dying may have been greater than that o f the elderly
person, who was at no significant risk.2
T o overcome the confusion about the term ‘attempted suicide’,
64
Attempted Suicide: Facts and Figures
clinicians and researchers have, over the past twenty years, adopted the
term ‘parasuicide’ or ‘deliberate self-harm’. Parasuicide was, for a time,
the most widely used term for all such self-harm (whatever the explicit
or implicit intention). Most clinicians, however, now use the terms
‘deliberate self-harm’ and ‘attempted suicide’ interchangeably (and I
shall follow that convention in this book, without making any assump
tions about whether the behaviour was intended to be fatal). In any
event, the definition is the same:
An act with non-fatal outcome, in which an individual deliberately initiates a

non-habitual behaviour that, without intervention from others, will cause
self-harm, or deliberately ingests a substance in excess o f the prescribed or
generally recognized therapeutic dosage, and which is aimed at realizing changes
which the subject desired via the actual or expected physical consequences.
Defining attempted suicide is one thing, collecting data about it another.

Most hospitals do not officially record attendance at accident and
emergency in these terms. For example, in the United Kingdom neither
‘attempted suicide’ nor ‘parasuicide’ exists in the classification system
used by the Department o f Health! The closest approximation, ‘adverse
effects o f medical agents’ , is used for statistical purposes instead. For a
complete picture o f attempted suicide rates and trends, we must rely
on data from those centres that have made special efforts to collect
information over the years. Fortunately, these have presented a remark
ably consistent picture.
Attempted suicide statistics
In 1961, when it ceased to be a criminal offence to attempt suicide, there

were some 20,000 attempted suicides in the United Kingdom each year.
During the 1960s and 1970s there was a steep rise in numbers o f people
harming themselves, reaching a peak around 1977 o f some 100,000.
Numbers then levelled off, and even began to fall. But during the 1980s
the rates turned upwards again. Current estimates are that there are at
least 100,000 episodes o f deliberate self-harm known to the hospital
service per year in the United Kingdom (17,000 o f whom are teenagers).
65
There are perhaps a further 30,000 (of all ages) who never reach hospital,
either telling no one about their self-harm, or seeing their family doctor,
who m ay not refer them to the hospital. Another slight fall at the end
o f the 1980s was followed by a rise at the beginning o f the 1990s. In the
United Kingdom this represents a rate o f 264 per 100,000 for men, and
368 per 100,000 for w om en.3 This ratio o f female:male o f 1.4:1 declined
from a ratio o f 2.1:1 in 1976, and 1.9:1 in 1984. During the 1990s the ratio
declined again, so that men almost caught up with w om en in attempted
suicide numbers. By the end o f the 1990s and the beginning o f the 2000s,
however, the ratio had begun to revert. Many more wom en than men
were again attempting suicide.
Until the 1990s, our understanding o f the problem o f attempted
suicide had been based on a pattern o f statistics that had not changed
for three decades. Overall numbers were rising, but the general charac
teristics o f the data remained relatively stable. W hereas suicide appeared
more a problem o f older men, attempted suicide was more a problem
o f younger wom en. T wo-thirds o f attempted suicide cases were wom en,
and two-thirds were under 35. Around 90 per cent o f such cases harmed
themselves by taking an overdose, with younger people more likely to
take analgesics such as aspirin or paracetamol, and older people more
likely to take sleeping pills or antidepressants. (This pattern o f age-
related overdose by different substances reflects the fact that all ages
take whatever is available to them. Younger people are less likely to
have sleeping pills and antidepressants prescribed, so take whatever else
comes to hand. If they do take psychotropic medication, it is often
because they have taken the medication prescribed for someone else in
the family.)
H ow far is this pattern o f data still true? T o answer this, w e need to
examine the statistics in more detail.
Oxford data in detail
The city o f Oxford contains a mix o f population, with a large university

population combining with areas o f great social deprivation. When the
pattern o f data from the city has been compared with other U K centres,
66
the trends are remarkably similar both for total numbers, balance
between the sexes, socio-economic status data and methods used. It is
therefore useful to examine the Oxford data to chart the U K trends,
since the Oxford Centre for Suicide Research has data from the 1970s
to the present. The picture o f suicidal behaviour begins in the mid
1970s, at a point where the rates (increased fourfold in the ten years to
1973) began to decline (see Figure 4.1). The decline was more marked in
wom en than men, corresponding to the earlier greater increase in
wom en. Young people are at most risk o f attempted suicide. The data
for 1999 are shown in Figure 4.2 where it can be seen that the peak age
for w om en remains at 15-19 years, whereas the peak age for men lies
at 25-29 years.
The data for marital and employment status help fill out the picture.
Most people who attempt suicide are people who are single or divorced
(56 per cent in 1999). H owever, despite the population o f those who
attempt suicide being weighted towards those who are single or di
vorced, the majority are living with parents or with a partner. Only 35
per cent o f men and 26 per cent o f wom en who attempted suicide in
1999 in Oxford lived alone. This is consistent with research that shows
that most non-fatal suicidal behaviour occurs in the context o f problems
in relationships.
The data also reveal that the majority are from low er socio-economic
status backgrounds. The rate for social classes I and II is around 50 per
100,000 for wom en and 20 per 100,000 for men. This rate goes up over
eight times for wom en and over 12 times for men in social classes III—V.
W hat proportion will harm themselves again? Estimates over the
years and from various centres have varied between 10 and 25 per cent,
but settled down to around 15 per cent. This is reflected in the Oxford
data, where 15.5 per cent o f men and 12.8 per cent o f wom en repeated
the attempt within the first year. Some go on to make further repeat
attempts, so that in any sample o f cases coming to hospital, around 44
per cent o f both men and wom en have a previous episode o f deliberate
self-harm.
In terms o f methods used in self-harm, although 9.5 per cent cut
themselves, the vast majority use self-poisoning (86 per cent medicines,
and 3 per cent other chemical substances), a pattern unchanged over
67
0001
oo
0\
68
00 h-*
oo oo oo
oo
saposidg
oo
oo
o
CNo
o o
o
Figure 4.1 Episodes of attempted suicide referred to the John Radcliffe Hospital, Oxford, 1976-99.
Males
J9quinf\i
69
<
3
H
u
O
G,
cbJD
bJO
Figure 4,2 Oxford City attempted suicides, 1999, by age and gender.
the years. H ow ever, the pattern o f substances used has changed (see
Figure 4.3). Tranquillizers and sedatives used to be the most common
method, but their use has declined from around 40 per cent o f all
overdoses to around 20 per cent, most likely because o f the marked
decline in prescribing these drugs over the period. Their place has been
taken by an increase in paracetamol (from around 14 to 42 per cent).
Often the tablets are either taken with alcohol (32 per cent o f men and
20 per cent o f wom en) or within a few hours o f consumption o f alcohol
(56 per cent o f men have taken alcohol in the previous six hours, as
have 37 per cent o f women).
Paracetamol
Other analgesics
(e.g. aspirin)
Tranquillizers
and sedatives
Antidepressants
1, , 1 1999
1 _ J 1976
• ';
•
Other
10 20 30 40 50
Percentage o f all overdoses
Figure 4.3 Methods of self-poisoning, Oxford: changes between

1976 and 1999.
70
Attempted suicide: international trends
H ow do these data differ from those o f other countries? Figure 4.4

shows the Oxford data (which can be taken as representative o f the
United Kingdom) alongside data from other European centres from the
W H O /E U R O Multicentre Study o f Parasuicide in Europe.4 The most
salient conclusion to emerge is that the United Kingdom has one o f the
highest rates o f attempted suicide in Europe. The rates for 1989-93 vary
from 46 per 100,000 for men in Guipuzcoa, Spain, to 327 per 100,000 for
men in Helsinki, Finland. Oxford's rate over this period was 264 per
100,000 for men (second highest) and 368 per 100,000 for wom en (the
Italy (Padua)
Italy (Emilia-Romagna)
—
___
Spain (Guipuzcoa)
Germ any (Wurzburg) B ll Female
.... ■
Austria (Innsbruck) Male
Netherlands (Leiden)
Sweden (Umla)
Switzerland (Berne)
France (Bordeaux)
Sweden (Stockholm)
IM M M H im ili
Norway (Sor-Trondelag)
Denmark (Odense)
Hungary (Szeged)
UK (Oxford)
Finland (Helsinki)
300 400
Rate per 100,000
Figure 4 .4 Rates of attempted suicide (per 100,000), 1989-93,

across European centres.
71
highest). In a direct comparison with Utrecht in Holland as part o f

another study, Keith H awton and colleagues have found that the higher
UK rates arise from the larger number o f teenage females in the U K
sample. H owever, this will not account for all the difference, and the
exact reasons remain unclear.
Data from North America are more difficult to obtain. In Canada,
attempted suicide has not been reportable since the 1960s, and in the
United States study o f attempted suicide is limited by the protection o f
individual privacy. H ow ever, where research has been done, it has
found rates o f deliberate self-harm comparable with the higher end o f
the European spectrum, with annual rates in Canada being estimated
as 304 per 100,000.5
The most comprehensive survey o f attempted suicide rates in the
United States comes from the National Institute o f Mental Health's
Epidemiologic Catchment Area study (1980-85). As part o f this survey,
over 18,000 people answered questions about suicide ideation and
previous suicidal behaviour. The results showed that 2.9 per cent had
made a suicide attempt at some point. The ratio o f men to wom en was
1:1.4; 59 per cent o f those saying they had ever attempted suicide were
w om en.6 This survey also found that the prevalence was significantly
higher for persons aged 25-44 years than for those over 44 years. With
regard to marital status, the highest rates were found among the
separated or divorced, and the lowest am ong the married and widowed.
There was an increase in the rate o f attempted suicide with decreasing
socio-economic status, but the most powerful risk factor was a lifetime
diagnosis o f psychiatric disorder. Many aspects o f these US data coincide
with conclusions reached from the European studies, but because they
were collected in the early 1980s they cannot inform us about recent
changes. There is some evidence that a major shift is taking place in the
gender distribution o f attempted suicide.
Trends in female:male ratio
Across Europe for the period up to the 1990s, the number o f wom en
attempting suicide exceeded that o f men in most countries. Only in
72
Helsinki, Finland, was the pattern reversed. H ow ever, reports from

various centres around the United Kingdom during the 1990s suggest
that the gap was closing. In Oxford, whereas the male to female ratio
had been 2:1 in the mid 1970s, the ratio changed rapidly in the 1990s and
in 1994 stood at 1.35:1. This came on top o f a general upward trend for
the total number o f attempted suicides over the period 1992-4. The rise
between 1993 and 1994 alone was 21.8 per cent, the increase occurring
in both sexes. Despite these changes, the age ratio did not change, with
71 per cent o f attempters in Oxford in 1994 being under 35 years old,
and the most vulnerable age for w om en remaining at 15-19 years, and
for men, 20-24 years.
In both men and wom en, the rate o f attempted suicide peaked in
1997, and since then has declined somewhat. H ow ever, the decline is
wholly accounted for by a reduction in the numbers o f young men
attempting suicide in the age category 15-24 years. There is no hint o f
such a decline in young w om en (or indeed w om en o f any age). This
changing pattern means that, since the mid 1990s, the gender ratio has
fallen back into something more like the older pattern with the female:
male ratio in 1999 being 1.5:1. The reversal o f the usual female/male
bias in attempted suicide that w e thought we had detected in the mid
1990s did not stabilize into a long-term trend.
Suicidal ideas and suicidal behaviour
What proportion o f people have thoughts about suicide at some point

in their lives? Studies vary in their estimates, from very low (3.5 per
cent) for those that merely ask about ‘recent’ thoughts, through 19 per
cent (if the question is asked about the past year) to 53 per cent (if people
are asked if they have ever thought o f suicide).7 Perhaps we should not
be surprised that over half the population have had suicidal thoughts at
some time in their lives, but that one in five have had such thoughts in
the past year is quite alarming. O f these, only 1 to 2 per cent go on
to harm themselves in some way, but when someone has harmed
themselves once, how can we assess whether the person is at future
risk o f harming themselves again?
73
There is no one single predictor, but the most com m only found
factors that predict repetition o f attempted suicide are:
1. Problems in use o f alcohol.

2. Previous diagnosis o f 'sociopathy' or other personality disorder.
3. Previous in-patient psychiatric treatment.
4. Previous out-patient psychiatric treatment.
5. Not living with relatives.
6. Previous deliberate self-harm.
In one o f the earliest studies o f prediction o f self-harm, it was found

that with o or only 1 o f these factors involved, risk o f repetition in the
following year was 5 per cent; with 5 or 6 o f them, the risk was 48 per
cent.8
A more recent study by Norman Kreitman in 19919 re-examined these
and other factors in people admitted to hospital following deliberate
self-harm. The following were found to be predictive:
I. P revio u s delib erate self-harm .

2. C linical diagn osis o f p erson ality disorder.
3 - A lco h o l co n su m p tion (m ore than 21 units p er w e e k in m ales;
m o re than 14 units p er w e e k in fem ales).
4 - P revio u s psych iatric treatm ent.
5 - U n em p lo ym en t.
6. Social class (V).
7* D ru g abuse.
8. C rim in al record.
9 - V io len ce (given o r received in the past five years).
10. A ged 25-54 years.
11. Single / w id o w e d o r divorced.
Those patients who scored 3 or less had a repetition rate averaging 4.9
per cent (5.1 per cent for females, 4.6 per cent for males). Those with a
score between 4 and 7 had a repetition rate o f 20.5 per cent (18 per cent
for females, 22.9 per cent for males). Those with a score o f 8+ had a
repetition rate o f 41.5 per cent (36.9 per cent for females, 46 per cent for
74
males). So that almost 60 per cent o f those with a score o f 8+ did not
attempt suicide again. Isaac Sakinofsky has pointed out that factors that
predict suicidal behaviour in one sub-group o f the population may not
generalize to another. Therefore, we must be cautious in making
predictions o f w ho is most at risk.10
Motivation for self-harm
If death is not the intended outcome o f a ‘suicide attempt', then what

is? The person's own report is one important consideration. A study by
John Bancroft and colleagues found that over half o f those who had
recently taken an overdose said they did not want to die at any stage,
and o f those who did say they wanted to die, a proportion said there
were other reasons, such as ‘showing how much I loved som eone'.11
The most com m on reasons given for taking an overdose are listed in
Table 4.1.
Note that the most commonly endorsed reason is, ‘The situation
was so unbearable that I had to do something and didn’t know what
else to do.' This gives an important clue to what is going on in the
mind o f the person w ho is desperate - up against a ‘brick wall' and
come to the end o f coping. This fits with the often reported feeling that
they didn't care whether they lived or died. They will, as it were, let
the Fates decide. It is more like Russian roulette than a considered act.
Suicidal intent
H ow can we tell if someone really intended to kill themselves? In the

past, a number o f terms have been coined to refer to people who, it is
assumed, did not mean to kill themselves, such as ‘gesture', ‘manipula
tive' and ‘cry for help'. These too easily assume that different m otiv
ations for deliberate self-harm are mutually exclusive, being either an
attempt to communicate (gesture), to influence others (manipulation)
or to die (suicide attempt).
Instead, it is important to build up a picture which includes a range
75
Table 4 .1 Proportions of people endorsing Bancroft & Hawton's

'reasons for overdose' (from Williams, 'Differences in reasons for
taking overdoses in high and low hopelessness groups', British
Journal o f Medical Psychology , 1986)
%
endorsement
The situation was so unbearable that I had to do something

and didn’t know what else to do 67
I wanted to die 61
I wanted to escape for a while from an impossible

situation 58
I wanted to get relief from aterrible state o f mind 52
I wanted to make people understand how desperate I was

feeling 39
I wanted to make things easier for others 36
I wanted to get help from someone 33
I wanted to show how much I loved someone 30
I wanted to try and get someone to change their mind 15
I wanted to try and find out whether someone really loved

me or not 12
I wanted to make people sorry for the way they have treated
me 9
I wanted to frighten someone 3
I wanted to get my own back on someone 3
Note : Each person w as allow ed to endorse as m any item s as they felt fitted th eir case, so
th e n u m bers do n o t add up to 100.
76
o f information, including details o f the circumstances surrounding the

episode (if necessary, obtained from relatives or friends) as well as
the patient's own report. The following issues need to be taken into
account:12
(a) External circumstances

1. How isolated was the person at the time? The greater the intention, the
more likely it is the person will choose to harm themselves in a place
aw ay from people.
2. Was the act timed so that intervention was likely or unlikely? The highly
suicidal person will calculate what time to harm themselves so the
likelihood o f interruption is minimal.
3. Were there precautions taken against discoveryi Some people will lock
doors, write notes saying they have gone away, travel to remote
locations without telling o f their whereabouts, or book into hotels
under other names, all to avoid discovery.
4. Did the patient do anything to gain help during or after the attempt? It
sometimes happens that someone will have second thoughts, either
just before or during a suicidal act, and telephone or tell someone what
they are about to do or have just done. The highly suicidal person is
less likely to do this.
5. Did the patient make any final acts anticipating they would die? Those
with greater suicidal intent write wills, cancel regular orders, clear their
desks, and do other things in the expectation they will die.
6. Did they write a suicide note? Although only 30 per cent o f people who
commit suicide leave a suicide note, it usually denotes a high degree o f
suicidal intent.
(b) Self-report
1. Did they believe what they did would kill them? People vary in how
much they know about the lethality o f suicidal acts, particularly over
doses. The amount o f drugs taken has been found to correlate with
suicide intent if a large enough sample is taken,13 although other studies
have not found this to be the case.14 Medical lethality is useful for
judging seriousness o f intent where it is known the person who has
taken the overdose is aware o f the relative lethality o f drugs. Otherwise,
77
the person's own report about what they thought the outcome would
be is the most important factor.
2. Do they say they wanted to die? Many studies show that over half o f
people who deliberate self-harm say they did not want to die. High
suicide intent is associated with a clear indication that death was the
intended outcome.
3. How premeditated was the actf Two-thirds o f patients have not thought
about it for more than an hour beforehand. The longer the idea o f
suicide had been in the mind, the greater the suicidal intent.
4. Is the patient glad (or sorry) to have recovered? Clinicians are always
concerned about the person who has harmed themselves, and after
wards says they are sorry to be still alive.
Research has found that high suicidality, defined in this w ay, predicts
future suicidal behaviour and future suicide.15 One study found that 21
per cent o f patients with high suicide intent later committed suicide.16
Another study, by Aaron Beck and colleagues,17 compared the Suicide
Intent Scale scores o f 194 patients who eventually completed suicide
against data from 231 other suicide attempters. They found that the
completers had higher scores on the 'Objective Circumstances’ items
o f the Suicide Intent Scale. Further, those attempters who had another
suicide attempt within a year o f discharge (n = 19) had a higher total
intent score for this last episode. Further studies from the same research
group18 have found that the ‘Precautions against Intervention’ factor
o f the Suicide Intent Scale predicted eventual suicide am ong suicide
attempters and that intent was higher just before completion o f suicide.
O f particular interest is how suicide intent relates to age, and whether
intent is related to the use o f alcohol, either in general or specifically
around the time o f the attempt. Results have shown that people who
harm themselves with high suicide intent are more likely to have had
previous episodes o f self-harm, to be single or divorced and to live
alone. However, they are no more likely to be depressed or personality
disordered, no more likely to have had psychiatric treatment in the past
and show no difference in the type o f drug used in overdose or in
whether alcohol was taken. By far the largest predictor is a chronic
78
problem with alcohol abuse. Since chronic alcohol abuse itself puts a
person at high risk o f suicide, the joint influence o f abusing alcohol and
harming oneself with high suicide intent should be taken as a very
serious indicator that a person is vulnerable to suicide.
Alvarez revisited
On the basis o f the above descriptions, was Alvarez’s deliberate self-

harm a serious attempt on his life? He very nearly died, so nearly
became another suicide statistic. Barbiturate sleeping pills were very
unforgiving, unlike more m odem equivalents. But taking account o f
the external circumstances, the episode would have to be judged as o f
low suicidal intent. He took the pills in his own house on Christmas
Day, after a row with his wife, not making any attempt to isolate
himself. The timing was more ambiguous in that she had just left
to sleep at their flat for the night, but there is little indication that
he believed she would not return the next morning. In any event,
there was a house guest, who brought him a cup o f tea in the morning,
and his wife returned at noon. The house guest was able to bring
the tea into his bedroom, so he had taken no precautions against
discovery, such as locking the door. He did not act to gain help before
or during the attempt, which might indicate high suicide intent, but he
had tried to contact his psychotherapist on Christmas Eve. When an
immediate appointment was difficult to arrange, nothing was done. He
made no final acts in anticipation o f death, and did not leave a suicide
note.
Alvarez’s internal feelings about what was going on indicate more
suicidality than the external circumstances. He seemed to think the pills
w ould kill him, seems to have wanted to die, and had been hoarding
pills for some time (indicating premeditation). Whether he was sorry
he had not died is less clear. He does express disappointment, but not
so much at being alive as from a sense o f being cheated by death. ‘Death
had let me dow n/ he says. There had been no moment o f cathartic
truth when the meaning o f his life was revealed. All I got was oblivion.’
79
‘As for suicide . . . it is not for me. Perhaps I am no longer optimistic

enough. I assume now that death, when it finally comes, will probably
be nastier than suicide, and certainly a great deal less convenient/
80
5
The Causes of Attempted Suicide
What causes people to harm themselves? I begin this chapter by taking

a temporal perspective, looking briefly at the three aspects (or phases)
in the build-up to a suicidal crisis: long-term vulnerability factors; short
term vulnerability factors; and precipitating factors.
Long-term vulnerability factors
The long-term vulnerability factors could be taken to include all those

factors in the person's past or current relationships and living conditions
which act as background to the shorter-term crises. Ronald Maris, for
example, found that 83 per cent o f attempters, as compared with 31 per
cent o f a control group, had experienced either early loss by death or
separation (from fathers more often than mothers) or other major
traumas within the families, such as alcohol or drug abuse, mental
illness, criminality or a sibling in a foster hom e.1
Other long-term vulnerability factors for suicidal behaviour are to
be found in the living conditions o f the patients. In a study o f attempted
suicide in 13-18-year-olds by Keith Hawton at Oxford,2 12 per cent had
been in care at some time in their lives, over half had problems with
schoolwork and relationships with teachers, and three-quarters had
difficulty with one or both parents. Indeed, 36 per cent were living with
only one parent, and 12 per cent with neither parent. Social isolation
was a fairly common background factor.
\
81
Parenting
Although there is an association between loss o f parents (through death
or divorce) and such suicidal behaviour, this does not imply that all
cases have this in their background. Nevertheless, several studies suggest
that poor parenting occurs unusually frequently. A study by Robert
Goldney3 in Adelaide examined the type o f care that young (18-30)
female overdose patients reported having received as children. They
used the Parental Bonding Instrument,4 a questionnaire that measures
how much people perceive their parents cared for them versus being
indifferent or rejecting. It also assesses how much parents w ere overpro-
tective versus encouraging independence. Several research studies have
established the reliability and validity o f this instrument.
Results showed that suicidal patients reported their parents as having
been more rejecting and more overprotective. It seems that these young
wom en, through their experience o f deficient parenting, perceived
themselves as less deserving o f care. They had poor self-esteem and
placed a low er value on their own life. The curious combination o f a
more rejecting style and overprotection is significant. Overprotection
undermines the grow ing child's sense o f autonomy and sense o f being
in control o f events in its own life.
Sexual abuse
Several authors have suggested that a high proportion o f suicide
attempters have an even more acute disruption in early social relation
ships: they suffer sexual abuse. Only recently has this been studied
systematically. A 1993 study by M aijan van Egm ond and colleagues5
from the Department o f Clinical and Health Psychology, University o f
Leiden, Netherlands, examined the extent o f sexual abuse in a sample
o f 158 female suicide attempters aged 20 years or older. The degree o f
sexual abuse was established by asking two questions: 'H ave you ever
in your life been forced by anyone to have sexual intercourse?' and,
‘Has anyone ever forced you to perform or allow sexual contacts other
than sexual intercourse?' Fifty per cent o f subjects (79) reported having
been sexually abused at some time in the past. The sexually abused
wom en made their first suicide attempt earlier than the non-abused
wom en: when they w ere 27 years old, on average, compared to an
82
average age o f 36 years. The abused w om en also had almost double the
number o f previous suicide attempts.
The 79 w om en had often been abused many times: 57 by an unknown
man or boy, 30 by a friend or acquaintance, 27 by relatives other than a
stepfather or father and 12 by a father or stepfather. Four had been
abused by a professional care giver. Only 20 per cent o f these wom en
had been abused once. The vast majority had been abused repeatedly
by the same abuser or by multiple abusers. For 75 o f the 79, the first
suicide attempt took place after sexual abuse.
As other research has found, once a person has been abused they
often find themselves being victimized again by other men in later
childhood or adulthood. One wom an had been raped and physically
abused by six boys at the age o f 17 years. A year later she was raped and
abused by two unknown men. She made her first suicide attempt when
21 years, but found herself hardly able to talk about the rapes, except to
say that on the second occasion her body was mutilated by the men
using cigarettes. The suicide attempt was precipitated by an innocent
remark by another m em ber o f her family about an article o f their
clothing being destroyed by burning cigarettes. The memories came
rushing back and shortly afterwards she gathered all the medicines she
could find at home and swallowed them.
Particularly w orrying in van Egm ond’s study is the finding that
w om en who have been abused have a much greater probability o f
multiple suicide attempts later. They followed up the w om en for a
year, and found that significantly more sexually abused w om en (48 per
cent) had made further suicide attempts during this period than wom en
with no history o f sexual abuse (29 per cent). They found no difference
within the group o f sexually abused w om en between those w ho had
been the victims o f child sexual abuse (under 16 years) and those first
abused when 16 years or older. Similarly, and somewhat surprisingly,
there was no difference in the pattern o f suicidal behaviour between
those who had one sexual abuse experience and those who had suffered
multiple experiences. Neither did the history o f sexual abuse affect the
characteristics o f the suicide attempt: for example, the method used, or
the reasons to explain the suicide attempt, or even the suicidal intent.
This is an important finding. It undermines the argument that a
83
suicide attempt by a wom an with a history o f such sexual abuse is a cry

for help - the view that people with a history o f such experiences come
to believe that only an extraordinary measure such as a suicide attempt
will gain attention. They did find that w om en who had been sexually
abused had, as others had found, poor relationships with people around
them, problems with achieving a sense o f fulfilment from their lives
and more problems in integrating with others. The difficulties with
self-fulfilment suggest a reduction in ability to see meaning in their
lives.
The fact that these wom en also predict (accurately) that they will be
suicidal in the future is worrying for those interested in reducing suicidal
behaviour. The pattern they show is consistent with the pattern o f
learned helplessness, where a person believes there will be nothing they
can do to achieve fulfilment in their life in the future, or to stop nasty
things happening to them. Even if small unpleasant events happen in
their daily lives, they are likely to explain these in terms o f stable factors
(something that w on't go away) and global factors (something that
affects all areas o f life). For example, if they get into an argument with
a friend, they are likely to believe it was caused by their long-term
difficulty in making a relationship with that person, and this same
problem in making relationships will also affect their relationship with
everyone else, including future boyfriends and girlfriends, future
employers, and so on. In future stressful situations this pattern o f
attribution is likely to produce an overwhelm ing sense o f helplessness
and hopelessness, even when there might have been something the
person could have done.
This is consistent with the results o f a study that Raym ond Jack and
I conducted in which w e found that w om en who attempted suicide
were much more likely to attribute negative events in their lives in this
stable and global way. They were followed up a few weeks later after
most o f the upset m ood had subsided. Although the w om en had become
less depressed, angry, anxious and confused, there had been little or no
change in the pattern with which they explained negative events in
their lives. W hat is now clear is that many o f these problems stem from
experience o f child sexual abuse. Since van Egm ond’s study, other
research (for example, by Coll, Law and colleagues, by van der Kolk
84
and colleagues and by Romans and colleagues) establish beyond doubt

a strong link between attempted suicide and a history o f such abuse.6
Around 50 per cent o f w om en w ho have attempted suicide report that
they suffered serious sexual abuse when they were children (involving
attempted or actual penetration by someone at least five years older
than they were before they themselves were 13 years old).
Short-term vulnerability factors
The short-term vulnerability factors could be taken to include all those

factors in the current situation which, against the background o f the
long-term factors, put an additional burden on, or actively reduce, the
person's coping ability in the month prior to an attempt. W hereas both
suicidal and depressed patients have increased incidence o f life events
compared with controls, the suicidal group suffer a steep increase.
Additionally, there is an increased incidence o f physical illness (especially
in wom en). Given the preponderance o f sources o f physical and em o
tional stress, it is not surprising that 57 per cent o f attempters, including
82 per cent o f adolescent attempters,7 contact some helping agency
(most o f them the G P ) during this crucial month.
Investigators have also found an increase in disturbance in relation
ships and w ork during this phase. Both male and female attempters
report equivalent levels o f relationship difficulties, though more males
than females report significant w ork difficulties.
Employment status
A large study in Edinburgh, conducted by Stephen Platt,8 examined a
two-year cohort o f male attempted suicide cases, comparing the
employed (n = 158) with the unemployed (n = 199). The unemployed
men were less likely to be married, and less likely to live with their
family; they were more likely to be o f a low er social class, to have been
given a diagnosis o f abnormal personality, to misuse drugs, to be in
trouble with the police and to have a criminal record. This conclusion
has important implications for studies that have found a relationship
between unemployment and deliberate self-harm. Unemployment is
85
undoubtedly an important additional stress on people, but w e must be

careful to be sure that the relationship is not caused by other factors (e.g.
drug abuse) that predict both unemployment and attempted suicide.
Substance abuse
Keith Hawton, in studying how many o f the cases o f deliberate self-
harm, which his team see in Oxford City, also abuse alcohol or drugs,9
found that 41 per cent o f men and 21 per cent o f w om en were abusing
alcohol. Sixteen per cent o f men and 6 per cent o f w om en were
habitually abusing drugs. In an earlier study o f predictors o f suicide in
15-24-year-olds, substance abuse emerged as a key predictor o f suicide
following an earlier attempted suicide.10 Habitual abuse o f alcohol and
drugs provides the person with a readily available means o f overdosing;
it also decreases the sense o f risk in doing so. Finally, the abuse can
affect judgem ent so that normal ability to solve problems (which may
already be suspect in this group) becomes even more impaired.
Precipitating factors
Precipitating factors are those events which occur in the few days prior
to the attempt. Disharm ony with 'key other' people in the person's life
is the most com m on event; disharmony with relatives, anxiety about
work/em ploym ent, financial difficulties and physical pain or illness are
other reasons. One study11 found that almost half their female patients
had had a quarrel with their spouse or boyfriend in the w eek (mostly
48 hours) prior to the attempts. Some clinicians say that, for those who
harm themselves repeatedly, different types o f events precede self-
harm o f differing levels o f lethality. This suggestion warrants further
research.
Special dates
Days in the calendar that are special for some are likely to be the
most difficult for others. W hen everyone else appears to be enjoying
themselves, those who are depressed and hopeless are at their most
vulnerable. W e have already seen how suicidal feelings are sensitive to
86
interpersonal disruption: boyfriend/girlfriend disputes are one o f the

major precipitating factors. Putting these facts together, one might
predict that a date such as St Valentine's Day, 14 February, will be a
time o f particular vulnerability. T o examine this, data were collected for
cases o f deliberate self-harm in Birmingham, England, for 1983-8.12 The
numbers on 14 February for the six years w ere 11, 11, 1 0, 10, 11 and 16 (a
total o f 69). Comparative figures for 7 February in each o f the same
years were 3, 6, 3, 7, 5 and 5 (a total o f 29). St Valentine's Day doubled
the number o f attempted suicides! Furthermore, the proportion o f cases
that were adolescents/young adults (aged 12-20) was 45 per cent on 14
February, but only 17 per cent (the national norm) on 7 February.
Public Events
The death o f Diana, Princess o f Wales, on Sunday 31 August 1997 was,

for many people in the United Kingdom (especially those who had not
lived through the wars o f the twentieth century), one o f the most tragic
public events they had experienced. In the w eek following her death, a
study by Keith Hawton, Lewis Appleby and colleagues found that the
number o f suicide attempts rose by 44 per cent (especially in wom en,
where the increase was 65 per cent).13 Even more striking, in the four
weeks following the funeral, deaths due to suicides and open verdicts
rose by 17 per cent (34 per cent in wom en, and especially in wom en
aged between 25-44 years where the increase was 48 per cent).
Why do events have such a catastrophic effect?
Events involving a key person, such as arguments with spouse or

partner, are very com m on immediately preceding attempted suicide.
Such events are particularly associated with suicidal feelings in the
context o f an ongoing difficulty, like a poor relationship. Relative to
others, suicidal individuals have w eak social support systems. W hen
asked, they report interpersonal situations as their chief problems in
living. The family background may be disturbed and current living
87
conditions fraught with difficulties. Yet, when such people come for
help, they often blame themselves. It sometimes seems as though the
very act o f seeking help confirms to them that everything is their fault.
W hy else, they think, would they be needing help?
Someone w ho has undergone long-term stress m ay suffer from a
number o f psychological difficulties. N ot all who have suffered such
stress will experience these, and there m ay be some who have these
difficulties without their having been brought about by a difficult past.
But the association between such difficulties and suicidal behaviour is
well-established and m ay need very little added stress to precipitate
suicidal ideas and behaviour.14
Interpersonal problem-solving
Difficulty in solving problems involving relationships is an obvious

effect o f past interpersonal problems, and also the cause o f further
problems. T o investigate deficits in problem-solving, the most com
m only used measure is the M eans-Ends Problem Solving Test
(M E P S).15 The M E P S includes a number o f different social scenarios.
For each scenario, people are given some initial circumstances in which
a problem has to be solved (e.g. argument with boy/girlfriend) and a
positive outcome projected (the friend likes him /her again). The task
is to complete the middle part o f the story, providing different ways in
which the initial problem can be solved. The M E P S is scored for the
numbers and quality of'relevan t means' (problem-solving steps).
Using this test with psychiatric in-patients who had expressed suicidal
ideas and a group o f equally depressed but non-suicidal in-patients, one
study16 found that suicidal patients were able to provide few er than half
as m any ways o f solving problems as the non-suicidal. This result might
have come about because the M E P S items did not fit these people's
experiences, so the investigators also devised a modified test based on
patients' personal situations. Patients were asked to provide a personal
problem which had led to their being in hospital. Here, too, suicidal
patients were m ore ineffective, their suggested ways o f solving a prob
lem often being irrelevant to its solution. Interestingly, they judged
88
their proposed solutions as effective, but were realistic enough to note

associated drawbacks.
Even when suicidal people come up with solutions, these differ from
the type o f solutions given by others.17 People who have made a suicide
attempt are more passive (relying on others for solutions) and less active
in their problem-solving than either people w ho have had unenacted
suicidal thoughts or non-suicidal psychiatric in-patients. This is one o f
the few studies to show differences between those with suicide ideation
and those who go on to harm themselves. Others have found that
attempted suicide patients' solutions differ on a range o f other qualita
tive aspects, including showing more avoidance, being less versatile and
being less relevant.18
Given the increased number o f stressful events known to precede a
suicidal episode, the reduced ability to solve personal problems is
especially problematic. An increase in such stresses, combined with a
reduced ability to think o f steps to solve the problems generated,
increases the likelihood o f suicidal behaviour. As w e saw in the previous
chapter, the person feels they have no further ideas about what to do
over the problems in his or her life.
In suicidal adolescents, anger and the severity o f initial suicidal
behaviour predict repetition. Studies have shown where this anger m ay
arise: repeated suicide attempts by adolescents are m ore likely if the
young person has suffered loss o f a parent and is living outside the
parental home. Stress seems to be cumulative, building from early loss,
through family disruption to interpersonal disputes occurring closer to
the time o f self-harm. W hether or not all such family backgrounds are
this disrupted, it remains true that suicidal adolescents view their family
as disengaged and inflexible. H ow ever, it is difficult to determine to
what extent this view arises from the depression, anger and hopelessness
o f the adolescent w ho is suicidal for other reasons. Adolescents often
judge their ow n internal and external problems in ways that differ from
the judgem ent o f either their parents or a health professional.
Nevertheless there is sufficient evidence o f a link with objectively
identified loss in the background to raise concern. Studies that extend
the age range from 16 to 30 show that the greater stress experienced by
younger people combines with a greater exposure to others that have
89
self-harmed. One such study in T exas19 found that 40 per cent o f the
suicide attempters had a family m em ber w ho had made a suicide
attempt (as opposed to 8 per cent o f the control group). The deliberate
self-harm patients w ere also more likely to have been arrested at some
time, and more likely to have had a recent break-up with a girlfriend or
boyfriend (38 per cent v. 10 per cent for both comparisons).
W hat is the basis for poorer problem-solving by suicidal individuals?
A possible clue comes from research on problems such people have in
their memories o f events from their past (the main topic o f Chapter 10,
but relevant to mention here briefly). Patients who have recently taken
overdoses tend to remember their past in a summarized, over-general
way. For example, in response to a cue word such as ‘happy', an
overdose patient might say, ‘W hen I'm out with friends'; that is, their
m em ory response does not single out a particular event. In contrast,
non-suicidal people retrieve specific, datable events, such as ‘Last Friday
when I w ent out for a meal with friends.' Attempted suicide patients
prefer to recall a general description o f a class o f events. They stop
short o f retrieving a specific m em ory contained within that general
description, though after further prompting often recall such an event.
As w e will see in Chapter 10, getting ‘stuck' at the stage o f recalling a
general category o f events has implications for problem-solving.
Future-directed thinking and hopelessness
Suicidal individuals differ in the w ay they think about the future.

