Psychological Medicine (2015), 45, 3441–3451.

© Cambridge University Press 2015 OR I G I N A L A R T I C L E

doi:10.1017/S0033291715001348

Cognitive–behavioural suicide prevention for male

prisoners: a pilot randomized controlled trial

D. Pratt1,2*, N. Tarrier3, G. Dunn4, Y. Awenat1, J. Shaw5, F. Ulph1 and P. Gooding1

1
School of Psychological Sciences, University of Manchester, Manchester, UK
2
Manchester Mental Health and Social Care NHS Trust, Manchester, UK
3
Department of Psychology, Institute of Psychiatry, King’s College London, London, UK
4
Institute of Population Health, University of Manchester, Manchester, UK
5
Institute of Brain, Behaviour and Mental Health, University of Manchester, Manchester, UK

Background. Prisoners have an exceptional risk of suicide. Cognitive–behavioural therapy for suicidal behaviour has
been shown to offer considerable potential, but has yet to be formally evaluated within prisons. This study investigated
the feasibility of delivering and evaluating a novel, manualized cognitive–behavioural suicide prevention (CBSP) therapy
for suicidal male prisoners.

Method. A pilot randomized controlled trial of CBSP in addition to treatment as usual (CBSP; n = 31) compared with
treatment as usual (TAU; n = 31) alone was conducted in a male prison in England. The primary outcome was self-injuri-
ous behaviour occurring within the past 6 months. Secondary outcomes were dimensions of suicidal ideation, psychiatric
symptomatology, personality dysfunction and psychological determinants of suicide, including depression and hope-
lessness. The trial was prospectively registered (number ISRCTN59909209).

Results. Relative to TAU, participants receiving CBSP therapy achieved a significantly greater reduction in suicidal
behaviours with a moderate treatment effect [Cohen’s d = −0.72, 95% confidence interval −1.71 to 0.09; baseline mean
TAU: 1.39 (S.D. = 3.28) v. CBSP: 1.06 (S.D. = 2.10), 6 months mean TAU: 1.48 (S.D. = 3.23) v. CBSP: 0.58 (S.D. = 1.52)].
Significant improvements were achieved on measures of psychiatric symptomatology and personality dysfunction.
Improvements on psychological determinants of suicide were non-significant. More than half of the participants in
the CBSP group achieved a clinically significant recovery by the end of therapy, compared with a quarter of the TAU
group.

Conclusions. The delivery and evaluation of CBSP therapy within a prison is feasible. CBSP therapy offers significant
promise in the prevention of prison suicide and an adequately powered randomized controlled trial is warranted.

Received 30 January 2015; Revised 15 May 2015; Accepted 18 June 2015; First published online 13 July 2015

Key words: Cognitive therapy, prison, randomized controlled trials, suicide prevention.

Introduction to be prioritized within national suicide prevention

strategies (Department of |Health, 2002, 2012).
The risk of suicide is particularly high amongst prison-
A meta-analysis of cognitive–behaviour therapies
ers. Male prisoners are five times more likely to die by
(CBT) for suicidal behaviour reported that such inter-
suicide than the general population (Fazel et al. 2005,
ventions were effective, when designed, tailored and
2011). Coping with an environment that engenders
implemented to focus on suicidality (Tarrier et al.
fear, distrust, and a lack of control, can leave indi-
2008) and CBT is now a recommended treatment for
viduals feeling overwhelmed and hopeless, lead-
suicidal behaviour (National Institute for Health and
ing some of them to choose suicide as a way to
Clinical Excellence, 2011). However, the potential
escape (Birmingham, 2003; Liebling & Maruna, 2005).
offered by CBT for suicidal patients located within
Suicide in prison is of considerable public and social
prison settings is unknown. Structured and systematic
concern (e.g. Bowcott et al. 2014) and prisoners continue
approaches to offender behaviour programmes have
already been established as effective in reducing
other types of prisoner behaviour (Gendreau, 1996;
McGuire, 2002; Landenberger & Lipsey, 2005); there-
* Address for correspondence: D. Pratt, Ph.D., School of
Psychological Sciences, University of Manchester, Oxford Road,
fore, it is important to investigate the possible benefits
Manchester M13 9PL, UK. of a CBT-informed structured intervention programme
(Email: [email protected]) specifically targeting suicidal ideation and behaviour.
3442 D. Pratt et al.

