www.thelancet.com Vol 368 July 8, 2006 119
Suicide in recently released prisoners: a population-based
Daniel Pratt, Mary Piper, Louis Appleby, Roger Webb, Jenny Shaw
Background Several studies have been undertaken on suicide in custody, but few on suicide after the release from
prison. We undertook a population-based cohort study to investigate suicide rates in recently released prisoners in
England and Wales.
Methods We used the database of the National Confi dential Inquiry into Suicide and Homicide by People with Mental
Illness for England and Wales to identify all individuals who died by suicide or who received an open verdict at the
coroner’s inquest between 2000 and 2002. These records were linked to a Home Offi ce register to identify all such
deaths in people within 1 year of release from prison in England and Wales. We compared suicide rates per
100 000 person-years in these released prisoners with rates in the general population by using the indirectly age-
standardised mortality ratio.
Findings We identifi ed 382 suicides occurring in 244 988 individuals within 1 year of release from prison; a suicide
rate of 156 per 100
suicide rates were higher in recently released prisoners than in the general population. The overall age-standardised
mortality ratio for recently released prisoners was 8·3 (95% CI 7·5–9·3) for men and 35·8 (25·4–50·2) for women.
000 person-years. 79 (21%) suicides occurred within the fi rst 28 days after release. In all age groups,
Interpretation Recently released prisoners are at a much greater risk of suicide than the general population, especially
in the fi rst few weeks after release. The risk of suicide in recently released prisoners is approaching that seen in
discharged psychiatric patients. A shared responsibility lies with the prison, probation, health, and social services to
develop more collaborative practices in providing services for this high-risk group.
Several studies have examined the risk factors for suicide
in custody,1–5 but little is known about the risk of suicide
after release from prison.6–8 A study in Finland7 showed
that the suicide rate in people within 1 year of release
from prison was almost three times the rate in the general
population. A study in England and Wales8 showed that
male off enders under criminal justice community super-
vision were at least ten times more likely to die by suicide
than men in the general population. This study was a
national study, but included only prisoners released under
the supervision of probation, thus excluding more than
half of all released prisoners. Here, we report data for a
comprehensive 3-year national sample of self-infl icted
death within 12 months of release from prison in England
This study was approved by the multicentre research
ethics committee of the Thames Valley National Health
Service (NHS), Reading, UK. Information on suicides
was taken from the database of the National Confi dential
Inquiry into Suicide and Homicide by People with Mental
Illness,9 which has records of individuals in England and
Wales who receive a verdict of suicide or an open verdict
at coroner’s inquest. This information was obtained from
the Offi ce for National Statistics for England and Wales.
Most open verdicts are conventionally defi ned as suicide.10
Therefore, we have classifi ed suicide and open verdicts
as suicides in this Article. From the National Confi dential
Inquiry database, we identifi ed all suicides that took place
between Jan 1, 2000, and Dec 31, 2002.
Person-identifi able data were obtained from the
discharge extract fi les of the Inmate Information System
(IIS), a Home Offi ce database containing details of all
sentenced and remanded prisoners released from prison
in England and Wales. Using names and dates of birth,
we linked the National Confi dential Inquiry list of
suicides with the IIS Discharge Extract fi le of prisoners
whose most recent release from prison was between
Jan 1, 1999, and Dec 31, 2002. Cases were selected from
the linked records if the person’s date of death occurred
within 12 months of their date of release from prison.
Study individuals were then cross-referenced against a
database recording information on all self-infl icted
deaths in prisons in England and Wales.5 These deaths
had been classifi ed by prison service staff as suicides and
recorded on a prison database before inquest. We found
that 13 self-infl icted deaths previously identifi ed as cases
actually took place on the date of release, but while the
prisoner was still in prison. These 13 deaths were
excluded from this study.