W hen they talk about the future, they use less elaborate descriptions.
Furthermore, they seem to thinkless far into the future, and psychological
testing reveals that they use few er future-tense verbs when asked to
finish incomplete sentences. Roy Baumeister calls such disengagement
from the future ‘cognitive deconstruction', and attributes it to an
attempt to avoid contemplating a painful future. In fact, Andrew
MacLeod, Gillian Rose and I have found that suicidal individuals do not
seem to spend time thinking o f negative things that might happen in
the future.20 Instead, their thinking has a marked absence o f any positive
events they are looking forward to. The ‘painful future' is not the
90
anticipation o f negative events, but rather a sense o f hopelessness arising

from the prospect o f few positive events.
This hopelessness about the future plays a central role in suicidal
behaviour. Many British and American studies report that hopelessness
is more closely related than depression to suicidal behaviour. People
seem able to bear depression so long as they are able to think the future
might improve, but if they begin to feel hopeless, the risk o f suicidal
behaviour rises. If a person harms themselves on one occasion, the
chance they will repeat the behaviour within the next six months is
higher if they are more hopeless at the time o f the first attempt. Even
more worryingly, such highly hopeless individuals are at greater risk o f
completing suicide over the following ten years.21
Most studies on hopelessness have relied on Aaron Beck’s Hopeless
ness Scale,22 a self-report questionnaire which measures global attitudes
towards the future, such as, ‘M y future seems dark to m e/ But what is
hopelessness? Our study examining the relative importance o f positive
and negative anticipation in hopelessness showed that attempted suicide
patients w ere less able to think o f future positive events, but showed
no difference from controls in being able to think o f future negative
events. Importantly, we also found that this problem with thinking o f
positive things to happen in the future was just as true for the immediate
future (the next day and week) as for the long-term future (a year and
ten years). The fact that highly hopeless people lack short-term routines,
as well as long-term plans and goals, is important for therapy. Further
research by Andrew M acLeod and colleagues has replicated this finding,
and shown the inability to imagine positive things that m ay happen in
the future is not simply due to the presence o f depression in attempted
suicide patients.23
Reasons for living
Some years ago Marsha Linehan and her colleagues developed a Reasons
for Living Inventory.24 Compared to both the general population and
psychiatric controls, attempted suicide patients endorse few er important
reasons for living. Some o f the main items from her scale can be seen
91
in Table 5.1. The importance o f knowing about such reasons for living
is that individuals m ay be very motivated to commit or attempt suicide,
yet not do so because they have reasons for staying alive. Reasons for
dying m ay vary independently from reasons for living, and both need
to be taken into account. There is particular danger when someone
who has long been suicidal suddenly finds he or she has lost their
reasons for living (e.g. their children, partner, religious faith).
W hy do suicidal people have problems thinking about positive things
that might happen in the future, which might give them a reason for
living? An obvious answer is that poorer circumstances and reduced
opportunities mean they actually have less to look forward to. Poorer
upbringing, stressful events, marital and family disputes, poverty and
unemployment all feature in the lives o f many such people. A poorer
outlook for the future m ay be realistic in many cases.
But our research has found that hopelessness adds significantly to
the burden. N ot everyone w ho has had such stress in their lives is
suicidal. People differ widely in the extent to which their mind translates
specific stress events into the general feeling that nothing can be done,
a sense o f helplessness that goes much wider than the original situation.
It is the over-generalization from one situation (that m ay indeed have
been impossible to do anything about) to other situations (that might
be solvable) which is at the root o f hopelessness. If a person is hopeless
that anything can be done about the second situation, they either do
nothing or they give up too early. There is then a self-fulfilling prophecy
('I said nothing w ould help, and nothing has') which serves only to
increase the sense o f helplessness about any new situations.
The result is that such people m ay disengage from thinking about
the future. This is adaptive in some ways. It m ay reduce how upset
they feel right now, but at a cost. They are less likely to become aware
o f any future possibilities for happiness or to make plans which could
bring about positive events. So whereas negative events appear to play
a major role in the onset o f suicidal feelings and hopelessness, whether
people are able to anticipate positive events is important in determining
how quickly they can recover from hopelessness.
92
Table 5.1 Sample of items from Marsha Linehan's Reasons for Living
Inventory
Survival and coping beliefs

I still have many things left to do.
I want to experience all that life has to offer and there are many
experiences I haven't had yet which I want to have.
I believe I can find a purpose in life, a reason to live.
No matter how badly I feel, I know that it will not last.
I am curious about what will happen in the future.
I believe killing m yself would not really accomplish or solve anything.
Responsibility to family
It would hurt my family too much and I would not want them to
suffer.
I have a responsibility and commitment to my family.
Child-related concerns
It would not be fair to leave the children for others to take care of.
I want to watch my children as they grow.
Fear of suicide
I am afraid o f the actual ‘act’ o f killing m yself (the pain, blood,
violence).
I am afraid o f the unknown.
Fear o f social disapproval

Other people would think I am weak and selfish.
I am concerned about what others would think o f me.
Moral objections
My religious beliefs forbid it.
I am afraid o f going to hell.
93
Emotional experience
There is a strong, convincing relationship between emotional experi

ence and suicidal behaviour. Interestingly, the emotions typical o f
individuals who attempt suicide m ay be different from those o f indi
viduals w ho commit suicide. W hile both attempted and completed
suicide are related to depression, parasuicidal behaviour appears to be
related to anger whereas suicide m ay be more related to apathy or an
absence o f strong emotions. Attempted suicide patients are more angry,
hostile and irritable compared to non-suicidal psychiatric patients and
the general population, both before the attempt and after. Often the
relationship with close family, friends and partners has deteriorated into
hostility, demandingness and conflict, especially in adolescence, as Alan
Apter's research in Israel shows.25 In contrast, persons w ho commit
suicide seem to have been less angry, appearing apathetic and/or
indifferent. Suicidal individuals appear unable to regulate their own
emotional responses or experiences o f emotional pain.
This does not mean that such suicidal behaviour is merely an
expression o f hostility or manipulation intended to harm or get back at
others. For many, the very experience o f intense anger, especially if it
is ongoing and they cannot seem to control it, is painful and intolerable.
In some cases, most notably in self-cutting (see below), parasuicidal
behaviour is itself used by the individual for emotion regulation. The
regulating mechanisms here m ay be several, e.g. sleep (following an
overdose), distraction from emotional stimuli or some sort o f direct
biological effect on the emotion system (e.g. by cutting, or ingesting
drugs).
Emotion regulation and self-cutting
Self-cutting as a form o f self-harm needs to be dealt with separately,

since people who harm themselves in this w ay appear to be a special
group. They are in the minority, accounting for about 10 per cent o f
parasuicide episodes. Even so, this means that in the United Kingdom
94
10,000 episodes o f self-cutting come to the attention o f accident and

emergency departments o f hospitals each year.
Although the cutting m ay not be serious, repeated attempts to
self-harm in this w ay may be associated with other personality problems:
impulsivity, instability both in emotion (especially anger) and in inter
personal relationships (experienced as very intense). Such individuals are
extremely sensitive to stress, especially that arising from interpersonal
problems. They are particularly prone to look for any signs that others
are abandoning them, and react to the situation as if the abandonment
were already complete.
If such unstable emotion and unstable relationships dominate the
person's life, they m ay be said to have a 'borderline personality’ . The
‘borderline’ category refers to those w ho show impulsive self-damaging
behaviour, unstable and intense interpersonal relationships, inappropri
ate and intense anger, problems with the experience o f self (frequent
crises o f self-identity), extremely unstable emotions, chronic feelings o f
emptiness or boredom and intolerance o f being alone. They are des
perate to have company, but m ay treat their companions with so much
intensity (clinging behaviour alternating with anger) that they are left
alone again very soon. Their physically self-damaging acts often occur
after dissociative states (e.g. out-of-body experiences, experienced very
negatively).
In the most extreme cases, all sense o f self (of being a voluntary
agent, having personal m emories and body) disintegrates. In the
resulting ‘out-of-body’ experience, the scene is viewed from another
angle as observer rather than as participant. Experiences o f alien control,
and o f hallucinations, m ay also occur following extreme or long-lasting
stress associated with such a failure to control emotion that it simply
escalates catastrophically. Such experiences occur spontaneously for the
first time under extreme stress (such as sexual abuse, often found in
the background o f the most severe cases). Once such a dissociative
experience has occurred, the threshold for the reappearance o f such a
phenomenon is lowered: it can be reactivated under conditions o f future
(and often lesser) stress.
People who cut themselves often report that the cutting is the only
w ay they know that stops the intense feelings they are experiencing.
95
One adolescent girl said it felt as if her head would explode unless she
did something to stop it. She knew that cutting would stop it and
put her back in touch with herself and her body. W hat explains the
catastrophic escalation o f emotion that often precedes self-cutting? T o
understand this w e need to understand that negative emotion often
results from breaking the rules one has set for oneself.
Consider the situation in which a person has been punished in the
past for the expression o f emotion. In relatively mild cases, this occurs
every time a parent tells a child to ‘pull yourself together'. The child
m ay learn the lesson, 'It is shameful to display your em otions/ Or, in
the most extreme case, a similar thing m ay happen in sexual abuse by
an adult, where a child has been threatened with punishment if he or
she becomes upset. In both the mild and the severe case, there is conflict
between the expression o f emotion and the suppression o f emotion to
avoid future punishment.
In later life, when the person feels emotional, the very fact o f feeling
emotion is breaking one o f their rules. But breaking the rules causes
more emotion, which then leads to attempts to suppress it. The result
is a rapid escalation o f negative emotion arising from such a feedback
loop: the expression o f emotion itself violates a goal (‘do not feel or
show emotion'), but the consequence o f violating any goal is increased
emotion. This leads to patients feeling upset but telling themselves they
should not feel upset (one o f the ‘invalidation strategies' that maintains
the very emotion it is designed to abolish, identified by Marsha Linehan
in her treatment for people who suffer from ‘borderline’ symptoms).
Linehan has also developed a number o f specific strategies for dealing,
in a non-harmful w ay, with the escalation o f emotion. As part o f her
therapy, she m ay advise a patient to hold an ice cube in each hand until
it melts. The pain o f the cold ice is often sufficient substitute for
self-cutting, but without the physical damage.
96
Concluding remarks
Several areas o f psychological functioning contribute to increased risk

o f non-fatal suicidal acts, often by adding to the burden o f already
stressful lives. Suicidal behaviour is associated with poor problem
solving ability, especially active problem-solving, and with difficulty in
recollecting events from the past in sufficient detail to help solve current
interpersonal crises. It is also associated with hopelessness, the inability
to imagine future positive occurrences (even for the next day, never
mind the following weeks, months or years). Finally, suicidal people
have difficulty regulating their emotion, a feature particularly strong in
those with 'borderline’ features, who m ay deliberately harm themselves
many times.
97
6
Rational Suicide, Euthanasia
and Martyrdom
In the last few days o f 1994, the State o f Oregon was blocked in its bid
to change the law to allow assisted suicide. From 8 Decem ber 1994, the
terminally ill w ere to be allowed to ask their doctors to prescribe a
lethal dose o f medication. The m ove, which followed a state referendum
held just over a year before, was always likely to cause controversy. Up
to the last moment, it looked as if Oregon's decision in favour o f assisted
suicide would stand. H ow ever, reports said that many doctors were
unprepared for the changes.
Agreeing with the m ove towards such a change, a 47-year-old man
suffering from Aids said, 'W hose life is it anyway? It should be between
me, m y God and m y doctor/ Elsewhere, a wom an w ho tw o years
previously had been told she would not live, argued against the measure.
She was now well again, but said that at the time she was so depressed
she would have asked a doctor to end her life.
The Catholic Church in the state vigorously opposed the change in
the law, and continue to campaign against it. They argued that it would
put the w eak and the vulnerable at risk. At least six other states
were considering similar legislation and were looking carefully at what
happened in Oregon.
In Britain it remains illegal for anyone to aid someone to take their
own life under any circumstances, but the courts are sometimes lenient.
In December 1994, the courts decided not to prosecute a man who had
administered a lethal dose o f morphine to his terminally ill wife. In this
case the Voluntary Euthanasia Society (Exit) were careful to distance
themselves from supporting the man. They pointed out that there was
98
Rational Suicide, Euthanasia and Martyrdom
no evidence that the wife had agreed to her life being ended in this
way.
Developments in this area are occurring at some speed. Euthanasia
was legalized in the Northern Territories o f Australia in 1996. Assisted
suicide was legalized in N ew York in April 1996 when two federal
appeals courts lifted the ban on doctors helping their patients to die.
Esther Fein o f The New York Times posed a stark question: 'W ill the
right to die becom e the duty to die?" Let us look more closely at the
arguments for and against.
The w ord euthanasia (from the Greek words eu and thanatos) literally
means 'easy* or 'gentle death’ . It was declared unethical by the W orld
Medical Association in 1950, and to understand w hy w e have to see
its background. T w o forms o f euthanasia are normally distinguished.
The first is compulsory euthanasia, or 'm ercy killing’ , done without the
consent o f the person concerned, and mostly applied to grossly
deformed children. For example, a Dutch gynaecologist in 1994 admitted
killing a three-day-old girl at her parents’ request. Her brain was only
partly developed, she had spina bifida and partial paralysis, her limbs
were malformed, and the doctor judged she faced a life o f constant
pain.
The usually cited problem with this is that it constitutes a 'slippery
slope’ and violates the fundamental right to life, this being backed
up by religious people’s belief in life as God-given. H owever,
a distinction is often made between an analgesic and lethal dosage o f
pain relief. An analgesic dose m ay have the unintended effect o f
shortening life, but a lethal dose has the direct intention o f shortening
life (the ethical doctrine o f 'double effect’ allows the first but not the
second).
The second form o f euthanasia is voluntary euthanasia. Here, indi
viduals in sound mind ask that their life should be ended in the event
o f their becom ing victims o f irreversible illness. Advocates (e.g. the
H em lock Society in the United States or members o f Exit) maintain
that, with safeguards, the law should permit it and people should be
supplied with the means to take their own life or a doctor should
be authorized to end their life provided the request is made before
witnesses.
99
The difficulty here is the assumption that w e have the right to decide
over our own life or death absolutely. The argument from these groups
suggests that w e ow n5 our own lives. This argument has a long history.
Historical background
Charles Gilden in 1695, searching to justify the death o f his friend

Charles Blount, attempted to put Blount’s death in the context o f Stoic
philosophy and dealt head-on with some o f the classical arguments
against suicide. Whereas Aristotle had argued that self-killing injured
the state and was illegal even when no decree specifically forbade it,
Gilden denied that society had the right to stop a man killing him self
any m ore than it could rightfully prevent him emigrating. ‘N ow if I can
leave any one particular Body Politick I have the same right to leave
another and so on through all those o f the W orld, and then by conse
quence I offend not, if by m y Death I take m yself aw ay from all/
Pythagoras’ view was that men were put on earth like soldiers at
their post. It was therefore ‘desertion’ to end one’s own life. Gilden
points out that, first, a simile is not an argument. Second, the simile is
inexact since soldiers enlist willingly in the army in which they serve.
Human beings do not choose to be bom .
In the Middle Ages, Aquinas (1225-74) argued that God had instilled
in humans an instinctual desire to preserve their ow n lives. Thus suicide
violated natural law. Gilden replied this was not so. Rather, suicide was
consistent with ‘the precepts o f Nature and Reason’ ; Hum an nature
cannot desire the continuance o f pain and suffering, so a life o f such pain
should be allowed to be ended. T o end one’s life in such circumstances is
thus a greater good than self-preservation. It cannot therefore be con
trary to natural law.
The arguments against suicide eventually lost their pow er to stop
the inexorable march o f change towards the secularization and medic-
alization o f suicide. The arguments o f the philosopher David Hume in
his essay ‘On Suicide’ , written in the early 1750s, were particularly
telling. The essay was suppressed and Hum e was forced to withdraw it
from a print run o f the Five Dissertations in which it was to appear.
100
Nevertheless it did circulate and proved to be a robust statement o f the

secular view.
Hume argued:
(a) that in as much as one believed in the providence o f a Deity,

this 'providence appears not immediately in any operation, but
governs everything by those general and immutable laws that
have been established from the beginning o f time’ ;
(b) human life and death do not depend on God, but rather on
'the general laws o f matter and motion';
(c) it is not valid to argue that men and wom en are 'ow ned' by
God, since people come into being through natural processes o f
reproduction. God no m ore owns us than do our parents;
(d) people w ho commit suicide cannot break the laws o f nature
since all our powers (including those that em power us to take
our ow n lives) are natural faculties. So, even employing them to
kill oneself, one cannot 'encroach upon the plan o f his providence
or disorder the Universe'.
Finally, Hum e conceded that humans display a 'natural horror o f

death', but this does not thereby imply that such horror represents a
'general law o f matter and motion'. Although such a horror might
explain w hy suicide is not more common, it does not have the authority
that natural-law theorists such as Aquinas implied.
Hum e was consistent with other thinkers o f the eighteenth century
w ho wished to base their ethics on social rather than religious grounds.
In this respect he needed to meet head-on the argument that somebody
w ho commits suicide commits a w rong act because they let down
society. Hume argued that 'a man w ho retires from life does no harm
to society; he only ceases to do good; which, if it is an injury, is o f the
lowest kind'. The important calculation is to balance the good to society
against the harm the suffering person sustains by continuing to live.
W here the suffering person has grow n so infirm as to become a burden
to society, Hume concludes that suicide is a service which has benefits:
'resignation o f life must not only be innocent, but laudable'.
W hat would Hume, as an empiricist, have made o f the evidence
101
which has now emerged about the aftermath o f suicide? In many

instances suicide appears to have such a devastating effect on the
survivors that, if it were only the balance o f suffering one w ere examin
ing, one could conclude that the suicide was not ethically justifiable. By
confining the argument to people under unbearable suffering, one m ay
to some extent clarify the issue about the ethical dilemma, but this m ay
not thereby be taken to imply that it covers all instances o f suicide.
Anxiety about assisted suicide also has a long history. Indeed, we
might say that the field o f medicine was founded on the need to separate
life-enhancing treatments from more superstitious beliefs that would
allow the person to die in some circumstances. The Hippocratic oath
arose from a situation in which Hippocrates and his disciples w ere not
clearly distinguished as physicians from magicians and witch-doctors.
Their oath that they w ould use treatment to help the sick according to
ability and judgem ent, that they would never use it to injure people or
w rong them, was devised at a time when the blurred distinction
between physician and magician meant that these people could and did
kill with impunity.
Physicians are taught to cure and to relieve suffering; psychologists
are taught to help people find their w ay through the intolerable burdens
o f living. But we have no ultimate pow er over our patients and clients.
In most cases the law will not require people to accept medical or psycho
logical treatment against their wills. Provided a sufferer has received
information from a physician and can make an informed consent, there
is no legal obligation on a person to accept medical treatment.
Recent exponents of Hume's position
In our own day, there are many people who, standing in the tradition
o f David Hume, take a more liberal view than the original Hippocratic
oath envisaged. Jack Kevorkian is such a person. He began to assist
terminally ill people with suicide in the United States in 1990. The
court’s reaction at first was to prosecute for murder, but the ju ry refused
to return such a verdict. The Michigan state senate then passed a
temporary Bill making it illegal to help anyone commit suicide. Sub
102
sequently, three local judges and a Michigan Appeals Court panel

declared the ban unconstitutional. In Novem ber 1994 Kevorkian assisted
with his twenty-first suicide hours after the temporary law lapsed. By
March 1996, he had assisted in 27 suicides. The state senate has not been
prevented, however, from voting overwhelm ingly to introduce a new
Bill to outlaw the practice o f assisting suicide.
Another who stands in the Hum e tradition is Derek Humphry,
author o f Final Exit and executive director o f the H em lock Society in
America since 1980. The H em lock Society, which believes in voluntary
euthanasia, has 40,000 members in the United States. Derek Hum phry
reasonably points out that it is no good for psychologists and psy
chiatrists to say that all 40,000 members are mentally ill or depressed.
H em lock members believe it is a basic civil liberty to be able to end
their lives at the point they choose. They are offended by the thought
o f having to make their exit in a secretive and ultimately violent
m anner.1 (H owever, this is expressed as a choice o f euthanasia versus
suicide rather than o f euthanasia versus another form o f living.)
Derek Hum phry suggests that euthanasia may actually extend the
life o f som eone who, on hearing a bad diagnosis, gets depressed and
commits suicide prematurely. With the option o f lawful medical eutha
nasia, they would realize that they need not end up a vegetable and
might hang on longer. He claims that in the Netherlands an estimated
4 per cent die from euthanasia and the families are involved and
consulted. Many have found it hard to disagree with Hum phry's con
clusions in many individual cases. He goes further and says that, if the
law continues to forbid assisted suicide, it is putting an intolerable
burden on the family.
Institutional support for assisted suicide:

the Netherlands
The Netherlands is one country where doctors are given greater free
dom to decide. In 1995, television audiences throughout the world were
able to watch a Dutch doctor featured in a television documentary,
Death on Request. Dr Wilfred van Oyjen carries out this ‘good death’
103
three or four times a year. The film featured a man who had developed
Lou Gehrig's disease, which brings about incurable degenerative wast
ing o f the muscles. His feet and legs became paralysed first, then his
right shoulder and arm, then his face. He and his family realized that,
unless some decisive action was taken, he faced death by suffocation,
as the weakening muscles o f his chest finally collapsed.
His requests for euthanasia were repeated as the guidelines in H ol
land prescribe (a voluntary, well-considered and lasting request to die).
Dr van Oyjen is one o f those w ho feels it w ould be letting his patient
down to refuse his request. The guidelines also prescribe that there
must be no other solution acceptable to the patient, that the time and
manner o f the death must not cause unnecessary suffering to others
(such as next o f kin) and the doctor must prescribe and administer the
right drugs.
This contrasts with the situation in the United Kingdom, where it is
accepted that doctors often use the 'doctrine o f double effect' to deal
with situations o f unbearable suffering. This permits doctors to use
drugs in sufficient quantities to relieve suffering, even if that hastens
death, so long as they do not intend to kill. According to the Voluntary
Euthanasia Society (Exit), one survey o f British doctors showed that 50
per cent had been asked by a patient for help to die. O f these, one-third
o f the doctors said they had complied.
This is not ethical, according to D r van Oyjen. Such surreptitious
euthanasia can take several weeks, and is dishonest. It also means that
the patient can die in delirium brought about by huge doses o f m or
phine. 'I am giving people the possibility to make choices. W hat kind
o f quality o f life, and death, do they want? Death is not always awful.
With a good doctor, death can be faithful, like a good friend.' Perhaps
most importantly, the doctor can choose the right moment, when he
judges that the person and his family are most at peace with the ending
o f life. Wilfred van Oyjen will take hours, days if necessary, waiting for
the right moment after the decision has been taken. He feels it essential
to get close to the fam ily if a 'good death' is to be brought about.
W hen the arguments are expressed in this way. they seem very
compelling. Yet many have continued to maintain that assisted suicide
is either intrinsically w rong or socially undesirable.
104
Arguments against euthanasia
Consent
One argument against euthanasia is that a terminally ill patient may
not be in a state, at the point near death, to be able to withdraw consent.
Suppose someone was to agree, before witnesses, family and doctor
that, in the event o f breakdown o f bodily function, active steps should
be taken to end their life. Once they have deteriorated, how could they
ever withdraw that consent should they change their mind? The right
(at any time) to withdraw consent from or change one's mind about a
medical procedure is seen by many to be paramount. Yet, with the very
frail, this right m ay be denied because o f their frailty.
On the other hand, this raises the question o f the prolongation o f
life, and with it a number o f imprecise issues. Should we withhold or
continue efforts to prolong life, for example by using m odem life-
support equipment? Under these circumstances, ethical theories usually
refer to the ordinary/extraordinary principles. That is, there is an obliga
tion to use ordinary means but no strict obligation to use extraordinary
means to prolong life. Extraordinary means refer to those means which
do not offer a reasonable hope o f cure or remission or those which
require excessive hardship to obtain. So most ethical theories assume
there are limits to preserving life, but that there is a presumption o f a
duty to preserve life in most circumstances. Ethical principles need to
be brought to bear.
The first principle cited is usually that o f autonom y and individual
conscience. The patient's attitude is primary, but if the patient cannot
be consulted, there is a secondary role for others on the grounds o f
principles o f beneficence and social justice. Proxy decisions are therefore
possible, indeed necessary in the case o f life-support machines. In these
cases there is a rank order o f people to be consulted, the family, the
physicians, a designated committee and the courts (in that order). These
make the decisions based on ‘substitutive judgem ent’ standards, which
involve asking what the patients themselves would have decided, based
if at all possible on the patients’ own expressed wishes.
105
Slippery slope arguments

>
The W orld Medical Association's decision in 1950 to declare euthanasia
unethical was made on the grounds that it would undermine trust
between patient and doctor, and between patient and family just when
they are at their most vulnerable. For every time a doctor assists
someone to die to ease their burden, or that o f their families, there
arises the possibility that someone, som ewhere will seek the death o f
someone for the w rong reasons. This is the slippery slope argument,
which asserts that a small change to allow assisted suicide would have
w ider and undesirable consequences. The first o f the tw o most often
cited undesirable consequences is where old and infirm people are seen
as a burden on their families, even more urgent at a time when there
is much in the news about old people being a burden on society because
o f increased life expectancy.
It is not possible to speak o f general policies o f euthanasia without
this being raised as a specific issue in m any individual households. Talk
o f euthanasia in a household where there is an older relative could fuel
talk about them being a burden. In his article on euthanasia,2 Hum phry
quotes the ex-govem or Richard Lam m o f Colorado, w ho made a speech
about the rationing o f healthcare costs, suggesting the elderly took
m ore than their fair share. Hum phry quotes Lam m to show that
economics will force this on the public agenda sooner rather than later.
But many would consider that where economic considerations are
brought in is when the euthanasia argument is at its weakest.
The second most often cited undesirable consequence is that people
w ho are depressed will make what appears to be a rational request to
end their life. In a controversial case in 1994 in the Netherlands, a doctor
assisted a wom an to die who appeared, on all accounts, to have been
depressed. This raised an important issue. People m ay consider it is
never justified to assist the suicide o f someone w ho has a condition, a
symptom o f which is itself a wish to die. Nevertheless, the Dutch
Supreme Court in 1994 ruled that, in exceptional cases, physician-assisted
suicide might be justified for patients with unbearable mental suffering
but without physical illness.
Are such slippery slopes likely to occur? Those who argue against
euthanasia say the 'slippery slope' is already taking place in the
106
Netherlands. Guidelines issued for doctors are being overstepped, they

claim, second opinions are regularly abandoned, and when a euthanasia
death is reported, there is rarely an investigation. Indeed, those who
oppose euthanasia allege that many euthanasia deaths are not recorded
even on the death certificate, all evidence in these matters falling under
the control o f the physician. This, it is suggested, does not provide
satisfactory safeguards against abuse. A study o f these issues in 1997
attempted to estimate the number o f assisted suicides in psychiatric
practice in the Netherlands. H alf o f all Dutch psychiatrists participated
in the study (n = 552). Thirty-seven per cent (205) had received requests
for assisted suicide during the period o f the study, though the m ajority
had not complied. Their study concluded that around 320 requests are
being made each year, and these result in between two and five assisted
suicides being carried out.3
No man is an island
Another argument against ‘rational suicide', assisted or not, is the ‘no
man is an island' argument (it is ironic that the quote comes from a
meditation by John Donne, who believed that suicide was justifiable
and argued so strongly in Biathanatos).
One version o f this argument is that by ending our life w e deny our
family and society the opportunity to fulfil their duty o f care to us.
Another version is m ore simple: those w ho would justify committing
suicide on the grounds that their life is their own property are failing to
take into account that each o f our lives is shared by parents, brothers
and sisters, partners, children, neighbours and friends, colleagues at
w ork and so on. Not to take account o f their reaction to our death, or
worse, to make assumptions (e.g. that they would be relieved o f a
burden) without checking, is hardly to take a ‘rational' decision.
In what w ay does suicide affect others? Research shows how suicide
affects family, friends and community. It can be the worst o f all deaths
in its impact on survivors. It causes grief they m ay never resolve, guilt
in a w ay no other death does, even raising the risk o f suicide in others.
A truly rational approach to self-killing or assisted suicide must take
account o f all these things. If a rational person says, ‘But no one would
mind if I ended m y life\ then it will be inappropriate to call it rational
107
unless they have sought evidence. It is irrational to make a ‘mind-reading

error , to presume opinions without checking.
Perhaps, though, in the case o f assisted suicide, especially o f people
who belong to societies such as Hemlock, the family are fully involved
so the death has less impact. H ow ever, Hum phry him self cites an
85-year-old wom an, a m em ber o f Hemlock, whose husband had died
the previous year. W hen she took an overdose and died, a neighbour
phoned him in distress, and H um phry felt he was able to reassure him.
H ow ever, tw o years later, when he was giving a talk in Hawaii, a young
man asked his views on suicide in the elderly and turned out to be a
grandson o f the wom an. At a meeting later in Dallas, a wom an asked a
similar question and turned out to be a daughter. Both said they had
come to the H em lock meetings to find out more about the organization
to which their relative had belonged. Hum phry claims they found it
cathartic to share their experience, since ‘they were still com ing to
terms with the manner o f the death'.4
This story supports the idea that such a death never occurs in
isolation but always has knock-on consequences. It is, as one physically
handicapped campaigner against euthanasia put it, like a delicate game
o f pick-up-sticks. You carefully pick up another stick, hoping not to
disturb the intricate web, but just when you think yo u ’ve succeeded,
your independent action ends up jiggling the fragile balance and every
thing falls apart.
Rational versus depressive suicide: Arthur Koestler
Perhaps the most difficult dilemma is to distinguish between those who

ask for suicide because they are depressed, and those who ask for it
when ‘o f sound m ind\ Take one often cited example o f someone who
decided rationally to kill himself: Arthur Koestler, w ho died by suicide
with his wife Cynthia in 1983. Because he was vice-president o f the
Voluntary Euthanasia Society (Exit) and w rote a preface to its publi
cation, Guide to Self-deliverance, he is com m only said to be the paradigm
case o f rational suicide.
Yet in his ow n prolific writings he shows that he suffered much from
108
depression, punctuated by bouts o f mania. His uncle had committed

suicide, and in his autobiographical book, Arrow in the Blue, he tells o f
his ow n troubled life.5 He was an only child, bom when his mother
was 35 years old. She was very possessive, yet capricious in her moods.
There were often abrupt changes from effusive tenderness to violent
temper. He talks about being tossed from 'the emotional climate o f the
Tropics to the Arctic and back again’. Plagued by guilt and often feeling
he deserved to be punished, he felt bewildered and rejected, and suffered
suicidal depressions and recurrent suicidal fantasies. He attempted
suicide twice, once at 29 years old by coal gas, the second time in prison
in Lisbon by an overdose.
At the end, suffering from leukaemia and Parkinson's disease, he
killed him self and his wife did likewise. H ow ever, one w eek before the
suicide, Cynthia had told a friend that Arthur was having hallucinations.
Cynthia herself had a family history o f suicide, her father having
committed suicide when she was 10 years old. It is also clear from her
writings that she had seriously contemplated suicide before. Professor
Robert Goldney o f Adelaide University, in a seminal paper in 1986,6
points out the irony that, had Koestler owned a copy o f the Guide to
Self-deliverance in the 1930s, the world would have been denied the vast
majority o f his writings.
Rational and irrational: can a distinction be made?
Joseph Richman, Professor Emeritus at the Albert Einstein College o f

Medicine, N ew York, argues that to drive a wedge between rational
suicide and other forms o f suicide is illusory.7 All suicides have features
in common. There is usually evidence o f a crisis and often the confron
tation o f loss and separation; there is the role o f others. T o make a
dichotomy between the sick and the healthy obscures the issue. W hen
there is a crisis around illness, loss and death, all unresolved problems
and conflicts o f the ill person come to the fore and can be seen as
overwhelming.
Many believe that being terminally ill is itself sufficient reason to
want to die. The terminally ill find that their depression is often seen as
109
natural and therefore goes untreated. Yet deep sadness is natural, while
depression is more than sadness. It brings with it a sense o f guilt, o f
being a burden, o f worthlessness and a feeling o f failure. Further, the
person often believes that no one will want to see them in their
condition, so there is no point in contacting anybody. They m ay feel
guilty about things they have done, or about situations or relationships
that remain unresolved, and the depression will ensure ‘it is not worth'
trying to resolve them.
Is there a role for therapy here? The evidence suggests there is.
People can, in fact, overcom e their depression even within the context
o f a terminal illness, evidence the hospice m ovem ent has been citing for
some years.8 Do people seeking assisted suicide need a psychotherapist?
Proponents o f rational suicide discourage people from seeking help on
the grounds they are not mentally ill. This is probably right, but you
don't have to be crazy to see a therapist any more than you have to be
stupid to go to school.9
All wishes to die represent in some degree a problem perceived as
insoluble except through death. In that situation one needs to respect a
person's autonomy. But personal autonom y should not be confused
with isolation and loss o f social cohesion. H ow can we judge whether
or not someone is in control o f their own actions? Depression leaves
little room for argument. W e know it affects chemical pathways in the
brain and results in lack o f energy, lack o f pleasure and lack o f interest
in social activities and social contacts. It also results in recurrent thoughts
o f death or suicide. Com pounding this is the sense o f hopelessness that
occurs in many depressions: a foreshortening o f the future and an
inability to see anything to look forward to. Depression therefore leaves
little choice. It pushes the person, impels them to certain actions. It is
not open to reasonable argument.
Proponents o f rational suicide say that people have chosen this
option and are not depressed. In that case they must expect to be able
to discuss rationally the pros and cons o f their actions. They must also
expect that their reasons will be questioned. The major presupposition,
that each o f us has an overall right to decide on our ow n life, can be
questioned, and only where the full social and family context is taken
into account can it ever be justified deliberately to shorten a life.
110
Martyrdom
The definition o f martyrdom is closely tied to the group that shares the
belief o f the one w ho dies. A suicide-bombing by an Islamic fundamen
talist m ay be called martyrdom by his or her own group, but is to the
non-Muslim m erely terrorism. Each group, religious or political, has its
martyrs. They differ in a number o f respects from the types o f suicide
so far considered.
Most people who commit suicide are depressed when they do so;
they see death as the end to their suffering. One o f tw o feelings usually
predominates in the mind o f the person w ho is suicidal in this depressive
sense, both stemming from hopelessness. The first is that they have
been abandoned by everyone; the second that they are a burden to
everyone, especially to those they love. Contrast this with the martyr.
They see hope, and believe in a cause. Although there are m any different
contexts for martyrdom, all martyrs believe that by their death they are
bringing about some combination o f the following gains:
(a) advancing a cause by inspiring members o f their own group;

(b) harming the enemy;
(c) delivering others from suffering (but not suffering that has
been caused by them, as in depressive suicide); and
(d) gaining entrance to an afterlife.
In this sense, martyrdom is a paradigm o f Durkheim 's altruistic

suicide. These people kill themselves because they are totally submerged
in their groups. Death for them is a duty. They are seen as making
heroic sacrifices for the group or its leadership. In altruistic suicide, the
individual's life is not their own property. His or her own goals o f
behaviour emanate from the external source, the group.
Is there a distinction between the suicide bom ber and others who
put their lives at risk for what they believe in? There seems to be a
continuum between altruistic suicide and other forms o f altruism in
which a person risks his or her life for their group and is prepared to
accept death as the unavoidable consequence o f performing some act
111
o f bravery, charity, justice, m ercy or piety. The difference is that, in

these cases, the person does not will their own death. If there w ere any
other w ay, they would take it.
Are suicide bombers, 'm artyrs' to their ow n cause, psychiatrically
disturbed - or is it simply belief in a cause that drives them on? Altruistic
suicide is not performed through madness - the person, as far as we
know, is in full possession o f his or her faculties. Indeed, it w ould not
have the impact it clearly does on the group if it w ere suspected that
the person was mentally ill.
Is martyrdom always associated with religious faith? Belief in God is
clearly an important component o f some acts o f martyrdom in that it
provides a moral context. Most martyrs m oreover have a belief in an
afterlife, even when the belief in a god is not conventional. Those w ho
took part in the mass suicide in Jonestow n, Guyana, in N ovem ber 1978,
led by Jim Jones, believed in reincarnation. But they also believed that
the revolutionary principles they stood for would achieve immortality
through their deaths. Jones said, 'W e are not committing suicide; we
are committing a revolutionary act/ The followers o f David Koresh
had also planned a mass suicide, with hand grenades, if he died during
the battle with law-enforcement officers at W aco, Texas. One witness
at the inquest said that following Koresh's death, 'W e were all to be
translated . . . come out o f our bodies and go to heaven.'
But not all the deaths that occurred or w ere planned in these contexts
were suicide. Some people in Jonestow n were murdered. Other suicides
there appear to have been not so m uch altruistic as fatalistic. Many
w ere malnourished after a poor diet, and their emotional disturbance
was heightened because the com m unity thought it was under attack.
In a chilling parallel, a witness from W aco said that if any follow er had
been afraid o f killing themselves, another could do it with a shotgun.
The kamikaze
Learning about the mind o f the m artyr seems impossible since they are
not here to study. H ow ever, there are some 'wartim e martyrs' who,
though they volunteered to die, escaped death: Japanese kamikaze
112
(literally ‘divine tempest') pilots who survived when their planes were
shot down on their w ay to a target in the Pacific and were awaiting
recall to another suicide mission when the w ar ended.
By the last year o f the war, the US N avy thought itself invincible in
the Pacific, but the kamikaze shook its confidence. So keen were the
young Japanese pilots to participate in this ‘spirit o f the Samurai' that
some signed their papers in their own blood to try and increase chances
o f selection. By the end o f 1944, over 500 kamikaze missions had been
flown against the US Fleet, but Japan was still losing. A new elite,
T h e Thunder Gods’ , was formed to pilot aircraft equipped with large
missiles. These turned out to be too cumbersome and the pilots were
shot down before they could reach their targets. The Thunder Gods
were therefore asked to undertake a different sort o f mission: to take
ordinary planes and crash them into the fleet.
On 1 April 1945 the United States invaded Okinawa, and on the 6th,
after an ominous silence, the US Fleet was attacked by up to 350 planes
at a time. By the end o f June, after about 2,000 kamikaze missions, 36
Allied ships had been sunk, ten times as many damaged; 500 Allied
personnel were dead and 4,800 wounded.
What was the motivation o f the kamikaze pilots? By the accounts o f
those who remain, and their officers who assigned them, they had an
attitude to death characteristic o f many martyrs. First, they believed
that obedience gave meaning to self-sacrifice. They had come to the
point when to give up one’s life for one’s country was the highest
honour. Second, they believed they were not going to die at all. Their
souls would go to Yasakuni, a special place for those who die fighting
for their country; a special place also reserved for the Emperor. As one
said, ‘Even if you did die, you felt you would still be alive to describe
it.’ The widows o f these men paid their respects and went to honour
their ‘god-like’ husbands.
But it wasn’t always quite so straightforward for the pilots. People
who interviewed them subsequently found that a different, more
sombre mood sometimes emerged. Some appear to have brooded
about their death, complaining that the waiting went on for ever.
Others said it felt as if they were being sentenced to death. Still others
questioned whether they had been in their right mind when they
113
volunteered. They rejoiced for the camera, but harboured doubts off
camera. They dealt with such feelings by reattributing any emotions to
things other than the fear o f dying. One pilot said that at the time he’d
find him self crying in the night, but not because he was frightened. He
couldn’t say he really ever knew w hy he was crying. Perhaps it was
because most kamikaze pilots never reached the age o f 20 years.
Concluding remarks
Can suicide be rational? There is little doubt it can. Durkheim argued

strongly in favour o f this position in 1897, and there is no reason
to disagree. Many endorse the aims o f those societies that promote
euthanasia as a rational, compassionate end to life. Many others have
no difficulty accepting that there are situations, as in wartime, when
there is a very thin line between those prepared to put their lives at risk
for the good o f their country and those who go out on a suicide mission.
As soon as a society, a nation or a subculture gives permission to its
members to kill on its behalf, it will find many volunteers, and, as we
shall see, violence against another and against oneself have similar
psychodynamic and biological roots.
114
7
Psychodynamics, Biology
and Genetics
There is no one reason w hy people kill themselves. Suicide is the

final common pathway for many human problems. But there are also
biological factors, and some o f these may have a genetic component.
I shall now build upon the discussion o f psychiatric, social and
psychological perspectives on suicide (Chapter 3) and attempted suicide
(Chapter 5) to review psychoanalytic and biological perspectives.
Psychoanalytic perspectives
The starting point for psychoanalytic w ork on suicide and attempted

suicide is Freud’s 1917 paper, Mourning and Melancholia, the aim o f which
was to draw a comparison between severe depression and the normal
experience o f mourning following loss. In suicide, the life instinct is
overcom e by more powerful forces, leading to a reaction that is an even
more extreme form o f the self-deprecation found in depressed people.
Mourning is a normal means o f coping following loss o f a loved person
(the ‘loved object’). H owever, people who have lost others on whom
they depended too much for their own sense o f self (a ‘narcissistic
object-choice’) find the experience o f loss impossible to tolerate. The
anger o f normal grief is, for these, a rage o f murderous proportions.
Identification with the lost object (internalization) means that the object
that is to be murdered becomes part o f the ego.
115
If one listens patiently to a melancholic’s many and various self-accusations,

one cannot in the end avoid the impression that often the most violent o f
them are hardly at all applicable to the patient himself, but that with in
significant modifications they do fit someone else, someone whom the patient
loves or has loved or should love . . . So we find the key to the clinical picture:
we perceive that the self-reproaches are reproaches against a loved object
(individual) which have been shifted away from it on to the patient’s own
ego . . .
There is no difficulty in reconstructing this process. An object-choice, an
attachment o f the libido to a particular person, had at one time existed; then
owing to a real slight or disappointment coming from this loved person, the
object-relationship was shattered. The result was not the normal one o f a
withdrawal o f the libido from this object and a displacement o f it on to a new
one, but something different . . . it was withdrawn into the ego. In this way,
the object loss was transformed into an ego loss.1
Since outwardly directed reproaches towards the lost object are not
possible, they turn into self-reproaches and the wish to harm oneself.
W hen the rage is sufficiently intense, it will lead to strong urges towards
self-destruction.
If the love for the object - a love which cannot be given up though the object
itself is given up - takes refuge in narcissistic identification, then the hate comes
into operation on this substitutive object, abusing it, debasing it, making it
suffer and deriving sadistic satisfaction from its suffering.
It is this sadism alone that solves the riddle o f the tendency to suicide which
makes melancholia so interesting - and so dangerous . . . No neurotic harbours
thoughts o f suicide which he has not turned back upon himself from murderous
impulses against others . . . The ego can kill itself. . . if it is able to direct against
itself the hostility which related to an object and which represents the ego’s
original reaction to objects in the external world.2
In a later discussion o f suicide, Freud introduces the concept o f the

death instinct (Thanatos). He defined this as a drive, commonly seen in
nature, to reinstate the form er state o f affairs, the return o f all organic
or living matter to its inorganic unorganized state. This view sees life
as a preparation for death, with the death instinct as a drive to its end.
116
Psychodynamics, Biology and Genetics
Freud uses the concept to explain w hy the super-ego develops such a

harshness towards the ego in melancholia.
We find that the excessively strong super-ego which has obtained a hold upon
consciousness rages against the ego with merciless violence, as if it had taken
possession o f the whole o f the sadism available in the person concerned.
Following our view o f sadism, we should say that the destructive component
had entrenched itself in the super-ego and turned against the ego. What is now
holding sway in the super-ego is, as it were, a pure culture o f the death instinct,
and in fact it often enough succeeds in driving the ego into death.3
At times Freud expressed uncertainty about the status o f the death

instinct he had invoked, saying he did not know how far he believed in
it and would not seek to persuade others. H owever, he continued to
find it a useful idea, eventually changing the concept o f masochism
from being simply the opposite o f dominant aspects o f sadism to a
primary instinct in its own right; a basic drive towards subjection and
submission, and in the most extreme form a drive to embrace death.
(Chapter 9 will show how many o f the same ideas have emerged in
sociobiology.)
Later psychoanalytic writers have built upon Freud’s observations.
Menninger asserted that every suicide had three elements: the wish to
kill (originating in the death instinct), the wish to be killed (originating
in the super-ego’s masochistic need for punishment), and the wish to
die (originating in a relatively fundamental desire to return to the
womb). Such approaches emphasize the aggressive nature o f suicide,
but careful attention to the fantasies o f suicidal patients shows that
aggressive impulses are not always the most prominent.
Patients speak about a wish to join a dead relative with whom they
identify strongly. These reunion fantasies are not readily explained
simply on the basis o f self-directed aggression. They appear to result
from libidinal rather than aggressive wishes, pleasurable rather than
masochistic fantasies. Other patients have fantasies about rebirth follow
ing their own death, a rebirth they can hasten by destroying this self
(compare the examples o f martyrdom in Chapter 6). Still others appear
to be looking for a sense o f mastery, or even omnipotence, over an
impossible situation, and their death is the only element in life that they
117
feel they have retained control over. Some believe they will still, in
some sense, be present after their death to ‘see what happens next’,
and, in a spirit o f revenge that is sometimes present as part o f suicidal
thoughts, able to experience pleasure in the distress their death might
cause others. The inconsistency between death as escape into oblivion
on the one hand and a continued existence in which the pain o f others
can be experienced on the other is not apparent to the suicidal person
in their confusion.4
Object-relations approach
Freud had suggested that suicide was an attack on the love-object, now
lost, that had been internalized. Melanie Klein extended these ideas,
suggesting that the person was motivated by the wish to preserve the
good aspects o f the internalized object (now a valued part o f the self ).5
The inwardly directed attack was therefore aimed mainly at the bad
part o f the object. Since aggression mainly arose from the death instinct,
there was a danger that such destructive forces might destroy a good
object. Guntrip developed these ideas further, distinguishing between
depressive suicide and schizoid suicide.6 Depressive suicide comes about
as the result o f hatred redirected towards the self from the hated and
loved object (as Freud had argued). By contrast, schizoid suicide comes
about when the person finds themselves facing the loss o f self, and
begins to fantasize about death as a path to rebirth.
Early in a child's development, it is in a symbiotic relationship with
its mother and the child’s idea o f self and o f mother are not yet
differentiated. Later there will come inevitable separation and individu
ation, where self and object are differentiated. W hat happens if this
transition is not completed satisfactorily? One outcome might be that
the person will tend to become involved in relationships where the
other party is not treated as a separate and unique individual, but as
part o f the self. This choice o f friend or partner has been called ‘symbiotic
object-choice' corresponding to the ‘narcissistic object-choice’ in Freud’s
writings.7 Suicide, according to this analysis, involves not only ridding
118
self o f bad internal objects, but the fusion with the earlier 'symbiotic
mother' o f pre-individuation infancy.
The hypotheses about suicidal behaviour that invokes fantasies about
return to infantile dependence on the mother have been extensively
developed by attachment theorists. According to Bow lby,8 infants
respond to even brief separations with distress, and if such separations
and returns are not predictable or controllable, there develops a style
o f anxious and insecure attachments. These styles, resulting from the
breakdown o f the reciprocal interaction between child and primary
care-giver, m ay persist, and can even affect the w ay the child, when
grown up, relates to his or her own children.
Secure attachments in infancy are the primary means by which the
child learns to regulate his or her own emotions. In the absence o f
mature self-regulatory structures, such children come to depend too
much on others, or ideas about others that have been internalized, for
comfort. Such a person will become over-sensitive to abandonment;
vulnerable to crises o f aloneness which give w ay to self-contempt and
murderous rage. Suicidal behaviour in this context can be seen as an
interpersonal act, an (albeit maladaptive) means o f procuring attach
ment, signalling distress to others in the environment and punishing
them for their actual or perceived rejection o f the person.
John Maltsberger9 summarizes the sorts o f ideas about death that
might be apparent in the suicidal person:
1. Suicide is a gateway leading into a dreamless sleep (nothingness).