International and national policies emphasize im- male prisoners. The PROSPeR study was registered
prisonment as an important opportunity to enhance as an International Standard Randomized Controlled
access to interventions aiming to reduce the risk of sui- Trial, ISRCTN59909209, and received approval from
cidal behaviour (Department of Health, 2007; Konrad the National Research Ethics Committee for Wales
et al. 2007). However, whilst prison settings may pre- (reference 11/WA/0002), the National Offender
sent an opportunity to engage with a ‘hard-to-reach’ Management Service’s National Research Committee
sector of society, a number of potential barriers at the (reference 16–11) and the Governor of the host prison.
contextual level and issues presented by the individual Inclusion criteria were male prisoners aged over
prisoners themselves have to be identified and over- 18 years; who had been identified within HM Prison
come to allow the acceptable and feasible delivery of Service’s Assessment, Care in Custody and
any preventative interventions. Teamwork (ACCT) (Ministry of Justice, 2013) system
The main aim of the Prevention Of Suicide in Prisons as being at risk of suicidal behaviour within the past
(PROSPeR) study, therefore, was to examine the feasibil- month. Eligible prisoners were excluded from the
ity of delivering and evaluating cognitive–behavioural study if they had insufficient knowledge of English
suicide prevention (CBSP) therapy for individuals iden- to enable adequate participation in the assessment pro-
tified as presenting a risk to themselves whilst in prison. cess; were deemed by prison staff to be too dangerous;
CBSP was selected as the psychosocial intervention for or unable to provide consent. Current or previous par-
this study as it is a suicide prevention intervention that ticipation in offending behaviour programmes was not
has been intentionally derived from an empirically vali- an exclusion criterion, as this was considered to be
dated theoretical model of suicide (Johnson et al. 2008; usual treatment for prisoners.
Tarrier et al. 2013), and has recently been shown to All potential participants were identified by the Safer
reduce measures of suicidality in community-dwelling Custody team within the host prison who held the re-
participants with a schizophrenia diagnosis (Tarrier sponsibility for administrating the ACCT system.
et al. 2014). Individuals identified under the ACCT system were
A further aim of the PROSPeR study was to develop provided with information about the PROSPeR study.
preliminary estimates of the impact of CBSP therapy After agreement to be contacted, prisoners expressing
over the usual care and support offered to suicidal an interest in the study were then invited to an initial re-
prisoners. We examined three specific hypotheses. search interview to confirm eligibility. This process of
First, those who received the CBSP therapy pro- identifying potential participants was conducted inde-
gramme would demonstrate significantly greater pendently of the research team. Those participants
reductions in the occurrence of suicidal behaviours meeting entry criteria were asked to provide written
compared with those receiving usual care and support. informed consent to take part in the study and to
Second, the CBSP group would achieve significantly agree to be subject to a ‘holding order’ to remain within
improved scores on psychological measures of suicid- the host prison for the duration of their participation in
ality, including suicidal ideation and hopelessness, the trial. Subsequent assessments were completed with
compared with the usual care group. Third, significant a research assistant, independently of trial therapists, at
improvements would be experienced by the CBSP 4 months (post-treatment) and 6 months (follow-up)
group, relative to the usual care group, on measures after the baseline assessment.
correlated with suicidal ideation, such as depression,
anxiety, self-esteem and psychiatric symptomatology. Interventions
Treatment as usual (TAU)

Method Participants randomized to the TAU group received

the usual care and support available to any prisoner
Study design and participants
identified under the ACCT system had they not parti-
In keeping with the guidance for the evaluation of cipated in the trial. The Prison Service Instruction de-
complex interventions (Medical Research Council, scribing the management of prisoners at risk of harm
2008), the PROSPeR study was a single-blind (rater) to self (Ministry of Justice, 2013) clearly prescribed
randomized controlled pilot trial. Recruitment into that all prisoners identified under the ACCT system
the trial was conducted between 4 January 2012 and received an assessment of risk when a risk to self
14 June 2013 at one UK site in the Northwest of was first identified, which then informed a risk man-
England with follow-up assessments taking place be- agement plan of how to keep the individual safe (e.g.
tween 2 July 2012 and 14 December 2013. The study levels of monitoring and observation). Subsequently,
sample was recruited from a closed prison establish- at least fortnightly, review meetings were arranged
ment with capacity to house approximately 1200 by prison officer staff until the risk was considered to
 Cognitive–behavioural suicide prevention for male prisoners 3443