We calculated suicide in the general population by using
the estimated aggregate national population base for
every year in 2000–02 combined as the denominator.11
Lancet 2006; 368: 119–23
See Editorial page 92
Centre for Suicide Prevention,
University of Manchester,
Manchester M13 9PL, UK
(D Pratt MSc,
Prof L Appleby FRCPsych,
R Webb MSc,
Prof J Shaw MRCPsych); and
Department of Health, London,
UK (M Piper FRCP)
Dr Jenny Shaw
www.thelancet.com Vol 368 July 8, 2006
Suicide for recently released prisoners was also calculated
by use of the aggregate yearly population of released
prisoners for 2000–02, with the person-years indicating
1-year post release information. These denominators were
used as proxies for the person-years at risk in the study
and general populations. Standardised mortality ratios
were defi ned as the ratio of the observed versus the
expected number of suicides, with adjustment for
diff erences in age distribution between the released
prisoners and general populations.12,13 This indirect
method of standardisation is robust if numbers of events
per stratum are sparse in the study population. We used
Poisson methods to calculate 95% CIs. Analyses were
undertaken with STATA version 8.0.
Role of the funding source
The sponsor of the study had no role in study design, data
collection, data analysis, data interpretation, or writing of
the report. The corresponding author had full access to all
the data in the study and had fi nal responsibility for the
decision to submit for publication.
In 244 988 released prisoners, we identifi ed 382 suicides
(34 women, 348 men) within 1 year of release. Suicides by
recently released prisoners accounted for 3% of the
14 097 suicides occurring between 2000 and 2002, while
released prisoners accounted for only 0·2% of the
population (table). Figure 1 shows the distribution of
these suicides, separated into 28-day periods over the
year after release. 79 (21%) suicides took place in the fi rst
28 days after release, and 195 (51%) within the fi rst
Figure 1: Frequency of suicides during 1 year of release from prison
364336308 280252224196 1681401128456 28
Time since release from prison (days)
Number of suicides
General population* Recently released prisoner population*Rate ratio (95% CI)
Rate (per 100 000
Rate (per 100 000
All women ≥18 years
All men ≥18 years
≥ 18 years
2 826 200
3 762 700
5 228 600
12 331 200
10 587 700
28 346 600
63 083 000
2 902 100
3 762 900
5 211 900
12 175 500
10 438 100
24 069 500
58 560 000
4·3 (2·9–6·1)‡ .... ....
121 643 000 14 097 11·6244 988 382155·9 13·5 (12·2–14·9)
*Data recorded for populations in England and Wales, 2000–02. †Standardised mortality ratio used. ‡Overall ratio compares the standardised mortality ratio for released
women with that for released men.
Table: Suicide rates and standardised mortality ratios by age and sex
www.thelancet.com Vol 368 July 8, 2006 121
The overall rate of suicide was 156 per 100 000 person-
years (table). The rate in recently released male prisoners
was 154 deaths per 100 000 person-years, which increased
with age, with the highest rates in those aged 40–49 years
(275 per 100 000 per year) and 50 years or older (234 per
100 000 per year; table). The overall rate of suicide in
recently released female prisoners was 180 per
100 000 person-years, although age-specifi c trends were
impossible to discern because of sparse data. In all age
and sex groups, suicide was substantially higher in
recently released prisoners than in the general
population (table). We directly compared the suicide
rate between male and female prisoners who were
recently released, using a Poisson regression model
adjusted for age. Suicide between recently released men
and women did not diff er signifi cantly (women vs men:
age-adjusted rate ratio 1·14 [95% CI 0·80–1·62], χ² test,
p=0·469), which contrasts with the three-fold increase
in risk recorded for men compared with women in the
According to standardised mortality ratios for recently
released prisoners, men were eight times and women
36 times more likely to die by suicide within 1 year of
release from prison than would be expected in their
respective sex groups in the general population. The
standardised mortality ratio for women was about four
times greater than that for men. This fi nding indicates
that a higher excess suicide risk was seen in female
released prisoners than in women from the general
population. We saw evidence of a highly signifi cant linear
trend in the age-specifi c rate ratios in recently released
men (χ² test, df=1, p<0·0001), with a rate ratio of about
5 for men younger than 25 years rising to more than 15 in
men aged 50 years and older (fi gure 2). Figure 3 shows the
age-specifi c rate ratios for suicide risk in recently released
This study has shown that, during the fi rst 12 months
after release, prisoners in our analysis were at a much
greater risk of suicide than the general population. Risk
was particularly increased during the fi rst 28 days, during
which about a fi fth of all suicides occurred. The risk was
similar in men and women but, compared with the
general population, the relative risk in women was
substantially higher than in men. This result seems to
support a previous fi nding that the rate of suicide for
female off enders is similar to that for men,3 although
female rates were based on small numbers of cases.