2. It will effect a reunion with someone or something which has
been lost.
3. It will be a w ay o f escaping from a persecutory enemy, interior
or exterior.
4. It will destroy an enemy w ho seems to have taken up a place in
the patient's body or some other part o f him or herself.
5. It will provide a passage into another, better world.
6. One can get revenge on someone else by abandoning him or her
or by destroying his or her favourite possession (one's own body),
and one can then watch him or her suffer from beyond the grave.
119
An underlying theme is the idea that one’s body is a prison from which
suicide promises an escape - particularly true o f those who suffer
physical pain. Maltsberger advises therapists to assess the extent to
which their suicidal clients have lost the capacity to tell whether or not
their own body is a part o f themselves. 'Does he feel at home in it and
take it for granted as an integrated self aspect, or does he experience it
as an alien cage in which he is confined, a cage belonging to, or even
identified with, someone else?' This aspect o f the psychodynamics o f
suicide takes us a long w ay from simply seeing such death as an
expression o f anger turned towards the self. Importantly, it raises the
theme o f escape. The notion o f escape is important because it builds
bridges to other approaches to suicide. M eanwhile inescapable stress
has biological consequences, which might be triggered by stress that is
believed to be inescapable (which introduces psychological mechanisms).
Biological approaches
From the brain, and from the brain only, arise our pleasures, joys, laughters
and jests, as well as our sorrows, pain, grief and tears . . . It is the brain which
makes us mad or delirious; inspires us with dread and fear, whether by night
or day; brings sleeplessness, mistakes, anxieties, absent-mindedness, acts that
are contrary to our normal habits. These things that we suffer from all come
from the brain, including madness.
So wrote Hippocrates, who lived between 460 and 377 b c . He had, as

this quote from his book On Sacred Disease shows, an approach to
mental disorder that feels very m odem .
T o understand the possible role o f the brain in general and suicidal
depression in particular, w e need to understand the role of neurotrans
mitters. Neurones (nerve cells o f which there are billions in the brain)
typically consist o f a cell body, dendrites (finely branched structures
connected to the cell body) and axons. Electrical impulses are trans
mitted from one neurone to another by means o f neurochemicals which
bridge the gap at the point o f contact between one cell and the next
(the synapse). Across this synaptic gap or cleft, neurotransmitters such
120
as dopamine or noradrenalin (norepinephrine) travel, released from one

neurone and triggering an electrical impulse at the ‘receptor' o f the
next neurone. Brain scientists have so far discovered over fifty different
types o f neurotransmitter, and each o f these m ay act alone or in
combination with others to serve different pathways in the brain.
One is at once able to see a number o f ways in which the delicate
balance o f brain activity could go wrong. First, there may be too
much or too little neurotransmitter present at the synapse. Second, the
receptor o f the ‘receiving' neurone m ay be too dense or too sparse, so
that even a normal amount o f neurotransmitter has too much or too
little effect. Third, there m ay be other neurochemicals that may block
or excite the action o f another, either inhibiting the neurone from firing
as it should or facilitating its action.
For many years it has been believed that depression is associated
with a depletion o f certain neurotransmitters. Most interest has focused
on norepinephrine and serotonin. A reduction o f the activity in brain
pathways which rely on these neurochemicals m ay be responsible
for the changes in appetite, sleep and energy levels in depression.
Antidepressant medication was thought to w ork by making more
neurotransmitter available at the synapse, by blocking the reabsorption
(‘reuptake') into the presynaptic neurones. This relatively simple theory
has always had a number o f difficulties, the major problem being that
the action o f the antidepressant in blocking reuptake is relatively quick,
but antidepressants take ten days to two weeks to have their therapeutic
effects. Another problem is that more careful measurement o f norepi
nephrine has shown it is not reduced in depressed patients, and may
even be increased in som e.10
Interest has therefore shifted away from simple (‘too little neuro-
transmitter') theories to consider instead the changes that take place in
the density and sensitivity o f the receptors - how well these are able to
respond to the amount o f neurotransmitter available. Such changes at
the receptor site take longer, and m ay explain w hy antidepressants take
so long to begin to alleviate mood. It is now believed that antidepres
sants alter the complex interaction between different inhibitory and
facilitatory pathways.
In suicidal behaviour, the neurotransmitter serotonin has been
121
thought to be involved. Evidence comes from a number o f sources. Low

levels o f the product o f the metabolism o f serotonin (the ‘metabolite’,
5-HIAA) have been found in the cerebrospinal fluid (C SF) o f suicide
attempters, measured after a lumbar puncture. Asberg and colleagues11
found this was true o f people who used more violent means to harm
themselves, consistent with parallel findings that reduced levels o f
serotonin, or reduced metabolism o f serotonin (not the same thing),
are found in violent offenders - those who take their anger out on
others. Low levels o f 5-H IAA in the cerebrospinal fluid are associated
with suicidality and impulsivity both on behavioural and self-report
measures, both in depressed people and in patients with personality
disorder, schizophrenia and alcoholism, and in impulsive violent
offenders.12 This has given rise to the suggestion that serotonin affects
tendencies to act with impulsive aggression rather than being associated
with the direction o f such aggression. Studies suggest that the sero
tonergic involvement is with trait (permanent predisposition) rather
than state (transient, situation-specific) features. Thus some people have
a life-long tendency to behave in an impulsive way. Then, if depression
produces a further temporary lowering in brain serotonin, there will be
increased tendency to act impulsively in response to negative events.
If violent suicide is associated with abnormalities in the serotonin
pathways in the brain, might this be detected at post-mortem? Estab
lishing the facts remains extremely difficult. The method by which
someone has died (overdose, carbon-monoxide poisoning) can affect
the physical state o f the brain at death, and the delay before post-mortem
also affects the chances o f discovering subtle differences between suicide
victims and controls. Error can also creep in through factors such as
abnormal diet, or drug and alcohol use.
H owever, most researchers agree that some sort o f association
between serotonin function and violence (either internally or externally
directed) exists13 as the review by Lil Traskman-Bendz and John Mann
indicates.14 Relevant to this is the finding that murderers have a suicide
rate several hundred times greater than people o f the same age and sex
who have not committed murder, and that people found guilty o f
violent and impulsive crimes, such as arsonists, have low serotonin
levels and a very high incidence o f violent suicide attempts. Crimes that
122
are premeditated are not so closely associated with serotonin levels,

emphasizing once again the importance o f im pulsivity.15 O f course, all
these groups tend to share other psychosocial risk factors such as
early loss, disrupted family background, alcohol and substance-abuse
problems. W hat has not been established yet is whether biochemical
studies have accounted for variance in the suicide statistics that cannot
be accounted for in these other ways.
The fact that biological findings have crossed diagnostic boundaries
(serotonin involvem ent is associated with violent suicidal behaviour in
depressed people, those with a diagnosis o f schizophrenia, alcoholics
and those with a personality disorder) implies that suicidality should be
assessed and treated in its own right, rather than relying on the treatment
o f the psychiatric disturbance in which it occurs. Another advantage o f
the biological approach is that it would provide a means by which
genetic influences might have their effects.
Genetic influences
Suicidal behaviour is more com m on in the relatives o f completed

suicides, suggesting a genetic component, but people related to each
other com m only share much more than genes. They also share a similar
environment, with all its stresses and difficulties. Furthermore, if a
suicide has occurred in a family, there is the possibility o f later imitation,
and a feeling that the family is 'tainted' with a suicide history which
makes other members more likely to see this type o f death as a way
out o f difficulties they themselves encounter.
Since identical twins share the same genetic make-up, such twins
should be more alike with respect to suicidal tendencies than non
identical twins. In the past the evidence was contradictory but further
research is clarifying the picture.16 Roy and colleagues17 found that
identical twins had a concordance rate for suicide o f 13.2 per cent (the
percentage o f twins who show the same behaviour as their other twin),
compared with a concordance rate o f 0.7 per cent in non-identical twins.
The balance o f evidence favours higher concordance rates for suicide
in identical twins.
123
A major single gene effect is unlikely. More likely is a multiple gene

which leads to an increased risk o f suicide, but needs environmental
triggers. These other effects are likely to outweigh the genetic effects.
Most twins, even in R oy’s study, were ‘discordant’ with respect to
suicidal behaviour. Also, most o f the suicides in the twins who were
concordant had psychiatric disturbances known to be associated with
suicide. If genes are involved, it remains to be seen whether what
is inherited is a suicide gene (most unlikely) or a predisposition to
psychological disturbance.
The studies most likely to be helpful are those that look for adoptees
who commit suicide in later life. It is then possible to tease apart
biological from environmental influences, by looking at the suicide rate
in biological versus adopted relatives. Such a study has been done
by Schulsinger and colleagues in Denm ark on a national sample o f
individuals adopted between 1925 and 1948.18 T hey identified 57 people
who were adopted and later committed suicide, and studied the suicide
rate in their adoptive and biological parents, siblings and half-siblings
(269 relatives in all). They selected a control group o f people who were
adopted but had not committed suicide and studied a total o f 269 o f
their relatives too.
The incidence o f suicide in the biological relatives o f the control
group was 0.7 per cent compared to an incidence o f 4.5 per cent (6.5
times higher) in the biological relatives o f those w ho committed suicide.
In the adoptive relatives o f both controls and suicides, there were
no suicides at all. Further study o f the sample showed that whereas
depression (unipolar and bipolar) was higher in the biological relatives
o f the suicide victims, the highest incidence o f suicide was in relatives
o f those who had a diagnosis o f brief affective reaction, a term given to
brief emotional disturbance often including an impulsive suicide
attempt precipitated by a stressful life event. This gives a clue to what
the genetic predisposition might be, adding weight to the idea that it is
the inability to inhibit an impulsive reaction to stress that is involved.
Such a personality predisposition crosses the boundaries between
different types o f psychiatric disorder.
Depression, schizophrenia, alcoholism or personality disorder might
all have been candidates as carriers o f a genetic loading for suicide, since
124
they all carry an increased risk o f self-harm behaviour. But the fact that
suicide does not discriminate between the diagnoses throws doubt on
this possibility. I have suggested that it was not the diagnosis per se but an
individual's perception o f their symptoms as aversive and uncontrollable
that was important.
The next generation o f research on serotonin is likely to have to
address several pressing issues. W hat is the effect o f chronic stress on
the serotonergic system and on impulsivity? W hat is the relationship
between alcohol and impulsivity? Alcohol disinhibits a person and
makes them act in a more dangerous way, but also lowers serotonin. It
remains unclear whether this disinhibition comes from the serotonergic
effect rather than the disinhibiting effect o f alcohol itself. It is possible
that alcohol’s effect on impulsiveness is simply short-term, whereas that
o f serotonin is longer-term.
Progress in this area is unlikely to be fast. First, changes in serotonin
function m ay only be a marker for another causal factor, and even if it
were decisively established that it was linked to certain forms o f suicide,
both suicide and altered serotonin function may be caused by this third
variable. Second, decreased serotonin function may represent the brain's
attempt to compensate in one pathway for an increase in serotonin
activity elsewhere, or to compensate for a problem in another neuro
transmitter system. Third, since all the data on serotonin function come
from correlations between known suicidal behaviour and biological
function, we do not yet know whether there are any changes in
serotonin function that precede such behaviour. Therefore no infer
ences can yet be drawn about the predictive value o f knowing about
serotonin function in an individual prior to suicidal behaviour.
A final problem is that biological studies have to date been done in
isolation. They have been concerned to establish links between brain
systems and behavioural tendencies such as violent or impulsive
behaviour, without asking what the psychological (e.g. information-
processing) and social (e.g. acute and chronic stress) factors involved in
such behaviour might be, or what the neuropsychological mediators o f
such behaviours are.
Following a review o f this material, Seym our Kety concludes:
125
W e cannot dismiss the possibility that the genetic factor in suicide is an inability
to control impulsive behavior, while depression and other mental illness, as
well as overwhelming environmental stress, serve as potentiating mechanisms
that foster or trigger the impulsive behavior, directing it toward a suicidal
outcome. In any case, suicide illustrates better than any o f the mental illnesses
. . . the very crucial and important interactions between genetic and environ
mental influences.19
Concluding remarks
A com m on theme emerges from both psychodynamic and biological

approaches to suicide: the theme o f escape from a situation that is seen
as inescapable in any other way. From the dynamic perspective, the
roots o f this inescapability are to be found in the type o f attachments
formed early in life. In some theories, these are seen specifically as
over-narcissistic, but it is common ground between many o f these
theories that the result o f these early experiences is that loss or threat
ened loss o f important relationships later in life produces an intolerable
feeling o f abandonment and interpersonal failure. Loss at any stage in
life produces a mixture o f protest, anger, anguish and despair. Such loss
is associated with biological changes, and the serotonin system in the
brain may act as an important link in the chain o f events connecting
uncontrollable events with impulsive and aggressive responses, whether
directed at others or at oneself.
Neither psychodynamic nor biological theories make a clear distinc
tion between impulses that drive someone towards suicide as an option
on the one hand, and factors that remove the obstacles to such behaviour
on the other. A wish that one could escape a hopeless life situation or
escape one's mental anguish might provide sufficient push towards
suicide, but it is equally important to understand the factors that take
away the usual reasons for living or remove the barriers to dying. One
such barrier might be the fear o f death. W e have seen how this can be
reduced by a feeling that death might not be the end. It might also be
reduced by exposure to real or fictional accounts o f suicidal behaviour
in books, magazines, television and radio.
126
8
The Effect of the Media
My teenage son and I were clearing up the back garden. I was clearing
parts o f a tree which had been trimmed some time before, and he was
sawing them into logs. He remarked that it was exactly a year ago since
he had last been sawing logs. W hen I expressed surprise at his accurate
m em ory, he said he remembered the date clearly because it was the
anniversary o f Kurt Cobain’s death by suicide. Kurt Cobain was the
lead singer o f the band Nirvana. He had harmed him self in the past and
finally shot himself, having left a long suicide note. When such a
high-profile suicide occurs, people want to know what effects the death
has on all the fans.
The debate about whether accounts o f suicides (whether fictional
stories or actual suicide) cause people to imitate them has a long history.
In the eighteenth century Goethe was accused o f encouraging suicidal
behaviour by the publication o f his novel The Sorrows o f Young Werther
(1774), in which a young man ends his life by shooting him self in the head.
The book had a tremendous influence on attitudes to romantic
suicide and coroners all over Europe began to believe that the W erther
effect was influential in bringing about suicides by lovelorn young
people. The book was accused o f sentimentalizing self-destruction, but
whether or not it produced imitative suicides, it certainly produced
imitative poetry and literature. The sentimentalization o f death which
followed such literature inspired pity for those who committed suicide
and helped to erode the revulsion suicide had form erly inspired. The
result was that many became concerned that suicide would spread by
imitation.
127
Emile Durkheim agreed that 'no other phenomenon is more readily

contagious', but he did not agree that imitative behaviour would affect
suicide rates when they were calculated over a longer-term period.1
H owever, evidence from several studies suggests that imitative suicidal
behavior can and does take place.
Evidence for imitative effects
In the first o f a series o f papers, David Phillips o f San Diego University,

California, examined the monthly suicide rates in the United States
for the period 1947-68.2 After correcting for the effects o f seasonal
fluctuations and linear trends, he found that suicide rates were higher
just after a heavily publicized suicide story. The more publicity devoted
to the story, the greater the increase in suicides, this increase occurring
mainly in the geographical areas where the story was publicized.
Similar w ork by Brian Barraclough and colleagues in the United
Kingdom examined 76 deaths by suicide or undetermined in the Ports
mouth area over the three-year period 1970-72.3 They also looked at
newspaper reports about suicide inquests recorded in a local newspaper
which reported 80 per cent o f inquests on suicides and was read by 70
per cent o f the adult population. Over the few days following the
newspaper report o f a suicide verdict, they found an excess o f observed
suicides which was statistically significant for males under 45 years.
Do television reports o f suicides have a similar effect? Phillips and
colleagues examined the US daily mortality figures for the period 1972-
6 in relation to national T V news coverage o f suicides.4 They identified
seven suicide stories which had been featured on national T V news
(e.g. Christine Chubbuck, the talk-show hostess who shot herself while
on the air on Monday, 15 Ju ly 1975). Following five out o f the seven
suicide stories, there was an increase in the daily mortality figures for
deaths by suicide.
A spate o f suicides occurred on the Viennese underground system
in the 1980s. Between 1980 and 1984 there had only been nine suicides
in total on the underground system. H owever, in 1986 one or two
suicides were followed by dramatic reporting o f the events, and several
128
other (apparently copycat) suicides followed, so that in 1986 alone 13

suicides occurred and a further nine in the first few months o f 1987.
Reporting restrictions were introduced and the number o f suicides on
the underground reverted to more normal levels (three in 1989 and four
in 1990). Such events appear to show that reporting o f real suicides can
have a contagious effect.5
Fictional portrayals
Does this effect generalize to suicides or attempted suicides shown in

fictional stories on television? T o examine this question, Phillips turned
to soap operas shown during 1977 on US television. Since the soap operas
attract a larger audience than any other sort o f television programme (40
per cent o f all US homes with televisions are turned to at least one soap
opera in any week), he assumed this was likely to be a large potential
source o f behaviour which people might attempt to model.
Phillips identified nine separate weeks in which one or more soap
operas carried a story in which a character committed or attempted
suicide. From the US daily mortality statistics, he then examined
whether there had been a rise in suicides in those particular weeks. As
control periods, he took the weeks before, correcting for trends and
eliminating public holiday periods and periods in which there had been
an actual suicide publicized in the media. The results showed a rise in
suicides in the latter part o f the weeks in which suicidal behaviour was
depicted in a major soap opera in eight out o f the nine periods. For
some fictional portrayals, however, the data is not always so clear cut.
The 'Angie' myth
An estimated 14 million watched the episode o f the soap opera EastEnders

on British T V on the evening o f Thursday, 27 February 1986, and
9 million watched the repeat showing during an omnibus edition the
following Sunday. The Thursday episode included an overdose by
one o f the main characters (Angie) together with alcohol (neat gin).
129
Although the overdose involved several bottles o f pills, the programme

had also shown this character buying one o f the bottles from a news
agent - a bottle o f aspirin. The supposition was that this was an overdose
with analgesics, and that, like many overdoses in real life, it arose out
o f a difficult interpersonal situation in which the ‘husband* had been
unfaithful. Following that programme and the omnibus edition on
Sunday (in which the actual overdose was not shown though all the
other elements were present), a large number o f attendances for deliber
ate overdose at Hackney Hospital and St Bartholom ew's Hospital in
London were claimed. A letter to the Lancet reported that the number
o f patients admitted to Hackney Hospital during that w eek was far in
excess o f the average for the previous ten weeks, and o f the average for
that w eek for six out o f the previous ten years.6
On close inspection the ‘copycat' effect was less certain. Stephen
Platt examined data from 63 hospitals for the period and for a compar
able control period the previous year.7 Although he found an increase
in all wom en, the increase was not especially prominent in wom en in
the 30-40 age group which he hypothesized might have been most
closely associated with a model in that age range. He also found an
increase in both sexes over 45 years. He concluded, however, that these
were freak results owing to the fact that the control period for 1985 had
witnessed a slight decline in the number o f attempted suicides within
these categories, so this was not a particularly good baseline period
against which to judge the ‘rise' in the 1986 figures.
Platt also found a significant correlation in the opposite direction to
that predicted between the viewing figures for different regions and the
increase in overdoses attending hospitals. The increase in overdoses
was lowest in London, which had the highest viewing figures, and
highest in Yorkshire and the north-east, in which a smaller proportion
o f the population watched the programme. For these reasons his con
clusion was ‘case not proven'.
With some colleagues, I also investigated the Angie phenomenon.
We examined in detail the numbers attending the departments that had
made the original claim at Hackney and St Bartholom ew’s Hospitals.
We examined data for the period up to and including the dates o f the
broadcasts and for the same periods for two control years.8 W e also
130
not only compared the number o f 'imitative' cases with the numbers
o f people attending in control years, but looked for any trends which
might have contributed to an apparently large number in the target
period. For both hospitals, overdose attenders in the 14 days after the
broadcast were higher than those in the corresponding period for
the tw o control years, but the rise in numbers began well before the
programme was shown, in December in one hospital, January in the
other. Analysis o f daily data showed no evidence o f unusual short-term
changes within the fortnight after the programme. W e concluded there
had been no copycat effect.
Do these findings mean that no imitation ever takes place? Clearly
that cannot be concluded on the basis o f this study. There remains
enough prima facie evidence to suggest caution when editors decide
what to print or to show. Other studies, which have provided a more
rigorous 'natural experiment’, indicate that fiction can be powerful.
Arnold Schmidtke showed how a fictional portrayal o f a young man’s
suicide on a railway line could have an imitation effect on view ers.9 He
studied the number o f railway suicides in the 70 days after the fictional
episode was broadcast and found an increase. The suggestion that this
was indeed imitation was reinforced by the finding that the suicides
most often occurred in people o f the same age and sex as the fictional
character. Furthermore, the numbers o f suicides correlated with the
audience figures. The broadcast was repeated by the television author
ities and Schmidtke found, once again, an increase in suicide. In total,
his report estimated that an extra 60 suicide deaths occurred as a result
o f this fictional episode.
Similarly, in 1999, Keith H awton and colleagues found a marked
increase in self-poisoning cases following a television programme show
ing a paracetamol overdose. Forty-nine accident departments around
the United Kingdom were put on alert, and monitored their attendances
before and after a fictional account o f an overdose on the B B C ’s Casualty
drama. Overdoses increased by 17 per cent in the w eek following the
broadcast. The use o f paracetamol in the overdoses doubled in those
who had actually seen the program m e.10
131
Criticisms of imitative research
Research on imitative suicidal behaviour has, how ever, been sur

rounded with controversy. There are three sorts o f criticism. First, that
there is very little theory which can explain some o f the anomalies in
the data. For example, some o f Phillips's w ork found a lag o f three
days before 'imitative' behaviour became apparent, but no theoretical
account has been given for this phenomenon. Furthermore, in the sort
o f ‘macro' research that looks only at the gross numbers o f suicides and
only at whether a suicide story has occurred or not, there is no attempt
to match the properties o f the media stimuli with the set o f behaviours
that qualify as an imitative response. Many o f the studies, for example,
have not followed Schmidtke and looked at whether the age and sex o f
the suicide victims correspond to the age and sex o f the person who
commits suicide in the news or soap opera story.
Perhaps more serious than a lack o f good theory is the possibility
that at least part o f the phenomenon is a mere statistical artefact. W hen
one group o f researchers re-analysed data from Phillips's 1982 soap
opera study,11 corrected some dating errors, increased the sample size,
and used a more reliable regression analysis, they failed to replicate
Phillips's results.12
Others have pointed out different sources o f bias in imitative suicide
research. These suggest that both the overall increase and the pattern
o f lags in the data are explained if one takes into account the fact that
m ore suicides happen on some days o f the week (Monday and Tuesday)
than occur on others. Introducing a control period which also includes
this day o f the w eek m ay not help if, along with the increased average
number o f suicides for this day, there is also an increase in the variance
o f the data. This aspect o f the data makes it more probable that one
will make a statistical error using conventional statistical significance
levels, claiming an effect when there really is none.
If one examines the same weeks that Bollen and Phillips used in their
1982 study, but looks at the following year when no suicide story was
published, one finds evidence for increases which mimic those found
by Phillips, even though it could not have been imitative behaviour.13
132
H owever, in reply to this criticism, Phillips and Bollen returned to their

data and eliminated these statistical problems. They found it did not
affect their results. Furthermore, in a subsequent study Phillips14 exam
ined the daily fluctuation o f a large number o f teenage suicides during
1973-9 in relation to news or feature stories on the m ajor T V networks.
Using statistical analyses that corrected for the day o f the week, month
o f the year, yearly trends and holidays, they found a mean increase per
story or feature o f 2.91 in the number o f suicides. They also found a
significant correlation between the number o f news bulletins that fea
tured a story and the increase in the subsequent seven-day period in
suicides. Taking into account all previous criticisms, they also looked
at similar time periods for the year before and the year after, and found
there were no effects when there had not been a suicide story or feature
article. Taken together with data from other researchers in very different
contexts (such as the railway suicides reported by Schmidtke), there can
be little doubt that imitation does occur.
Mechanisms of imitation
Much o f our behaviour is learned by copying the behaviour o f others.

This is especially true when w e are in situations where we do not know
the right thing to do. Such processes determine children’s behaviour
much o f the time, but are found later in life, too. Uncertainty m ay be
the key to when modelling is most likely to occur in suicidal behaviour.
It is when prior suicidal feelings are accompanied by feelings o f uncer
tainty about what to do that imitation will be most likely to increase
the probability o f suicidal behaviour. If the model is attractive or
famous, the imitative effect will be stronger because their behaviour
will resolve more o f the uncertainty about suicide as the right course
o f action. Indeed, Phillips has found that the increase in copycat suicide
is greater if more prominence is given to it. Other researchers have
found that celebrity suicides in the United States produced imitative
behaviour only if the celebrity was famous enough to rate the front
page o f The New York Times.
133
Concluding remarks
There is compelling evidence that imitation is a factor in increasing risk

o f suicide. H owever, there is little evidence that such effects occur in
isolation from all the other vulnerability factors. Imitation m ay take the
form o f suicide or o f attempted suicide, and may follow real or fictional
accounts in the media. Seeing others commit suicide, especially som e
one with whom there is a strong identification, m ay low er the barrier
to suicide in certain circumstances. H ow ever, the overall effects on the
suicide rate are relatively small, and we need therefore to ask w hy
imitation does not have a larger effect.
The first, most obvious reason is that most people are not in
the ‘vulnerability w indow ' that gives them a strong m otive to die.
Decreasing barriers to suicide are simply irrelevant for the majority.
But the second reason takes us back to the example o f Kurt Cobain.
Although there have been some suicides by his fans that appeared to
imitate his death, m any o f these occurred a considerable time after
wards, so the role o f direct imitation is questionable. The puzzle is w hy
there have not been more. I believe the answer lies in the fact that the
death o f such an idol, though it upsets many fans, can unite them in
mourning. At the very least, many followers o f his music have a major
topic o f conversation, which acts as a source o f bonding. The increased
social support countermands the tendency to imitate. It follows that
imitative suicide is more likely to happen in those individuals w ho are
loners, or are temporarily outside their social group. Obsessing about
the death in isolation, without the social support o f others, is most
likely to result in imitative suicidal behaviour.
Do any recommendations follow from this for media presentation
o f suicide? The following suggestions are based on those o f G ould:15
1. Detailed descriptions o f methods o f committing suicide or

deliberate self-harm should not be given, nor shown in fictional
portrayals, since these will demonstrate possible means for those
who are suicidal.
2. Physical consequences o f suicide attempts, where this has
134
occurred (e.g. paralysis, brain damage), should not be minimized,

but neither should such harmful consequences be dwelt upon,
which might attract those who seek victim status.
3. If the victim had prior mental health problems, these should not
be ignored. Many stories refer to life circumstances or philo
sophies as if these were sufficient precipitants. (See Chapter 6,
where Arthur Koestler’s suicide is discussed (pages 108-9), and it
is pointed out how reports o f his death failed to focus on his
long-standing psychological problems.)
4. Simplistic psychological notions, such as ‘pressure5, should not
be used. Many people have such pressure, but few commit
suicide, and such references belittle the complexity o f the situ
ation. Similarly, simple motives such as ‘getting even’ or ‘becom
ing fam ous’ should be avoided.
5. W here possible, stories should avoid emphasizing the attractive
qualities o f the deceased in such a w ay as to make the suicide a
part o f this attractiveness.
6. Such stories should be accompanied by details o f what help is
available, together with hotline numbers (such as the Samaritans
or other befriending agencies).
7. Extensive or unnecessary repeated coverage o f such events
should be avoided.
8. There should be ongoing discussion between those working in
the media and mental health experts in which the evidence for
such imitative effects is kept under review.
135
9
The Cry of Pain
In the United Kingdom each August teenagers receive their end o f

school examination results from schools and colleges. Often the grades
obtained are critical in determining whether the person is able to go to
the college or university course o f their choice. In 1994, when the
Advanced level (the end o f school public examination) results came
out, a talented pupil from a village in Norfolk killed herself. She felt
that her grades (a B, a C and two D ’s) w ere not good enough to win
her a place on the physiotherapy course she hoped to do. On the day
she collected her results, she told friends she wanted to be alone, but
then let herself into her cousin’s empty cottage and hanged herself.
What can explain such a rapid escalation o f suicidal depression? She
was popular and talented, had hobbies and appeared happy-go-lucky. It
particularly surprised people since she had fought back after a riding
accident two years before in which she broke her back and seemed to
be a survivor. Hindsight in such cases can be very deceptive, and it is
unwise to imagine that such a rare event might have been predicted.
H owever, such a tragedy serves as a reminder to be vigilant for those
who may be:
(a) especially sensitive to feelings o f failure; and

(b) likely to conclude they are trapped if they do fail.
At the time, several commentators suggested that her case should signal
a danger that the education system had become too dominated by a
league-table culture that emphasized academic performance as the sole
136
The Cry of Pain
criterion o f success. The example o f Japan was thought to be another

warning.
The high rate o f suicide among teenage students in Japan is often
linked to the highly competitive nature o f the education system. Pupils as
young as 12 years often attend cramming schools in attempts to improve
marks so that, eventually, they will win a job with a top company. A
further consequence o f such a dominance-subordinance environment is
the high rate o f bullying, or ijime, between pupils. Such bullying is often
the actual precipitant o f a suicide. In Decem ber 1994, an emergency Cabi
net meeting in Japan was called to discuss bullying in schools. It was
becoming clear that it could go to extreme lengths, but was being under
reported by schools, who feared criticism for allowing it to occur. The
meeting took place against the background o f the suicide o f two 13-year-
old boys. One o f them, from Nishio, in the prefecture o f Aichi, had hanged
himself. He left a four-page suicide note giving details o f the beatings and
extortion he had suffered for over a year. He had been repeatedly dunked
in a river until he feared he would drown, and forced many times to give
m oney. ‘These days they bully me so hard and demand large sums o f
m oney although I have none. I can’t stand it any m ore.’
The Education Ministry was forced to admit that a large part o f the
problem stemmed from the highly competitive nature o f the education
system. In a special report in December 1994, they called on teachers to
be more aware o f pupils’ stress levels and competition between them
in examinations. Little guidance was given, however, in how teachers
could assess whether students were suffering in this way.
The use o f exam success as the sole criterion o f self-worth produces
individuals who feel ashamed o f their ‘weakness’ and trapped by their
apparent failure. The feeling o f humiliation and ‘weakness’ acts as a
barrier to sharing the fear o f failure with peers or family. The feeling o f
humiliation is exacerbated by feelings o f loneliness, and each mood
feeds on the other in a vicious spiral o f despair.
These themes are illustrated by another situation that came to light
relatively recently. Around 50 suicides in the United Kingdom each year
are am ong people found shoplifting and awaiting a court appearance.
They are most often people who had no financial need to shoplift.
There is, for them, an overwhelm ing feeling o f shame and humiliation.
137
If the case comes to court, the verdict will often reflect the judgem ent
that the person is suffering a psychological problem that needs
psychological intervention rather than punishment, but many commit
suicide before the case comes to court.
Interpersonal disputes can involve combinations o f the same em o
tions. A wom an in her mid-forties took a massive overdose soon after she
discovered that the man she planned to m arry in a few weeks (having lived
with him for several years) had unilaterally cancelled the arrangements.
Without her knowledge, he had sent notes to all the guests to say the
wedding was cancelled. He came to visit her in hospital, but only to give
her ring back, and tell her to collect her stuff, which he had left in suitcases
on the front doorstep. Her overdose was not lethal, and she lived. But her
crisis threatened to last for some time, and while it did so, she remained
at serious risk o f further suicidal behaviour.
Biological, psychodynamic and social aspects m ay be combined in
explaining such behaviour. Individuals are sensitive to signals from their
social environment o f threats both to their rank within the group and
their acceptance as part o f the group. Events signalling threat to rank or
group similarity are seen, by the person, as signs that they are a ‘loser’.
Like the wom an whose wedding was cancelled, such people have lost a
battle. Such ‘loser’ status we know as reduced self-esteem, and it triggers
evolutionarily ‘old’ biological patterns in the brain. This affects m ood and
produces further hypersensitivity to social information, resulting in a
vicious circle that appears to confirm the individual as a ‘loser’ . In the early
stages a ‘reactance’ pattern is shown, in which the individual ‘protests'
against the threatened loss. At this stage, ‘low intent’ suicidal behaviour
m ay occur. In the later stages o f response, despair and apathy m ay be
seen, and serious attempts at suicide may occur. Occasionally, however,
a person may experience a catastrophic increase o f suicidal impulse in
response to loss or failure, and an unexpected suicide occurs.
Knowledge o f brain function is not all we need to explain such
behaviour. W e need rather to understand what environmental and
psychological ‘signals’ bring about biological changes in brain function
(see pages 120-6), and, in turn, what effects such biological changes will
have on the w ay we think and act. For example, psychosocial stress is
a critical factor in bringing about biological changes. Long-term stress
138
The Cry of Pain
in both animals and humans can have profound effects on patterns o f

endocrine and neurotransmitter function. There is increasing evidence
that repeated stress in life can bring about changes in messenger R N A
and D N A , and there is a renewed interest in mapping the effect o f such
changes on psychological function.1
Depression can be seen as the response to the long-term threat posed
by constant frustration and disappointment, by the loss o f people, things
and status. Such frustrations and losses m ay result from constantly
being put down by society, or receiving insufficient help or support. In
considering how we can explain suicidal behaviour in a w ay that takes
full account o f biological, psychological and social processes, a helpful
starting point is ethology. Ethology is the study o f animals in their
natural habitat. The approach has been influential in such important
theories as that o f Bow lby on attachment. There are a number o f
different w ays animals and humans respond normally to losses:
(a) by energizing behaviour to try and secure the rewards in

other ways;
(b) by switching to alternative sources o f reward; and
(c) by giving up.
These can become distorted in catastrophic w ays that cause the indi
vidual to believe he or she is trapped, and to begin to think o f ways o f
escaping. Suicidal behaviour is best seen as a cry o f pain - a response
elicited by this situation o f entrapment - and only secondarily as an
attempt to communicate or change people or things in the environment.
Conservation-withdrawal
The Australian psychiatrist Robert Goldney was one o f the first to take
an ethological approach to suicidal behaviour, employing Engel’s ideas2
about ‘conservation-withdrawal’.3 This refers to a biological system
which responds to stress by reducing activity, raising barriers against
stimulation and conserving energy: a sort o f hibernation. This pattern
m ay follow a prolonged period o f heightened arousal where the
139
organism has attempted to cope by over-engagement and over-activity.

W hen the organism falls into the state o f conservation-withdrawal, the
signs that are visible in the social group have a signal function, warning
others o f a loss o f supplies and o f exhaustion. The intended outcome is
to ‘ensure the supply' and ‘retain the object’.
This, Goldney notes, has striking parallels with the reports o f those
who take overdoses: they want to escape, they do not know whether
they want to live or die, they cannot stand the pressure any longer.
Schneidman (1966) has pointed out how many see a connection between
death and sleep, and this can be found in the reasons people give for
suicidal acts.4 They want to ‘pull the bedclothes’ over their heads, or
‘go to sleep for a long time’. The fact that alcohol abuse and suicidal
behaviour often go together can be linked to the innate mechanism o f
‘conservation-withdrawal’. Alcohol also represents an attempt to react
to the stresses o f life by reducing the demands on consciousness by
shutting out external stimuli. W hen such attempts to bring about
temporary oblivion with alcohol fail, then other mechanisms for conser
vation-withdrawal come into play. The fact that alcohol reduces the
fear o f the possible negative consequences o f any behaviour, when
applied to suicidal acts, renders them all the more likely.
Many suicide attempts can thus be seen as a wish for temporary
relief. The attempt may represent a conservation o f energy, a raising o f
the barriers to the outside world, or at least an attempt to re-exert
control over it. Although the idea o f conservation-withdrawal fits in
some respects, it seems to apply best to the later stages o f response to
uncontrollable stress, when the individual has given up hope, become
despairing and apathetic. It does not capture the more active ‘protest'
aspects o f some suicidal behaviour, which, I shall argue, arise at an
earlier point in the process.
Evolutionary approach
Recent research helps us to go into more detail about the types o f

biological mechanisms that m ay be involved in depression and suicidal
behaviour. This has been drawn together by Professor Paul Gilbert,5
140
The Cry of Pain
who suggests w e look to the animal kingdom for hypotheses about

the evolutionary primitive mechanisms that might influence human
behaviour. He is particularly concerned to explore the concept o f
defence and safety.
People differ widely in what they find threatening: the fear o f a
physical illness such as a heart attack, the fear that others will criticize
them, the fear o f being overwhelm ed by emotion when describing past
trauma, the fear that people will ‘find out’ one’s weaknesses. Gilbert
suggests that defence and safety m ay be fundamental psycho-biological
organizing systems that guide the development o f interpersonal sche
mata and strategies. Non-threatening environments promote creativity,
co-operation and affiliation. Threatening environments inhibit flexibility
and exploration, producing more stereotyped, automatic responses.
This has implications for suicidal behaviour, and how it is linked with
inflexibility in cognitive processes such as m em ory and problem solving.
The most important ‘unsafe’ environment for most humans, however,
is not exposure to predators or physical harm, but exposure to unfavour
able social comparison. Thus we need to explore further the evolution
ary significance o f social comparison in determining rank and status
within social groups.
The animal kingdom shows many examples where social groups
survive because o f the smooth operation o f social processes whose
functions are:
(a) so that sexual partners are chosen (intersexual selection -

where one sex attracts and chooses a partner); and
(b) so that potential rivals are excluded (intrasexual selection -
where members o f a single sex prevent other members o f the
same sex from gaining access to potential partners).
An important aspect is rank and status within a social group that shares
the same territory. High-ranking individuals explore more, have more
erect posture and are less timid than low-ranking individuals across a
number o f species.
Changes in rank can bring about changes in biological systems (e.g.
in levels o f testosterone), but more importantly, for understanding
141
depression and suicidal behaviour, in levels o f the neurotransmitter

serotonin. What evolutionary function do such systems have? An
important study in 1984 by Michael Raleigh and his colleagues at the
U C L A School o f Medicine, Los Angeles,6 investigated the connection, in
colonies o f adult male vervet monkeys, between social rank and blood
levels o f serotonin. They found that dominant animals ( judged by obser
vation o f how many encounters a male m onkey w on over its peers) had
much higher levels. If the dominant animal lost its position, however,
there followed a reduction (of some 40 per cent) ofblood serotonin levels.
This occurred whether loss o f status happened naturally or was experi
mentally produced by removal o f the dominant male. In either case, the
male that then became dominant experienced an increase in serotonin to
the same levels characteristic o f the previous dominant male.
Such biological systems seem to be involved in regulating fights and
challenges in a group by establishing a hierarchy o f rank. It is important
for an animal that is likely to lose encounters not to compete continually,
so expending valuable resources in challenges that cannot be won. On
the other hand, some contests m ay be worth fighting. The ranking
determines who is likely and who unlikely to be beaten. There has
evolved a large number o f signals used by animals to communicate
challenge, attack and submission. Submission, when triggered, involves
a menu o f behaviours signalling that no further threat to the victor is
intended. Further challenges are inhibited by this mechanism. Neither
party need waste further resources on establishing relative rank. Fur
thermore, the signalling that goes on in animal groups means that
animals m ay not actually have to fight it out.
Out o f this research comes the study o f analogous behaviour in
humans. Although such extrapolation has to be done with extreme care,
the analogy between observations o f social groups in animals and humans
is justified so long as they are used to set up hypotheses rather than coming
to conclusions about human behaviour. H ow best to proceed? W e can
observe that humans appear sensitive to, and make rapid judgem ents
about, one another’s rank and status. The social comparison process may
be done on the basis o f another person’s size o f house, or make o f car,
or quality o f clothes. Indeed, we have invented games which explicitly
ascribe rank, such as the role-playing games played by increasing numbers
142
The Cry of Pain
o f people during the 1980s and 1990s. Players are allocated characteristics
in various dimensions: size, strength, skill, previous success, weapons,
allies, etc. The outcome o f encounters between competitors is deter
mined by their ability to use and to conserve these ‘attributes’.
Social biologists refer collectively to these attributes, when observed
in animals, as Resource Holding Potential (RH P). Ritual encounters
between two members o f the same species result in comparison o f
relative R H P , and the loser will back off. As the hierarchy becomes
established, the amount o f fighting reduces. Gilbert suggests that the
human equivalent o f R H P is self-esteem. Self-esteem is an estimate o f
one’s ability to secure important goals such as a desired job, or partner.
It is damaged by unemployment, by failure in love and by aggression
from others, as in school by bullies or in the home by family members
who use psychological or physical violence.
Paul Gilbert suggests that dominance hierarchies have evolved
through stages. First, they were territorial, where fighting was designed
to create space for territorial ownerships. Subsequently, territories gradu
ally dissolved and success in the dominance hierarchy became associated
with other forms o f social success; it was a dispenser o f useful resources.
Dominance, how ever established, creates inhibition in subordinates.
Since that time there have been large changes in the importance o f
affiliative behaviour, particularly alliance building (see next section),
which requires approach rather than inhibition. This results in the fact
that, today in humans, rank is often (though not always) determined
by things such as perceived popularity, beauty, talent, etc. Gilbert main
tains that when individuals lose rank, then the defensive systems o f
escape, etc., become activated. The means o f acquiring rank and status
m ay have changed, but the feeling when one loses has not. Submission
in humans continues to involve giving up claims on resources.
Being 'in' with the group
In addition to helping to determine rank by looking for signs o f R H P,