be lowered, at which time the individual was dis- case supervision was provided throughout the trial
charged from the ACCT system. on a weekly basis, by an experienced clinical psycholo-
Within the risk management plan, a referral could be gist who was independent of the research team.
made to the prison’s Mental Health In-Reach team that
offered psychosocial assessment, ongoing monitoring Measures
of psychiatric symptoms, medication therapies and
nursing support. All participants within the current Adherence measures
trial were referred to the In-Reach Team for a psycho- To assess participants’ adherence to the CBSP interven-
social assessment. At the time of the study, TAU did tion, the trial therapists were asked to rate the partici-
not include access to a psychological therapy. We did pant’s active involvement in the therapy programme
not register use of psychiatric medications, but previous on a bespoke rating form based upon existing assess-
studies have reported that at least a third of suicidal pris- ment tools from the offending behaviour programmes
oners are routinely prescribed antidepressant medica- (Hollin & Palmer, 2006). For each participant, therapist
tion (Humber et al. 2010; Rivlin et al. 2010). (s) assessed attendance, promptness, level of participa-
tion, mastery of programme content, disruptive behav-
Cognitive–behavioural suicide prevention (CBSP) therapy iour, completion of homework tasks, and an overall
evaluation of therapy success. Each of these items
In addition to TAU, participants randomly allocated to was rated on a Likert scale (1 = poor to 5 = excellent).
the CBSP group also received a cognitive therapy- Additionally, participant attendance was recorded to
informed intervention. The CBSP therapy is a struc- provide an indicator of engagement in the treatment.
tured, time-limited psychosocial intervention devel- Similarly, reasons for non-attendance were monitored.
oped to treat individuals experiencing suicidal Finally, the therapist(s) maintained a record for each
ideation and/or behaviour (Tarrier et al. 2013). CBSP participant of the content of each session, with specific
is informed by the Schematic Appraisals Model of reference to the modules within the treatment protocol.
Suicide (SAMS; Johnson et al. 2008), which identifies
(i) information-processing biases, (ii) appraisals and
Outcome measures
(iii) a suicide schema to be the main components con-
tributing to an individual’s experience of suicidality. Completed suicide occurs too infrequently to be a use-
CBSP draws from established clinical techniques to re- ful outcome measure; however, suicidal behaviours,
structure the three aspects of the SAMS model, includ- thoughts and feelings are common, distressing and,
ing the use of cognitive techniques to encourage therefore, legitimate outcome measures. This is stand-
participants to evaluate some of their appraisals about ard practice as used in previous trials (Tarrier et al.
themselves, their situation and their future, as well as 2008; Tarrier et al. 2014). In accordance with the trial
the use of behavioural techniques to identify and re- protocol, the primary outcome measure was the num-
hearse more helpful responses to distressing situations. ber of episodes of suicidal or self-injurious behaviour
The intervention for the current study was manualized (SIB) in the past 6 months assessed by examination
and developed from our previously published treat- of participants’ prison records. Secondary outcome
ment manual (Tarrier et al. 2013). Briefly, the interven- measures included scores on the Beck Scale for
tion was modularized into five components: Suicidal Ideation (BSSI; Beck & Steer, 1991) to assess
suicidal ideation, and the Suicide Probability Scale
Attention broadening
(SPS; Cull & Gill, 1982) to provide an overall estimate
Cognitive restructuring
of suicidal potential. Both measures have established
Mood management and behavioural activation
reliability and predictive validity within prisoner
Problem-solving training
populations (Senior et al. 2007; Naud & Daigle, 2010;
Improving self-esteem and positive schema.
Perry et al. 2010).
Delivery of the CBSP therapy programme consisted Further measures were completed to assess key psy-
of up to 20 sessions, delivered twice weekly during chological and psychiatric variables relevant to suicide.
the initial phases of therapy, reducing to once-weekly Perceptions of pessimism and hopelessness were mea-
sessions when therapeutic engagement had been estab- sured using the Beck Hopelessness Scale (BHS; Beck &
lished. Each session typically lasted for up to 1 h. Steer, 1988), levels of depression and anxiety were
Treatment sessions were provided by two trial thera- assessed using the Beck Depression Inventory (BDI-II;
pists, of whom both were clinical psychologists (doc- Beck et al. 1996) and the Beck Anxiety Inventory (Beck
toral level) with 2–5 years’ experience of CBT. Both et al. 1988), respectively, and self-esteem was measured
therapists received initial training to familiarize them using the Robson Self-Concept Questionnaire (Robson,
to the specifics of the CBSP programme. Ongoing 1989). The 24-item version of the Brief Psychiatric
3444 D. Pratt et al.