Few population-based studies have reported on suicides
in released prisoners. Our main fi nding that off enders
are at a raised risk of suicide after release from prison
supports previous research.5,6 Furthermore, the high rate
ratios of suicides in released prisoners compared with the
general population accords with previous studies
reporting an increase of 3–10 fold in the risk of suicide.7,8
The Inmate Information System used by this study is a
database maintained by the Home Offi ce and has been
used by other research studies5,14 as well as to produce
yearly prison statistics.15 Similarly, the database of suicides
for England and Wales, to which we had access via the
National Confi dential Inquiry, has been extensively used
to describe suicide in the general population and in
people with mental illness.9 However, we have linked the
two databases in our study.
Our sample size of suicides in male prisoners is adequate
for drawing reliable conclusions, even when stratifi ed by
age. For suicides in female prisoners, a sample size of 34 was
suffi cient to calculate the standardised mortality ratio, but
comparisons between diff erent age groups were not
reliable because of small numbers per stratum.
Although the standardised mortality ratios adjust for
diff erences in age distributions between the study and
18–20 21–24 25–2930–39 40–49 ≥50
Figure 2: Age-specifi c rate ratios for suicide risk in recently released male
prisoners compared with general population
Dots and error bars represent point estimates of the rate ratios and 95% CIs,
Figure 3: Age-specifi c rate ratios for suicide risk in recently released female
prisoners compared with general population
Dots and error bars represent point estimates of the rate ratios and 95% CIs,
respectively. Age categories have been collapsed for the fi gure.
www.thelancet.com Vol 368 July 8, 2006
general populations, prisoners have proportionally more
risk factors for suicide, which include increased prevalence
of mental illness, substance misuse, and socioeconomic
deprivation.16–20 Consequently, prisoners are already at a
high risk of suicide and it is not surprising to fi nd similar
standardised mortality ratios for suicide in released and
current prisoners. A study of suicide in male prisoners
reported a standard mortality ratio of 5·1 (95% CI
4·8–5·3).21 However, this fi nding does not explain the
increased occurrence of suicide in the fi rst 28 days after
A few of the deaths might have been drug-related. Drug
misusers returning to their previous level of drug use after
release from prison could have fatal consequences because
of a loss of tolerance.14 This event is unlikely to account for
any of our recorded number of deaths in this study, since
unintentional drug-related deaths are usually recorded as
The immediate postdischarge period is also a time of
increased suicide risk in psychiatric patients after
discharge from hospital, with a risk of suicide 50 times
that expected in the general population.23–25 The early
stages after release are daunting and prisoners often
face exclusion by the communities to which they are
returning, as well as mutually re-enforcing barriers.26,27
For example, in 2002, about a third of prisoners were
reported to have had no accommodation, almost half
had lost contact with their families since entering
prison, and three-quarters did not have paid employment
to go to on release.28 In the UK, prisoners sentenced to
more than 14 days are eligible for a GBP£46 discharge
grant, and an additional payment of up to £50 can be
made to a legitimate housing provider before the
prisoner’s release, to secure accommodation, although
these individuals often have to wait for several weeks
before receiving any further benefi ts payments.28
Remanded prisoners do not receive any discharge grant
despite the length of time in custody.