social comparison can also judge the extent to which one is succeeding in
gaining favourable attention from other members o f the group. Unlike
143
R H P , which is to do with ranking, this is concerned with judgements

about sameness v. difference from the reference group. Gilbert calls it Social
Attention Holding Power. It conveys status by attractiveness rather than
by aggressive displays o f strength over weakness. It is concerned with
receiving signals that one is attractive to others and/ or needed by them.
In clinical practice, one m ay observe tw o types o f problem when indi
viduals are over-concerned with whether their company is rewarding to
others. First there are those who appear to be sustained exclusively by
receiving attention from others. They are sensitive to any signs that these
sources o f attention m ay be under threat o f being switched off. They feel
abandoned when this happens, and are likely to become demanding o f
others’ attention, seeking reassurance that they are still loved. In therapy,
they may say, ‘No one really loves m e/ or ‘All m y friends are sick o f m e/
Second, there are those who appear to be sustained by their ability
to give attention to others and be needed by them. People w ho seek
helping roles in various organizations m ay fit into this category. They
are often held up as examples to the group, as people w ho ‘can’t do
enough for others’. They are welcom e to most groups they join, since
they often appear to have boundless energy. There is little w rong with
this except when ‘being needed by others’ is the sole source o f their
satisfaction in life. If so, they become vulnerable when other members
o f the group or fam ily make it clear they no longer need their help. In
this case, the person m ay feel depressed, for they know no other w ay
o f relating to people other than to help them. W hen depressed, they
do not doubt that others love them, but rather they say, ‘I am a burden
to others’; and in the most extreme cases, ‘They would be better off
without m e.’ Their ‘black and w hite’ (dichotomous) thinking makes
them conclude that there are only two possibilities: either they are a
help to others or they are a burden to them.
Social comparison and psychopathology
W hether attempting to compare resources for ranking using R H P or

to evaluate the extent to which one fits in with the group (the amount
o f Social Attention Holding Power), the effects o f social comparison
144
The Cry of Pain
run deep. It is the comparative effect, not the absolute effect, that
determines the amount o f stress or depression. During the Thatcher/
Reagan years, politicians used to answer the accusation that a large gap
had opened up between rich and poor by saying that, although the rich
had indeed become richer, the poor had little cause to complain since
they had not become much poorer, if at all. H owever, this takes no
account o f the fact that people's m ood and their psychological (and
sometimes physical) health are determined by social comparison. Abso
lute levels o f deprivation are higher in war, but cases o f suicide and
depression fall because all are perceived to be suffering together - social
cohesion goes up. Similarly, absolute levels o f deprivation are higher in
underdeveloped countries than in most Western countries, but there is
no excess o f suicides and depression in these poorer countries.
This explains some o f the discrepancy in the debate about suicide
and unemployment. There is little doubt that being unemployed is an
extra risk factor for suicidal behaviour, but overall rates are sometimes
found not to fluctuate with the unemployment rate as one might expect.
Research by Stephen Platt and colleagues in Edinburgh shows that this
confusion arises partly from the fact that unemployment has less
capacity to cause depression and suicide if the person perceives everyone
else to be in the same boat.7 It is not absolute levels o f poverty and
employment which are important in physical and mental health, but
distribution o f inequality.8 It is the conclusion that, compared to others,
one is a failure, unwanted or powerless, which increases vulnerability
to emotional distress.
Involuntary subordination
In Paul Gilbert’s Depression: The Evolution o f Powerlessness,9 he draws an

important distinction between voluntarily giving up and m oving on,
and involuntarily being defeated. In the latter case, the individual is
trapped and feels they have no alternative. Involuntary subordination
is an important aspect o f depression, which may arise from a person
being forced to be subordinate and take a submissive role. This takes
us back into social rank theory. The dimension o f rank (whether one is
145
dominant or subordinate in a group) m ay be more significant in the

cause o f depression than previously thought. W hen depressed, people
seem to feel an increased need for recognition and prestige. They w orry
about how they are succeeding in relation to others, evaluate themselves
as inferior or second rate, feel weak and ashamed about such weakness.
The usual summ ary term for such a constellation o f symptoms is ‘low
self-esteem’ , which sums up the sense o f subordination, the perception
or reality o f negative social comparison.
The value o f the concept o f involuntary subordination is that it sums
up what many see as the cause o f increased depression in wom en
relative to men: that it is often a consequence o f enforced subordinacy.
T o protect themselves (and often their children), wom en are forced to
take a submissive role, increasing self-blame so as to reduce the likeli
hood o f counter-aggression from their partner. Consistent with this,
researchers in London (Bernice Andrews and Chris Brewin) found that
wom en being abused by their partners blamed themselves for their
partner’s violence. H ow ever, this was only true if they were asked
about it while still in the relationship. If asked after they had m oved
out o f the relationship, they blamed their partners. So far as w e can see,
their self-blame while still in the relationship felt genuine to them; they
were not consciously telling lies as a protective strategy. As Gilbert
suggests, in a context where one partner is criticized and disempowered,
an evolutionary primitive inhibitory system comes into play and this
system gives rise to the symptoms we know as a depressive state.
Such an evolutionary primitive mechanism may come into play even
if there is the perception o f low status, failure and weakness.10 A person
m ay be pursuing unrealistically high goals, so their perfectionism itself
m ay give rise to the perception that they are failing, w eak or subordinate.
They invent for themselves a punitive environment. They feel they
have failed, then feel trapped by such failure, since there seems no way
out o f their situation. Occasionally a single failure can activate these
depressive mechanisms very suddenly with devastating effect, as we
see by the cases o f school ‘failure’ and suicide reported at the start o f
this chapter. Suicide following bullying is an example where the more
long-lasting stress o f being trapped gives way, eventually, to this most
extreme form o f escape.
146
The Cry of Pain
Cry for help or cry of pain?
In the past, many authors have seen suicidal behaviour (especially

attempted suicide) as a ‘cry for help’.11 That is, the suicidal behaviour
is seen as motivated by impulses towards preservation rather than
self-destruction. Stengel12 wrote that suicidal behaviour functions as an
alarm system, possibly even as a social releaser in the Lorenz/Tinbergen
sense.
Schneidman spoke o f ‘cessation’, breaking this concept into sub
categories o f ‘intentioned, subintentioned, unintentioned, and con-
traintentioned’.13 Each categorization attempts to capture the prevailing
sense that suicidal behaviour varies in its ‘seriousness’ (hence a general
classification used by many clinicians into ‘gestures’, ‘ambivalent’ and
‘serious’). Certainly in the past it seemed that classification into two
main groups appeared to be justified: first, the depressed, alienated
group with a high risk o f death; second, a group whose act is highly
operant (behaviour maintained and modifiable by its consequences).
This is consistent with Beck’s conclusions that the two most salient
dimensions along which such behaviour can be judged are the ‘escape’
motive and the ‘communication’ motive.
Unfortunately, even within this overall classification, the motivation
for suicidal behaviour is complex. W riting in 1966, Kessel took exception
to those who thought that all suicidal behaviour was consciously or
unconsciously motivated by the wish to die.14 He accused suicidologists
o f deceiving themselves: ‘in the fact o f the obvious and the simple, they
cleave to the complicated, fortified against the evidence by the concept
o f unconscious am bivalence’. Kessel is right to be suspicious o f any
theory which assumes, a priori, that self-harm behaviour must always
have suicidal intent, and that, if not found, then it is ‘latent’. However,
such sweeping statements take little account o f the complexity o f
motivations for those who do kill themselves. The fatality o f the
outcome should not be seen as the only determinant o f the factors that
motivate the behaviour.
It may be better to assume that most suicidal behaviour, whether
the outcome is fatal or not, has some element o f the Janus face - mixed
147
feelings about the will to live and the will to die. As w e know, even
within the general population, many people have at some time felt
suicidal. Given that the lifetime risk for clinical depression is 20 per
cent, and that suicidal feelings are a com m on accompaniment to such
depression, it makes little sense to dismiss those who actually harm
themselves as not serious, or to dismiss the pressure to escape that the
person feels.
For these reasons, the cry for help, which many have misinterpreted
as a lack o f genuineness, is better seen as a cry o f pain. Suicidal behaviour
can have a communication outcome without communication being
the main motive. The behaviour is elicited by a situation in which the
person feels trapped. As with the animal in a trap that cries in pain, the
fact that the behaviour affects the behaviour o f other members o f
the species does not mean that the only motive for the cry was to seek
help. Suicidal behaviour m ay be overtly communicative in a minority
o f cases, but mainly it is ‘elicited’ by the pain o f a situation with which
the person cannot cope - it is a cry o f pain first, and only after that a
cry for help.
The ‘cry for help’ idea, though originally intended to be a neutral
theory about suicidal behaviour, has outlived its usefulness. It has
become limited to non-fatal suicide attempts and thus contributed to a
widening o f the gap between fatal suicide and attempted suicide. Also
it is almost always used pejoratively, or at least to imply that a certain
suicidal act was not so serious, but ‘m erely’ a cry for help. Such
behaviour is never ‘m erely’ anything.
Some self-harm m ay not be motivated by a wish to die, but most
shares with suicide the wish to escape from an otherwise unbearable
situation. The difficulty many authors have got into is to define com
pleted suicide as the core behaviour that needs to be explained, and
attempted suicide as its pale reflection. Instead, if one defines entrapment
and helplessness in theface of actual or threatened loss as the basic dimension,
it becomes easier to see that all self-harm falls som ewhere along the
dimension, w hatever the outcome. Such feelings o f being trapped are
fuelled and maintained by biological and psychological changes. They
appear completely impervious to intellectual argument, as Alvarez
found:
148
The Cry of Pain
. . . mere intellectual recognition did no good, and anyway, my clear moments

were few. My life felt so cluttered and obstructed that I could hardly breathe. I
inhabited a closed, concentrated world, airless and without exits. I doubt if any
o f this was noticeable socially: I was simply more tense, more nervous than
usual, and I drank more. But underneath I was going a bit mad. I had entered
the closed world o f suicide, and my life was being lived for me by forces I
couldn’t control.15
Entrapment
By the word 'entrapment', I mean anything that stops an animal or

human from getting away when it wants to flee. One striking example
o f the consequences o f entrapment from the animal world is the
behaviour o f birds establishing their territories. If birds meet within a
disputed territory, they engage in aggressive displays. One wins, the
other loses. The loser flies aw ay to find another territory. Note that in
this case, although it has been defeated, because it can escape it suffers
little ill effects from its encounter. But if this meeting occurs in a limited
territory, in a cage or other circumstance in which the defeated bird
cannot escape, it is a different story. Here is one early description o f
what happens in this case:
Its behaviour becomes entirely changed. Deeply depressed in spirit, humbled,

with drooping wings and head in the dust, it is overcome with paralysis,
although one cannot detect any physical injury.
The bird’s resistance now seems broken, and in some cases the effects o f the
psychological conditions are so strong that the bird sooner or later comes to grief.
(discussed in Paul Gilbert’s (1989) Human Nature and Suffering) 16
Note, first, that the ‘depressed’ behaviour occurs in the absence o f

physical damage. Second, it is striking that the defeat itself is not
sufficient to trigger the response. If the bird can escape to another
territory, it shows no ill effect. It is the combination o f defeat and the
lack o f escape (entrapment) that is needed for the reaction to occur (what
Gilbert calls ‘arrested flight').17
I suggest the sense o f entrapment is central to suicidal behaviour. In
149
the case o f the A level student with which the chapter began, there was
an unpredicted and sudden surge o f such feelings o f ‘no escape'. In
most cases, however, the entrapment is a longer-term state, first because
external causes o f stress are themselves long-term (e.g. the person is
trapped by bullying peers or partners). But it m ay be prolonged also
because the person has grown up in an environment where he or she
could exercise very little control, where others had all the power, so
the person has learned over a long time period that the only possible
response option is to submit. The person is thus extra-sensitive to social
threats, and their world constantly seems to present fewer alternatives
for action, whatever the reality. The result is long-term demobilization,
a biological state involving chemical changes.
The depressed person effectively ‘takes themselves off” the list" o f
those who might threaten the more powerful. This is akin to the
low-risk low-gain strategy o f some animals, a strategy that involves
internal inhibition o f any behaviour that would appear to challenge
those higher in the dominance hierarchy. Gilbert suggests that the
downturn into depression may involve the activation o f these inhibitory
mechanisms o f ‘no challenge'. At the extreme, they are associated with
not only reduced aspiration but also abnormally low aspiration: feelings
o f worthlessness, uselessness, powerlessness; and lack o f interest in
engaging in any social behaviour that might involve taking even min
imal risks. Depression is the biological assignment o f ‘loser’ status. It
signals to oneself and others that one is not prepared to take on
challenges or fight for resources.
The benefit in the animal kingdom is that the risk o f losing fights is
minimized, but the cost is that the chance o f dominating others is lost.
In the case o f bullying in schools or armed services or prisons, however,
the competitive nature o f the institution appears to encourage the bully
to continue to torment the subordinate long after the weaker one has
given up. The inhibitory mechanisms designed to protect the loser from
further losses fail to work. At such times the risk o f suicide is increased.
150
The Cry of Pain
Entrapment - evidence from history
One benefit o f the entrapment model is that it helps to explain historical

data. Take, for example, the increased risk o f suicide that children and
young people suffered in the late Middle Ages and early m odem period.
It appears the best explanation for this is the violence with which
children were mistreated and the fact that such brutalized boys and
girls had nowhere to turn. They therefore killed themselves to escape.
MacDonald and M urphy18 give the example o f a 12-year-old girl
called Agnes Addam who went horse riding with a girlfriend in 1565
and dirtied her clothes. She started to return home but then became
seized with terror over what her father would do to her when he saw
she had spoilt her clothes. She rushed into a pond and drowned herself.
Another example was a boy who in 1729 was frightened after he threw
a piece o f glass at his brother. He knew he would receive a severe
beating from his father and hanged him self in the outhouse. A third
example was the 13-year-old son o f a Hackney tradesman who was
victim o f a cruel practical joke. He lost his hat after his father had
threatened to beat him within an inch o f his life if he lost his new one.
The boy’s sister had overheard the threat and as a joke hidden the hat.
W hen the boy found it was missing, he hanged him self for fear o f the
beating.
The situation was compounded because when young people left
home in their early teens to w ork as servants or apprentices, the law
forbade them to flee the homes o f their masters and mistresses. This
only encouraged many such masters to mistreat them and abuse them
in a miserable way. Only the most flagrant brutality was regarded as
sufficient grounds for leaving a master. Neither could those boys or
girls who ran away expect much sympathy from their parents if they
returned home. A 12-year-old lad called Daniel Rose was miserable after
having been apprenticed to a w eaver and returned home. They warned
him and sent him back to his master. He left the house that evening
and was found dead the next morning. He had hanged him self in his
master’s garden.
Girls who found themselves pregnant would also be likely to kill
151
themselves. In some cases girls made pregnant by their master found

themselves isolated from their master and mistress as well as their
parents. Since there was no w ay out, suicide was common.
Running through these stories are the elements o f helplessness and
the impossibility o f escape, and particularly fear o f punishment. W e see
the same in contemporary societies with the fear o f punishment that
comes in a school system which is institutionalized in the penalties
society exacts for lack o f success at exams. Similarly we see it with the
fear o f bullying which is often the origin o f suicidal thoughts in people
serving in the army, navy or air force. Finally we see it dramatically
illustrated in the suicide o f those in prison, especially young people on
remand, who can escape neither the prison nor the terrible fear that
they will be bullied there.
Self-harm and suicide: connections
High suicide-intent self-harm and completed suicide do not require

a different theory from that needed to explain less serious suicidal
behaviour. The less serious behaviour represents the early active ‘pro
test' stage o f response to threatened loss o f rank, response to threatened
entrapment. It is later, after repeated exposure to social comparison
resulting in ‘loser’ status or ‘non-belonging/abandoned' status, that
the more severe manifestation o f conservation-withdrawal emerges:
self-denigration, worthlessness and despair, and with them the high
suicide intent. Low-intent suicide attempts can thus be seen as the
‘reactance' (the increased activity in response to threatened loss) which
precedes the ‘helplessness' (the decreased activity in response to actual
loss) associated with high intent-to-die suicidal behaviour.
The recent rise in younger males attempting and committing suicide
can be related to reduced role both through loss o f employment
opportunities and loss o f opportunities o f long-term relationship. This
latter effect arises from the increases in divorce and family breakdown.
In later life, separation or divorce or living alone can reactivate such
feelings o f being ‘less-favoured' together with the feeling o f being a
‘loser'. Such are the components o f Durkheim's ‘anomie' that he saw
152
The Cry of Pain
as closely associated with suicide. W e can now see how such anomie,
a sociological concept, relates to the psychological concept o f social
comparison, which itself relates directly to evolutionary old biological
sub-systems, each having a generally adaptive function, but sometimes
producing a maladaptive outcome. Social comparison has two import
ant elements: ranking (upwards v. downwards) and similarity to others
(like v. unlike). Those who are depressed and 'anomic' feel both inferior
to and also different from others.
Thus low self-esteem is a biologically old tendency that m ay be
activated relatively early in life, and reactivated at times when indi
viduals are, or imagine themselves to be, defeated, powerless and
failing to meet the challenges o f the world o f w ork or o f interpersonal
relationships. H ow is the gender effect to be explained within the
'entrapment' model? Historically, w om en have been given roles in
society in which they have less control, e.g. in earning power, control
over resources, in jobs even when working. In Western society, wom en
are likely to attribute their failures to themselves, and attribute their
successes to luck or to other people or circumstances. W om en have
thus explored the boundaries o f ‘no control' more than have men.
W hen the trap begins to close, they are more likely to see it closing.
The result is that they take action earlier than men. The cry o f pain
comes earlier in the entrapment process. At this early stage, people are
m ore ambivalent about dying, and use less lethal methods. The historic
tendency for w om en to be more likely to attempt suicide, but less likely
to commit suicide, is explained by this 'time course' model. In the past,
men did not show self-punitive behaviour at this stage. H owever,
having not reacted to threatened loss and entrapment early on in the
sequence, they miss out on the important benefits that earlier cries o f
pain may bring, and are more likely to m ove, in the extreme cases, to
helplessness and social isolation. W hen the trap closes, the effect is
more catastrophic.
For men, with less opportunity for receiving care (fewer social
support networks) and less opportunity for care giving (fewer caring
roles), the expectation o f no escape feeds directly into a sense o f
hopelessness and the possibility o f suicidal behaviour as the escape
route. As men become increasingly marginalized in the job market,
153
they may experience, as w om en have always done, more uncontrollable

aspects o f life at a younger age. The effect will be that they, too, show
suicidal behaviour at an earlier stage in the entrapment. This m ay be
w hy the numbers o f young men are beginning to match the numbers
o f young wom en in the attempted suicide statistics.
Concluding remarks
If social signals switch on old biological scripts o f defeat and submission,

then antidepressants m ay switch off these processes to give the person
more energy to explore other response options, with reduced sensitivity
to signals o f social loss and failure. Equally, psychotherapy m ay help
the person recognize their patterns o f thinking that see every situation
as potentially threatening, every encounter as win-or-lose. In this way,
the person m ay learn to modify their aspirations, to accept things how
they are so as to see better how they may change. The limits o f such
interventions are set by the persistence o f a real state o f powerlessness
arising from, say, bullying by other students, or domination by a violent
partner. H owever, psychotherapy has often been able to em power
people to change relationships, to m ove aw ay from non-supportive or
threatening individuals, to change the external factors that normally
conspire to keep them trapped.
H ow far a person finds alternative escape routes m ay depend on an
important psychological factor, a person's m em ory o f their own past
history. There is evidence that autobiographical m em ory may be so
affected that it reduces the ability to solve current problems and restricts
their view o f the future. The next chapter will consider this aspect in
depth, for it promises to provide a w ay in to break the vicious circle o f
biological and social influences on suicidal feelings and behaviour.
154
10
Memory Traps
I have suggested that a major factor in suicide and attempted suicide is

the feeling o f being trapped - trapped by both circumstances and one’s
own thoughts and feelings. The psychological aspect, the thoughts and
feelings, were seen as important in worsening the effects o f the external
realities. Though people rarely think themselves into a hole, once there,
thoughts about their own helplessness often maintain the depression
and prevent them climbing out. One response to these thoughts is to
view them as the symptoms o f underlying depression. W hen the
depression clears up, or is treated, they will stop.
In one sense this is true. W e know that such negative and hopeless
thoughts go up and down with depressed mood. W e know that, when
a person takes antidepressants, the thoughts may lose their grip, given
time. H ow ever, if the antidepressants are stopped, the risk o f becoming
depressed remains. Once a person has been clinically depressed, they
remain at an increased risk o f becom ing depressed again. In 50 per cent
o f cases, the depression returns within two years. For those who have
been depressed more than once in the past, the risk o f recurrence is 70
per cent. The thoughts return, the helplessness returns. The pills have
clearly not dealt with the underlying vulnerability to becoming
depressed and suicidal, much o f which m ay stem from a person’s
m em ory o f the past.
The importance o f m em ory is hard to overstate. M em ory provides
us with all our knowledge about who w e are, what we have been
through in the past: happy times and sad times. It is on the basis o f our
m em ory that we make predictions about the future. If our m em ory is
155
biased or faulty, then our predictions are also likely to be biased and
faulty. Our self-esteem also depends on m emory. Self-esteem is based
upon our past successes and failures, and how successfully we have
navigated a path through the world so far.
Depression is one o f the major vulnerability factors for suicidal
behaviour and depression does not only affect mood, it affects m em ory
as well. Memory provides the key to understanding how, when som e
one feels under pressure from their life circumstances, they begin to
feel trapped in a mental cage from which they appear unable to escape.
First, m em ory can be biased so that it tends to retrieve only negative
events. Second, m em ory can be over-general. Instead o f recording
specific events, a person tends to lump together events o f the same type
or category without distinguishing between them. The story o f how
this m em ory problem was discovered is interesting in its own right.
Memory bias - the horse race
I am writing this paragraph on the train on the w ay to a meeting o f the

British Association for the Advancement o f Science. As I gaze out o f
the w indow at the mountains o f North W ales on m y right and the sea
on m y left, m y mind wanders. I see a sandy beach, empty but for a lone
figure and his dog. It triggers a m em ory o f me walking along a similar
beach many years ago with my own collie dog. Such triggering o f
memories happens constantly, whether we are daydreaming or engaged
in conversation. Many o f the triggers are fairly neutral and ambiguous.
A man, out with his dog, might be ‘peaceful’ or ‘lonely’. In other
words, seeing something like that might trigger memories o f positive
or negative events from our past.
W hen something triggers or ‘cues’ m emory, the event seems to
come to mind immediately. This is not the case. There is a small
time-lag between the cue and the event com ing to mind. During this
brief interval there is a race between a number o f events which the
same cue could trigger. This same cue can activate a range o f m em ory
fragments, each o f which could be completed by a number o f different
memories. Some psychologists have likened m em ory retrieval to a
156
Memory Traps
horse race. W hichever event gets to the line first wins, and when it
wins, that is the event that comes to mind.
It is now known that such a ‘horse race' occurs between positive
events and negative events from the past. One can imagine positive
memories as a white horse, and negative memories as a black horse.
My seeing the man and his dog activated a pleasant m em ory o f long
ago; the white horse won. But one effect o f depressed mood is to bias
m em ory so that the black horse wins much o f the time. Had I been
depressed, the sight o f the man and the dog might have brought to
mind the day I had to take m y dog to the vet to be put down, when he
was very old and grey and ailing. Further, the depression would have
suppressed any happy memories o f subsequent dogs, thus allowing me
to conclude, unhappily, that I have never had a dog as wonderful again.
Some o f the first experiments on depressed m ood m em ory bias
phenomena were performed by Professor Alwyn Lishman at the Insti
tute o f Psychiatry in London and by John Teasdale and colleagues at
the W am eford Hospital in Oxford. Lishman concluded that patients
who were more depressed in m ood found it easier to recall negative
events, but a problem with this early study was that the apparent mood
m em ory ‘bias’ might have arisen because the people who were more
depressed had fewer positive events in their lives.
T o avoid this problem, John Teasdale took student volunteers who
were not depressed and experimentally manipulated their mood using
a Mood Induction Procedure. All students started by being randomly
allocated to different m ood groups, so that results could not be explained
by differences in the number o f positive and negative events in their
lives. The earlier effect was replicated, with the additional finding that
negative m ood did not so much speed negative events as slow down
the recall o f positive events.1
Taking the analogy o f a horse race, the finding that depression does
not seem to w ork by speeding up the black horse is important. In
psychotherapy for depression, it will not be sufficient for a client to
w ork to make negative events less accessible. The therapy will need
also to increase the availability o f more positive aspects o f the past.
There is abundant evidence that depressed people have suffered a great
deal o f genuinely negative life events and chronic difficulties. Given the
157
real difficulties o f these people's actual experience, the effect o f their

m ood on m em ory is to place an additional burden on them by making
it difficult for them to rem em ber those positive events, the recall o f
which might alleviate some o f the distress.
W hen I and m y colleagues began our research on suicidal behaviour
in the early 1980s in Cambridge, we wanted to find out what turns a
crisis into a suicidal crisis and thought bias in m em ory a good candidate.
Perhaps some people, at certain points, became so dominated by nega
tive memories that their past seemed nothing but a string o f failures,
disappointments and arguments. W e set out to see if this bias operates
in people who are feeling suicidal, even if they are not clinically
depressed. W e w ere to find something about their memories more
significant than the m em ory bias we expected.
Memory in suicidal patients
In research o f this sort, it is important to obtain as accurate a measure

as possible o f how fast people are in recalling events from their past.
W e also wanted to measure the time people took to recall events, so
we followed previous research in giving patients words one at a time,
a task reminiscent o f the word-association test pioneered by C. G. Jung.
But instead o f responding with the first w ord that came to mind,
participants were asked to respond with the first m em ory that came to
mind. By giving some positive and some negative words as 'cues', it is
possible to examine how long people take to recall positive and negative
events.
The first patient I gave the task to was a young wom an in her early
twenties who had been kept in hospital following an overdose. She was
still quite sad and hopeless, but had no hesitation in volunteering, saying
she was glad o f someone to talk to. I had carefully prepared my
questionnaires and m em ory tasks, choosing five positive words and
five negative words to use as m em ory cues. The words - happy, safe,
interested, successful and surprised; and sorry, hurt, clumsy, angry and lonely
- were chosen because a colleague at Cambridge had used them in a
m em ory experiment on a large number o f non-depressed people. These
158
Memory Traps
had found no difficulty in responding to the words with events from

their lives.
I had typed each word on the top o f a piece o f paper, and printed
the instructions on the top o f the first page. These asked the person to
look at each w ord in turn, and to write down an event from their past
which the w ord reminded them of. The event could be recent or have
happened a long time ago. It might be important or trivial.
I timed how long it took this first participant to start writing in each
case. She responded to each word. Later I found that to the cue word
‘happy’ she had written ‘m y father’; to the cue w ord ‘sorry’ she had
written ‘when I do things w rong’ ; to ‘surprised’ she had written ‘when
m y brother plays tricks on m e’ . This was not what was supposed to
have happened. She was supposed to be recalling specific events, but
one response was a person (her father), and the rest were summaries
o f many events.
W e tried the task with one or two other patients (who, like all the
others, had recently taken an overdose), and seemed to get similar
results. Perhaps asking them to write down their memories was the
problem. W e changed to reading out the words and asking for a verbal
response, which we then wrote down. Then we could prompt them if
they started to produce such general responses. Things did not improve.
Perhaps the instructions were not being understood. W e started to give
more detailed instructions and asked people to recall specific events,
defining them clearly as events that happened at a particular place and
time and lasted less than a day. W e gave some practice words until the
person was able to produce a specific m em ory and repeat the instruc
tions back.
The first consistent result to appear2 was that the suicidal individuals
do indeed take much longer to retrieve positive events from their lives,
though they are not much quicker than non-depressed controls at
retrieving negative events (see Figure io.i). In this respect they behaved
like the depressed subjects in the studies mentioned earlier. But the
feature that particularly interested us was the cause o f the delay in
retrieving a specific positive m em ory. It was because they persisted,
despite all our prompting, in retrieving an inappropriately general
m em ory when the cue word was first given.
159
Overdose Hospital Panel
Cue-word valence
Figure 1 0 .1 Time taken to recall specific autobiographical memories
(from Williams and Broadbent, 'Autobiographical Memory in
Attempted Suicide Patients', Journal o f Abnormal Psychology , 1986).
Table io .i shows some o f their responses compared with the

responses o f control participants, who were either people from the
Applied Psychology Unit Subject Panel in Cambridge or non-psychiatric
patients from the same hospital wards as the overdose patients. Control
participants respond to the cues by giving specific memories as asked,
but overdose patients are much more likely to respond by giving a
general m em ory which summarizes a number o f events. They do this
equally for positive and negative memories.
The mechanism of memory retrieval
M em ory for events in one’s life is hierarchically organized, with the

‘upper’ layers containing general m em ory information that can act as
pointers to the more specific and detailed ‘low er’ layers. Thus these
upper layers act as intermediate stages in the laying down and later
recollection o f events. When we try to recollect an event, we first find
160
Memory Traps
Table 10.1 Responses to cue words given by overdose and matched

control subjects (from Williams and Broadbent, 'Autobiographical
Memory in Attempted Suicide Patients', Journal o f Abnormal
Psychology, 1986)
Overdose patients
Cue Latency (secs) Response
Happy 3 Being with John

9 The night he told me he loved me
Sorry 26 Sorry if I've hurt anyone, any time

35 Arguments
53 This weekend, Friday
Safe 3 Being in my flat

12 Just when I'm sitting there
Angry 10 A lot o f the time

23 A lot o f people make me angry
35 Racial prejudice
46 I went to a party once - this Nigerian
wasn’t allowed in
Control participants
Happy 18 When I went to see my daughter in her

new house
Sorry 42 When I went to see my sister after her

husband had had a heart attack
Safe 6 After reaching home after driving a long

way (from Yorkshire)
Angry 7 Very angry after I found that my older

son had been misbehaving
161
an 'upper layer’ general description. This is then used to search the

'low er layer’ m em ory database for an appropriate candidate m emory.
For example, in response to the cue ‘happy’, people generate an inter
mediate description based on the implicit question, ‘W hat sort o f people,
activities, places make me happy?’ This ‘upper layer’ description may
be such things as ‘gardening’ or ‘m y girlfriend’ or ‘drinking in pubs with
friends’ . It appears that suicidal and depressed patients get stuck at
that intermediate stage, and cannot use the general descriptions they
generate to help them retrieve specific examples.
W hy do they abort the search for a specific m em ory at this intermedi
ate stage? Further research suggests that this may be partly because o f
the traumatic nature o f some past event or events, so that the search is
stopped as a w ay o f defending against the pain o f remembering.3
H owever, this would not explain w hy such people are as likely, or even
more likely, to give a generic m em ory in response to positive cues, i.e.
when the m em ory system is searching for a positive event.
Maybe patients become caught up at the intermediate description
level - a phenomenon called mnemonic interlock’ . Evidence suggests
that whenever the m em ory system attempts to retrieve an event using
a personal description, the description itself simply tends to activate
other general self-descriptions (as illustrated in Figures 10.2 and 10.3). In
these figures, one can see the cue word ‘sorry’ elicits the intermediate
description, ‘W hen I’ve hurt som eone.’ Instead o f proceeding to gener
ate possible memories which fit the general description (e.g. the specific
m em ory o f the time he received a letter telling him his partner was
leaving), the description activates further self-descriptive summaries
such as 'arguments’ and 'I always hurt people I love.’
W e begin to see how an event can become a ‘final straw ’ for suicidal
behaviour. The final straw is not simply one extra source o f stress. It is
rather any extra source which activates this network o f intermediate,
self-referent descriptions, causing mnemonic interlock.
162
Memory Traps
Intermediate
Figure 10.2 Stages in retrieval of specific autobiographical memory.
Figure 10.3 Effect of aborting search for specific memory. This

results in many repetitions of intermediate stage, resulting in over
elaborated self-descriptive categories ('mnemonic interlock').
163
Table 10.2 Scenarios used to cue memories (from Moore, Watts and
Williams, The Specificity of Personal Memories in Depression', British
Journal of Clinical Psychology, 1988)
My brother/sister criticized me.

A neighbour helped me with some practical problem.
My partner caused some practical problem.
A neighbour caused some practical problem.
My best friend reassured and encouraged me.
My partner reassured and encouraged me.
My best friend criticized me.
My brother/sister helped me with some practical problem.
My partner criticized me.
My best friend helped me with some practical problem.
My brother/sister caused some practical problem.
A neighbour reassured and encouraged me.
My best friend caused some practical problem.
My brother/sister reassured and encouraged me.
A neighbour criticized me.
My partner helped me with some practical problem.
Mnemonic interlock in depression
T o return to depression, to see whether depressed people who are not

currently suicidal have similar difficulties, recall that the early w ork on
m em ory in depression was concerned only with the probability or
latency o f negative versus positive memories; with the ‘horse race*
model o f memory. It was not concerned with the quality o f the mem ory
produced. Some indication that depressed people might have the same
difficulty was found in a study carried out by Richard Moore as part o f
his doctoral thesis at Cambridge.4 He asked depressed people questions
about their past, focusing on times when they felt supported (either in
practical terms, or emotionally) by a neighbour, friend, family member
or partner. Eight positive and eight negative scenarios were presented
(see Table 10.2) to depressed people and matched volunteers who were
not depressed.
The non-depressed participants gave general m emories 21 per cent
164
Memory Traps
o f the time. By contrast, the depressed participants were almost twice

as likely to give general memories (38 per cent o f the time).
Moore had used depressed volunteers. W ould the same phenom
enon be found in people who met criteria for m ajor depressive disorder?
Jan Scott and I tested the autobiographical m em ory o f twenty in-patients
with m ajor depressive disorder.5 W e compared depressed patients’
m em ory with twenty controls, matched for age, educational level and
performance on a semantic processing speed task, which had been
found sensitive to drugs. W e discovered that the speed with which they
recalled positive and negative memories from their past was biased.
They were much slower to recall positive events (confirming the
earlier researches on m em ory and depression). As we had anticipated,
depressed patients were much more likely to respond with over-general
memories. Other research clinics have replicated this result.6
The origins of over-general memory
Developmental psychologists have shown that retrieval o f events in a

summary, over-general form is a normal developmental phase before
specific-event m em ory emerges at the ages o f three to four years.7
Before this time children are likely to answer questions about what has
happened, say, at their nursery school or kindergarten, with a general
reply (such as, ‘On Tuesdays we have orange juice’). This occurs even
when unique events are arranged for children at playgroup (e.g. a
person visiting dressed as a witch). Following such an interesting event,
questioning on the subsequent day shows the children do not have very
good specific m em ory for it.8
Further research reveals that the children do have some mem ory
for the event, but eliciting details takes very careful questioning. It is
not the young child’s preferred method o f retrieving events in their life.
Could it be, then, that some children never adequately get beyond this
method o f retrieval? It is possible that stressful events at around that
time have the two effects alluded to earlier: making specific events too
traumatic to remember in detail and causing the child to be self-focused.9
Such self-focus means that attempts to recollect events (even positive
165
events) causes mnemonic interlock and he or she aborts the search. A

small amount o f circumstantial evidence points in this direction.
Willem Kuyken and Chris Brewin, working at the Institute o f Psy
chiatry in London, studied autobiographical recall in 58 depressed
w om en.10 Comparing their retrieval patterns with a matched control
group, they were able to replicate our earlier findings on the effects o f
depression on generality o f recall. H ow ever, o f even greater significance
was their finding that many o f these wom en (64 per cent) had been
sexually or physically abused in childhood and adolescence. The wom en
who had been sexually abused had even greater difficulty in recalling
specific events from their past, despite the fact that the abused situations
were not the events being asked about nor the events the wom en
retrieved.
The average age o f the wom en in this sample was 37 years, and the
traumatic events they experienced occurred prior to 17 years. Despite
this large gap in time, many were still having flashbacks o f the physical
and sexual abuse. Kuyken and Brewin used a questionnaire, the Impact
o f Events Scale, to assess the extent o f the intrusiveness o f thoughts
relating to the abuse and how much the person tried to avoid these
thoughts and images. Dividing those wom en who had been abused
into high and low scorers on the Impact o f Events Scale showed
that those still suffering intrusive thoughts, and trying to avoid such
flashbacks, had particular difficulties in retrieving specific events from
their past, either positive or negative.
These data suggest that general m em ory m ay serve as an indicator
that emotionally disturbing events have not yet been emotionally
processed. In a sense, this is unsurprising. Emotional processing involves
being able to return voluntarily to specific details o f events, sometimes
to give a greater sense o f control over them, sometimes to generate
alternative explanations and accounts o f them (e.g. to shift the blame
away from oneself). Generic m em ory prevents this emotional pro
cessing from taking place.
166
Memory Traps
The consequences of over-general memory
Therefore what sort o f variables are likely to make a person suffering

from chronic problems decide to make an attempt on their own life?
W hy this person; w hy now? The mnemonic interlock means that any
one small event (the final straw) can activate a network o f summary
self-descriptive memories. W hether these are positive or negative, they
are still likely to have damaging consequences. In the case o f negative
memories, the person's mind is quickly dominated by global self
referent descriptions such as T v e always been a failure', 'Nobody's ever
really liked m e’. But even mnemonic interlock around positive events
may have damaging effects. The person has no speedy access to specific
positive events which would allow him or her to generate specific
ideas for how to bring about similar positive events in the future.
This is likely to undermine attempts to solve current problems, to in
crease hopelessness and thereby finally to lengthen any episode o f
depression.
G eneral m emory and problem -solving

If the w ord ‘happy' simply brings back a sum m ary m em ory - ‘There
were lots o f happy times when I lived at hom e’ - the summary mem ory
does not deliver especially good hints for what to do about current
unhappiness. The much-loved home may have broken up, and the
resulting conclusion may be, ‘I cannot be happy.’ Over-general positive
mnemonic interlock can be as damaging as negative mnemonic inter
lock. By contrast, those able to retrieve specific events, positive or
negative, are more likely to be able to generate alternative problem
solving strategies for current problems. Those who, when given the
cue w ord ‘happy', are able to remember a specific event when they
lived at home, such as, ‘Going out with a friend to the cinema, where
we met Jan and Pete and went back to their house for a coffee', are
more likely to find in this some hints for what to do about their current
unhappiness: the event, the cinema, the friends, the coffee, etc.
The hypothesis that over-general m em ory impairs problem-solving
ability has been tested in a study by Julie Evans, working with the
167
Table 10*3 Situations used from the Means-Ends Problem Solving

Test (used by Evans, Williams, O'Loughlin and Howells,
'Autobiographical Memory and Problem Solving Strategies of
Parasuicide Patients', Psychological Medicine, 1992).
1. Someone who loves his/her partner very much, but they have
many arguments, after one o f which the partner leaves.
2. Someone who had just moved to a new neighbourhood and didn’t

know anyone, though he /she wanted to have friends in the
neighbourhood.
3. Someone who sees a person o f the opposite sex, eating in a

restaurant, and to whom they are attracted.
4. Someone who is having trouble getting along with the foreman on

his/her job.
5. Someone who comes home after shopping and finds that he /she
had lost a watch.
parasuicide counselling team at Kidderminster General Hospital in

England.11 They contrasted the m em ory performance o f people who
had recently taken an overdose with carefully matched control patients
who were in hospital for surgery. As well as using the autobiographical
m em ory tests, they assessed problem-solving ability using the M eans-
Ends Problem Solving Test (M EPS). In this test, individuals are given
the beginning and end o f a story and asked to fill in the middle section.
The five items used in the study are shown in Table 10.3.
Earlier w ork using this task with depressed patients12 had found that
problem-solving was more difficult for depressed patients, but had
not examined the link between problem-solving abilities and m em ory
function. W hat it had done, how ever, was to propose a scoring
system o f the M E P S task to take into account the effectiveness o f the
solutions people generated. An example o f this scoring system is given
in Table 10.4.
Not only were depressed patients unable to produce as many alterna
tive means o f solving the problems as were non-depressed controls, but
when they did generate a possible solution, it was less effective, i.e. less
likely to produce the desired outcome. Julie Evans found both that
168
Memory Traps
Table 10,4 Ratings of effectiveness of problem solutions (from Marx,

Williams and Claridge, 'Depression and Social Problem-solving',
Journal o f Abnormal Psychology , 1992)
In order to illustrate the scoring for effectiveness, and to illustrate the range o f
this dimension, two examples representing the extreme poles o f the scale,
from
i = not at all effective' to

7 = ‘extremely effective’
These ratings are made by independent judges.