Rating Scale (BPRS; Ventura et al. 1993) was adminis- independent-groups t test with a two-sided significance
tered to assess the presence and severity of psychiatric level of 0.15. Since this was a preliminary trial, we were
symptoms, and the Standardised Assessment of prepared to accept a higher type 1 error rate in order to
Personality – Abbreviated Scale (SAPAS; Moran et al. avoid missing promising effects (Stone et al. 2007).
2003) was used to briefly measure the level of personal- All analyses were conducted using Stata version 11
ity dysfunction/disorder. Other assessments were also (StataCorp, 2009). Estimation of treatment effects was
administered to investigate potential predictors of based on the intention-to-treat principle. Random-
outcome within further secondary analyses, which are effects (i.e. random-intercepts) models for repeated-
not reported within this paper. Additionally, a range measures data were fitted to both 4- and 6-month
of demographic, clinical and criminological details outcome variables, with the baseline value of the out-
were collected from participant interviews, clinical and come variable being used as a covariate. Stata’s xtreg
prison records, and the host prison’s management command was used. After preliminary examination of
information system, subject to participant consent. the summary statistics for the outcome variables, treat-
ment effects (differences in outcomes between the two
Clinically significant recovery arms of the trial) were assumed to be the same for
both follow-up times. Fitting the appropriate random-
Clinically significant change was calculated on the sec- or mixed-effects model provides an estimate of this
ondary outcome measures of suicidal ideation and sui- common treatment effect. Missing outcome data were
cide probability. Using standardized procedures assumed to be missing at random, using the termin-
(Jacobson et al. 1999), clinical significance was indi- ology of Little & Rubin (2002), i.e. conditional on the
cated by an improvement in scores from the clinical data used in the model, whether an observation is miss-
to the non-clinical range for the measures. ing or not does not depend on its actual value.
Since the primary outcome (SIB) was positively
Random assignment and masking skewed, confidence intervals (CIs) for the standard
Immediately following completion of baseline assess- errors, and confidence for the treatment effects were
ments, participants were randomly allocated to one estimated by applying a bootstrap procedure (Efron
of the two treatment groups: TAU or CBSP. & Tibshirani, 1993) using the percentiles based on the
Randomization of participants to the two treatment results of 1000 replications (using the trial participant
groups was achieved by referring to a sequence of as the sampling unit).
sealed envelopes provided by the research statistician
(G.D.). Treatment allocated was based on pseudo-
random number generation, and based on a randomly Results
permuted blocks algorithm (with block sizes randomly Recruitment and retention
varying between 4 and 8).
The randomization schedules were generated and During the 2-year period of recruitment, 267 prisoners
provided to the study by the research statistician, be- were assessed for suitability for the PROSPeR trial (see
fore being kept securely and confidentially by the Fig. 1). Of the 205 who were excluded from the trial, 56
trial administrator who contacted the trial therapists, (27%) failed to meet entry criteria (i.e. deemed by
as appropriate, with the participant’s details for those prison staff as too dangerous or too unwell to partici-
allocated to the CBSP group. Thus, randomization pate), and 131 (64%) declined to participate, with 79
was independent and the research assistants complet- (39%) expressing a lack of interest in the trial and 39
ing the assessments were blind to group allocation. A (19%) refusing to be placed on a ‘holding order’ to re-
number of strategies were developed to achieve and main in the host prison for 6 months. Of the prisoners,
maintain the masking of assessors, such as removing 15 (7%) were prevented from participating due to legal
any research assistant involvement in the random as- reasons pertaining to immigration orders, and
signment process, research assistant and trial therapist three (1%) were unexpectedly transferred out of the
to avoid simultaneous use of allocated interview/ther- prison whilst undertaking the baseline assessment.
apy rooms to preserve blindness to allocation, and par- Recruitment into the study was successful, with a
ticipants were encouraged at each assessment not to final sample size slightly larger than the original re-
refer to treatment group allocation. cruitment target, with 62 participants randomized to
the CBSP plus TAU group (n = 31) or the TAU-alone
group (n = 31). The follow-up rates for the study sam-
Statistical analysis
ple as a whole was 40 out of 62 (65%) at the 4-month
A sample size of 30 per group gave the trial 80% assessment and 35 (56%) at the 6-month assessment.
power to detect an effect size of 0.60 using an Of the participants, five (8%) withdrew from the
 Cognitive–behavioural suicide prevention for male prisoners 3445

Fig. 1. CONSORT (Consolidated Standards of Reporting Trials) flow diagram of participant progress through the Prevention
Of Suicide in Prisons (PROSPeR) trial. CBSP, Cognitive–behavioural suicide prevention; TAU, treatment as usual.