In addition to the social problems facing prisoners on
release, many also have to cope with issues associated
with mental-health problems. High psychiatric morbidity
is well-known in prisoners, especially for women,16–20 and
studies2,4 have shown mental illness to be a strong risk
factor for prison suicide. Therefore, the rate of suicide in
released prisoners would be expected to be higher than
that in the general population. Furthermore, the raised
prevalence of mental illness and personality disorder in
female prisoners might have been indicated in the
increased risk of postrelease suicide.29 Therefore, another
possible explanation for high suicide risk in recently
released prisoners could be the lack of adequate care and
treatment given to prisoners for their suicidal ideation
and associated mental-health problems, after release from
We recorded that older men were at a greater risk in
this study, which is unexpected because aggressive
impulsive traits, an important risk factor for suicide, can
become less severe in off enders of older age. However,
the challenges of reintegration into society after release
from prison could aff ect older off enders more acutely
than their younger counterparts. Also, suicides might
have been more likely to take place in older individuals if
mental illness was more prevalent in older off enders
than in the general population.1 However, previous
fi ndings are contradictory and inconclusive, with older
off enders over-represented in prison suicides in some
studies, and under-represented in others.1,5 The problems
of small sample sizes and the few comparable control
groups make any interpretation diffi cult.
Our fi ndings highlight the need for shared responsibility
of the prison, probation, health, and community services
including social services, housing, and benefi ts to provide
good quality, integrated, mental-health care and social
support for prisoners, both before and after release from
prison. The eff ective resettlement or release of people
from prison is thus vitally important. It should be a
multidisciplinary function addressing the full range of
resettlement needs, and recognising the multiple barriers
that ex-prisoners face. Such needs include: the
maintenance or rebuilding of family ties; referral and
access to community-health care and treatment pro-
grammes, where appropriate; identifi cation of housing
needs and assistance to access accommodation, including
supported accommodation; and training or employment
oppor tunities. Individual care plans should be developed
and monitored by multidisciplinary teams consisting of
those professionals caring for the individual from both
before and after release from prison. The team would be
responsible for a structured assessment of prisoners’ risk
of suicide on their release and for ensuring that those
prisoners identifi ed as being at risk of suicide are provided
with regular and routine contact with a community
mental-health professional from the moment the
individual leaves prison and during the fi rst few months
after release, depending on the needs of the individual.
Confl ict of interest statement
L Appleby is National Director for Mental Health in England and
overseas England’s National Suicide Prevention Strategy. M Piper is
Senior Public Health Adviser of Prison Health, UK Department of
Health. The other authors declare that they have no confl ict of interest.
Daniel Pratt, Jenny Shaw, and Louis Appleby designed the study.
Daniel Pratt obtained the data and, with assistance from Roger Webb, did
the statistical analysis, supervised by Jenny Shaw. Daniel Pratt prepared
the paper and all investigators contributed to the fi nal manuscript.
The study was funded by the National Forensic Mental Health Research
and Development Programme in the Department of Health, UK. We
thank Nigel Hancock (Safer Custody Group, HM Prison Service) for his
assistance in the collection of data.
1 Dooley E. Prison suicide in England and Wales 1972–87.
Br J Psychiatry 1990; 156: 40–45.
2 Fruehwald S, Frottier P, Matschnig T, König F, Bauer P. Suicide in
custody: a case-control study. Br J Psychiatry 2004; 185: 494–98.
3 Towl G, Snow L, McHugh M, eds. Suicide in prisons. Oxford: BPS
Articles Download full-text
www.thelancet.com Vol 368 July 8, 2006 123
4 Shaw J, Baker D, Hunt IM, Moloney A, Appleby L. Suicide by
prisoners: a national clinical survey. Br J Psychiatry 2004; 184:
Shaw J, Appleby L, Baker D. Safer prisons: a national study of
prison suicides 1999–2000 by the National Confi dential Inquiry into
Suicide and Homicide by People with Mental Illness. London:
Department of Health, 2003.