The following examples are taken from the situation: ‘trying to make new
friends in a new neighbourhood’:
I wouldn’t know what to do in that situation. I have always had

someone to go out and make friends for me. I can’t imagine what
to do. (Rated i - Not at all effective)
First thing is to introduce herself to the immediate neighbours,

explaining that she had just moved in, possibly inviting the people
for a coffee anytime, and also if they seem interested in her, if
they invite her, making it clear that she intends to take it up
anytime. Chatting to people in the local shops, joining clubs,
offering to be helpful in some ways, e.g. baby-sitting, gardening
for old people. One tactic would be to get a dog and take it for
walks - that’s easy to get in contact. Similarly, if she has children,
it’s easy to get in contact. Inviting people round for dinner or
drinks. (Rated 7 - Extremely effective)
overdose patients produced few er w ays o f solving the problems, and

that what they did produce were generally less effective solutions.
W ere these difficulties associated with difficulty in retrieving specific
memories? The result confirmed, as predicted, a significant correlation
between the effectiveness o f solutions and the generality o f autobio
graphical memories.
The results were consistent with the hypothesis that deficits in
m em ory may play an important role in blocking access to effective
solutions to current difficulties, increasing the sense o f helplessness and
entrapment. These results have recently been replicated in two studies.
169
The first, by Gary Sidley and colleagues in Manchester, found a signifi

cant association between more over-general autobiographical m em ory
and poorer problem-solving effectiveness in 35 overdose patients.13 The
second, by Leslie Pollock and m yself in Powys, Wales, found not only
the link between m em ory and problem solving in people w ho had
recently attempted suicide, but that this group had more severe prob
lems in their m em ory than a carefully matched group o f psychiatric
patients.14 T o understand how all-pervasive this effect o f m em ory can
be, one only has to realize that such m em ory problems also feed forward
into a person’s view o f the future.
General memory and hopelessness

Hopelessness has been seen as the critical factor mediating between
depression and suicidality. Hopelessness about the future appears to
combine so lethally with depression as to produce suicidal ideas and
behaviour. Hopelessness has been found to predict repetition o f
attempted suicide six months later and completed suicides up to ten
years later. If over-general m em ory is seen as a significant effect on
future suicidal behaviour, then w e need to examine its effect on people’s
attitude to the future.
With m y colleague Andrew MacLeod, I began to investigate the
components o f hopelessness (see page 90). W e were particularly inter
ested in looking at the effect o f over-general memories on how specific
ally or vaguely a person imagined the future. W e reasoned that an
important element in hopelessness was that people cannot imagine any
future with certainty, either positive or negative.
One experiment asked people to rem em ber times in their life when
they had been either unhappy or happy. W e cued them with sentences
such as, T r y to remember an event in your life when you were in
tears’, or, ‘T ry to remember an event in your life when you were
laughing.’ Similarly w e asked the participants on another occasion, ‘T ry
to think o f a time in the future when you might be in tears’ , or, ‘T ry to
envisage a time in the future when you will laugh.’ W e coded the
specificity o f future images into three levels (Table 10.5).
Patients admitted to hospital following an overdose were com
pared with matched medical patients and matched non-hospitalized
170
Memory Traps
volunteers. The results were as predicted. Overdose patients were

more vague than controls in descriptions o f future and past. Further
more, subjects who were less specific about past and future were more
hopeless about the future. General m em ory can undermine the very
process by which a person constructs a specific future for themselves,
and thereby allows hopelessness to grow and develop without hin
drance.15
General memory and persistence of depression

Finally, if these m em ory deficits affect both problem-solving and levels
o f hopelessness in this way, then one might expect that people who
have such a problem will have unusually prolonged episodes o f
depression, since they will not have the resources to take advantage o f
any breaks in m ood that other treatments (e.g. antidepressants) might
bring about. Can w e therefore improve our prediction o f how long
depression lasts by taking account o f how patients perform when asked
about their autobiographical memory? Is it the case that two people
equally depressed will have different prognoses, depending on the level
o f specificity in their memories? This is an important question. We
Table 10.5 Ratings of specificity of future images (from Williams

et a l., The specificity of autobiographical memory and imageability
of the future', Memory and Cognition , 24, 116-25)
Try and picture a situation in the future where:
. . . ‘you make a mistake"
general intermediate specific
T il always be making Perhaps giving a friend My law exams in October

mistakes' the wrong advice
. . . ‘someone pays you a compliment’
general intermediate specific
‘A friend could’ ‘Someone at work may ‘Next week from my

say I’ve lost weight’ husband when I have my
hair cut again’
171
know it is the longer-lasting depressions which are most likely to lead

to suicide.
Some results from the early literature on psychotherapy process are
consistent with the hypothesis that specificity o f recall affects rate o f
progress in overcom ing emotional problems. One early investigation
examined taped transcripts from over a hundred sessions o f group
psychotherapy with hospitalized patients o f various diagnoses. It found
that clients who described current feelings and life situations in a
concrete and specific w ay w ere more likely to make progress in
therapy.16 Does non-specificity therefore have a predictive effect on the
course o f depressive illness?
With the help o f colleagues, Andy Brittlebank, Jan Scott and Nicol
Ferrier, we examined this question in some seriously depressed patients
admitted to a psychiatric unit in Newcastle upon T y n e.17 W hen admit
ted, patients completed the Hamilton Rating Scale, widely used to
measure the severity o f depression, and the Autobiographical M em ory
Test. Each patient was followed up after three and seven months. The
more over-general in m em ory the patients were at admission, the worse
their Hamilton Depression Scale scores at both three and seven months.
Dividing patients at admission into high and low over-generality in
response to positive cues (found to be the m ore sensitive indicator), we
examined the outcome at seven months for each sub-group. O f nine
patients who were 'over-general to positive cues', only one had re
covered. O f ten patients who were ‘specific to positive cues’, eight had
recovered. These results suggest that this aspect o f m em ory is a pow er
ful determinant o f how long the depression lasts. It appears that for
depressed people who have this additional problem, their m ood will
appear more pervasive and durable, with the result that these individuals
will feel more helpless, and less motivated to engage in activities which
might otherwise lift their mood. Further research by Allison Harvey
and colleagues has shown that Acute Stress Disorder following a road
traffic accident lasts longer if the victim has few er specific m emories for
the event.18 In addition, research by Tim Dalgleish and W illem Kuyken
shows that people who suffer from Seasonal Affective Disorder who
have fewer specific memories (assessed in autumn when they are
depressed) are more likely to remain depressed the following sum m er.19
172
Memory Traps
These results show the importance o f m em ory in contributing to

the trap from which the suicidal person wishes to escape. The past is
dominated by generalities, which undermine the ability to see an
effective w ay to solve current problems. If this were not bad enough,
the m em ory problem has an additional effect on how the future is
viewed, contributing to the feeling o f vagueness about the future.
Non-specificity about past and future is significantly associated with
hopelessness, that aspect o f depression which most points towards
suicidal behaviour as an option. As we will see in Chapter 12, a major
characteristic o f successful treatments for suicidal feelings and behaviour
is that they induce the person to be very specific about the past, often
with the use o f diaries to record daily events.
Concluding remarks
This chapter gives an account o f w ork in progress. Our aim is ultimately

to explain w hy some people take drastic action to harm themselves or
end their life. W e need a theory which can relate to other perspectives:
sociological, psychiatric and biological. W e need a theory which
promises to specify the nature o f the final common pathway leading
from depression to hopelessness, then to suicide. W e now feel in a
better position to understand w hy some people are more vulnerable to
suicidal behaviour than others.
An inability to be specific in retrieval o f personal memories may
signal a history o f negative events, a history which is unprocessed and
still has pow er to interrupt the present with intrusive thoughts. Such
general memories lead to a situation in which people are vague about
the future and fail to produce good effective problem-solving alterna
tives. It sets the context against which the final straw, that event which
produces a global summ ary o f the emptiness o f one’s life, may have its
devastating consequences. If life circumstances are the factors that put
a person in a cage, it is m em ory that springs the door closed.
173
11
The Prevention of Suicidal
Behaviour
In assessing what scope exists for preventing suicide, one o f the most
obvious measures would be the more sensitive assessment o f suicide
by health professionals. Depression is a frequently occurring psychiatric
disorder and most depressed patients are treated in general practice.
Training primary-care professionals to recognize depressive symptoms
and treat them appropriately is a potentially important prevention
strategy. Evidence suggests that 50 per cent o f sufferers are not recog
nized by their general practitioner as suffering m ajor depression. A
further 10 per cent are subsequently recognized, and o f the 40 per cent
not recognized half will remain depressed. Tw enty per cent o f the
original sample, therefore, are still depressed and not recognized six
months later.
There are a number o f reasons w hy depression is missed in the
context o f a G P ’s clinic. The patient m ay present somatic symptoms
and physical problems; m ay feel there is a stigma about presenting
psychological problems, compounded by a belief that the doctor does
not have time to listen to their psychological problems. There is also
some evidence that depression is more likely to be missed if it is o f
recent origin, where the symptoms are atypical, the depressed m ood is
less severe and the patient has less insight into their own symptoms. A
picture emerges o f a complex interplay o f factors which result in the
primary-care worker not spotting depressed symptoms.
Depression and hopelessness are clearly associated with suicidal
ideation and behaviour. If depression is going unrecognized, then sui
cidal ideation is also likely to go unrecognized. Implicit blame is often
174
The Prevention of Suicidal Behaviour
put on the physician or mental-health professional who has failed to

pick up the signs, yet it is never quite so simple. Data from 20 years ago
seemed to suggest that between 50 and 80 per cent o f people who
commit suicide had seen a doctor up to one month prior to death.
More recent reviews suggest that only 20 to 25 per cent o f those
committing suicide have seen their G P or other health-care professional
in the w eek before death, and 40 per cent in the month before. It is
likely that the decrease in suicide among the older population and the
increase in younger men accounts for this change. Young men are
much less likely to visit their doctors than young wom en or older
men.
Identifying scope for improvement in the recognition o f depressed
symptoms in general, and suicidal symptoms in particular, need not be
associated with blaming the primary-care physicians for not spotting
them in the first place. W e need to understand more about the interper
sonal processes that govern the interaction between doctor and patient.
These were studied by David Coom bs and colleagues at the University
o f Alabam a,1 who examined the consultations between care-givers
(physicians, psychiatrists, psychologists and mental-health counsellors)
and patients who had subsequently attempted suicide up to three
months later. Thirty-six o f the 50 patients studied (72 per cent) had, by
their own accounts, sought professional help for depression or suicidal
thinking in the three months prior to the episode. H owever, according
to the patients’ own reports, they made direct reference to suicidal
thinking in only 23 per cent o f all consultations. H ow much this was
because o f reluctance to talk about these feelings, and how much a
result o f the professional failing to ask the right questions, is an open
question. According to the professionals’ reports, they said they asked
about suicidal feelings in 48 per cent o f cases.
Coom bs also wanted to see whether people were less likely to report
depression and suicidal feelings to physicians than to mental-health
professionals. They found they were much less likely to disclose symp
toms to the physicians. This confirms that a major determinant o f what
is talked about is the context o f the visit. Even when a person intends
to communicate emotional and psychological distress, once they enter
the context o f the physician’s surgery, the evidence o f physical medicine
175
all around them elicits talk about their physical symptoms rather than
their psychological difficulties.
Another indication o f how the interpersonal context determines
what takes place between health professional and patient is a finding
about how the age o f the patient determined outcome. Patients o f 55
or older were not asked at all about suicidal thinking. Given that these
patients were the most vulnerable for completed suicide, it seemed a
surprising result. One possible explanation is that older people are more
likely to express problems through physical symptoms. Another is that
care-givers feel more reluctant to ask about psychological symptoms o f
people older than themselves. Social mores that dictate respect for older
people and their privacy m ay be a more powerful determinant o f
professionals' behaviour than previously thought.
Suicide prevention through keeping in touch

Diego De Leo and colleagues conducted an interesting study, reported
in the American Journal of Psychiatry in 1995. They were investigating
the usefulness o f a TeleH elp/TeleC heck service for older adults. The
TeleHelp was a portable device that allowed patients to send an alarm
signal if they needed help. The TeleCheck service was a regular (twice
a week) check by telephone to see if people were all right. Over four
years there was only one suicide in a population o f 12,135, where the
expected number would have been seven or more. They also found
that patients required fewer home visits, fewer hospital admissions and
that their mood was significantly better.2
Can simply keeping in contact with people have such a dramatic
effect? An earlier study by Jerom e Motto seemed to suggest that this
could happen.3 The study was conducted in San Francisco, and con
cerned people who had attempted suicide. They were contacted one
month after the event and asked if they had taken up offers o f help.
Those who had not done so were randomly allocated to one o f two
groups: a contact group and a no contact group. The contact group
received a note every month (for four months), every tw o months for
the next eight months, and every three months thereafter for the next
four years (a total follow-up o f five years). O ver this five-year period
there was a significantly low er number o f suicides in the contacted
176
group. A later report stated that these differences later disappeared, so

that after four years there was no difference in the groups. This seems
at first sight to indicate that the contact had not worked in the long
term. On the other hand, since the study was stopped after five years,
if it had been the regular contact that had made the difference, one
might expect a rise in risk when the contact itself came to an end.
Motto's study concerned those who did not seek, or who refused
treatment. But what o f those who come for help?
The Gotland study
If people do seek help, what evidence is there to suggest that any

appropriate help at all can be given? In 1989 an important study by Rutz
and colleagues in Sweden investigated the possibility that recognizing
and treating depressed patients in the primary-care setting would affect
rates o f suicidal behaviour.4 They conducted their study on the island
o f Gotland, which has a population o f 56,000, a single psychiatric
department and 18 G Ps. In 1983 and 1984, all the G P s attended two
education programmes given by the Swedish Prevention and Treatment
o f Depression Committee. The aims were to increase their knowledge
o f diagnosis and treatment o f depressive disorders. Interestingly, the
study monitored suicide rates not because it was thought these would
decrease, but because o f w orry that they might increase if more cases
o f depression were inappropriately treated in general practice rather
than being sent to the psychiatric department.
The resulting suicide rates for 1982-5, shown in Figure 11.1, reveal a
significant decline on the island o f Gotland compared with the rates for
the same time period on mainland Sweden. H owever, a 1992 report
showed that this effect was only temporary, owing to the fact that some
o f the doctors who had been trained subsequently left the island.5
The authors concluded that there was a significant effect from their
intervention, but that the educational programme needed to be main
tained if it was to continue to have an impact. They also showed that
training G P s increased the appropriate use o f antidepressant medi
cation, with a parallel decrease in the use o f hypnotics, sedatives and
177
1982 1983 1984 1985
Figure 11.1 Suicide rates in Gotland and the mainland of Sweden,

during and after systematic education of Gotland GPs on diagnosis
and treatment of depressive disorders.
major tranquillizers. There was also a decrease in patients taking sick

leave from work. These results suggested that the benefits o f such a
training programme were likely to generalize beyond simply the treat
ment o f the symptoms o f depression which were its primary focus.
Criticism of the Gotland study

The interpretation o f the Gotland study has been controversial. Because
it was not a randomized trial (with a control group o f G P s w ho did not
receive the training), some have said its results should be treated with
extreme caution. The reduction in suicide rate m ay have happened by
chance. Other critics have gone further, and said that even the published
results give a misleading impression. A letter to the British Journal o f
Psychiatry, claiming that the numbers committing suicide on Gotland
had started to fall well before the education programme was intro
duced,6 included a graph showing a reanalysis o f the data. The graph
seemed to show a long, slow decline in the suicide rate before the
education programme was begun. Because the study is the only one o f
178
its type, it is important to consider this last criticism with particular

care.
Reply to criticism
In fact, the writer o f the letter had made a simple but critical error in
his reanalysis o f the Gotland data. Instead o f examining the suicide rate
year by year, he had calculated a five-year m oving average. Taking a
m oving average o f such data is a well-recognized method o f ‘smoothing
a graph so that trends m ay be seen more clearly. In the procedure, a
‘w indow ' o f consecutive points in a series (between two or five points
usually) is taken, and the average used as a new data point. So, a
five-year m oving average takes data points from years i, 2, 3, 4 and 5
for the first data point; years 2, 3, 4, 5 and 6 for the second; 3, 4, 5, 6 and
7 for the third, and so on. The five-year ‘w indow ' is m oved across the
data and the graph is smoothed.
H owever, if an intervention is introduced, the m oving w indow must
stop at the last data point before the intervention. Otherwise the window
is gradually contaminated by more and more data points from the
intervention phase. T o illustrate this, consider the dum my data in
Figure 11.2 which clearly show the effect o f an intervention. Figure 11.3,
however, shows what happens to this graph if a five-year m oving
average is applied which does not stop when the intervention starts.
The baseline data are contaminated increasingly by the data from the
intervention phase, giving the w rong impression that the rate o f the
behaviour in question started to decline before the intervention -
precisely the mistake made in the critical letter.
Figure 11.4 shows the smoothed means o f the Gotland data retaining
a five-year m oving average but stopping the window prior to the
intervention. As can be seen, the intervention follows three successive
drops in the smoothed graph (o f 1.9 ,1.2 and 2.7 per 100,000 respectively).
In this sense, the author o f the critical letter was right to point out some
reduction in the suicide rate prior to the intervention phase starting,
but his error made this reduction seem much greater than it was. The
first post-intervention data point is 7.0 per 100,000, low er again than the
last point prior to the intervention - a much larger drop than that
between any two prior data points.
179
30 1-
20 -
10 -
o i 1 1 1 ___ 1_____1__ 1_____ 1_1_____ 1____ 1_________ 1_1_1____1----- 1-------- 1--1----- 1------1----- 1—
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time
Figure 11.2 Dummy data illustrating effect of hypothetical
intervention to reduce death-rate.
Time
Figure 11.3 Graph illustrating effect of applying five-year

moving average to data in Figure 11.2 if moving average is not
stopped prior to intervention.
O f course, without a controlled study, the Gotland data remain

preliminary. But the implications for suicide prevention by G P s cannot
be ignored. Nevertheless, it is right to point out that such intervention
m ay be extremely difficult to achieve.
180
Years
Figure 11.4 Graph showing the smoothed means of actual Gotland
data retaining a five-year moving average but stopping the window
prior to intervention.
Primary care and primary prevention: some cautions
W e have seen that physicians sometimes miss the seriousness o f

depression, treating the single symptom o f insomnia or agitation with
out checking for other symptoms. H owever, it remains painfully true
that the chances o f a G P successfully spotting a suicidal patient are
relatively low, given that there is likely to be a suicide only once every
four to five years among his or her patients. Because o f the 40 per cent
rate o f contact o f suicide victims with their doctors in the month prior
to suicide, this means that a G P is likely to receive a consultation from
somebody in the month before they commit suicide only once every
eight to ten years. This makes it an extremely rare event compared
with the large proportion who consult their doctors with emotional
problems.7
Some have suggested that just because suicide is rare does not mean
the doctor should not test for it. If a G P sees a child with a fever, he or
she will test for meningitis, although this is also very rare. By analogy,
it is suggested that the G P should spend some time testing for suicide
risk. H ow ever, physical tests for rare diseases are not a fair analogy.
Physical tests often produce a greater certainty o f diagnosis, and even
if they do not, an 'at risk’ result m ay warrant further, more specialist
testing. Yet even if a health professional used every available measuring
181
•
instrument to assess suicide risk, the evidence suggests that he or she

would successfully detect only half the people who are going to commit
suicide. For this number, m oreover, he would ‘detect* an enormous
number o f false positives. His clinic would be full o f people whom he
suspected might kill themselves, but who, the statistics show, will not
actually harm themselves. Taking the analogy o f a radar screen, the
doctor is being expected to perceive an image on the screen which will
occur three or four times in a professional life, against the background
o f many similar images occurring at every daily surgery.
The general problem of prediction: sensitivity

and specificity
W hy is prediction so difficult? If w e take the com m on predictors o f

suicide (see Table ii.i), w e find a number o f factors that w e feel should
help identify those most at risk.
There are two problems in predicting such behaviour: sensitivity
‘not having many misses') and specificity (‘not having false positives').
An assessment device needs to have adequate sensitivity (i.e. a high hit
Table 1 1 .1 Predictors of suicide
Mental disorder, especially depression

Alcohol or drug abuse
Suicide ideation, talking and planning
Prior attempted suicide
Lethal methods
Isolation, living alone, few social supports
Hopelessness
Being white male
Suicide in the family (modelling, genetic loading?)
W ork problems, unemployment, high-risk occupation
Family and martial problems
Stress and negative life events
Problems o f affective control (anger, impulsivity)
Physical illness
182
rate for ‘true positives', predicting suicide and getting suicide). It must
also have adequate specificity, predicting the true negatives (predict no
suicide and get no suicide).8 The issue o f specificity is particularly
significant because o f the low base rate o f suicidal behaviour. Thus,
even in a highly selected group, there will be more people falsely
identified as at risk (false positives) than correctly identified as at risk
(hits). W hen trying to predict suicide the problem is even greater.
Pokom y, in a prospective study,9 followed up almost 5,000 in
patients. The predictive model identified 35 out o f the 67 subsequent
suicides but at the cost o f over 1,000 false positives. W hichever analysis
Pokom y used, he found the result was a false positive rate o f 25-30
per cent. O ver 1,000 people would have been put on the vulnerable list,
but not have committed suicide. Perhaps more worrying was the 44
per cent false negative rate. This means that o f the people who did
com m it suicide, 44 per cent would have been allocated to the ‘low risk'
category.
This pessimistic message is reinforced by the suicidologist, Ronald
Maris.10 He points out that so-called 'high risk' groups are not, after all,
at that high a risk o f suicide. Take the statistic that 15 per cent o f those
who have been patients in psychiatric units with a diagnosis o f major
depression will commit suicide. They are at a greatly increased risk
compared with the general population. Yet even they will commit
suicide at a group rate o f some 1 per cent per year over a period o f
about 30-35 years. For every 100 patients in the ‘suicide high risk'
category, only one will actually commit suicide in any one year, and
we cannot be sure which one, or when. The question o f ‘when’ is one
o f the most difficult to answer.
Timing
It is clear that assessment o f suicidal risk cannot merely be done at a

single point in time, on the basis only o f predisposing (and pre-existing)
vulnerability factors. Changes over time need also to be taken into
account. Are there any indications about what the vulnerable times
might be?
183
Perhaps the greatest indication that change in circumstances puts

people at greater risk is the finding that the period following discharge
from psychiatric hospital is the time o f most suicide risk. This is one o f
the major conclusions o f Lewis Appleby and his colleagues who
conducted the U K ’s National Confidential Inquiry into Suicide and
Homicide by People with Mental Illness.11 The inquiry found that, o f
some 10,000 cases o f suicide (or cases where open verdicts had been
returned) over two years, some 2,400 had been in contact with mental
health services in the year before their death, and details were available
for 2,177 o f them. O f this sub-sample, 358 (16 per cent) had occurred
while the person was an in-patient (this represents 3.6 per cent o f all the
suicides over that period). Five hundred and nineteen (24 per cent)
occurred within three months o f discharge from hospital. The statistics
confirmed that the risk o f suicide is highest immediately after discharge,
and falls day by day during the first week, and w eek by w eek thereafter.
In terms o f prevention, it seems that better aftercare immediately
following discharge is a major priority.
On the other hand, w e need to bear in mind that over the two-year
period o f the study, there would have been around 300,000 admissions
to psychiatric units and wards. Knowing who is at most risk might seem
to be the best w ay o f planning aftercare, but this study found, as others
have done, that the usual characteristics that indicate higher risk o f
suicide in a com m unity sample (being male, single, living alone,
unemployed, having a history o f alcohol or drug abuse) do not identify
a high risk group in a sample o f people who have been admitted to a
psychiatric unit.12 Further analysis o f the data showed that 45 per cent
o f suicides by people who have been in touch with a psychiatric service
in the year before their death will be by those who have harmed
themselves before, and either have a history o f alcohol and drug abuse
or a previous admission to a psychiatric unit. The inquiry also raised
concerns about the fact that many o f the patients had not been taking
the medication that had been prescribed for them.
Finally, a further study by the Manchester group has begun to
investigate whether there is any aspects o f a person’s behaviour while
they are in hospital that might indicate a special risk o f suicide later.13
Perhaps unsurprisingly, they found that previous admission under the
184
Mental Health Act, together with actually expressing suicidal ideas

while in hospital, predicted suicides within three months o f discharge.
More seriously, given Lewis Appleby’s model o f suicide14 that sees such
a tragedy as being the result o f a cumulative build up o f risk factors
over time that are not counterbalanced by a corresponding rise in
protective factors, a further study found post-discharge suicide to have
often taken place following a reduction in care. This reduction o f care
had been decided at the final appointment with the health professional
responsible for the patient. It seems that only in retrospect is one able
to see w ho was at risk. Beforehand, it remains extremely difficult. In
the absence o f such predictive power, the protective factors that might
be available to the health professional (for example, to w ork collabor-
atively with the person to review difficulties they may be having with
medication, or in other areas o f their lives) are simply not mobilized,
for there seem to be few w ays in which scarce resources can be allocated
rationally.
These studies all confirm that the time following discharge is the
most vulnerable. But other studies show that any time o f change can
increase risk. For example, in prison populations, the first twenty-four
hours are the most risky time. O ver 50 per cent o f prison suicides occur
in the period immediately following incarceration. As Maris points out,
the vulnerable times for psychiatric patients are also times o f change: a
m ove out o f hospital, whether on a week-end pass or at discharge, or
even a m ove within the hospital to another unit; or change in symptoms,
such as an im provement in activity, sleep and appetite that may precede
a change in m ood and hopelessness; changes in social, family or living
arrangements.15
This is an under-researched topic, so any theory must be speculative.
Nevertheless we can narrow the possibilities. I suggest that any disrup
tion to a settled routine triggers a state in which the individual reviews
goals and plans. They start to compare current circumstances to goals
- what they would like to be the case. The review includes an assessment
o f their own energy and ability levels needed to achieve these goals. If
the individual estimates they do not have the energy or resources to
reduce the gap between current reality and future goals, they are in
danger o f feeling even more hopeless. . y*'.v
The change in circumstances that can trigger such a review m ay be

subtle. It may be a change in the external environment, or a change in
the ‘internal’ environment, i.e. in a person’s m ood state. But how ever
it has been triggered, once the review has started the individual may
find it very difficult to switch it off. The ruminative habits o f thought
in such circumstances can produce, for the individual, extremely pessi
mistic self-assessment, and further catastrophic downward spirals in
mood. These are the critical periods for suicidal thoughts and behaviour,
when the availability o f the means o f suicide becomes an important
factor.
Availability of means
The devastating effect o f the availability o f lethal substances is illustrated

by the case o f a brilliant Cambridge postgraduate student who killed
him self with a lethal injection early in 1995. He had been the best A
level biology student at his home school. At Cambridge, he performed
as well as his early school career had promised, com ing top in all
subjects in the university examinations out o f all the medical students
in his first and second years. At 24 years old his tutors thought him one
o f the brightest students o f the past decade. It was clear he was heading
for a brilliant academic career, and he started on a course that would
eventually give him not only a medical qualification but a PhD as well.
But his private life was not going so well. According to friends and
tutors, he loved a wom an who did not love him. His w ork seemed no
compensation for such unrequited love. He was quite a private person,
but nobody expected the devastating effect this would have. At a party
in Cambridge one Saturday night he appeared cheerful, and may have
gone home briefly before going to the laboratory. It was there that the
availability o f lethal substances had its effect, for none o f his friends or
tutors doubted that, had he been able to get over his acute feeling o f
desperation, the weekend would have passed and he could have recov
ered his composure. But the laboratory contained chemicals used on
animals in research - substances lethal if used in an unsafe manner.
They were at hand when he felt at his worst, and suicide was the result.
186
Clinicians and researchers have had little doubt for some time that
if help and protection are available during a period o f suicidal crisis, and
lethal methods are not to hand, the crisis m ay pass and the person not
commit suicide. It m ay not even be help that is needed, but simply
distraction at a critical moment. One person who put the muzzle o f a
revolver into his mouth and was feeling for the trigger, suddenly heard
his children laughing and running through the hall. 'That snapped me
out o f it. The suicide impulse lasted only a moment - but that's all it
takes,' he reported later.
The basis for expecting that availability o f lethal methods makes a
difference is this: if the preferred method is not there, the motivation
to search for an alternative m ay not be high enough to prompt such a
search. The suicidal feelings m ay pass without being acted upon. If this
seems difficult to understand, we only need recall that suicide occurs in
the context o f hopelessness and despair. If a single opportunity appears
to present an escape, yet that escape route is blocked, the despair may
turn into hopelessness about suicide as an effective solution. The
‘Russian roulette' aspect o f suicidal motivation produces a sense o f not
caring ‘whether I live or die'. If a person has effectively ‘allowed the
Fates to decide', then their verdict in favour o f staying alive may be
passively accepted. But if, at that moment, the person has access to
lethal methods, the outcome will be suicide.
O f course, a person determined to kill him or herself may take many
steps to ensure success, including taking themselves away to a place
where they will not be interrupted. Even where this does not occur,
the family or friends o f a suicidal person cannot be on hand 24 hours a
day. T o try to be constantly present in this w ay would put so much
pressure on a relationship it would be unsustainable. Nevertheless it is
possible to try to ensure that lethal means o f suicide are unavailable for
those attempts that are more impulsive. The problem is knowing which
threats to remove.
In Savage God, Alvarez quotes the views o f Seneca, who said that the
means o f committing suicide are everywhere: each precipice and river,
each branch o f each tree, every vein in the body will set a person free.
Alvarez disagreed: ‘No one is promiscuous in his w ay o f dying. A man
w ho has decided to hang him self will never jum p in front o f a train.
187
And the more sophisticated and painless the method, the greater the
chance o f failure: I can vouch, at least, for that.’ If true, this means that
a health professional may ask someone who is suicidal what they have
thought o f doing, and try to ensure that these means are removed.
Research evidence
The preventative effect o f rem oving such lethal means has been demon
strated many times. The most com m only cited example is the fall in
suicide rates during the 1960s and 1970s in the United Kingdom as
domestic gas was detoxified. In 1948-50, poisoning by domestic gas
accounted for 41 per cent o f male suicides and 60 per cent o f female
suicides. By 1970, only 16 per cent o f males and 9 per cent o f females
used domestic gas in suicide. Death by this method had completely
disappeared by 1990. It has been estimated that the detoxification o f
domestic gas has prevented approximately 6,700 deaths by suicide.
Similarly, there was a decline in the suicide rate in Australia in the
late 1960s and early 1970s, an effect directly attributable to legislation to
reduce amounts o f barbiturate and other sleeping pills. Before the mid
1960s, drug overdose was the most common form o f suicide in Australia.
Up to this time, 100-300 tablets or capsules o f sedative sleeping pills,
such as barbiturates, had been available, making suicide by their use
extremely easy. In Ju ly 1967 it became illegal to prescribe sedative
hypnotic drugs, particularly barbiturates, in greater quantities than
25 tablets or capsules. Statistics for Australia showed that restricting
barbiturates in this w ay made the suicide rate fall in ensuing years.
H owever, the reduction in overall rate from the unavailability o f
one method may gradually bottom out and alternative methods become
more common. In England and W ales this occurred with the use o f car
exhaust replacing domestic gas. Car exhaust now accounts for 35 per
cent o f male suicides in England and Wales, compared with 31 per cent
w ho hang themselves and 14 per cent who take poisons. The pattern
for females is slightly different, with self-poisoning remaining the most
common method (44 per cent), followed by hanging (23 per cent) and
vehicle exhaust (13 per cent).
The increase in vehicle exhaust deaths parallels the increase in m otor
vehicle use in the past twenty years. In fact it slightly exceeds the rate o f
188
increase in car usage, but this may be owing to the fact that within the
car market there has been an increase in the proportion o f hatchback
models, which makes suicide by this method easier. Some have advo
cated changes to exhaust-pipe design, to make it more difficult to attach
a hose, though ultimately it m ay be more useful to install safety cut-out
switches in cars to switch o ff the engine in response to high levels o f
fumes. In any event, as emission controls on car exhausts are adopted
by more and more countries, we can expect to see a reduction in suicide
using car exhaust in the coming years.
Dangerous weapons
Suicide rates using firearms are, o f course, another indicator that avail
ability o f means is important. Guns are used in over 50 per cent o f cases
in the United States compared to only 3 per cent in Great Britain. In the
United States the impact o f availability has focused on the strictness
o f gun legislation from state to state. Lester (1989) found significant
correlations between ‘gun control statute strictness' and rates o f suicide
over the 48 continental states.
Although stricter gun laws correlated with few suicides by firearms
(suggesting that making the method less available will reduce fatalities),
there was also an association between low er firearm availability and
the use o f alternative methods (though not poisons or hanging). Despite
this, there are reasons to think that availability is important. The positive
correlation between strict gun laws and alternative methods does not
tell us the overall level o f suicide by these other methods. Indeed, other
w ork by Lester16 has shown that the total suicide rate is lower in states
with strict gun control laws. It appears Alvarez is right: only a few
suicidal people switch to an alternative method for suicide. People seem
to have a preferred method they would use to kill themselves, and are
unlikely to deviate from this when actually suicidal. This is important
clinically, for depressed patients, when not very suicidal, may agree to
get rid o f the means o f killing themselves that they know they might
use later when feeling worse.
189
Dangerous medication
Other evidence o f a link between suicide and availability o f method is
shown by the findings o f a marked correlation between the number o f
prescriptions given out for sedatives, sleeping pills and tranquillizers,
and the rate o f suicide by self-poisoning.
Prescribed mood-altering drugs, such as tricyclic antidepressants,
account for 15 per cent o f all suicides. Some suicides may therefore be
prevented by the prescribing o f relatively non-toxic antidepressants.
One problem with prescribing safer antidepressants is that they are
more expensive, and difficulty in predicting who is most vulnerable to
suicide at what point in time means there might have to be a general
switch to these newer substances to bring about a reduction in the
suicide rate. Such a switch would mean that the cost per life saved
would be very high indeed. On the other hand, use o f these safer
antidepressants in those known to have a higher risk o f suicide (psychi
atric patients, those with a history o f self-harm, etc.) appears an im por
tant potential development.
A clear relationship between prescribing patterns and the deaths by
these drugs can be found. The clear implication is the need to give
patients at risk medication for a few days only and/or give the medi
cation to another family m em ber to keep. Neither o f these preventative
strategies has been systematically studied.
There are even more deaths attributable to analgesics, anti-fever
and anti-rheumatic medications, many o f which are available without
prescription. Ten per cent o f all suicide deaths by overdose are caused
by an overdose o f paracetamol and 5 per cent from aspirin. H ow ever,
nearly all patients (94 per cent) who take an overdose take the first drug
they can obtain. Few (only 20 per cent)17 are found to know about the
toxicity o f a drug or what quantity would be lethal. Many wrongly
believe that aspirin and paracetamol, because freely available from
the pharmacist, are relatively non-toxic. In France paracetamol is not
allowed to be sold in greater quantities than 8 grams (16 x 500 milligram
tablets) and there are very few fatal paracetamol overdoses. Similar
limits have now been introduced in the United Kingdom.
190
Dangerous jo b s
A final piece o f evidence that availability o f lethal means is important
comes from the sort o f jobs people do. The only thing the professions
most vulnerable to suicide (veterinary surgeon, dental practitioner,
pharmacist, farmer and medical practitioner) have in common is access
to lethal means o f self-killing. Recent evidence suggests that the nursing
profession is also vulnerable. Between 1988 and 1992, nurses accounted
for a greater number o f suicides than any other occupation and over
5 per cent o f all female suicides.18 This is w holly consistent with the
‘dangerous jobs' explanation, since nurses know about the lethality o f
drugs and have ready access to them.
Samaritans and suicide prevention centres
Between 4 and 7 per cent o f people who commit suicide in the United
Kingdom have had past contact with the Samaritans, but how recent
these contacts are is not clear. For many years there has been a debate
about whether suicide prevention centres and organizations significantly
affect the suicide rate. W hat is not in dispute is that there is a great need
for such centres and organizations, as shown by the extensive use made
o f them. H owever, even though people know where to turn for help,
when they are very suicidal they m ay not act on this knowledge. One
study conducted by Greer and Alderson in 197919 found that o f those
patients who attempted suicide, 72 per cent had sufficient knowledge o f
the Samaritans to be able to contact them yet less than 2 per cent had
actually sought their help on that occasion. Barraclough and colleagues
found that those towns which had established a Samaritans branch did
not differ in suicide rate from those towns which had not.20
A more encouraging report came from an analysis o f the changes of
the suicide rates in 226 cities in the United States between 1968 and 1973.
Some had developed suicide prevention centres over the period, others
had not. There was a reduction in suicide rates am ong white females
below the age o f 25 years in those cities that had introduced such
centres. Given that this was just the sort o f client who most frequently
used those centres, here indeed was an encouraging finding.
191
A recent study by David Lester has examined the change in suicide

rates in the United States from 1970 to 1980, state by state, and the
number o f suicide prevention centres in each state in 1970 (both the
absolute number, and the number per capita o f the population).21 He
finds a significant negative correlation: that is, the greater the availability
o f such centres, the greater the reduction in suicide rate. The largest
association was between the change in suicide rate among w om en and
the number o f suicide prevention centres (—.54), but there was some
indication o f a preventative effect for both sexes in certain age bands
(15-24; 45-64; and 75+). If these results are generalizable to other time
periods and other places, they will represent an important advance.
The pattern o f use o f such services itself provides interesting insights.
Callers to the Samaritans in the United Kingdom were more likely to
be female until 1995, when the number o f male callers exceeded that o f
females for the first time. The suicide statistics show that males are
much more likely to commit suicide than wom en, but for many years
this ratio was in the opposite direction for non-fatal suicidal episodes,
with young wom en more likely than men to attempt suicide. H owever,
for the first time in the mid 1990s, hospitals around the United Kingdom
and Ireland independently reported that men were more frequently
attempting suicide than they had in the past.
Parallel data from the Samaritans are unlikely to be coincidence. It
is much more likely that w e were witnessing a m ajor change in the
socio-demographic characteristics o f suicidal behaviour in the 1990s.
School-based intervention
Following the rapid increase in youth suicide in the United States in

the 1980s, many school-based suicide intervention programmes were
introduced. Their aims were to increase awareness o f the problem o f
suicide, to provide information about the help available and to encour
age suicidal teenagers to come forward and seek it. W herever such
programmes were evaluated, however, they had little or no effect.22
Such education programmes target a relatively low-risk audience.
The suicide rate among teenagers in school, though increasing, is still
192
low compared with the general population. Indeed, such programmes

might not reach those adolescents most at risk, such as regular truants.
More seriously, the risk profile o f a potential teenage suicide patient is
still not fully known. This means that the warning signs as taught may
not be representative or accurate.
A further study23 has assessed the impact o f suicide prevention
programmes on teenagers' attitudes and knowledge about suicide,
with evaluation based on a questionnaire completed before and after
exposure to the programme. Even before the programme, most stu
dents had a sound knowledge o f the issues relating to suicide: the
warning signs; the fact that suicide threat should be taken seriously;
and that vulnerable individuals should be helped in consultation with
responsible adults. But where the research found attitudes that would
be considered inappropriate (e.g. that suicide could be a reasonable
solution to problems), the education programmes did not change them.
Chapter 8 reviewed the evidence suggesting an increased risk o f
copycat suicidal behaviour if vulnerable people were exposed to
examples o f people who harmed themselves. Given this risk, school-
based suicide awareness programmes would need to be shown to be
clearly effective in reducing the risk o f suicidal behaviour for them to
be justified. Such effects have not been shown. It is possible that any
benefits o f such programmes are offset by the fact that they de-
stigmatize suicide. By portraying it as an understandable response
to stress, the intervention m ay actually backfire by encouraging an
unrealistic, romantic view o f suicide, increasing the chances o f imitation.
Concluding remarks
Given all this information about who is most vulnerable and what are
the most dangerous circumstances, can we estimate the possible effects
o f making changes? Gunnell lists all the potential effects o f suicide
prevention strategies for the United Kingdom.24 Most strategies, he
concludes, would have uncertain effects. H ow ever, he estimates that a
4 per cent reduction in suicide rates could be achieved by G P s prescrib
ing safer antidepressants; that increased care around the time o f
193
discharge from psychiatric hospital could reduce the rate by 2 to 3 per

cent; that improved safety measures around places such as bridges and
undergrounds/subways where there is a great risk o f suicide would
save between 1 and 3 per cent o f lives; that taking steps to prevent car
exhaust being used for suicide, together with the general introduction
o f catalytic converters, would probably reduce the suicide rate by 7 per
cent. Reduction o f access to guns (which account for 3 per cent o f all
suicides in England and Wales) might reduce the suicide rate by a
further 1 per cent. The total, a reduction o f 15 to 18 per cent, would
represent over 700 lives saved in England and W ales alone each year.
The evidence strongly suggests that, for any individual, suicidal
impulses come in waves. This implies that if such impulses can find no
ready expression, they m ay pass without the person having harmed
themselves. The precise timing o f such compelling suicidal urges
remains unpredictable, but further research should focus on times o f
change and how such changes can be prepared for so they do not bring
about a life-review process that sends the person's m ood spiralling out
o f control. Meanwhile, if suicide occurs, there will always be some
who feel they should have seen it coming: family, friends, health
professionals. Hindsight is a painful but pointless source o f grief. Instead,
primary prevention needs to focus on general measures to reduce the
number o f ‘invitations to suicide’ in the environment.
194
12
Therapy for Suicidal Feelings and
Behaviour
W hat more m ay be done to prevent farther suicidal behaviour once a

person has harmed him or herself? Several studies have confirmed that
the following six factors predict repetition o f attempted suicide in the
year following an episode: previous attempted suicide, previous out
patient and in-patient psychiatric treatment, not living with relatives,
previous diagnosis o f personality disorder and problems in the use o f
alcohol. W hat emerges is the fact that even having a large number o f
psychosocial difficulties still does not predict further suicidal behaviour
in over 50 per cent o f the sample. Even in this highly selected group,
there will be more people falsely identified as ‘at risk’ than correctly
identified. Nevertheless, therapists are faced with an increasing number
judged to be at risk o f suicidal behaviour. W e need to know what skills
are required and what special issues should be raised in therapy with
suicidal clients. A systematic review o f all the randomized controlled
trial or psychotherapeutic approaches to reduce suicidal behaviour
concluded that problem-solving treatments had the best evidence in
their favour, though other related approaches seemed promising.1
Vigilance for suicidal expression
Although, as we have said, depression is often associated with suicidal

thoughts, not all depressed people are suicidal: it is when depressed
people also become hopeless that they are most likely to feel suicidal.
Therefore a primary goal o f therapy with suicidal clients must be:
195
(a) accurate assessment o f their state o f hopelessness;

(b) vigilance for further changes in level o f hopelessness;
(c) reduction o f current state o f hopelessness; and
(d) reduction o f vulnerability for future states o f hopelessness.
T w o other important factors must also be assessed:
(e) how stable is the client’s life situation (both objectively and
subjectively to the client), especially their interpersonal relation
ships (by far the most com m on precipitant)?;
( f ) H ow impulsive is the client? Given that two thirds o f episodes
o f self-harm are contemplated for less than an hour beforehand,
impulsiveness m ay be considered an important vulnerability
factor.
The client may arrive at the conclusion that life is intolerable through
combinations o f the errors in reasoning described in depression. In
dichotomous thinking (black/white, all-or-nothing thinking), there is no
perceived middle path - just the extremes. Selective abstraction consists
o f the selecting out o f small parts o f a situation and ignoring others,
e.g. a tutor’s report on an essay gives much praise, but mentions at one
point that the introduction is too long: ‘He doesn’t like m y essay’ would
be selective abstraction. In arbitrary inference, a conclusion is inferred
from irrelevant evidence, e.g. an individual phones a boy/girlfriend and
no one answers. The conclusion: ‘(S)he’s probably out with another
partner’ would be an arbitrary inference (if inferred on those grounds
alone). Over-generalization is concluding from one specific negative event
that other negative events are therefore more likely, e.g. failure at
maths means failure at everything. Finally, catastrophizing is to think
the very worst o f a situation.
Many o f the errors o f logic which underlie such beliefs and assump
tions are exactly those found in depression. The danger with suicidal
clients is that they act so decisively and violently on their beliefs. Often
this process takes place in only a few minutes. First, their thinking about
their problems is dominated by the distortions described; then they
react to these thoughts as inescapable facts and take proportionately
196
Therapy for Suicidal Feelings and Behaviour
drastic action. Because this is a process - how ever swift - intervention

to prevent acts o f deliberate self-harm may be aimed at different stages.
If someone expresses suicidal ideas, it is important to determine
what method is contemplated, how familiar the client is with the
lethality o f medicines, and the availability o f methods (e.g. firearms).
The therapist will need to be vigilant for verbal or m ood cues which
might indirectly signal suicidal intent. Verbal expressions o f hopeless
ness provide the best clue (see Chapter 5). Not all depressed clients are
hopeless, and there is an accumulating body o f evidence to suggest that
hopelessness is the factor that turns depression into suicidal depression.
Sudden changes in emotion in either direction m ay also signal impend
ing suicidal behaviour.
A framework for therapy
Suicidal intent is a continuum. There is a balance between the intention

to live and the intention to die, and even relatively insignificant chance
factors m ay tip the balance. Tipping the balance against suicide involves
building a bridge to the next session, if possible, by getting the client to
see the next episode o f suicidal feelings as an opportunity to note in
detail how they feel, so as to bring this information to the next session.
The therapist might encourage the client to agree to make explicit the
pros and cons o f living and dying.
Dealing with hopelessness will involve careful assessment o f the
contribution o f the reality o f the life situation o f the client and o f the
interpretative biases the client may be using to evaluate it. For many
there will have been real failures and/or real rejection experiences
which must not be minimized by the therapist. But the depression may
also have made them select the most catastrophic interpretation o f
these life situations and o f their implication for the future. In this case
the questions are: what biased conclusions m ay be blocking hope?
W hat alternative behaviours and choices are realistically available?
‘Alternative therapy’ can be used, in which the person uses their
imagination o f a crisis situation to generate within the session some o f
the same hopelessness and despair typically felt outside the therapy
197
situation. Under these conditions, the person attempts to generate some

alternative coping responses.
Assessing suicidal intent and the probability

of repetition
If a therapist is seeing a client immediately following a suicidal episode,

he or she will need to assess suicidal intent and the probability o f
repetition. Assessment o f suicidal intent is best made on the basis o f the
circumstances surrounding the episode and the client's own report (as
outlined in Chapter 4). The more the behaviour approximates to suicide
(e.g. precautions taken against discovery, etc.), the greater the assumed
intent. W hether the actual medical risk should be taken into account
remains a controversial issue. O ver a large number o f cases there is a
significant correlation between actual lethality and suicidal intent, but
it may be difficult to infer intent from the medical lethality in an
individual case. Some clients (especially those not used to taking pills)
may believe that relatively few pills are lethal. Actual physical risk
would then be no guide to what m ay in fact be a very serious suicidal
attempt. There is also, however, the obvious but often overlooked
point that the correlation between actual physical lethality and intent
is much stronger in cases where the client is knowledgeable about the
lethality o f drugs available to them.
Finally, the therapist ascertains how actively suicide is being con
sidered. Modifying some o f the items from the Suicide Ideation Scale
o f Beck is helpful as a basis. The following questions are useful in that
they converge relatively quickly from general issues about problem
solving to the specific issue o f suicide. The therapist asks:
(a) whether they feel their problems can be overcome;

(b) whether they have any current ideas or fantasies about dying;
(c) whether they find ideas about suicide going through their
mind;
(d) whether they expect to make an attempt on their life;
(e) whether they are actually planning such an attempt;
198
( f ) whether they expect to put their plan into action, and how
imminently;
(g) whether anyone would care if they carried out such a plan.
Some common problems
A number o f questions arise during therapy with suicidal clients. Two,

are worth considering at this point. First, since many o f these clients
have severe real-life problems, is not their hopelessness understandable?
Second, might not consideration o f reasons for dying precipitate a
suicide attempt?
Reality-based hopelessness is indeed often found. Many clients have
real problems that must not be minimized by the therapist. On the
other hand, some people appear able to cope with apparently unbearable
problems without becoming suicidal. Why? Possibly because depressive
hopelessness is not the same as normal sadness. Financial hardship
and interpersonal chaos m ay reasonably cause a great deal o f anger,
frustration and sadness. H owever, a person becomes depressed and
suicidal when their sadness changes to a situation in which the person
tells themselves, T m to blam e’, T v e never succeeded at anything in
m y life’, ‘If m y love leaves, I am nothing.’ People under great stress -
stress which has understandable consequences on mood - need all the
coping resources they can muster. W hat they do not need is a constant
stream o f negative thoughts and images to convince them they are a
bad or worthless person who could never be forgiven. Therapy may be
conceived as enabling the person to discriminate between realistic
and depressive hopelessness so that the real problems can be faced
realistically.
As to whether focusing on problems in therapy precipitates suicidal
wishes by increasing reasons for dying, making these factors explicit is
more likely to be therapeutic. If the client believes there are overwhelm
ing reasons for dying, it is not beneficial to proceed with therapy as if
these reasons did not exist. Objective definition o f the problems (reasons
for dying) is only the first step to deciding which are solvable and which
are not; what might be done about those that are; evaluating the
199
evidence for and against each reason for living and for dying; and
discussing w hy some reasons are weighted in the client's mind more or
less heavily than others. There is no evidence that explicit discussion o f
such issues increases suicidal intent. Indeed, clinical experience suggests
the converse.
Problem-solving therapy
In reviewing the evidence on factors leading up to suicidal behaviour,

a common factor has been the presence o f stress factors with which the
person feels unable to cope, and from which they wish to escape. This
has led many to adopt a generic problem-solving approach in therapy
for suicidal clients.2 Poor problem-solving has been targeted because o f
the belief that such deficits m ay be a final com m on pathway for many
psychological difficulties. This assumes that, w hatever the psychological
problem, w hatever the diagnosis, poor problem-solving will be present
as a contributor, whether as a vulnerability factor, a precipitating factor,
or most importantly for therapy, a maintenance factor.
Problem-solving therapy has certain key characteristics. First, it has
a well-planned rationale which provides an initial structure that guides
patients to the belief that they can control their own behaviour and
thereby their own emotional problems. Second, it provides clients with
the motivation and the training in skills to feel more effective in solving
problems in their life. Third, it emphasizes the independent use o f
these skills by the individual outside the therapy context, and provides
sufficient structure so that he or she can attain the independent use o f
them. Finally, it allows patients to attribute im provement in their mood
to their own increased skilfulness and not that o f the therapist.
T o proceed with such therapy, the first need is to specify which
aspects o f problem-solving are impaired. Problem-solving involves sev
eral steps,3 the first being general orientation to a problem. People have
to be able to recognize that a problem exists. The second is problem
definition: they have to be able to articulate the problem as precisely as
possible, and what their goals are for each problem. This may involve
breaking down the problem into different parts. The next stage is to
200
generate as many alternative solutions as possible. During this stage,

the person has to inhibit the tendency to prejudge which potential
solutions might w ork and deal with the tendency for their hopelessness
to overwhelm them. Then comes the need to weigh up the advantages
and disadvantages o f implementing each alternative, and to devise ways
to test out some o f them. Later the person has to be able to evaluate
the effects and learn to congratulate him or herself if some progress has
been made, even if such progress is the discovery that some aspects o f
the problem cannot be solved at present.
There is little doubt that depressed and suicidal patients find such
problem-solving difficult. They do not generate as many alternative
solutions as non-depressed and non-suicidal people, and what alterna
tives they do generate tend to be less effective and more passive. During
a crisis, when a person faces obstacles to important life-goals that
seem insurmountable through usual methods o f problem-solving, there
follows a period o f intense disorganization. The upset caused can result
in many abortive attempts at solutions, which only serve to increase
the feelings o f crisis.
Keith Hawton and Joan Kirk have developed a b rie f‘Problem Solving
Treatm ent' that has been used in several studies in Oxford.4 It starts
with careful assessment o f the situation as the therapist:
1. Identifies the client's problems.

2. Identifies the client's resources - assets and supports.
3. Obtains information from other sources.
4. Decides whether problem-solving is appropriate.
5. Decides on practical arrangements - who will be involved, the
likely number o f sessions, duration, timing, etc.
6. Establishes a therapeutic contract - including the client's and
therapist's responsibilities in problem-solving.
The first step is probably the most important. During a suicidal crisis,
everything in a client’s life can seem overwhelm ing to them and to
their therapist. It m ay be difficult therefore to collaborate in articulating
as accurately as possible all the things that are going wrong. Yet the
aim remains to draw up a problem list. On the list, agreed between
201
client and therapist, each problem is clearly described. Often this will
be very difficult. People's problems may not seem easy to describe
clearly. The therapist will then need to help the client be more specific
(e.g. ‘W hat is it about x which is difficult?' ‘Could you go into more
detail about the problem you are having with yY). Only if sufficient
specificity is achieved at this stage can concrete strategies to cope with
the problem be identified.
Hawton and Kirk offer the following helpful check-list o f potential
problem areas:
I. Relationship with partner or spouse.

2. Relationship with other family members, particularly young
children.
3- Em ploym ent or studies.
4- Finances.
5. Housing.
6. Legal.
7. Social isolation and relationships with friends.
8. Use o f alcohol and drugs.
9 - Psychiatric health.
10. Physical health.
11. Sexual adjustment.
12. Bereavement and impending loss.
The next step in assessment is identifying the client's resources, their

assets and strengths, using other sources o f information where relevant.
This is much easier if the problems have been clearly articulated at the
outset. The central question is: ‘H ow have you coped in the past when
this sort o f thing has happened?’ The therapist will discuss some o f the
reasons why previous ways o f dealing with the crisis are no longer
available, evaluating as objectively as possible what resources remain,
despite the feelings that all options have been closed off. As in all such
therapies, the therapist's role is to help the client clarify things for him
or herself, not to enter into arguments about availability o f resources.
Having agreed a problem list, there are several steps to be taken in
problem-solving treatment, steps made explicit to the client so that the
202
process is entirely transparent. (The statements in brackets have been

suggested as useful aide-memoires for clients to use in such treatments.)
1. Decide which problem(s) to be tackled first. (What is my main

concern at the moment?)
2. Agree goal(s). (What do I want?)
3. W ork out steps necessary to achieve goal(s). (What can I do?)
4. Consider the possible consequences. (What might happen?)
5. Decide tasks necessary to tackle first step. (What is my decision?)
6. Carry out task as hom ework. (Now do it!)
7. Review progress at next therapy session, including difficulties
encountered. (Did it work?)
During the next session, client and therapist make use o f the information
gained from the last hom ew ork to decide on the next step. This clearly
depends on what progress was made, and will lead to further agreement
about subsequent tasks. Further goal(s) m ay need to be agreed, or
problems redefined for further loops around the sequence.
There are a number o f other techniques used in problem-solving
therapy, including some that are common to cognitive therapy. These
include:
Brainstorming techniques: an approach in which a client suggests as many

solutions as possible, without evaluating the potential usefulness o f any.
The aim is to generate a list without premature dismissal o f alternatives
as ‘useless’, which simply inhibits further problem-solving.
The therapist may suggest some possibilities if the client finds it
difficult to generate solutions. Hawton and Kirk advocate the deliberate
suggestion o f some extreme or even humorous solutions to facilitate
the involvement o f the client in the process. They suggest that including
extreme solutions can lead the client into unexplored avenues and
produce other novel solutions.
They give the example o f a client called Mary, who was helped to
brainstorm possible solutions to her problem o f her mother, who she felt
was intrusive, constantly visiting or telephoning at inconvenient times.
This was the list, which included some deliberately extreme solutions:
203
1. Ask her not to visit or telephone any more.

2. Ask her to reduce her visits and telephone calls.
3. Leave the country.
4. Change telephone number and go ex-directory.
5. Discuss the problem with the mother.
6. Do nothing and accept the status quo.
After Mary examined the advantages and disadvantages o f each solution

in detail, she decided that discussing the problem with her mother, pre
viously seen as impossible, was the most appropriate thing to try first.
Examining alternatives: using ‘two-colum n’ or ‘pros and cons’ technique

(simply writing down the advantages and disadvantages, including likely
outcomes, o f each possibility, giving relative weightings if appropriate).
Sometimes the best course o f action becomes clearer as a result o f this
technique, or it shows how much further information is needed to give
further weight to one or another alternative.
Cognitive rehearsal: the detailed rehearsal in imagination o f carrying out

a particular task, including details o f steps taken, the consequences o f
each step and any accompanying thoughts and feelings. Since the person
in this technique imagines completing the task, it can be useful for
helping a client develop confidence in attempting a task, in identifying
possible blocks to progress not previously obvious.
Activity scheduling: the deliberate setting aside o f time to attempt particu

lar tasks or actions. Tasks can be split up or graded into simpler sub-tasks
early on in therapy, gradually building up to the assignment o f whole
tasks. While completion o f a task is the main goal, the person decides
in advance how much time can be allocated, and keeps to that time,
even if a task is not completed. In this way, the person may make some
helpful observations about how well they are able to estimate how long
different tasks may take. Clients are also encouraged to make time for
breaks in their schedule, to give time for reviewing progress and for
self-reward. Subsequent treatment sessions will involve examining to
what extent the person was able to carry out assigned tasks, learn from
204
any problems encountered, then plan a new schedule in the light o f

what is learned.
Hawton and Kirk point out that problem-solving may not be appropriate
in every case, or at every stage in the crisis:
The fundamental decision which determines if problem-solving is currently

applicable for a patient concerns whether the person is so severely disabled by
psychiatric symptoms or disorder that he or she cannot at present be expected
to take responsibility, even with the support o f the therapist, for managing the
problems that require attention. For example, a patient with marked retarded
or agitated depression is very unlikely to be able to engage in the steps necessary
for problem-solving until there has been some reduction in the severity o f the
affective disturbance. Similarly, when a person is in a severe state o f crisis,
especially if suicidal, problem-solving will usually be inappropriate until the
level o f disorganization and helplessness which often characterizes such a state
has been reduced. Attention to exacerbating factors (e.g. sleep disturbance, lack
o f supports) can often bring this about, following which problem-solving may
then be very appropriate.
Evaluation of problem-solving approaches
Early studies using these types o f approach have used both out-patient
treatment sessions and sessions conducted in the client’s own home. In
these early studies, repetition rates were not reduced. O f course, other
benefits o f such treatments have emerged. One such study5 used task-
centred case-work (nine sessions over three months) and compared it
with normal psychiatric out-patient visits. At four months’ follow-up,
the psycho-social treatment group showed significantly fewer social
problems and were better at handling personal relationships and social
transitions than the other group. In addition, they were less lonely and
better at coping with their own emotional distress as well as with
practical difficulties in their lives.
H ow ever, this study, and others like it, was not able to show an effect
on repetition. This is for two reasons. First, the treatment produces most
205
positive benefits in those less likely to repeat the attempts as judged by

the vulnerability indices (page 74). Simply put, one can divide suicidal
individuals into those likely to respond to psycho-social intervention and
those unlikely to respond (because o f their chronic psychological / psychi
atric and social problems). The problem in these therapy studies is that
those most likely to respond are also more likely to have been suffering
an acute crisis which has now passed and are therefore unlikely to repeat
parasuicide. Second, this leaves the ‘non-responder’ group, who are always
going to be difficult to treat using short-term psycho-social treatments.
This pessimistic conclusion m ay be prem ature.6 One possible future
strategy for therapy research is to examine those clients w ho are at
greatest risk for repetition. This is important on tw o counts. First, it is
targeting those most in need o f help. Second, with an increased prob
ability o f repetition o f self-harm, one m ay more clearly see whether a
treatment is working. Might a larger scale, more complex treatment
package w ork more effectively? Such a treatment has been developed
by Marsha Linehan in the University o f W ashington.7
Dialectical Behaviour Therapy (DBT)
Linehan’s Dialectical Behaviour Therapy (D B T ) combines w eekly

individual and weekly group therapy over a one-year period. It uses
treatment strategies from behavioural, cognitive and supportive psycho
therapies. Thus behavioural skills training, contingency management,
cognitive modification and exposure to emotional cues are combined
and balanced with supportive techniques such as reflection, empathy
and acceptance. The behavioural/problem -solving component focuses
on enhancing capability, generating alternative w ays o f coping, clarify
ing and managing contingencies, all with the emphasis on the ‘here and
n ow ’. Second, D B T focuses on the dialectical aspect both o f the client’s
experience and the therapy, e.g. the client m ay be experiencing feelings
o f anger very intensively, yet may continuously invalidate those feelings
by saying he or she ought not to feel angry. This only increases the
amount o f emotional disturbance in a vicious circle, which quickly
escalates to the point where the client feels their affect is completely
206
out o f control. Within the general orientation o f the therapy, the

dialectical focus lies in its emphasis on balancing acceptance o f the
sources o f stress that exist in the environment on the one hand with
the need to change them on the other. The theme is encouraging clients
to see things clearly, as they are, and to step back from them temporarily
to see what things may be changed.
The individual therapy (lasting one hour each week) stays within a
strict agenda and prioritizes themes related to suicidal behaviour. Thus,
for example, if any suicidal act, threat or thought has occurred since
the previous therapy session, that will be the priority for the current
session. Therapist and client will then investigate in great detail the
circumstances surrounding each suicidal ‘episode’ , even if there was no
overt suicidal behaviour. T o help this process, the client keeps a record
o f the use o f prescribed or non-prescribed drugs, and o f the intensity
and frequency o f suicidal urges and bouts o f depression. Finally, the
client records whether there has been self-harm in the previous week.
No suicidal behaviour, no matter how apparently insignificant, is
ignored. Linehan believes this is an important keynote o f the therapy.
She asserts that newly trained therapists too easily avoid explicit dis
cussion o f these suicidal thoughts, threats and behaviour. The client
needs to learn that help for other problems can only be discussed when
the self-harm behaviour has been brought under control.
Once the suicidal behaviour or thinking has been discussed in
therapy, or if there has been no episode during the previous week, the
second priority within a therapy session is to deal with ‘therapy interfering
behaviours’ such as non-adherence to agreed goals, not filling in the
hom ew ork records or non-attendance. These are again subject to a
detailed functional analysis within the context o f therapy. The emphasis
throughout is on teaching clients how to manage emotional crises rather
than reducing the trauma or withdrawing from them.
These individual therapy sessions are supplemented by weekly group
therapy, which lasts for two and a half hours a session, focusing on:
(a) interpersonal problem-solving skills,

(b) distress tolerance and reality acceptance skills, and
(c) the skills o f emotional regulation.
207
Discussion o f individual crises is discouraged during group sessions,

clients being referred to their individual counsellors for help with these.
In a study to assess the effectiveness o f the therapy, forty-four clients
were treated. Each met criteria for borderline personality disorder (see
page 98). In addition, each had had at least two episodes o f deliberate
self-harm in the previous five years, at least one o f which occurred in
the previous eight weeks. Clients were randomly allocated to receive
either D B T or Treatm ent as Usual (TA U ). The client group were
wom en, aged 18-45 years, and were excluded only if they met criteria
for schizophrenia, bipolar disorder, substance dependence or learning
disability (mental handicap); or if they declined to agree to the study
conditions, including the agreement to give up psychotherapy if they
were assigned to the D B T group. Linehan’s 1991 report gave the results
for the period while the clients were in therapy. A later report gave the
details o f what happened in the following 12 months.
Looking first at the data for the period o f the study (twelve months),
the results showed that D B T significantly reduced episodes o f self-harm
over the year (see Figure 12.1). The T A U control group had a median
o f nine such episodes over the year, compared to the D B T group’s 1.5.
During the final four-month period, over 60 per cent o f the T A U were
still having episodes o f self-harm, whereas only 25 per cent o f the D B T
group were showing such behaviour. Ninety-six per cent o f the T A U
group had an episode o f self-harm during the year in question, compared
to 64 per cent o f the D B T .
It is always possible that some o f these differences emerged because
o f the D B T group’s unwillingness to report suicidal episodes that had
in fact been taking place. H ow ever, there was independent evidence
from hospital records that these results were not subject to reporting
biases. For example, D B T clients spent a median o f 17 days whereas
the control group spent a median o f 51 days in psychiatric hospital over
the year. Furthermore, the suicidal episodes o f the control group were
likely to incur greater medical risk than those o f the D B T group.
Com paring risk for just those control subjects (n = 10) and D B T
(n = 5) who needed medical treatment following self-harm, the differ
ence in medical risk was significant.
208
159
15
DBT
14
TAU
13
12
11
10
0-4 months 5-8 months 9-12 months
Phase o f treatment
Figure 12.1 Number of episodes of self-harm in borderline person

ality clients at three phases of treatment (in Linehan, Armstrong,
Suarez, Allmond and Heard, 'Behavioural Treatment of Chronically
Parasuicidal Borderline Patients', Archives o f General Psychiatry, 1991).
209
r "V""-
_ J DBT
TAU
2.1
1.06
O.I
0-6 months 7-12 months

Post-treatment period
Figure 12«2 Mean number of episodes of self-harm in year following

DBT for borderline personality disorder clients (from Linehan, Heard
and Armstrong, 'Naturalistic Follow-up of a Behavioural Treatment',
Archives o f General Psychiatry , 1993).
Follow-up of DBT for suicidal behaviour
The promising data o f the 1991 report only provide details o f the effect
o f treatment while it continued. Remarkable as these treatment gains
are, it was still an open question whether any o f them would be
maintained once the therapy finished. The follow-up study was there
fore o f great interest.8 It was a naturalistic follow-up, and therefore did
not prevent people from returning to treatment if they wished. Control
clients were able to continue to receive psychotherapy in the com
munity. D B T clients, although required to take a ‘two-month vacation'
from their individual D B T therapists, could continue in psychotherapy
thereafter if the therapist was agreeable. Seven subjects (35 per cent)
took this option. This contrasts with 55 per cent who continued with
psychotherapy in the control group.
Follow-up assessments were made at 6 months and 12 months
210
16
___: DBT
14
TAU
12
10
o
Figure 12.3 Mean number of days in psychiatric hospital in year

following DBT for borderline personality disorder clients
(Linehan et a L , art. cit., 1993).
following the end o f the experimental treatment year. Patients were

asked about the frequency o f their suicidal behaviour and any medical
treatment that resulted. Results showed that over the entire year the
repetition rate for self-harm was low er in the D B T group (26 per cent)
than the control group (60 per cent). The difference in number o f
suicidal episodes was mainly ow ing to differences in the first six months
o f the follow-up phase (Figure 12.2).
There was also a difference in the number o f psychiatric hospitaliza
tions between D B T (11 per cent) and control (40 per cent), mostly
because o f a difference that emerged between the groups in the second
six-month follow-up phase (Figure 12.3). During this period none o f the
D B T groups were hospitalized, contrasting with a mean o f 5.3 psychi
atric hospital days for the control group. Subsequent analysis confirmed
there were also differences between groups in that D B T clients were
experiencing less anger and had greater interviewer-rated social adjust
ment. Consistent with this, the employment performance o f subjects
in the D B T group was significantly better than that o f the control
211
o. 8
m D BT
TAU
0.6 0.56
0.2
0.1
0.1 0.05
o
Figure 1 2,4 Mean number of medically treated episodes of self-harm

in year following DBT for borderline personality disorder clients
(Linehan et a i , art. cit., 1993).
group. Furthermore, there was a difference in the number o f suicidal

episodes requiring medical treatment (Figure 12.4). Thus, while the
results o f this follow-up can only be seen as preliminary because o f the
small numbers involved, they are clearly very encouraging for this
structured psycho-social approach to chronically suicidal clients.
Further research to evaluate D B T is continuing.9 One small study
o f 10 female patients with a diagnosis o f borderline personality disorder
in a high security hospital showed a dramatic reduction in rates o f
deliberate self-harm from before the start o f treatment to the end o f the
12 months that continued at a six-month follow-up.10 Another study is
underway examining the extent to which D B T and a cognitive-
behavioural problem-solving approach can be combined in a very short
(six-session) treatment, assisted by a booklet that is given to patients.11
Preliminary data is encouraging, but the sample is too small to draw
definite conclusions.
212
Psychological mechanisms underlying

therapeutic change
D B T and other such treatment approaches involve many different

techniques, and it is important to attempt to understand what the
critical aspects o f therapy may be so as to give therapists a theoretical
fram ework to orientate themselves within the therapy. Research on the
psychological processes mediating hopelessness and suicidal behaviour
would suggest that any successful therapy will need to affect the global
and undifferentiated cognitive style in such clients, which tends to be
activated w henever any change triggers a ruminative review o f their
lives. The tendency o f clients to recall events in their lives in a gen
eralized, undifferentiated w ay is associated with an inability to problem-
solve, both in terms o f the number o f alternatives people can generate
and the effectiveness o f those alternatives. Furthermore, the study by
Julie Evans12 found that the more angry the clients, the less they could
retrieve specific events from their past, and the less effective were the
problem solutions they produced on the M eans-Ends Problem Solving
Test. A similar pattern o f correlations was found for the hopelessness
levels o f the clients.
These results suggest that non-specificity o f mem ory and difficulties
in emotional regulation are closely related in producing and exacerbat
ing suicidal crises. The therapy appears to be successful in making the
link between thoughts, feelings and behaviour more explicit. Might the
therapy be helping the person become more specific in their encoding
and retrieval o f events in their lives? Examining Marsha Linehan’s
description o f her therapy, this appears likely. She points out that it
involves ‘an exhaustive description o f the moment-to-moment chain
o f environmental and behavioural events that preceded the suicidal
behaviour . . . Alternative solutions that the individual could have used
are explored, behavioural deficits as well as factors that interfere with
more adaptive solutions are examined, and remedial procedures are
applied if necessary.' If true, future research will find that those
clients who respond least well to treatment will be those who are not
able to recall events in this more specific way. Such mem ory problems
213
prevent them generating effective alternative solutions to their

problems.
Another important feature o f all these therapies is that they give the
person a sense o f control over their lives. Since a major problem with
suicidal individuals is the tendency to generalize from one idea that is
uncontrollable to other situations that are potentially solvable, and
simply assume there is nothing they can do about a new situation, a
self-fulfilling prophecy ensues. They do not try because they ‘know ’
there is nothing they can do. When the problem does not resolve, they
take this as evidence that they were right. O f course, after a lifetime o f
stressful events, such a conclusion may be quite understandable for
some, but it is often applied too generally. Any therapy able to give
clients a sense o f control over events in their lives is likely to be helpful.
W hat is the minimum intervention one might make with those who
are vulnerable to suicide to give them a sense o f control? A possibility
is to let such people know they have control, at least over the help they
can receive; that they are able to contact the emergency services at any
time, and they will be seen. This was the aim o f an innovative U K
study in Bristol: the Green Card Study. The approach, pioneered by
Professor Gethin Morgan and colleagues, involves giving people follow
ing a first episode o f self-harm a Green Card. It is explained that
possession o f the Green Card gives the person the right to telephone
or personal contact, even admission to hospital, at any time (day or
night) should the need arise, provided that the person possessing the
card had not repeated the suicidal behaviour on the same occasion. A
total o f 212 patients were randomly allocated, either to receive the usual
psychiatric and social aftercare (n = in ) or to receive a card in addition
to the usual help (n = i o i ) . 13 T w elve people in the control group (10.8
per cent) harmed themselves again over the following year. By contrast,
only five (5 per cent) who had the Green Card did so. Although these
results are not statistically significant, they do offer hope that a larger
study might find the Green Card a helpful adjunct to existing treatments.
O f considerable interest was that, whereas many might have predicted
that making such an open commitment to suicidal people was danger
ous since services might have been inundated with requests for help,
this did not occur. Only 15 o f the 101 who had the Green Card made
214
contact with the psychiatric service over the year, the majority by
telephone. Some got in contact more than once, but even the total
number o f contacts was only 19, o f which 4 involved face-to-face
interviews. In half such contacts, a single episode o f discussion was
enough to deal with the problem. Only in one case was there a request
for readmission to hospital.
There needs to be some caution in generalizing these results. W e
have seen (page 72) that the pattern o f non-fatal suicidal behaviour
changed during the 1990s, with more men being admitted following
self-harm than ever before. W hether a Green Card will be effective for
this new group o f suicidal clients remains to be seen. More recent
studies o f the Green Card have been more uncertain o f its effects.
H ow ever, the studies referred to in the previous chapter, by Motto
(1976) and De Leo (1995), are more optimistic. M erely keeping in touch
with vulnerable people by correspondence or by telephone appears to
be able to reduce the suicide rate, at least for the time that the contact
is kept up.14
Concluding remarks: the future of secondary

prevention research
W hat does the future o f research on secondary prevention hold? Clearly

there will need to be replication and extension o f previous studies. First,
Linehan’s pioneering w ork needs to be extended, to see how well it can
be applied in other settings throughout the world, with a focus on both
male and female clients (Linehan’s study consisted o f wom en only).
Second, studies o f problem-solving and the Green Card need to be
extended to include large enough numbers to be able definitively to
assess their impact on repetition. Third, the use o f such approaches in
combination with medication could be investigated. Although psycho
social treatments have developed in isolation from improvements in
drug treatments for various psychiatric conditions, there is evidence
that the combined use o f pharmacological and psychological treatments
might be useful. (This has been studied for treatments o f depression,
where clients clearly do not find the combination o f biological and
215
psychological models confusing.) Third, since depression remains the

most important final com m on pathway to suicide, the quest for new
and better treatments for depression itself is an important part o f the
agenda in attempting to deal with suicide. Future w ork will need to
concentrate on the prevention o f relapse as a problem in its own right.
Secondary prevention demands a multi-modal approach. Mental
health professionals need to be aware o f the ease with which suicidal
feelings may be missed in a therapy consultation. They need to be
aware o f the preferred method by which clients would com m it suicide
if they had the chance, and able to estimate the lethality o f the method
as well as the probability o f carrying out the action. They need to be
aware o f the risk factors for repetition following episodes o f deliberate
self-harm, but also aware that h alf the clients who have all these risk
factors do not actually repeat within the following year.
The data help us to understand w hy early studies o f treatment
interventions were pessimistic. Those clients who might have
responded well to treatment tended to be those less likely to repeat the
attempted suicide episode anyway. That left those who have had a
great many psychological, psychiatric and social problems in their lives.
Only as studies have started to look at this difficult sub-group have we
begun to make advances. In particular, the structured psychological
approach o f Marsha Linehan, Dialectical Behaviour Therapy, offers an
interesting range o f methods to be investigated in future research.
216
13
Final Thoughts
Anguish is known to everyone since childhood, and everyone knows that it is

often blank, undifferentiated. It rarely carries a clearly written label that also
contains its motivation; when it does have one, it is often mendacious. One can
believe or declare oneself to be anguished for one reason and be so due to
something totally different: one can think that one is suffering at facing the
future and instead be suffering because o f one’s past; one can think that one is
suffering for others, out o f pity, out o f compassion, and instead be suffering for
one’s own reasons, more or less profound, more or less avowable and avowed;
sometimes so deep that only the specialist, the analyst o f souls, knows how to
exhume them. Primo Levi, The Drowned and the Saved1
I have suggested that the key to understanding suicidal behaviour is to

view it as a cry o f pain. Suicide comes out o f mental anguish. It is a
response to uncontrollable stresses that arise from the environment,
or from the uncontrollability o f the mental anguish itself. When an
individual first becomes aware that they lack control over important
areas o f their circumstances or o f their mental life, the cry o f pain may
be one o f anger and rage: a protest against the feelings o f entrapment.
As the person becomes more and more convinced they have failed, or
that they have been rejected or abandoned, the anger becomes mixed
with hopelessness and despair. A tunnel vision ensues, in which normal
escape routes are not noticed. Offers o f help are rejected or misinter
preted. The person feels more alienated, increasing his or her feelings
o f anger and hopelessness, and begins to seek alternative w ays o f escape.
The likelihood o f suicide at such times depends on how overwhelm ing
217
such feelings are, whether there have been models among family or
friends or in the media, whether a suitable method is readily available,
how violent or impulsive the person is, and whether drugs or alcohol
are available which reduce fear o f death and impair judgem ent.
Completed suicide and non-fatal attempted suicide can be under
stood as different responses to these circumstances, occurring at differ
ent points in the downward spiral into hopelessness. For m any years it
was thought that suicide and attempted suicide had to be qualitatively
different behaviours, partly because many w ho harm themselves say
they do not want to die. Their motivation seemed much more complex
than a simple wish to die. But motivation for completed suicide is also
complex. Because the tragedy o f a death by suicide is so extreme, we
tend to assume that, when death is the outcome, death has been the
predominant motive.
Yet I have also suggested that the predominant motivation in suicidal
behaviour is escape. The person feels trapped. They can see no w ay
out o f their prison, and take little account o f the possibility that some
o f their feeling o f entrapment comes from a biased view o f their own
past life that feeds into their hopelessness about the future.
Getting aw ay from seeing completed suicide as motivated only by a
wish to die frees us from a punitive view o f attempted suicide. In the
past, we have allowed the question o f how much a person wished to
die to define the w ay in which w e understand all who harm themselves.
W e ask ourselves, ‘H ow suicidal was this behaviour really?' If we decide
it was not, then w e are inclined to dismiss it, get angry about it, see it
as ‘manipulative', and so on. (O f course, the truth is that those whom
we call manipulative have actually failed to manipulate us. For w e all
manipulate each other all the time. W e are simply so good at it that no
one notices. It is only when w e meet someone in w hom such attempts
are more obvious that we notice it. It is ironic that we accuse them
o f being manipulative when, in fact, they have failed adequately to
hide the social performance in which we all engage.) The problem with
the idea o f the ‘cry for help’ is that it fed this dismissive view o f
self-harm.
O f course, there were other reasons for thinking that suicide and
attempted suicide were different things. The most important reason
218
Final Thoughts
was that those who completed suicide on the one hand, and those who
attempted suicide on the other, seemed to differ in a number o f respects.
For example, while suicide, until recently, was predominantly a feature
o f older males, attempted suicide seemed predominantly a feature o f
younger females. Furthermore, while suicide rates were coming down
in the 1960s in the United Kingdom, attempted suicide was rising. Such
differences between suicide and attempted suicide seemed to suggest
that different explanations for each were needed. Yet, if we are to see
age and sex differences as pointing to different underlying causes, what
are we to make o f the recent data from the United States and the United
Kingdom that show male suicide is rising while female suicide is falling?
Here w e have the same outcome, suicide deaths, m oving in different
directions in different sub-groups o f the same population.
Instead it is possible that the same underlying set o f causes might
produce the age and sex differences in suicidal behaviour that are
sometimes observed. If suicidal behaviour is best seen as a cry o f pain
- a response to feeling trapped by uncontrollable external circumstances
and uncontrollable internal anguish - then different types o f people,
older or younger, male or female, take action on these feelings at
different points in the sequence o f events as the trap is perceived to
close. Internal and external stresses give rise to differences in perceived
escape potential (i.e. their ability to see a w ay out). This, in turn, gives
rise to stronger or weaker wishes to die.
At one end o f the continuum, a weaker wish to die may not be
expressed as a wish to die at all, but rather be seen as seeking temporary
oblivion. But such people are best seen as one end o f a dimension o f
lethality that cannot be ignored. The protest and anger that most often
produce non-fatal suicidal behaviour represent a response early in the
sequence o f events when escape still seems possible. The despair and
apathy that produce more lethal suicidal behaviour represent the
response to loss that comes later when the person sees no hope in their
situation at all. Even at this stage, however, the presence o f social
support can ameliorate the intensity o f feelings o f hopelessness. Men
differ from w om en in the extent to which they perceive or make use
o f such supports. This is not simply a question o f biology, but o f the
different ways in which boys and girls are taught to view emotional
219
expression. W herever children have been allowed to develop the belief

that it is shameful to display one’s weaknesses, they will find it difficult
later in life to seek support from others when in a crisis. Traditionally
boys were taught this attitude m ore than girls, but we need to guard
against the possibility that increased opportunities for w om en require
adoption o f this attitude.
Finally, we have seen that all the important contributors on the
causal pathway from stress to suicidal behaviour involve a judgem ent
by the individual (see Figure 13.i). The person whose life is in crisis
needs all the resources they can muster. But people most vulnerable to
depression and suicide suffer an additional burden which makes their
real problems seem even more aversive, uncontrollable and unsolvable.
Biases in their m em ory prevent them from recalling positive aspects o f
their past and present life. Even worse, when they do recollect events,
these are recalled in such non-specific ways as to suggest few ways o f
dealing with current problems. Furthermore, they induce a non-specific
view o f the future. Making concrete plans for the future (even for
tom orrow or next week) is extremely difficult against the background
o f such vague and over-general recollection o f the past.
The aftermath of suicide
My aim has been to explain something o f what I understand o f suicide

and self-harm. Looking for unifying themes is important, but I do not
deny that each individual circumstance will have some elements that
do not fit the overall pattern. In seeking an explanation o f suicide, I
have been conscious o f the need to explain things not only to those in
the helping professions, but also to the large number o f people whose
lives are touched by suicide and self-harm. My last word must be for
them and those who care for them.
Clinicians who w ork with suicidal people and their families estimate
that for every person who commits suicide there will be at least six
survivors strongly affected by the event - the bereaved family, friends
and close workmates. For the survivors there are enormous problems
in the aftermath o f such a death. Even if the deceased has been
220
Final Thoughts
Figure 1 3.1 At important points in the causal pathway to suicidal

behaviour a judgement by the individual allows scope for bias in
estimates of the aversiveness of stress, its controllability, and
how much social support is available.
depressed, even if they have talked o f suicide, the actual death comes
as a huge shock. And, o f course, suicide sometimes comes right out o f
the blue.
The huge trauma o f any sudden death is compounded in a suicide
by a number o f factors: many still feel a social stigma about suicide, and
there is often a ‘conspiracy o f silence’ . This is especially true where
children are involved, and children whose parent or brother or sister
has committed suicide may thereby have to deal, with little support,
with their own misconceptions and irrational guilt about what has
happened. The outcome for children when their parents have separated
before one parent commits suicide is worse than for those whose
parents are still together.2
Even when social support is available, the suicide survivor is less
likely to use it than those bereaved by other kinds o f sudden death,
221
because o f the social stigma and the presence o f feelings o f personal

guilt and shame surrounding the event.3 Suicide survivors have to cope
with their own grief against the background o f many ‘w h y’ questions:
'W hy did they not ask me for help?5 ‘Could I have done more?' ‘W hy
did I (or someone else) not take the suicide threat more seriously?’
Some survivors will blame themselves for not intervening, and endlessly
ruminate on what they could have done to prevent it happening. In
addition, many feel extremely angry with the deceased. W hy couldn’t
they understand the effect it would have on those left behind?
No answers to these questions can be completely satisfactory, but it
m ay be helpful at some point for survivors to know how suicidal despair
reduces a person’s ability to ask for help, to accept help if offered, or to
understand the effect their action would have on others. W e have seen
in Chapter 9 that suicide arises from an overwhelm ing urge to escape
feelings o f being trapped, and that such entrapment is closely linked
with feelings that one has failed. In such states, an individual feels shame
and humiliation, so is highly sensitive to anything that might increase
such shame. Asking for or accepting help from others will seem to
many just such a shameful situation. Even the thought o f receiving help
from others may increase suicidal thoughts in such circumstances.
Further, the deficit in m em ory outlined in Chapter 10 further exacer
bates the situation by blocking recollection o f previous times when help
was available.
But how can w e understand the w ay a suicidal individual appears to
take such little account o f the effect o f their action on others? Like
someone trying to escape from a blazing house-fire, the suicidal person
is focused on escape. He or she has tunnel vision, which prevents them
imagining what the act would do to others. They are completely self-
absorbed. The feelings o f other people do not appear in their calculations.
Such catastrophic failure o f empathy, the complete breakdown in
understanding how others will react, is not well understood. One
possibility is that it occurs because, in states o f extreme despair, the
normal capacity to understand other people’s minds is impaired. Recent
research in developmental psychology shows that understanding the
feelings o f others depends on understanding other people’s beliefs and
intentions, and how they will be affected by things that happen in the
222
Final Thoughts
world, including one’s own actions. Small children develop the ability
to take full account o f other people’s feelings and beliefs as part o f their
developing language and social communication in the second and third
years o f life.
I believe that the radical loss o f empathy that accompanies suicidal
impulses in some people results from impairment o f these same psycho
logical mechanisms. Suicidal despair switches o ff the processes respon
sible for understanding the beliefs, intentions and feelings o f others.
Though this will often appear to be a callous disregard for others’
feelings (especially close family and friends), it is more likely to be
something over which the person has little control.
Making peace with despair
The quotation from Primo Levi with which this short chapter began is
taken from his last book, The Drowned and the Saved. Levi had survived
Auschwitz, and had written o f these experiences in his books, I f This Is
a Man and The Truce. He was found dead at the bottom o f the stairwell
o f his house in Turin in 1987, and is presumed to have committed
suicide. In the quotation, Levi speaks o f an anguish sometimes so deep
that only the specialist, the analyst o f souls, knows how to exhume it.
But what help can an analyst o f souls provide? W e cannot opt out
o f trying to answer such questions, no matter how much we would
wish to remain silent. W hen therapies bring hope to those in despair,
what is it that they are doing? O f course, they are offering support.
They are helping the person to see that perhaps some o f their problems
are solvable. They are helping the person to gain some distance from
the constant propaganda o f the mind that would persuade them they
are a failure. But I believe these are helpful only to the extent that they
allow the person to give up wanting things to be different. As Marsha
Linehan points out in her therapy (see page 208), it is about balancing
acceptance and change. For when people are in such distress, they
become trapped by the idea that, if things were different, then all would
be well: if their partner was different; if their job was different; if their
house was different; if they themselves were different.
223
Often, then, the anguish is exacerbated by the search for an escape.