Harding-Pink D. Mortality following release from prison.
Med Sci Law 1990; 30: 12–16.
Joukamaa M. The mortality of released Finnish prisoners: a 7 year
follow-up study of the WATTU project. Forensic Sci Int 1998; 96:
Sattar G. The death of off enders in England and Wales. Crisis 2003;
Appleby L, Shaw J, Sherratt J, et al. Safety fi rst: a 5-year report of the
National Confi dential Inquiry into Suicide and Homicide by People
with Mental Illness. London: Department of Health, 2001.
10 Neeleman J, Wessely S. Changes in classifi cation of suicide in
England and Wales: time trends and association with coroners’
professional background. Psychol Med 1997; 27: 467–72.
11 Offi ce for National Statistics. Mortality Statistics—cause: series
DH2 no 30. London: Offi ce for National Statistics, 2004.
12 Rothman KJ. Modern epidemiology. Boston: Little Brown, 1986.
13 Breslow N, Day N. Statistical methods in cancer research. Vol II: the
design and analysis of cohort studies. Oxford: Oxford University
14 Singleton N, Pendry E, Taylor C, Farrell M, Marsden J. Drug related
mortality among newly-released off enders. London: Offi ce for
National Statistics, 2002.
15 Home Offi ce. Prison Statistics: England and Wales 2002. London:
Home Offi ce, 2003.
16 Fazel S, Danesh J. Serious mental disorder in 23 000 prisoners: a
systematic review of 62 surveys. Lancet 2002; 359: 545–50.
17 Joukamaa M. Psychiatric morbidity among Finnish prisoners with
special reference to socio-demographic factors: results of the Health
Survey of Finnish Prisoners (Wattu Project). Forensic Sci Int 1995;
18 Brooke D, Taylor C, Gunn J, Maden A. Point prevalence of mental
disorder in unconvicted male prisoners in England and Wales. BMJ
1996; 313: 1524–27.
19 Bland RC, Newman SC, Thompson AH, Dyck RJ. Psychiatric
disorders in the population and in prisoners. Int J Law Psychiatry
1998; 21: 273–79.
20 Singleton N, Meltzer H, Gatword R, Coid J, Deasy D. Psychiatric
morbidity among prisoners in England and Wales. London:
Home Offi ce, 1998.
21 Fazel S, Benning R, Danesh J. Suicides in male prisoners in
England and Wales, 1978–2003. Lancet 2005; 366: 1301–02.
22 Christophersen O, Rooney C, Kelly S. Drug-related mortality:
methods and trends. Popul Trends 1998; 93: 29–37.
23 Goldacre M, Seagroatt V, Hawton K. Suicide after discharge from
psychiatric in-patient care. Lancet 1993; 342: 283–86.
24 King EA, Baldwin DS, Sinclair JM, Baker NG, Campbell MJ,
Thompson C. The Wessex recent in-patient suicide study I: case-
control study of 234 recently discharged psychiatric patient suicides.
Br J Psychiatry 2001; 178: 531–36.
25 Sohlman B, Lehtinen V. Mortality among discharged psychiatric
patients in Finland. Acta Psychiatr Scand 1999; 99: 102–09.
26 Visher CA, Travis J. Transitions from prison to community:
understanding individual pathways. Ann Rev Socio 2003; 29: 89–113.
soc.29.010202.0959 (accessed June 21, 2006).
27 Fox A, Khan L, Briggs D, Rees-Jones N, Thompson Z, Owens J.
Home Offi ce on-line report 01/05—throughcare and aftercare:
approaches and promising practice in service delivery for clients
released from prison or leaving residential rehabilitation. London:
Home Offi ce, 2005.
28 Social Exclusion Unit. Reducing re-off ending by ex-prisoners.
London: Social Exclusion Unit, 2002.
29 Baxter D, Appleby L. A case register study of suicide risk in mental
disorders. Br J Psychiatry 1999; 175: 322–26.