In the same w ay as an animal caught in a trap will only tighten the trap
with its struggles, so the desperate search for a w ay out only drives the
suicidal person deeper into despair. The first step in dealing with anguish
is therefore to make peace with despair; to take a step back and take
stock o f the situation. T o give permission to oneself to feel w hatever it
is that one is feeling. This reflects the important insight that it is not
the initial feelings o f depression and anxiety that cause people the most
problems but how they react to them. If a person becomes depressed,
they m ay either see it as a temporary m ood which will pass or as
evidence that they are worthless. In the form er case, the depression
may well lift, but in the latter, the self-denigration will cause further
depression. There is the danger o f a vicious spiral downwards into
despair.
Some, such as Ajahn Sucitto, have seen this as a profoundly spiritual
question:
It’s one that cannot be resolved by trying to make the world into a different
place - which tends to be the normal approach. To make peace with despair is
a matter o f understanding not just where difficulties such as sickness and
violence arise, but also how the feeling o f being bound to and oppressed by
those problems occurs.4
Therapies that are helpful are those which allow the person to see their
moods as normal, rather than as evidence o f their inherent deficiency
as a person. They encourage the person to ask not, 'H ow can I make
everything different?' but rather, 'H ow can I take care o f m yself right
now?' In short, they encourage the person to be gentle with themselves.
224
Notes and References
Introduction
1. A recent example is Sue Chance, Stronger than Death, W. W. Norton, New

York, 1992. The author is a psychiatrist and a columnist for The New York Times,
and the book is based on her own experiences when her son committed suicide.
It contains extracts from her journal, starting one week after his death, together
with poems and other reflections.
2. E. Stengel, Suicide and Attempted Suicide, revised edition, Penguin Books,
Harmonds worth, 1964.
3. Stengel, Suicide and Attempted Suicide, p. 115 (1975 edition).
4. Many authors now refer to any non-fatal suicidal behaviour as ‘parasuicide’
(with its deliberate agnosticism about motive), since most who deliberately
harm themselves say they do not want to die. In this book I use the terms
‘parasuicide’, ‘attempted suicide’ and ‘self-harm’ interchangeably.
1 A Brief History of Suicide
1. H. G. Morgan, Suicide Prevention: The Assessment and Management of Suicide

Risk, Health Advisory Service, Bristol, 1993.
2. For the data in this chapter I am grateful for Michael MacDonald and Terence
Murphy’s fascinating study, Sleepless Souls: Suicide in Early Modem England,
Oxford University Press, 1990, which shows how attitudes and responses to
suicide have shifted over the past two thousand years. I have attempted to
summarize their review as faithfully as possible, but the interested reader
should refer to their book for the data and the careful analysis on which their
conclusions are based.
225
3. MacDonald and Murphy, Sleepless Souls, p. 21.

4. Diary of Samuel Pepys, vol. 111, Swan Sonnenschein, London, 1906, pp. 344-5.
5. Montaigne, Essayes, ii, pp. 26-8, 41.
2 Suicide: Facts and Figures
1. E. Durkheim, Le Suicide, Alcan, Paris, 1897; Suicide, trans. byjoh n A. Spaulding

and George Simpson, Free Press, New York, 1951.
2. P. Sainsbury, J. Jenkins and A. E. Baert, Suicide Trends in Europe, World Health
Organization, Copenhagen, 1981, IC P /M N H 036.
3. Until the 1960s it remained possible that the different overall rates merely
reflected different levels o f stringency by the authorities in different countries.
In an important paper, P. Sainsbury and B. M. Barraclough showed this was
not the case: see ‘Differences between Suicide Rates’, Nature, 220 (1968), p. 1252.
They examined the suicide rates o f immigrants to the United States from eleven
countries, and found that such immigrant populations retained the suicide rate
from the country o f origin.
4. S. Platt, ‘Epidemiology o f Suicide and Parasuicide’,Journal of Psychopharmacol
ogy, 6 (1968), pp. 291-9.
5. M. Weinstein and P. Satumo, Economic Impact of Youth Suicides and Attempted
Suicides, D H H S publication no. (ADM ) 89-1264, Department o f Health and
Human Services, Rockville, Maryland, 1989, pp. 4-89.
6. J. Charlton, J. S. Kelly, K. Dunnell, B. Evans, R. Jenkins and R. Wallis, ‘Trends
in Suicide Deaths in England and W ales’, Population Trends, O P C S, London,
1992, pp. 10 -16 and 34-42.
7. M. Buda and M. T. Tsuang, ‘The Epidemiology o f Suicide: Implications for
Clinical Practice', in S. Blumenthal and D. Kupfer (eds), Suicide over the Life
Cycle: Risk Factors, Assessment and Treatment of Suicidal Patients, American Psychi
atric Press, Washington, D C , 1990.
8. The only exception to this is in China, where there are reports o f higher rates
in women than men, especially younger women in the more rural areas. See
A. T. A. Cheng and C. -S. Lee, ‘Suicide in Asia and the Far East’, in K. Hawton
and K van Heeringen (eds), International Handbook of Suicide and Attempted
Suicide (2000), John Wiley, Chichester, pp. 29-48.
9. MacDonald and Murphy, Sleepless Souls, p. 247 (see note 2, Chapter 1).
10. J. Harry, Sexual Identity Issues, Report o f the Secretary’s Task Force Report
on Youth Suicide, vol. 2, D H H S publication no. (ADM ) 89-162, Department o f
Health and Human Services, Washington, D C , 1989, pp. 131-42.
11. C. L. Rich, R. C. Fowler, D. Young and M. Blenkush, ‘San Diego Study:
226
Comparison o f Gay to Straight Males’, Suicide and Life Threatening Behavior, 16

(1986), pp. 448 - 57 -
12. D. Shaffer, P. Fisher, R. H. Hicks, M. Parides and M. Gould, ‘Sexual
Orientation in Adolescents Who Commit Suicide’, Suicide and Life Threatening
Behavior, 25 (1995), Supplement, pp. 64-71.
13. See P. Muehrer, 'Suicide and Sexual Orientation: a Critical Summary o f
Recent Research and Directions for Future Research’, Suicide and Life Threatening
Behavior, 25 (1995), Supplement, pp. 72-81 and J. Catalan, ‘Sexuality, Repro
ductive Cycle and Suicidal Behaviour’, in K. Hawton and K. van Heeringen
(eds) (op. cit), pp. 293-308.
14. See S. Platt and K. Hawton, ‘Suicidal Behaviour and the Labour Market’, in
K. Hawton and K. van Heeringen (eds) (op. cit.), pp. 309-34.
15. M. Boor, ‘Relationship between Employment Rates and Suicide Rates in
Eight Countries: 1962-1967’, Psychological Reports, 47 (1980), pp. 1095-101.
16. K. Hawton, J. Faggs, S. Simkin, L. Harriss and A. Malmberg, ‘Methods Used
for Suicide by Farmers in England and W ales’, British Journal of Psychiatry, 173
(1998), pp. 320-24.
17. E. Dooley, ‘Prison Suicide in Wales 1972 to 1987’, British Journal o f Psychiatry,
156 (1990), pp. 40-45.
18. E. N. Stenager and E. Stenager, ‘Physical Illness and Suicidal Behaviour’, in
K. Hawton and K. van Heeringen (eds) (op cit.), pp. 405-20.
19. A case illustrating how important it is not to ignore depression and other
accompaniments o f terminal illness is cited in J. Scott, ‘Cancer Patients’, in J.
Scott, J. M. G. Williams and A. T. Beck (eds.), Cognitive Therapy in Clinical
Practice, Routledge, London, 1989.
20. For a study o f psychiatric diagnoses as they differed according to age in suicide
victims, see C. L. Rich, R. C. Fowler, L. A. Fogarty etal, ‘San Diego Suicide Study
III: Relationships between Diagnosis and Stressors’, Archives of General Psychiatry, 445
(1988), pp. 589-92. See also A. Apter, D. Gothelp, I. OrbachetaL, ‘Correlates ofSuicidal
and Violent Behaviour in Different Diagnostic Categories in Hospitalised Patients’,
Journal of the American Academy of Child and Adolescent Psychiatry 34 (1995), 912-18.
21. G. Murphy, ‘Psychiatric Aspects o f Suicidal Behaviour: Substance Abuse’, in
K. Hawton and K. van Heeringen, (eds) (op. cit.), pp. 135-46.
22. For detailed statistics, see M. Shafii, J. Steltz-Lenarsky, A. M. Denick et a i,
‘Co-morbidity o f Mental Disorders in the Post-mortem Diagnosis o f Completed
Suicides in Children and Adolescents’ , Journal of Affective Disorders, 15 (1988),
pp. 227-33; Rich, Fowler, Fogarty et al., ‘San Diego Suicide Study III’ (op. cit.);
and, for a British follow-up survey in Edinburgh (1968-85), K. Hawton, S. Platt,
J. Fagg and M. Hawkins, ‘Suicide following Parasuicide in Young People’, British
Journal of Psychiatry, 152 (1993), PP- 359-66.
227
23. G. E. Murphy and R. D. Wetzel, 'Suicide Risk by Birth Cohort in the United
States, 1949-1974’, Archives of General Psychiatry, 37 (1980), pp. 519-2.3; M. J.
Solomon and C. P. Hellon, ‘Suicide and Age in Alberta, Canada, 1951-1977: a
Cohort Analysis’, Archives o f General Psychiatry, 37 (1980), pp. 511-13; R. D.
Goldney and M. Katsikitis, ‘Suicide Rates in Australia’, Archives of General
Psychiatry, 40 (1983), pp. 71-4.
24. Charlton, Kelly, Dunnell et a i, ‘Trends in Suicide Deaths in England and
Wales’ (op. cit. see note 6).
25. See U. Bille-Brahe, ‘Sociology and Suicidal Behaviour’ , in K. Hawton and K.
van Heeringen, (eds) (op. cit.), pp. 193-208. Bille-Brahe discusses evidence
suggesting that cohort size is not always associated with suicide, but only in
those countries that have poorer welfare provision so cannot make up for
limited resources when a large cohort competes for them (pp. 24-5).
26. Charlton, Kelly, Dunnell et al., ‘Trends in Suicide Deaths in England and
W ales’ (op. cit., see note 6).
27. N. Kreitman, V. Carstairs and J. Duffy, ‘Association o f Age and Social Class
with Suicide among Men in Great Britain’,Journal o f Epidemiology and Community
Health, 45 (1991), pp. 195-202.
28. For the Cross National Collaborative Group reference, see ‘The Changing
Rate o f Major Depression: Cross-national Comparisons’, Journal of the American
Medical Association, 268 (1992), pp. 355-73.
3 Psychiatric and Social Factors in Suicide
1. Durkheim, Le Suicide (see note 1, Chapter 2).

2. B. Barraclough, J. Bunch, B. Nelson and P. Sainsbury, ‘A Hundred Cases o f
Suicide: Clinical Aspects’, British Journal o f Psychiatry, 25 (1974), pp. 355-73.
3. G. W. Blair-West, G. W. Mellsop, M. L. Eyeson-Annan, ‘Down-rating
Lifetime Suicide Risk in Major Depression’, Acta Psychiatrica Scandinavica (1997),
pp. 259-63.
4.J. Fawcett, K. A. Busch, D. Jacobs, e ta i, ‘Suicide: A Four-pathway Biochemical
Model’, Annals of New York Academy of Sciences, 836 (1997), pp. 288-310.
5. See study by M. Birchwood, Z . Iqbal, P. Chadwick and P. Trower, ‘C og
nitive Approach to Depression and Suicidal Thinking in Psychosis. 1. Ontogeny
o f Post-psychotic Depression’ , British Journal of Psychiatry, 177 (2000),
pp. 516-21.
6. C. Allgulander, ‘Anxiety Disorder’, in K. Hawton and K. van Heeringen,
(eds) (op. cit.), pp. 179-92.
7. S. Platt, R. Micciolo and M. Tansella, ‘Suicide and Unemployment in Italy:
228
Description, Analysis and Interpretation o f Recent Trends’ , Social Science Medi

cine, 34 (1992), pp. 1191-201.
8. P. Sainsbury, J. Jenkins and A. Levey, ‘The Social Correlates o f Suicide in
Europe’, in R. Farmer and S. Hirsch (eds.), The Suicide Syndrome, Croom Helm,
London, 1980, pp. 38-53.
9. For recent definitions o f ‘social support’, see G. Brown and T. Harris, The
Social Origins of Depression - a Study o f Psychiatric Disorder in Women, Tavistock,
London, 1978.
10. Z. Iqbal, M. Birchwood, P. Chadwick and P. Trower, ‘Cognitive Approach
to Depression and Suicidal Thinking in Psychosis. 2. Testing the Validity o f a
Social Ranking Model’ , British Journal of Psychiatry, 177 (2000), pp. 522-28.
U .K . Hawton and J. Fagg, ‘Suicide, and Other Causes o f Death, Following
Attempted Suicide’, British Journal of Psychiatry, 152 (1988), pp. 359-66.
12. J. K. Lonnqvist, ‘Psychiatric Aspects: Depression’, in K. Hawton and K. van
Heeringen, (eds) (op. cit.), pp. 107-20.
13. See Z . Iqbal et al.y (op. cit.) in which the authors consider the evidence for
an entrapment model o f depression and suicidal thinking in schizophrenia.
4 Attempted Suicide: Facts and Figures
1. A. Alvarez, Savage God, A Study of Suicide, Random House, New York, 1972.
2. Although true in the individual case, studies o f larger numbers have found a
correlation between suicide intent and the number o f pills taken. More than
twenty pills has been found to indicate higher suicide intent. See R. D. Goldney,
‘Attempted Suicide in Young Women: Correlates o f Lethality’, British Journal
o f Psychiatry, 147 (1981), pp. 382-90; K. G. Power, D. J. Cooke andj. S. Gibbons,
‘Life Stress, Medical Lethality, and Suicidal Intent’, British Journal of Psychiatry,
147(1985), pp. 655-9; and D. J. Pallis, J. S. Gibbons and D. W. Pierce, ‘Estimating
Suicide Risk among Attempted Suicides’, British Journal o f Psychiatry, 144 (1984),
pp. 139-48.
3. K. Hawton, J. Fagg, S. Simkin, E. Bale and A. Bond, ‘Attempted Suicide in
Oxford, 1994’, unpublished report from University Department o f Psychiatry,
Oxford.
4. K Michel, P. Ballinari, U. Bille-Brahe et a l, ‘Methods Used for Parasuicide:
Results o f the W H O /E U R O Multicentre Study o f Parasuicide’, Social Psychiatry
and Psychiatric Epidemiology, 35 (2000), pp. 156-63.
5. For details o f the Canada data, see I. Sakinofsky, ‘The Epidemiology o f Suicide
in Canada’, in A. Leenaars, S. Wenckstem, I. Sakinofsky, M. Krai, R. Dyck and
R. Bland (eds), Suicide in Canada, University o f Toronto, 1996.
229
6. E. K. Moscicki, P. O’Carroll, D. S. Rae, B. Z. Locke, A. Roy and D. A. Regier,

‘Suicide Attempts in the Epidemiologic Catchment Area Study’, Yale Journal of
Biology and Medicine, 6i (1988), pp. 259-68.
7. R. D. Goldney, ‘Suicidal Ideation in a Young Adult Population’ , Acta Psychi-
atrica Scandinavica, 79 (1989), pp. 481-9.
8. D. Buglass and J. Horton, ‘A Scale for Predicting Subsequent Suicidal
Behaviour’, British Journal o f Psychiatry, 124 (1974), pp. 573-8.
9. N. Kreitman and J. Foster, ‘The Construction and Selection o f Predictive
Scales, with Special Reference to Parasuicide’ , British Journal of Psychiatry, 159
(1991), pp. 185-92.
10. I. Sakinofsky, ‘Repetition o f Suicidal Behaviour’, in K. Hawton and K. van
Heeringen, (eds) (op. cit.), pp. 385-404.
11. J. Bancroft, K. Hawton, S. Simkin et a l, ‘The Reasons People Give for Taking
Overdoses’ , British Journal of Medical Psychology, 52 (1979), pp. 353-65.
12. A. T. Beck, D. Schuyler and J. Herman, ‘Development o f Suicidal Intent
Scales', in A. T. Beck, H. L. P. Resnick and D. J. Lettieri (eds), The Prediction of
Suicide, Charles Press, Maryland, 1974.
13. D. W. Pierce, ‘Suicidal Intent in Self-injury’, British Journal of Psychiatry, 130
(1977), PP- 377- 85 .
14. R. Plutchik, H. M. Van Praag, S. Picard, H. R. Conte et al., ‘Is There a
Relationship between the Seriousness o f Suicidal Intent and the Lethality o f the
Suicidal Attempt?’ , Psychiatry Research, 27 (1989), pp. 71-9.
15. D. W. Pierce, ‘Predictive Validation o f a Suicide Intent Scale’, British Journal
of Psychiatry, 139 (1981), pp. 445-6.
16. J. Suokas an d j. Lonnqvist, ‘Outcome o f Attempted Suicide and Psychiatric
Consultation: Risk Factors and Suicide Mortality during a Five-year Follow-up’,
Acta Psychiatrica Scandinavica, 84 (1991), pp. 545-9.
17. R. W. Beck, J. B. Morris and A. T. Beck, ‘Cross-validation o f the Suicidal
Intent Scale’, Psychological Reports, 34 (1974), pp. 445-6.
18. A. T. Beck and R. Steer, ‘Clinical Predictors o f Eventual Suicide: a 5-10-year
Prospective Study o f Suicide Attempters’, Journal of Affective Disorders, 17 (1989),
pp. 203-9; D- Lester, A. T. Beck and S. Narrett, ‘Suicidal Intent in Successive
Suicidal Actions’, Psychological Reports, 43 (1978), p. no.
230
5 The Causes of Attempted Suicide
1. R. W. Maris, 'Deviance as Therapy: the Paradox o f the Self-destructive

Females’, Journal o f Health and Social Behaviour, 12 (1981), pp. 113-24.
2. See K. Hawton and J. Catalan, Attempted Suicide, second edition, Oxford
University Press, 1987.
3. R. D. Goldney, ‘Parental Representation in Young Women Who Attempt
Suicide’, Acta Psychiatrica Scandinavica, 72 (1985), pp. 230-32.
4. G. Parker, H. Tupling and L. B. Brown, ‘A Parental Bonding Instrument’,
British Journal of Medical Psychology, 52 (1979), pp. 1-10 .
5. M. van Egmond, N. Gamefski, D. Jonker and A. Kerkhov, ‘The Relationship
between Sexual Abuse and Female Suicidal Behavior’, Crisis, 14 (1993),
pp. 129-39.
6. See X. Coll, F. Law, A. Tobias and K. Hawton, ‘Child Sexual Abuse in Women
W ho Take Overdoses: A Study o f Prevalence and Severity’, Archives of Suicide
Research, 4 (1998), pp. 291-306; B. A. van der Kolk, J. C. Perry andj. L. Herman,
‘Childhood Origins o f Self-destructive Behavior’ , American Journal of Psychiatry,
148 (1991), pp. 1665-71; S. E. Romans, J. L. Martin, J. C. Anderson, G. P. Herbison
and P. E. Mullen, ‘Sexual Abuse in Childhood and Deliberate Self-harm’,
American Journal o f Psychiatry, 152 (1995), pp. 1336-42.
7. Hawton and Catalan, Attempted Suicide (see note 2).
8. S. Platt and N. Kreitman, ‘Trends in Parasuicide and Unemployment among
Men in Edinburgh, 1968-1982’, British Medical Journal, 289 (1984), pp. 1029-32.
9. Hawton, Fagg, Simkin, Bale and Bond, ‘Attempted Suicide in Oxford, 1994',
unpublished report (see note 3, Chapter 4).
10. K. Hawton, J. Fagg, S. Platt and M. Hawkins, ‘Factors Associated with
Suicide after Parasuicide in Young People’, British Medical Journal, 306 (1993),
pp. 1641-4.
11. J. Bancroft, A. Skrimshire, J. Casson, O. Harvard-Watts and F. Reynolds,
‘People Who Deliberately Poison Themselves: Their Problems and Their
Contacts with Helping Agencies’, Psychological Medicine, 7 (1977), pp. 289-303.
12. S. M. Davonport, ‘Association between Parasuicide and St Valentine’s Day’,
British Medical Journal, 300 (1990), pp. 783-4.
13. K. Hawton, L. Harriss, L. Appleby et a l, ‘The Effect o f the Death o f Diana,
Princess o f Wales, on Suicide and Deliberate Self-harm’, British Journal of
Psychiatry, 177 (2000), pp. 463-6.
14. For further discussion o f these difficulties, see A. K. MacLeod, J. M. G.
Williams and M. M. Linehan, ‘New Developments in the Understanding and
Treatment o f Suicidal Behaviour’, Behavioural Psychotherapy, 20 (1992), pp. 193-218.
231
15. J. J. Platt, G. Spivack and W. Bloom, Manualfor the Means-Ends Problem-solving

Procedure (MEPS): A Measure of Interpersonal Problem-solving Skill, Hahnemann
Community M H /M R Center, Department o f Mental Health Services, Hahne
mann Medical College and Hospital, Philadelphia, 1987.
16. D. E. Schotte and G. A. Clum, ‘Problem-solving Skills in Suicidal Psychiatric
Patients’, Journal of Consulting and Clinical Psychology, 55 (1987), pp. 49-54.
Similarly poor problem-solving ability has also been found in younger suicidal
people; see M. J. Rotheram-Borus, P. D. Trautman, S. C. Dopkins and P. E.
Shrout, ‘Cognitive Style and Pleasant Activities among Female Adolescent
Suicide Attempters’, Journal o f Consulting and Clinical Psychology, 58 (1990),
pp. 554- 6 i.
17. M. M. Linehan, P. Camper, J. A. Chiles, K. Strohsal and E. N. Shearin, ‘Inter
personal Problem Solving and Parasuicide’, Cognitive Therapy and Research, 11
(1987), pp. 1-12 .
18. I. Orbach, H. Bar-Joseph and N. Dror, ‘Styles o f Problem Solving in Suicidal
Individuals’, Suicide and Life Threatening Behavior, 20 (1990), pp. 56-64.
19. J. F. Simonds, T. McMahon and D. Armstrong, ‘Young Suicide Attempters
Compared with a Control Group’, Suicide and Life Threatening Behavior, 21 (1991),
pp. 134-51.
20. A. K. MacLeod, G. Rose andj. M. G. Williams, ‘Components o f Helplessness
about the Future in Parasuicide’, Cognitive Therapy and Research, 17 (1993),
pp. 441 - 55-
21. A. T. Beck, G. Brown and R. A. Steer, ‘Prediction o f Eventual Suicide in
Psychiatric Inpatients by Clinical Ratings o f Helplessness’, Journal of Consulting
and Clinical Psychology, 57 (1989), pp. 309-10.
22. A. T. Beck, A. Weissman, D. Lester and L. Trexler, ‘The Measurement o f
Pessimism: The Hopelessness Seal t Journal of Consulting and Clinical Psychology,
42 (1974), PP- 861-5.
23. A. K. MacLeod, B. Pankhania, M. Lee and D. Mitchell, ‘Parasuicide,
Depression and the Anticipation o f Positive and Negative Experiences’, Psycho
logical Medicine, 27 (1997), pp. 973-7.
24. M. M. Linehan, J. L. Goodstein, S. L. Neilsen an d j. A. Chiles, ‘Reasons for
Staying Alive when You Are Thinking o f Killing Yourself: the Reasons for Living
Inventory’, Journal o f Consulting and Clinical Psychology, 51 (1983), pp. 276-86.
25. See A. Apter and O. Freusenstein, ‘Adolescent Suicidal Behaviour’, in K.
Hawton and K. van Heeringen (eds) (op. cit.), pp. 261-74.
•
232
6 Rational Suicide, Euthanasia and Martyrdom
1. D. Humphry, ‘Rational Suicide among the Elderly’, Suicide and Life Threatening
Behavior, 22 (1992), pp. 125-9, see p. 127; D. Humphry, Final Exit: the Practicalities
o f Self-deliverance and Assisted Suicide, Carol Publishing, Secaucus, NJ, 1991.
2. Humphry, ‘Rational Suicide among the Elderly’ (op. cit., see note 1).
3.J. H. Groenewoud, P .J. van der Maas, G. van der Wal eta i, ‘Physician-Assisted
Death in Psychiatric Practice in the Netherlands’, New England Journal of
Medicine, 336 (1997), pp. 1795-801.
4. Humphry, ‘Rational Suicide among the Elderly’, (op. cit.), p. 126.
5. A. Koestler, Arrow in the Blue, Collins and Hamish Hamilton, London, 1952.
6. R. D. Goldney, ‘Arthur Koestler: Was His Suicide Rational?’ , Crisis, 7 (1986),
pp- 33- 53-
7. J. Richman, ‘A Rational Approach to Rational Suicide’ , Suicide and Life
Threatening Behavior, 22 (1992), pp. 130-41.
8. Scott, ‘Cancer Patients’, in Scott, Williams and Beck (eds), Cognitive Theory
in Clinical Practice, (op. cit., see note 19, Chapter 2).
9. Richman, ‘A Rational Approach to Rational Suicide’ , (op. cit., see note 7).
7 Psychodynamics, Biology and Genetics
1. S. Freud, Mourning and Melancholia (1917), Standard Edition, vol. 14, Hogarth
Press, London, 1957.
2. Ibid.
3. S. Freud, The Ego and the Id (1923), Standard Edition, vol. 19, Hogarth Press,
London, 1961.
4. For psychoanalytic observations on suicide as a ‘reversible’ and ‘magical’ act,
see K A. Menninger, ‘Psychoanalytic Aspects o f Suicide’, International Journal
o f Psychoanalysis, 14 (1933), p. 376; C. W. Wahl, ‘Suicide as a Magical Act’, Bulletin
o f the Menninger Clinic, 21 (1957), p. 91.
5. M. Klein, ‘A Contribution to the Psychogenesis o f Manic-depressive States’
(i935), in Contributions to Psycho-Analysis 1921-1945: Melanie Klein, Hogarth Press,
London.
6. H. Guntrip, Schizoid Phenomena, Object Relations and the Self, Hogarth Press,
London, 1968.
7. S. Asch, ‘Suicide and the Hidden Executioner’, International Review of Psycho
analysis, 7 (1980), pp. 51-60.
8. J. Bowlby, Maternal Care and Mental Health, Columbia University Press, 1951.
233
9. J. T. Maltsberger, Suicide Risk: the Formulation of Clinical Judgement, New York

University Press, New York and London, 1986.
10. P. W. Gold, F. K. Goodwin and G. P. Chrousos, ‘Clinical and Biochemical
Manifestation o f Depression', New England Journal of Medicine, 319 (1988),
pp. 348-420.
11. M. Asberg, P. Thoren, L. Traskman etal., ‘Serotonin Depression: a Biochemi
cal Subgroup within the Affective Disorders?’, Science, 191 (1986) pp. 478-80.
12. See also E. F. Coccaro, L. J. Siever, H. M. Klar et al, ‘Serotonergic Studies in
Patients with Affective and Personality Disorders’, Archives o f General Psychiatry,
46 (1989), pp. 587-99; K. M. Malone, E. M. Corbitt, L. Shuhua and L. Mann,
‘Prolactin Response to Fenfluramine and Suicide Attempt Lethality in Major
Depression’, British Journal of Psychiatry, 168 (1996), pp. 324-9; G. N. Pandey,
S. C. Pandey, Y. Dwivedi et al., ‘Platelet Serotonin-2A Receptors: a Potential
Biological Marker for Suicidal Behavior’, American Journal of Psychiatry, 152
(i 995 ), pp. 850-55.
13. See R. M. Winchel, B. Stanley and M. Stanley, ‘Biochemical Aspects o f
Suicide’, in S. J. Blumenthal and D. J. Kupfer (eds.), Suicide over the Life Cycle:
Risk Factors, Assessment and Treatment of Suicidal Patients, American Psychiatric
Association Press, Washington, DC, 1990.
14. L. Traskman-Bendz an dJ.J. Mann, ‘Biological Aspects o f Suicidal Behaviour’,
in K. Hawton and K. van Heeringen, (eds) (op. cit.), pp. 64-77.
15. M. Linnoila, M. Virkkunen, M. Scheinin et al., ‘Low Cerebralspinal Fluid
5-HIAA Concentration Differentiates Impulsive from Nonimpulsive Violent
Behavior’, Life Science, 33, pp. 2609-14.
16. A. Roy, D. Nielson, G. Rylander and M. Sarchipione, ‘The Genetics o f
Suicidal Behaviour’, in K. Hawton and K. van Heeringen, (eds) (op. cit.),
pp. 209-21. For the relationship with personality and mood variable, see A.
Apter, H. van Praag, R. Plutchik et al., ‘Interrelationships Among Anxiety,
Aggression, Impulsivity and Mood: A Serotonergically Linked Cluster?’ Psy
chiatry Research, 32 (1990), PP- I9 I - 9 9 *
17. A. Roy, N. L. Segal, B. S. and Robinette Centralwall, ‘Suicide in Tw ins’,
Archives o f General Psychiatry, 48 (1991), PP- 29-32.
18. F. Schulsinger, S. S. Kety, D. Rosenthal and P. H. Wender, ‘A Family Study
o f Suicide’, in M. Schou and E. Stromgren (eds), Prevention and Treatment of
Affective Disorders, Academic Press, N ew York, 1979.
19. S. Kety, ‘Genetic Factors in Suicide: Family, Twin and Adoption Studies’, in
Blumenthal and Kupfer (eds), Suicide over the Life Cycle, (op. cit., see note 13).
234
8 The Effect of the Media
1. Durkheim, Le Suicide, (op. cit., see note i, Chapter 2).

2. D. P. Phillips, 'The Influence o f Suggestion on Suicide: Substantive and
Theoretical Implications o f the Werther Effect’, American Sociological Review, 39
(1974), PP- 340 - 54 .
3. B. Barradough, D. Shepherd and C. Jennings, T w o Newspaper Reports of
Coroners’ Inquests Incite People to Commit Suicide?’, British Journal o f Psy
chiatry, 131 (1977), PP- 528-32.
4. K. A. Bollen and D. P. Phillips, ‘Imitative Suicides: a National Study o f the
Effects o f Television News Stories’ , American Sociological Review, 47 (1982),
pp. 802-9.
5. For evidence o f suicide by burning as an imitative phenomenon, see J. R.
Ashton and S. Donnan, ‘Suicide by Burning as an Epidemic Phenomena [sic]:
an Analysis o f 82 Deaths and Inquests in England and Wales in 1978-9’,
Psychological Medicine, 11 (1981), pp. 735-9.
6. S. J. Ellis and S. Walsh, ‘Soap May Seriously Damage Your Health’ (letter),
Lancet (1986), p. 686.
7. S. Platt, ‘The Aftermath o f Angie’s Overdose: Is Soap (Opera) Damaging to
Your Health?’ , British Medical Journal, 294 (1987), pp. 954-7; S. Platt, ‘The
Consequences o f a Televised Soap Opera Drug Overdose: Is There a Mass
Media Imitation Effect?’ , in R. F. W. Diekstra, R. W. Maris, S. D. Platt, A.
Schmidtke and G. Sonneck (eds), Attitudinal Factors in Suicidal Behaviour and Its
Prevention, Swets 3c Zeitlinger, Brill, 1987.
8. J. M. G. Williams, C. Lawton, S. Ellis, S. Walsh and J. Reed, ‘Imitative
Parasuicide by Overdose’ , Lancet, 8550 (1987), pp. 102-3 (report o f findings o f
research commissioned by the Independent Broadcasting Authority).
9. A. Schmidtke and H. Hafner, ‘The Werther Effect after Television Films: New
Evidence from an Old Hypothesis’, Psychological Medicine, 18 (1988), pp. 665-76.
10. K. Hawton, S. Simkin, J. Deeks et a l, ‘Effects o f Drug Overdose in a
Television Drama on Presentations to the Hospital for Self-poisoning: Time
Series and Questionnaire Study’, British Medical Journal, 318 (1999), pp. 972-9.
11. D. P. Phillips, ‘The Impact o f Fictional Television Stories on American Adult
Fatalities’, American Journal of Sociology, 87 (1982), pp. 1340-59.
12. R. C. Kessler and H. Stipp, ‘The Impact o f Fictional Television Suicide
Stories on US Fatalities: a Replication’, American Journal of Sociology, 90 (1984),
pp. 151-67.
13. J. N. Barron and P. C. Reiss, ‘Mass Media and Violent Behavior’ , American
Sociological Review, 131 (1985), pp. 528-32.
235
14. D. P. Phillips and L. L. Carstensen, ‘Clustering Teenage Suicides after

Television News Stories about Suicides’, New England Journal o f Medicine, 315
(1986), pp. 685-9-
15. M. Gould, ‘Suicide Clusters and Media Exposure’ , in Blumenthal and Kupfer
(eds), Suicide over the Life Cycle (op. cit., see note 13, Chapter 7).
9 The Cry of Pain
1. R. M. Post, ‘Transduction o f Psychosocial Stress into the Neurobiology

o f Recurrent Affective Disorder’, American Journal of Psychiatry, 149 (1992),
pp. 999-1010.
2. G. L. Engel, ‘Anxiety and Depression-withdrawal: the Primary Effects o f
Unpleasure’, International Journal o f Psychoanalysis, 43 (1962), pp. 89-97.
3. R. D. Goldney, ‘Attempted Suicide: an Ethological Perspective’ , Suicide and
Life Threatening Behavior, 10 (1980), pp. 131-41.
4. E. S. Schneidman, ‘Orientations towards Death: a Vital Aspect o f the Study
o f Lives’, International Journal of Psychiatry, 2 (19 66), pp. 167-200.
5. P. Gilbert, J. Prince and S. Allan, ‘Social Comparison, Social Attractiveness
and Evolution: H ow Might They Be Related?’, New Ideas in Psychology, 13 (1995),
pp. 149 - 65 .
6. M. J. Raleigh, M. T. McGuire, G. L. Brammer and A. Yuwiler, ‘Social and
Environmental Influences on Blood Serotonin Concentrations in Monkeys’,
Archives o f General Psychiatry, 41 (1984), pp. 405-10.
7. S. Platt, ‘Unemployment and Suicidal Behaviour: a Review o f the Literature’,
Social Science and Medicine, 19 (1984), pp. 93-115.
8. R. G. Wilkinson, ‘Income Distribution and Life Expectancy’, British Medical
Journal, 304 (1992), pp. 165-8.
9. P. Gilbert, Depression: The Evolution of Powerlessness, Lawrence Erlbaum
Associates, Hove, and Guilford, New York, 1992.
10. P. Gilbert, ‘Defence and Safety: Their Function in Social Behaviour and
Psychopathology’, British Journal of Clinical Psychology, 32 (1993), pp. 131-54.
11. N. L. Farberow and E. S. Schneidman (eds), The Cry for Help, McGraw-Hill,
N ew York, 1961.
12. Stengel, Suicide and Attempted Suicide (op. cit., see note 2, Introduction).
13. Schneidman, ‘Orientations towards Death’, (op. cit., see note 4).
14. N. Kessel, ‘The Respectability o f Self-poisoning and the Fashion o f Survival’,
Journal of Psychosomatic Research, 10 (1966), pp. 29-36.
15. Alvarez, Savage God (op. cit., see note 1, Chapter 4).
16. P. Gilbert, Human Nature and Suffering, Lawrence Erlbaum Associates, Hove
236
and London, 1989. See also J. S. Price and L. Sloman, ‘Depression as Yielding
Behaviour: An Animal Model Based on Schjelerup-Ebb's Pecking Order',
Ethology and Sociobiology, 8 (1987), pp. 85-98.
17. See P. Gilbert and S. Allan, ‘The Role o f Defeat and Entrapment (Arrested
Flight) in Depression: An Exploration o f an Evolutionary View', Psychological
Medicine, 28 (1998), pp. 585-98. There is now increasing evidence that biological
processes are triggered by the perception and expectation o f inescapability and
that a vicious circle involving biological, social and psychological factors can
occur. See K. van Heeringen, K. Hawton an d j. M. G. Williams, ‘Pathways to
Suicide: An Integrative Approach', in K. Hawton and K. van Heeringen (eds)
(op cit.), pp. 223-36.
18. MacDonald and Murphy, Sleepless Souls (op. cit., see note 2, Chapter 1).
10 Memory Traps
1. J. D. Teasdale and S. J. Fogarty, ‘Differential Effects o f Induced Mood on

Retrieval o f Pleasant and Unpleasant Events from Episodic Memory', Journal of
Abnormal Psychology, 88 (1979), pp. 248-57; G. G. Lloyd and W. A. Lishman,
‘Effect o f Depression on the Speed o f Recall o f Pleasant and Unpleasant
Experiences’, Psychological Medicine, 5 (1975), pp. 173-80.
2. J. M. G. Williams and K. Broadbent, ‘Autobiographical Memory in Attempted
Suicide Patients', Journal of Abnormal Psychology, 95 (1986), pp. 144-9; J. M. G.
Williams and B. H. Dritschel, ‘Emotional Disturbance and the Specificity o f
Autobiographical Mem ory', Cognition and Emotion, 2 (1988), pp. 221-34.
3. J. M. G. Williams, ‘Depression and the Specificity o f Autobiographical
Memory', in D. C. Rubin (ed.), Remembering Our Past: Studies in Autobiographical
Memory, Cambridge University Press, 1996.
4. R. G. Moore, F. N. Watts and J. M. G. Williams, ‘The Specificity o f
Personal Memories in Depression', British Journal of Clinical Psychology, 27 (1988),
pp. 275-6.
5. J. M. G. Williams and J. Scott, ‘Autobiographical Memory in Depression',
Psychological Medicine, 18 (1988), pp. 689-95.
6. A. Puffet, D. Jehin-Marchot, M. Timsit-Berthier and M. Timsit, ‘Autobio
graphical Memory and Major Depressive States’, European Psychiatry, 6 (1991),
pp. 141-5; H. Mackinger, M. Pachinger, M. Leibetseder and R. Fartyacek,
‘Autobiographical Memories in Women Remitted from Major Depression’,
Journal o f Abnormal Psychology, 109 (2000), pp. 331-4.
7. K. Nelson, ‘The Ontogeny o f Memory for Real Events', in U. Neisser
and C. E. Winograd (eds), Remembering Reconsidered: Ecological and Traditional
237
Approaches to the Study o f Memory, Cambridge University Press, 1988, pp. 244-76.
8. J. Morton, T h e Development o f Event M em ory’, Psychologist, 1 (1990),
pp. 3-10.
9. Although no studies have yet been designed to answer questions about non
specificity in children's memory, there are indications that early attachment
problems are seen in biases in memory in children as young as three years old.
See J. Belsky, B. Spritz and K. Cmic, ‘Infant Attachment Security and Affective-
cognitive Information Processing at Age 3’ , Psychological Science, 7 (1996),
pp. 111-14 .
10. W. Kuyken and C. R. Brewin, ‘Autobiographical Memory Functioning in
Depression and Reports o f Early Abuse’, Journal o f Abnormal Psychology, 104
(1995), pp. 585-91.
11. J. Evans, J. M. G. Williams, S. O’Loughlin and K. Howells, ‘Autobiographical
Memory and Problem Solving Strategies o f Parasuicide Patients’, Psychological
Medicine, 22 (1992), pp. 399 - 405 *
12. E. M. Marx, J. M. G. Williams and G. S. Claridge, ‘Depression and Social
Problem-solving’, Journal of Abnormal Psychology, 101 (1992), pp. 78-86.
13. G. L. Sidley, K. Whitaker, R. M. Calam and A. Wells, T h e Relationship
Between Problem Solving and Autobiographical Memory in Parasuicide
Patients’, Behavioural and Cognitive Psychotherapy, 25 (1997), pp. 195-202.
14. L. R. Pollock andj. M. G. Williams, ‘Effective Problem Solving Depends on
Specific Autobiographical Recall’, Suicide and Life Threatening Behaviour (in
press).
15. For the work that led to these conclusions, see J. M. G. Williams, N.
Ellis, C. Tyers, H. Healy, G. Rose and A. K. MacLeod, T h e Specificity of
Autobiographical Memory and Imageability o f the Future’, Memory and Cog
nition, 24 (1996), pp. 116-25.
16. C. B. Traux and R. R. Carkhuff, ‘Concreteness: a Neglected Variable in
Research in Psychotherapy’, unpublished MS, Psychotherapy Research Pro
gram, Universities o f Kentucky and Wisconsin, 1967. For a similar observation,
made in studies o f young mothers who were having problems in relationships
with their children, see R. G. Wahler and A. D. Afton, ‘Attentional Processes
in Insular and Non-insular Mothers: Some Differences in Their Summary
Reports about Child Problem Behaviours’, Behaviour Therapy, 2 (1980), pp. 25-41.
17. A. D. Brittlebank, J. Scott, J. M. G. Williams and I. N. Ferrier, ‘Autobiographi
cal Memory in Depression; State or Trait Marker?’ , British Journal o f Psychiatry,
162 (1993), pp. 118-21.
18. A. G. Harvey, R. A. Bryant and S. T. Dang (1998), ‘Autobiographical Memory
in Acute Stress Disorder’, Journal of Consulting and Clinical Psychology, 66,
pp. 500-506.
238
19. T. Dalgleish, H. Spinks, J. Yiend and W. Kuyken (in press), Autobiographical

Memory Style in Seasonal Affective Disorder and Its Relationship to Future
Symptom Remission’, Journal of Abnormal Psychology, no (2001), pp. 335-40.
11 The Prevention of Suicidal Behaviour
1. D. W. Coombs, H. L, Miller, R. Alarcon, C. Herlinhy, J. M. Lee and

D. P. Morison, ‘Presuicide Attempt Communications between Parasuicides
and Consulted Caregivers’, Suicide and Life Threatening Behavior, 22 (1992),
pp. 289-302.
2. D. De Leo, G. Carollo and M. Dello Buono, ‘Lower Suicide Rates Associated
with TeleH elp/TeleCheck Service for the Elderly at Home , American Journal
of Psychiatry, 152 (1995), pp. 632-4.
3. J. Motto, ‘Suicide Prevention for High-risk Persons Who Refuse Treatment’,
Suicide and Life Threatening Behaviour, 6 (1976), pp. 223-30.
4. W. Rutz, L. von Knorring and J. Walinder, ‘Frequency o f Suicide on Gotland
after Systematic Postgraduate Education o f General Practitioners’, Acta Psychi-
atrica Scandinavica, 80 (1989), pp. 151-4.
5. W. Rutz, L. von Knorring and J. Walinder, ‘Long-term Effects o f an Edu
cational Program for General Practitioners Given by the Swedish Committee
for the Prevention and Treatment o f Depression’, Acta Psychiatrica Scandinavica,
85 (1992), pp. 83-8.
6. A. MacDonald, ‘The Myth o f Suicide Prevention by General Practitioners’,
British Journal of Psychiatry, 162 (1993), p. 260.
7. As an Office o f Population Census and Surveys report found, one in seven
people in Britain between 16 and 64 has some sort o f neurotic health problem
at any one time; see H. Meltzer, B. Gill and M. Pettigrew, OPCS Survey of
Psychiatric Morbidity in Great Britain, O P C S, London, 1994.
8. Sensitivity o f a test is defined as the number o f true positives divided by the
sum o f true positives plus false negatives, then multiplied by 100. The numerator
is the ‘positive hit rate’ (suicides successfully predicted) and the denominator is
the total number o f such hits that are possible - the total number o f actual
suicides in the population being studied. Specificity o f a test is defined as the
number o f true negatives divided by the sum o f false positives and true
negatives, then multiplied by 100. The numerator is the negative ‘hit rate’ (lack
o f suicide successfully predicted) and the denominator is the total number o f
such hits that are possible - the total number o f people who do not commit
suicide in the population being studied.
9. A. D. Pokomy, ‘Prediction o f Suicide in Psychiatric Patients: Report o f a
239
Prospective Study’, in R. W. Maris, A. L. Berman, J. T. Maltsberger and R. I.

Yufit (eds), Assessment and Prediction of Suicide, Guilford Press, N ew York, 1992.
10. R. W. Maris, T h e Prediction o f Suicide’ in M. J. Kelleher (ed.), Divergent
Perspectives on Suicidal Behaviour, Report o f Fifth European Symposium on
Suicide, Cork, 1994, pp. 28-41.
11. L. Appleby, J. Shaw, J. Amos et al., Safer Services: Report o f the National
Confidential Inquiry into Suicide and Homicide by People with Mental Illness,
Stationery Office, London, 1999.
12. L. Appleby, J. A. Dennehy, C. S. Thomas et a l, Aftercare and Clinical
Characteristics o f People with Mental Illness W ho Commit Suicide: A Case
Control Study’, Lancet, 353 (1999), pp. 1397-1400.
13. J. A. Dennehy. L. Appleby and B. Faragher, ‘A Case Control Study o f Suicide
by Discharged Psychiatric Patients’, British Medical Journal, 312 (1996), p. 1580.
14. L. Appleby, ‘Suicide in Psychiatric Patients: Risks and Prevention’, British
Journal of Psychiatry, 161 (1992), pp. 749-58.
15. R. W. Maris, T h e Prediction o f Suicide’ (op. cit., see note 10)
16. For further discussion o f this work, see D. Lester, Can We Prevent Suicide?,
AM S Press, New York, 1989.
17. D. I. R. Jones in British Medical Journal, 282 (1977), pp. 28-9.
18. K. Hawton and L. Vislisel, ‘Suicide in Nurses’, Suicide & Life Threatening
Behaviour, 29 (1999), pp. 86-95.
19. S. Greer and M. Alderson, ‘Samaritan Contact among 325 Parasuicide
Patients’, British Journal of Psychiatry, 135 (1979), pp. 263-8.
20. C. Jennings, B. M. Barraclough and J. R. Moss, ‘Have the Samaritans
Lowered the Suicide Rate: a Controlled Study’ , Psychological Medicine, 8 (1978),
pp. 413-27.
21. D. Lester, T h e Effectiveness o f Suicide Prevention Centres', Suicide and Life
Threatening Behavior, 23 (1993), pp. 263-7.
22. For sceptical evaluations, see Suicide and Parasuicide, M R C Topic Review,
Medical Research Council, London, 1995; D. Shaffer, M. A. Garland, M. Gould
et al., ‘Preventing Teenage Suicide: a Critical Review ’, Journal of the American
Academy of Childhood and Adolescent Psychiatry, 27 (1988), pp. 675-87.
23. D. Shaffer, A. Garland, V. Vieland et al., T h e Impact o f Curriculum-based
Suicide Prevention Programs for Teenagers’, Journal of the American Academy of
Childhood and Adolescent Psychiatry, 30 (1991), pp. 588-96.
24. D. Gunnell, The Potential for Preventing Suicide, Health Care Evaluation Unit,
University o f Bristol, 1994.
240
12 Therapy for Suicidal Feelings and Behaviour
1. K. Hawton, E. Arensman, E. Townsend et al., ‘Deliberate Self-harm: System

atic Review o f Efficacy o f Psychosocial and Pharmacological Treatments in
Preventing Repetition’, British Medical Journal, 317 (1998), pp. 441-7.
2. One o f the best introductions to such an approach can be found in K. Hawton
andj. Kirk, ‘Problem Solving Treatment’, in K. Hawton, D. Clark, P. Salkovskis
and J. Kirk (eds), Cognitive Therapyfor Adult Psychiatric Patients, Oxford University
Press, 1989.
3. T. J. D ’Zurilla and M. R. Goldfried, ‘Problem Solving and Behaviour Modifi
cation’, Journal of Abnormal Psychology, 78 (1970), pp. 107-26: M. R. Goldfried
and A. P. Goldfried, ‘Cognitive Change Methods’, in F. H. Kanfer and A. P.
Goldstein (eds), Helping People Change, Academic Press, New York, 1975.
4. Hawton and Kirk, ‘Problem Solving Treatment’, in Hawton et al. (eds),
Cognitive Therapy for Adult Psychiatric Patients (see note 2).
5. J. S. Gibbons, J. Butler, P. Urwin an d j. L. Gibbons, ‘Evaluation o f a Social
W ork Service for Self-poisoning Patients’, British Journal of Psychiatry, 133 (1978),
pp. 111-18 .
6. For two small-scale studies that have given provisionally more encouraging
results, see B. C. McLeavey, R. J. Daly, J. W. Ludgate and C. M. Murray,
‘Interpersonal Problem Solving Skills Training in the Treatment o f Self Poison
ing Patients’, Suicide and Life Threatening Behavior, 24 (1994), pp. 382-94; P. M.
Salkovskis, C. Atha and D. Storer, ‘Cognitive-Behavioural Problem Solving in
the Treatment o f Patients Who Repeatedly Attempt Suicide’, British Journal of
Psychiatry, 157 (1990), pp. 871-6.
7. M. M. Linehan, H. E. Armstrong, A. Suarez, D. Allmond and H. L. Heard,
‘Behavioural Treatment o f Chronically Parasuicidal Borderline Patients’,
Archives of General Psychiatry, 48 (1991), pp. 1060-64.
8. M. M. Linehan, H. L. Heard and H. E. Armstrong, ‘Naturalistic Follow-up
o f a Behavioural Treatment for Chronically Parasuicidal Borderline Clients’,
Archives of General Psychiatry, 50 (1993), pp. 971-4.
9. For overview, see K Koemer and M. Linehan, ‘Research on Dialectical
Behaviour Therapy for Patients with Borderline Personality Disorder’, Psychi
atric Clinics of North America, 23 (March 2000), pp. 151-67.
10. G. Low, D. Jones, C. Duggan, M. Power and A. MacLeod, ‘The Treatment
o f Deliberate Self-harm in Borderline Personality Disorder using Dialectical
Behaviour Therapy: A Pilot Study in a High Security Hospital, Behavioural and
Cognitive Psychotherapy, 29 (2001), pp. 85-92.
11. K. Evans, P. Tyrer, J. Catalan, U. Schmidt, K. Davidson, J. Dent, P. Tata,
241
S. Thornton, J. Barber and S. Thompson, ‘Manual-assisted Cognitive-behaviour

Therapy (M A C T ): A Randomised Controlled Trial o f a Brief Intervention with
Bibliotherapy in the Treatment o f Recurrent Deliberate Self-harm’, Psychological
Medicine, 29 (1999), pp. 19-25.
See also A. K. MacLeod, P. Tata, K. Evans, P. Tyrer, U. Schmidt, K. Davidson,
S. Thornton and J. Catalan, ‘Recovery o f Positive Future Thinking within a
High-risk Parasuicide Group: Results from a Pilot Randomised Control Trial’,
British Journal o f Clinical Psychology, 37 (1998), pp. 371-9.
12. Evans, Williams, O’Loughlin and Howells, ‘Autobiographical Memory and
Problem Solving Strategies’ (op. cit., see note 11, Chapter 10).
13. H. G. Morgan, E. M. Jones an d j. H. Owen, ‘Secondary Prevention o f Non-
fatal Deliberate Self-Harm: the Green Card Study’, British Journal of Psychiatry,
163 (1993), pp. 111-12 .
14. J. Motto (op cit., see note 3, Chapter n) and D. De Leo, et a l, (op cit., see
note 2, Chapter 11).
13 Final Thoughts
1. P. Levi, The Drowned and the Saved, Michael Joseph, London, 1988.
2. D. M. Shepherd and B. M. Barraclough, ‘The Aftermath o f Parental Suicide
for Children’, British Journal o f Psychiatry, 129 (1976), pp. 267-76.
3. N. L. Farberow, ‘Adult Survivors after Suicide: Research Problems and
Needs’, in A. A. Leenaars (ed.), Life Span Perspectives o f Suicide, Plenum Press,
New York, 1991.
4. Ajahn Sucitto, ‘Making Peace with Despair’, in Peace and Kindness, Amaravati
Publications, Hemel Hempstead, 1990.
242
Index
Numbers in italics refer to figures; those in bold refer to tables.
Abbott, George, 5 analgesic dose, 99

adolescents analgesics, 66, 70, 130, 190
parasuicide, 89 Andrewes, Lancelot, 5
suicide, 37, 137 Andrews, Bernice, 146
African Americans, 43 anger, xv, 47, 89, 94, 95, 115, 116, 126,
age 161, 182, 217
and parasuicide, 69, 74 anomic suicide, 54
and suicide rate, 20, 22, 23 anomie, 56, 152-3
age and sex ratio, xv antidepressants, 48, 61, 66, 70, 121, 154,
Aids, 35-6, 98 , ,
155 177 190
alcohol abuse/alcoholism, 56 antisocial personality disorder, 51
and drug overdose, 70 anxiety disorders, 50
feeling at the mercy o f craving, 60 Appleby, Lewis, 87, 185
and genetic influences, 124-5 Apter, Alan, 37, 94
and low levels o f 5-HIAA, 122 Aquinas, St. Thomas, 3, 100, 101
and parasuicide, 74, 86, 195 arbitrary inference, 196
and peptic ulceration, 35 Aristotle, 2, 100
and psychiatric disorder, 37 Arrow in the Blue (Koestler), 109
as risk factor for suicide, 38, 46-9, arsonists, 122
60, 61-2, 78-9 Asberg, M., 122
suicide and life events, 6 1-2 Asian women (suicide rate in the
Alderson, M., 191 UK), 31
alternative therapy, 197-8 aspirin, 66, 190
altruistic suicide, 54, 112 assisted suicide, xiii, 98-9
Alvarez, Al, 63-4, 79-80, 148-9, anxiety about, 102
187-8 institutional support for, 103-4
243
Index
attachment, 139 gradually changing (seventeenth

attempted suicide, 63-80 century), 7-10
causes, 81-97: catastrophic effect madness as illness, 11-13
o f events, 87-8; emotional secularization of, 7, 8
experience, 94; emotional the situation in other countries,
regulation and self-cutting, 13-14
94-6; future directed thinking the Tudor revolution in attitudes,
and hopelessness, 90-91; 4 -7
interpersonal problem solving, Augustine, St, 2, 10
88-90; long-term vulnerability Australia, 24, 99, 188
factors, 81-5; precipitating Austria, 24, 71
factors, 86; reasons for living, Autobiographical Memory Test, 172
91-2, 93; short-term automatic suicide, 52-3
vulnerability factors, 85-6
criminalized, 16 ‘baby boom ’ generation, 39
and the ‘cry for help’, xv Bancroft, John, 75
decriminalized, 16 barbiturates, 188
definitions, 64-5 Barraclough, Brian, 46, 128, 191
international trends, 7 1-2 Baumeister, Roy, 90
motivation for self-harm, 75 BBC, effect o f Casualty drama, 131
Oxford data, 66-70 Beck, Aaron, 78, 91, 147, 198
parasuicide statistics, 65-6 Belgium, 23, 24, 71
repetition rate, 74 bereavement, 51, 59
the reversal o f the female:male Biathanatos (Donne), 9, 10, 107
ratio, 73 Bible, the, 2-3, 9
separation from completed suicide not explicitly condemned
suicide, xiv in, 2, 9
suicidal ideas and suicidal ‘Thou shalt not kill’
behaviour, 73-5 commandment, 2-3, 9-10
suicidal intent, 75-9 biological aspects, 120-23, 138
and suicide, 218-19 bipolar affective disorder, 39, 48, 208
attitudes, 1-17 Birchwood, Max, 60
ambivalence within the Church, Blount, Charles, 100
14-15 Blumenthal, Susan, 45-6
ambivalent attitude towards Bollen, K. A., 132-3
suicide, 1 Book o f Common Prayer, 14
in classical and medieval times, Boor, M., 34
2-4 ‘borderline’ personality disorder,
the end o f the crime o f self- 50-51, 95, 97, 208
murder, 15-16 Bowlby, J., 119, 139
244
Index
brain damage, and uncontrolled conservation-withdrawal, 139-40

limbs, 7 Coombs, David, 175
Brewin, Chris, 146, 166 criminal record, 74
brief affective reaction, 124 Cross National Collaborative Group,
Bristol, 55, 214 42
British Association for the ‘cry for help’ idea
Advancement o f Science, 156 misunderstood, xv, 148
British Journal of Psychiatry, 178 and parasuicide, xv, 147
Brittlebank, Andy, 172 used pejoratively, xv
bullying, 137, 146, 150, 154 cry o f pain, xiv-xv, 136-54, 217
burial o f suicides being ‘in’ with the group, 143-4
denied Christian burial, 3, 14-15 connections between self-harm
rituals o f desecration, 3-4, 14, 15 and suicide, 152-4
conservation-withdrawal, 139-40
Calvinism, 5, 14 cry for help or cry o f pain, 147-9
Canada, 23, 24, 25, 33, 34, 72 entrapment, 149-52
cancer, 35 evolutionary approach, 140-43
car exhaust, 25,30, 188, 194 involuntary subordination, 145-6
catastrophizing, 196 social comparison and
Catholics, 32, 98 psychopathology, 144-5
cerebrospinal fluid (CSF), 122 cutting, 29, 67, 94-6
Charlton, 41
chemotherapy, 35 Dalgleish, Tim, 172
Chubbuck, Christine, 128 D BT see Dialectical Behaviour
Church, the Therapy
ambivalence towards, 14-15 death instinct (Thanatos), 116-17,
and the commandment ‘Thou De Leo, Diego, 176, 215
shalt not kill', 2-3, 9-10 dendrites, 120
some suicides seen as martyrs, 9,10 Denmark, 23, 24, 28, 71
see also Devil, the Department o f Health, 65
churchgoing, 32 depression
Cobain, Kurt, 127, 134 and assisted suicide, 106
cognitive deconstruction, 90 and the ‘baby boom ’ generation, 39
Coll, Xavier, 84 and the Devil metaphor, 6, 17
communication endogenous, 48, 58
and completed/attempted suicide, and general practice, 174
xiv-xv and genetic influences, 124-5
completed suicide and hopelessness, 90, 173, 196-7
separation from attempted suicide, and involuntary subordination,
xiv, xv 145-6
245
Index
depression - cont. Diaries (Pepys), 8-9

long-term, 60 dichotomous thinking, 196
and lo se r’ status, 150 disempowerment, 39
and low levels o f 5-HIAA, 122 dissociative states, 95
medical illness and, 36, 109-10 divorce
and melancholy suicide, 52 and male suicide, 23, 39, 41
and memory, 156-7, 171-3, 220 and parasuicide, 67, 72, 74, 78
mnemonic interlock in, 162-5 and suicide, 57, 58
more common in women, 47, 51 D N A (deoxyribonuleic acid), 139
and mourning, 115 Donne, John, 9, 10 -11, 107
as a normal mood state, 47 Dooley, E., 35
reactive, 58 dopamine, 121
recurrence, 155 Dr Faustus (Marlowe), 5-6
as a response to frustration and Drowned and the Saved, The (Levi),
losses, 139 217, 223
and risk o f suicide, 37, 46, 48, 59, drowning
60, 62, 148, 182 likelihood o f being called suicide,
and suicidal behaviour, xiv, 6, 122, 18
123, 173, 174 suicide rates (UK), 27, 29
symptoms, 48, 60 drugs and other substances
and the Hamilton Rating Scale, 172 in Australia, 188
vulnerability on special dates, 86-7 correlating with suicide intent,
Depression, the, 20, 23 7 7 -8
depressive suicide, 118 knowledge o f lethality, 191
deprivation, 55, 145 as major risk factors for suicide,
deterrence 38-9
families' forfeit o f property, 3, 8, 11 and parasuicide, 67, 70, 74, 85, 86
rituals o f desecration, 3-4, 13, 14, 16 and the state o f the brain at death,
Wesley and, 15-16 122
Devil, the UK statistics, 27
and depression, 6, 17 US statistics, 30
and melancholy, 7 Durkheim, Emile, 18, 45, 46, 52-9,
suicide brought about by the, 3, h i , 114, 128, 152-3
5-6, 7, 8, 12, 17 Dutch Supreme Court, ruling on
Dialectical Behaviour Therapy physician-assisted suicide, 106
(D BT), 206-12, 216
follow-up for suicidal behaviour, EastEnders (television programme),
208-10, 210, 211, 212 12 9 -3 1
Diana, Princess o f Wales, effect of economic circumstances, and suicide,
her death, 87 7, 8
246
Index
economic productivity, suicide as a Evans, Julie, 167-8, 213

loss of, 20 Evolution of Powerlessness, The
ECT, 49, 61 (Gilbert), 145
education Exit, 98-9, 104, 108
competition within, 39, 136-7
Japanese, 137 family
wom en’s tertiary, 56 and alcoholics, 6 1-2
Egbert (eighth-century writer), 3 changes in structure, 39, 56
Egmond, Marjan Van, 82, 83 decreased family cohesion, 41
ego, 115, 116, 117 and a relative’s suicide, 220-21
egoistic suicide, 54 seizure o f goods from families o f
empathy, 223 suicides, 3, 8
employment Fein, Esther, 99
large cohorts, 41 felo de se (a felon o f himself), 4, 5, 8,
loss of, 55, 61 11, 12,13, 16
opportunities, 39, 56 female attempted suicide
and parasuicide, 85-6 and entrapment, 153
endocrine function, 139 female:male ratio, 73
Engel, G. L., 139 peak age, 67
England and Wales female suicide rate, 21, 22, 24, 26,
cohort data, 39-40, 21 40
methods o f suicide, 25, 27, 29 female suicide rates
prison suicides, 35 alcoholics, 47-9
suicide rates, 20, 21, 22, 23, 24, 25 decrease in, 20, 23, 28, 219
Enlightenment, the 11, 14 international comparisons, 24
entrapment, 61, 149-52, 169, 217, 218 male rates exceed those for
Epicureanism, 2, 7 women, 19, 20, 21, 22, 23, 24, 26,
Epidemiologic Catchment Area 28-9
study (National Institute o f methods o f suicide (UK), 25, 27,
Mental Health), 72 28, 29, 30, 188
epilepsy, 35 in the UK, 19-20, 21, 22, 24
Epworth Workhouse, 13 Ferrier, Nicol, 172
ethnic groupings, 31 Final Exit (Humphry), xiii-xiv, 103
ethology, 139 Finland, 23, 24, 71
Europe firearms, 25, 27, 29,30, 187, 189
loss o f economic productivity UK, 29, 34, 189, 194
through suicide, 20 US, 30, 43, 189
euthanasia, xiii, 98, 114 First World War, 20
arguments against, 105-8 5-HIAA, 122
voluntary, 99, 103-4 Five Dissertations (Hume), 100-101
247
Index
France, 13, 14, 23, 24, 34, 71,190 Harvey, Allison, 172
Freud, Sigmund, 115-17, 118 Hawton, Keith, 34, 72, 81, 86, 131,
201-5
gambling, 56 Health Advisory Service, 1
gases and other vapours Hellon, C. P., 39
domestic gas detoxified, 25, 188 helplessness, 60, 61/92, 152, 169
and the state o f the brain at death, Helsinki, Finland, 71, 73
122 Hemlock Society, 99, 103
UK statistics, 25, 27 Hippocrates, 102, 120
US statistics, 30 Hippocratic oath, 102
gender, 28-9 Hollis, John, 34
and attempted suicide, 68, 69 homosexuality, 31
suicide method by, 23, 24, 27,31 hopelessness, xiv, 48, 60, 61, 86,
see also female suicide rates; male 90-91, no, 155, 167, 170-71, 173,
suicide rates 174, 182, 187, 195, 196-7, 199 , 217
genetic influences, 123-6 Hopelessness Scale, 91
Gentlemen's Companion, The humanism, 9, 10
(Ramesey), 12 Hume, David, 100-101, 102, 103
geographical mobility, 39 Humphry, Derek, xiii-xiv, 103, 106
Germany, 13, 23, 24, 34, 71 Hungary, 71
Gilbert, Professor Paul, 140-41, 143, Huntington’s chorea, 35
145-6, 149-50 hypnotics, 177, 188
Gilden, Charles, 100
Gilpin, John, 6 identical twins, 123-4
Goethe, Johann W olfgang von, 127 I f This Is a Man (Levi), 223
Goldney, Professor Robert D., 39, 82, Impact o f Events Scale, 166
109, 139-40 impulsive suicide, 52
Gotland study, 177-9 ,180 impulsivity, 122-3, 125-6, 182, 196
Gould, M., 134-5 insomnia, 48, 181
Greece, 24 internalization, 115
Green Card system, 214-15 interpersonal relationships
Greer, S., 191 and parasuicide, 86-7, 88-90, 138
Guide to Self-deliverance (Exit), 108-9 and suicidal behaviour, 196
Gunnell, D., 193 invalidation strategies, 96
guns see firearms involuntary subordination, 145-6
Guntrip, H., 118 Ireland, 23, 24
Italy, 13, 24, 34, 71
Hackney Hospital, 130
Hamilton Rating Scale, 172 Jack, Raymond, 84
hanging, i8, 25, 27,30,136, 137,151,18 8 Japan, 34, 24, 38,137
248
Index
Jefferson, Thomas, 13 suicide ceases to be a criminal

Jesus o f Nazareth, 10 offence, 2, 14 ,15 -17 , 65
Jew s, 32 Lennard, T. B., 15
Jones, Jim , 112 Lester, David, 189,192
Jonestown, Guyana, 112 lethal dose, 99
Joyce, Anthony, 8 Levi, Primo, 217, 223
jumping, 28, 29 life span issues, 36-40
Jung, C. G., 158 Linehan, Marsha, 91, 92, 96, 206-8,
215, 216, 223
kamikaze, 112-14 Lishman, Professor, Alwyn, 157
Katsikitis, M., 39 Locke, John, 11
Ker, Sir Robert, 10 London, 29
Kessel, N., 147 Lonngvist, Jouko, 61
Kety, Seymour, 125-6 Lorenz, Konrad, 147
Kevorkian, Jack, 102 Luxembourg, 23
Kidderminster General Hospital,
168 MacDonald, Michael, 151
K ings Bench, 4, 5, 11, 29 Macintosh, Sir James, 15
Kirk, Joan, 201, 203-5 MacLeod, Andrew, 90,170
Klein, Melanie, 118 Magnyficence (Skelton), 5
Koestler, Arthur, 108-9, *35 male attempted suicide
Koestler, Cynthia, 108-9 and entrapment, 153-4
Koresh, David, 112 female:male ratio, 73
Kreitman, Norman, 41, 74 peak age, 67
Kuyken, William, 166, 172 male suicide rate
and bereavement, 51
Lamm, Richard, 106 cohort effect, 39-40
Lancet, The, 139 male suicide rates
Law, Fergus, 88 alcoholics, 47, 49
law, the, 98-9, 102-3 increase in, 20, 22, 23, 28, 34, 219
and assisted suicide, 16 -17 international comparisons, 23, 24,
attempted suicide as a crime, 65 2.5
Court o f the Star Chamber, 4 male rates exceed those for
felo de se verdict, 4, 5, 8, 11, 12, 13, women, 19, 20, 21, 22, 23, 24, 25,
16 26, 28-9, 29
goods seized, 3, 8 ,11, 16 methods o f suicide, 27, 29, 30, 188,
juries, 3, 4-5, n , 12, 13 rise in young male suicide, 20, 23,
King's Bench, 4, 5, 11, 29 35, 38-42, 43: changes in rates of
non compos mentis verdict, 4, 5, 6, 7, depression, 42; cohort effect
11 explained, 39-40, 43; decreased
249
Index
male suicide rates - cont. lowering mood and impairing

family cohesion, 41; period judgement, 37
effect explanation, 40-41, 43 melancholy
risk highest in elderly men, 36 Durkheim’s view of, 58
and social class (UK), 32 as proof that a person’s balance o f
in the UK, 19, 20, 21, 22, 23, 24, 29, mind was disturbed, 11
35 as a sign that the Devil had taken
Maltsberger, John, 119 ,120 over a person’s soul, 7, 11
maniacal suicide, 52 melancholy suicide, 52, 53,
manipulative behaviour, 218 memory, impaired, 48
Mann, John, 122 memory traps, 154, 155-73, 220
Maris, Ronald, 81, 183, 185 the consequences o f over-general
marital breakdown, 59 memory, 167-73
see also divorce the mechanisms o f memory
marital status retrieval, 160-62
and male suicide, 23, 39, 41 memory bias, 156-8
and parasuicide, 67, 72 memory in suicidal patients,
Marlowe, Christopher, 5-6 158-60
married people mnemonic interlock in depression,
and parasuicide, 72 164-5
and suicide, 23 the origins o f over-general
martyrdom, 9, 111- 1 2 ,1 1 7 memory, 165-6
masochism, 117 Menninger, K A., 117
Means-Ends Problem Solving Test mental illness see psychiatric illness
(MEPS), 88, 168, 168, 213 methods o f suicide
media males use more lethal methods
impact on suicidal behaviour, xiv than women, 25, 43
media, the effect o f the, 127-35 see also cutting; drowning; drugs
the ‘Angie’ myth, 129-31 and other substances; firearms;
Casualty drama, 131 gases and other vapours;
criticism o f imitative research, hanging; jumping
132-3 Michigan, 102-3
evidence for imitative effects, mnemonic interlock, 164-5, ^ 7
128-9 Montaigne, Michel de, 10
fictional portrayals, 129 mood disorders
mechanisms o f imitation, 133 and the ‘baby boom ’ generation,
medical intervention 39
and probability o f suicidal Mood Indication Procedure, 157
behaviour, 2 Moore, Richard, 164
medication morality plays, 5
Index
More, Thomas, 7, 9, 10 Office o f Population Census and

Morgan, Professor Gethin, 1, 214 Surveys (O PCS), 20, 39
motivation, xv older people
Motto, Jerom e, 176, 215 decrease in suicides, xiii, 43
mourning, 115 increase in numbers, 56
Mourning and Melancholia (Freud), suicide risk, 37
115-16 On Sacred Disease (Hippocrates), 120
multiple sclerosis, 35 ‘On Suicide’ (Hume), 100-101
murderers, 122 Orbach, Israel, 37
Murphy, Terence, 151 Oregon, State of, 98
Murphy, G. E., 39 out-of-body experiences, 95
Murphy, George, 38 over-generalization, 196
overcrowding, 55
narcissistic object-choice, 116 Oxford Centre for Suicide Research,
National Confidential Inquiry into 70
Suicide and Homicide by people Oxford data, 66-70, 68, 69, 70, 73,
with Mental Illness, 184 86
National Institute o f Mental Health, 72 Oyjen, Dr Wilfred van, 103-4
native Americans, 31, 43
negative life events, 59-60 panic disorder, 50
Netherlands, 23, 24, 28, 57, 103-104, paracetamol, 66, 70, 70, 190
106,107 parasuicide see attempted suicide
neurones, 120-21 Parental Bonding Instrument, 82
neuroses, 49 parenting and parasuicide, 82
neurotransmitters, 37, 120-21, 125, peptic ulceration, 35
139, 14 1-2 Pepys, Samuel, 8-9, 12
Nevada, 34 personality disorder, 49, 50-51, 74,
New York, 99 122, 195
Nirvana, 127 personality traits, 37
Non compos mentis (not o f sound Phillips, David, 128-9, 132, 133
mind), 4, 5, 6, 7, n - 1 2 physical illness
non-fatal suicide attempts see conditions that increase the risk o f
attempted suicide suicide, 35-6, 37
noradrenalin (norepinephrine), 121 and depression, 36, 37
Northern Ireland, 23 and parasuicide, 85
Norway, 23, 24, 71 suicide as the right course o f
Norwich, 12, 29 action, 2, 7
Platt, Stephen, 33, 85, 130, 145
object-relations approach, 118-120 pleasure, loss of, 48
obsessive suicide, 52, 53 PMR see proportional mortality ratio
251
Index
poisoning proportional mortality ratio, 32

death-rates (England and Wales), Protestants, 32
190 psychiatric illness
and gun control strictness, 189 as an explanation o f suicide, 11-13,
as a leading cause o f suicide (UK), 36
25, 188 Durkheim’s view o f psychiatric
and parasuicide, 67-70 causation, 52-3
Pokomy, A. D., 183 evidence for psychiatric illness as a
Pollock, Leslie, 34, 170 cause o f suicide, 46-51
Portsmouth, 128 gender and age effects not
Portugal, 24 explained by, 5 1-2
Powys, Wales, 34 non compos mentis verdicts, 4, 5, 6,
prevention 7, 11- 12
disagreement over whether and the risk o f suicide, 37
suicide can / should be statistics o f suicides, 2, 45
prevented, 1 psychiatric intervention
strategies, xiv and probability o f suicidal
see also primary prevention o f behaviour, 2, 195
suicidal behaviour psychoanalytic perspectives, 115-18
primary prevention o f suicidal psychological autopsy, 46
behaviour, 174-94 psychosis, 55
availability o f means, 186-91 psychotherapy, 17, 58, 154, 172, 219
the general problem o f prediction, Pythagoras, 2, 100
182-3
the Gotland study, 177-9 Raleigh, Michael, 142
primary care and primary Ramesey, William, 12
prevention, 18 1-2 rational suicide, 108-10, 114
Samaritans and suicide prevention renewed fascination with, xiii
centres, 19 1-2 Reasons for Living Inventory, 91-2,
school-based intervention, 192-3 93
timing, 183-4 religious affiliation, 32, 39, 57
prison suicides, 35, 185 Renaissance, the, 7, 9
Problem Solving Treatment, 201-5 renal dialysis, 35
problem-solving Resource Holding Potential (RHP),
evaluation o f problem-solving 143-4
approaches, 205-6 Richman, Joseph, 109
some common problems, 199-200 R N A (ribonucleic acid), 139
therapy, 200-205 Romans, S. E., 85
professional workers, 32 Rose, Gillian, 90
property, forfeiture of, 3, 8, 11, 16 Roy, A., 123-4
252
Index
rural areas, 34, 56 self-neglect, 48

Rutz, W ., 177 semi-skilled workers, 32, 32 , 55
serotonin, 37, 121-2, 125, 126, 14 1-2
sadism, 117 sexual abuse, 95, 96
St Bartholomew’s Hospital, 130 sexual abuse and parasuicide, 82-5
Sakinofsky, Isaac, 75 sexual orientation, 31
Samaritans, 19 1-2 shop-lifting, 137-8
Savage God (Alvarez), 63-4, 187 Sidley, Gary, 170
Scandinavia, 36 single people
schizoid suicide, 118 and male suicide, 41
schizophrenia and parasuicide, 67, 74, 78
and D B T , 208 single-parent families, 57
external stresses, 61 Skelton, John, 5
and genetic influences, 124-5 skilled workers, 32, 32
and low levels o f 5-HIAA, 122 sleeping pills, 60, 63, 66, 79, 188, 190
and maniacal suicide, 52 soap operas, 129-31
risk o f suicide, 37, 46-7, 48- 9 , Social Attention Holding Power,
49-50, 61, 62 144-5
and suicidal behaviour, 122 social class, 32-3, 32, 43, 56, 67, 74
symptoms, 49, 60, 61 social comparison, 144-5
and uncontrolled limbs, 6-7 social factors, xiv
Schmidtke, Arnold, 131, 133 evidence to support the social
Schneidman, E. S., 140, 147 theory, 55-7
school-based intervention, 192-3 importance of, 53-4
Schulsinger, F., 124 and the individual, 57-9
Scott, Jan, 165, 172 and suicide rates, xiv, 39
Seasonal Affective Disorder social support, 221, 221
overgeneral memory and society
persistence o f depression, 172 Aristotle and, 2, 100
Second World War, 20 discourages people from suicide,
secondary prevention o f suicidal 13
behaviour see therapy for a person’s contract with, 2
suicidal feelings and behaviour view o f itself as successful
Secretary's Task Force on Youth challenged, xiii
Suicide,The (US Department o f sociobiology, 117
Health and Human Services), 31 sociopathy, 74
secularization o f suicide, 8 Solomon, M. J., 39
sedatives, 70, 70, 177, 188, 190 Sorrows o f Young Werther, The
selective abstraction, 196 (Goethe), 127
self-deliverance see rational suicide Spain, 13, 23, 24 , 71
253
Index
spinal-cord injuries, 35 as a sin, 1-3, 5

status, 57 survivors affected by, 220-23
Stengel, Erwin, xiv, xv, 147 Suicide and Attempted Suicide
Stoicism, 2, 9, 10, 100 (Stengel), xiv
stress suicide deaths
aversiveness of, 221 the socio-demographic pattern of,
effects of, 138-9 xvi
submission, 142, 143 Suicide Ideation Scale, 198
substance abuse, 37-9, 208 Suicide Intent Scale, 78
Sucitto, Ajahn, 224 suicide notes, 77
suicidal behaviour, 74 suicide prevention centres, 19 1-2
as an alarm system, 147 suicide rates, 18-44
‘appeal function' of, xiv, xv factors affecting, 28-36: ethnic
associated with poor problem group, 31; gender, 28-9;
solving ability, 97 religious affiliation, 32; sexual
and attachment, 119 orientation, 31; social class, 32-3;
biological approahes, 120-23 unemployment, 33-4
as a ‘cry for help’, xv, 139 female see female suicide rates
as a ‘cry o f pain’ first, xiv-xv, 139 international, 23, 24 , 25
o f depressed in-patients, 60 male see male suicide rates
and depression see depression and social factors, xvi
and entrapment, 149-50 U K statistics, 20, 21, 22 , 23, 25
genetic influences, 123-6 World Health Organisation and,
‘low intent’, 138 xiii
object-relations approach, 118-20 super-ego, 117
primary prevention see primary Sweden, 24 , 28, 34, 38, JU 177
prevention o f suicidal behaviour Swedish Prevention and Treatment
psychoanalytical perspectives, o f Depression Committee,
115-18 177
therapy see therapy for suicidal Switzerland, 13, 14, 24 , 71
feelings and behaviour symbiotic object-choice, 118
suicidal ideation, xv-xvi, 174-5, J82 synapses, 120-21
suicidal thoughts, 73-5
suicide Teasdale, John, 157
the aftermath of, 220-23 Telehelp/TeleCheck, effect on
and attempted suicide, 218-19 suicide rate o f keeping in touch
and ‘dangerous jobs’, 191 with old people, 176
defined, 18 testosterone, 141
and doctors, 174-5 therapy for suicidal feelings and
ethical arguments, 100-102 behaviour, 195-216
254
Index
assessing suicidal intent and the United States o f America, 13

probability o f repetition, alcohol dependency in, 47, 49
198-9 ethnic groupings, 31
Dialectical Behaviour Therapy and firearms, 25,30, 43, 189
(D BT), 206-8, 210 loss o f economic productivity
evaluation o f problem-solving through suicide, 20
approaches, 205-6 methods o f suicide, 25,30
follow-up o f D B T for suicidal rural areas, 34
behaviour, 210 -12 social class, 33
a framework for therapy, 197-8 suicide rates, 23, 24 , 26
the future o f secondary prevention unskilled workers, 32 , 55
research, 215-16 US Department o f Health and
problem-solving therapy, 200-205 Human Services, 31
psychological mechanisms Utopia (More), 7, 10
underlying therapeutic change, Utrecht, Holland, 72
213-15
some common problems, 199-200 van der Kolk, B. A., 84-5
vigilance for suicidal expression, Vesey, Patrick, 34
195-7 Viennese underground system, 128-9
Tinbergen, Nikolaas, 147 violence, 74, 122
tranquillizers, 70, 70, 178, 190 voluntary euthanasia, 99, 103, 104
Traskman-Bendz, Lil, 122 Voluntary Euthanasia Society see Exit
Treatment as usual (TAU ), 208
Truce, The (Levi), 223 Waco, Texas, 112
Wallington, Nehemiah, 6
unemployment, 33-4, 43-4, 55, 56, 74, Wesley, John, 15
85-6, 145, 182 Wetzel, R. D., 39
United Kingdom W H O /E U R O Multicentre Study o f
and the detoxification o f domestic Parasuicide In Europe, 71
gas, 25, 188 widowed people
and firearms, 25,30, 34, 189, 194 and parasuicide, 72, 74
one o f the highest rates o f World Health Organization
parasuicide in Europe, 71, 71 collates data on suicide rates, 19
prison suicides, 35 health care targets, xiii
rural areas, 34 World Medical Association, 99,
suicide rates, 19-20, 21, 23, 24 106
unemployment rates and suicide
rates, 33-4 year o f death
see also England and Wales; and distribution o f suicides^U^),
Northern Ireland 25,27
Index
young people sexual orientation, 31

rapid increase o f suicides (since suicide verdict less likely, 18-19
the mid 1970s), xiii UK suicide rates, 19, 25, 35
risk o f suicide, 36, 66 US suicide rates, 25
A CRY FOR HELP, OR A CRY OF PAIN?
Exploring the whole issue of suicide and attempted suicide, Mark

Williams’s thought-provoking book is essential reading for anyone -
whether parent, friend or health professional - who has been touched by
this trauma. He discusses the many factors - biological, psychological,
social and economic - that may drive a person to end their life or draw
attention to their plight. In the light of the latest research findings, he
reviews the changing patterns of suicidal behaviour, including the close
link between attempted suicide and child sexual abuse in women and the
rapid increase in suicide in rural areas. And, asking whether suicide is
ever justified, he considers the moral issues surrounding euthanasia.
Putting forward the theory that suicide is usually a way of expressing one’s
otherwise inexpressible pain, Mark Williams looks at practical ways to help
those at risk; greater vigilance, more sensitive medical assessment and pre
ventative therapies. Informative and clearly written, this wide-ranging book
wilt help anyone seeking a deeper understanding of the enigma r \
‘Mark Williams has articulated a language of u

fessional care givers and the lay public that
MBI
suicidal behaviours comprehensible ... a remarkable achi
Dr Lanny Berman, Executive Director, American Association of
to
i
‘A valuable resource for a very wide readership - in fact to
a greater understanding of suicidal behaviour’ Professor Kei
Director, Centre for Suicide Research, Oxford University
Previously published as Cry of Pain
(6)
PENGUIN
Get closer to Penguin at
Psychology/Health/Medicine
www.penguin.com Cover photo by Richard Hamilton / Millennium Images

Mark Williams - Suicide & Attempted Suicide - Understanding The Cry of Pain-Penguin UK (2002)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mark Williams - Suicide & Attempted Suicide - Understanding The Cry of Pain-Penguin UK (2002)

Uploaded by

Copyright:

Available Formats

SUICIDE AND

‘An important book that

Mark Williams was bom in 1952 and educated at Stockton-on-Tees

Published by the Penguin G ro u p

Penguin B ooks Ltd, R egistered O ffices: 80 Strand, Lon don W C 2R o r l , England

First published 1997

C o p yrigh t © J . M . G . W illiam s, 1997, 2001

T h e m oral righ t o f the au thor has been asserted

Set in 10 /12 .5 pt M on o typ e Dante

Except in the U nited States o f A m erica, this b o o k is sold subject

List o f Text Figures vi

1 A Brief History o f Suicide i

Notes and References 225

2.1 Number o f suicide deaths by gender, 1911-99, England and

and age, 1983-95, England and Wales. 22

I I .2 Dum m y data illustrating effect o f hypothetical intervention to

2.1 Suicide rates for a sample o f countries reporting to the W orld

A number o f people have helped with this book at various stages. I am

Pollock ('Factors Mediating Suicide B eh avio u r) that appeared in the

by the pain, but m ay nevertheless communicate distress in a w ay which

often does not lead to suicidal behaviour) m ay be dominated by the

A Brief History of Suicide

A psychiatrist was talking about suicide risk in her patients. ‘I can

Those at risk o f suicide come to us in our professional capacity to get help.

This represents a disagreement about whether suicide is preventable,

someone wearing a sandwich-board thrust a leaflet into m y hands. It

Attitudes in classical and medieval times

in the Bible. Christian thinkers and philosophers have always had to

The Tudor revolution in attitudes: government

During the early sixteenth century, the governm ent tightened up on a

As in contemporary accounts o f ‘reasons for living’ v. ‘reasons for dying’,

to be out o f the control o f the person, are not unknown in schizophrenic

The seventeenth century: gradually changing attitudes

During the seventeenth century an increasing range o f motives began

Shifts in opinion about suicide during the seventeenth century might

should rather be objects o f our greatest pity than condemnation as murderers,

w ay o f stigmatizing the dead for some offence committed previously.

The situation in other countries

Ambivalence within the Church

Although the Church had roundly condemned self-murder for most o f

In the light o f such confusion, clergymen looked for a compromise

The end of the crime of self-murder

The history o f attitudes to suicide could be interpreted as a journey

Suicide: Facts and Figures

problems is that a verdict o f suicide m ay be less likely when a young

The scope of the problem - UK statistics

In males within the United Kingdom, suicide is the second most

suicide o f a young person in the United States (aged 15-24 years)

1983 1985 1987 1989 1991 1993 1995

Figure 2.2 Recorded suicide deaths: rate per million population by

International suicide rates

Table 2.1 Suicide rates for a sample of countries reporting to the

Male Female Male and

Southern Greece (1996) 5-7 1.2 345

Western Austria (1997) 30.0 10.0 20.0

Scandinavia Denmark (1996) 24.3 9.8 17.05

Old World United Kingdom (1997) 11.0 3.2 7I

Asia Japan (1994) 17-3 7-2 12.2

New World Australia (1995) 19.0 51 12.05

There have been changes in age distribution in the United States

Figure 2.4 shows the distribution o f recorded suicides by sex, year o f

wmk .................... .............

uopEjndod u o i j j i u i j^d aiE'y

Researchers have come to the conclusion that suicide is closely associ

Despite these apparently strong relationships between psychiatric diag

Before Durkheim the connection between suicide and psychiatric dis