study and we were unable to follow-up 22 (35%) partici- single. According to participants’ self-reports, the
pants who had been unexpectedly released early or mean age of their first custodial sentence was 25.7
transferred to other prisons for security reasons during (S.D. = 11.91) years with an average of 5.4 (S.D. = 8.95)
the course of the trial. Participants that were lost to previous imprisonments. Of the participants, 34
follow-up did not differ significantly from participants (55%) were currently serving a prison sentence, 25
that completed the 4-month or 6-month assessment on (40%) had been remanded into prison custody and
any of the sociodemographic or custodial characteristics. the custodial status of three (5%) participants was not
known.
To meet entry criteria to the study, all participants
Demographic and clinical characteristics
had been managed under the ACCT process during
Baseline characteristics for each group are presented in the month prior to entry to the study. Of these, 44
Table 1. For the overall sample, participant ages ran- (71%) were under ACCT at the start of their participa-
ged from 21 to 60 years with a mean of 35.2 (S.D. = tion in the study, 11 (18%) had been under ACCT less
11.10) years. Of the participants, 53 (85%) described than 2 weeks prior to starting the study, and the
themselves as white (UK), four (6%) as black (UK), remaining seven (11%) up to a month prior. There
three (5%) as white (non-UK), and three (5%) as was a substantial proportion of previous suicide
other/not stated. Of the participants, 35 (57%) were attempts within the sample, with only nine (15%)
3446 D. Pratt et al.

Table 1. Sociodemographic and custodial characteristics Table 2. CBSP individual treatment modules: number and
percentage of participants receiving the module
CBSP plus TAU TAU alone
(n = 31) (n = 31) Mean no. of
sessions in which No. of participants
Sociodemographic variables module was used receiving module
Age, years 38.5 (11.3) 32.0 (10.1) CBSP module (S.D.) (%)
Ethnicity: white, 26 (84) 27 (87)
British, n (%) Attention 6.1 (2.6) 8 (26)
Marital status: single, 19 (61) 16 (52) broadening
n (%) Cognitive 4.9 (2.3) 13 (42)
Custodial variables restructuring
Age first imprisoned, 25.5 (12.7) 25.9 (11.3) Mood 4.0 (2.7) 3 (10)
years management
Number of previous 7.5 (11.3) 3.4 (5.1) Problem-solving 3.0 (1.2) 7 (23)
imprisonments training
Custodial status, n (%) Improving 4.4 (2.2) 5 (16)
Sentence 20 (65) 14 (45) self-esteem
Remand 9 (29) 16 (52)
Other/not known 2 (6) 1 (3) CBSP, Cognitive–behavioural suicide prevention; S.D.,
standard deviation.
Data are given as mean (standard deviation) unless other-
wise indicated.
CBSP, Cognitive–behavioural suicide prevention; TAU,
immediately after randomization and prior to the
treatment as usual.
first treatment session (see Fig. 1).

participants self-reporting no lifetime history of a sui- Adherence to treatment protocol

cide attempt, whereas 18 (29%) participants had
made a single previous attempt and 35 (57%) had pre- Table 2 shows the list of treatment modules and the
viously attempted suicide on two or more occasions. frequency of use of each module across the CBSP par-
On both measures of suicidality, the mean scores ticipants, as judged by the therapist. To foster engage-
[BSSI = 13.8 (S.D. = 10.9), SPS = 87.1 (S.D. = 21.2)] indi- ment with the participant, the prioritization of
cated a severe level of suicidal ideation and risk modules was collaboratively agreed between the par-
amongst the overall sample of participants (Cull & ticipant and therapist. The module most frequently
Gill, 1982; Beck & Steer, 1991). Similarly, the mean delivered was cognitive restructuring, with almost
scores for depression [BDI = 34.7 (S.D. = 12.5)] and hope- half (42%) of participants expressing an interest in en-
lessness [BHS = 11.1 (S.D. = 6.4)] were in the severe gaging in this work, with an average of five sessions
ranges (Beck & Steer, 1988; Beck et al. 1996). focused on directly challenging unhelpful or inaccurate
appraisals. The attention-broadening technique was
used by a quarter (26%) of participants, with those en-
Intervention feasibility gaging in this work completing an average of six ses-
sions. Problem-solving training was delivered to a
Engagement and retention in the trial
quarter (23%) of participants and less than a fifth of
In total, 276 CBSP therapy sessions were voluntarily participants received techniques to improve self-
attended by participants, with an average of 8.9 (S.D. esteem (16%) or mood management (10%).
= 7.42, range 0–20) sessions per participant. Only 16
sessions were refused (participant mean = 0.52, S.D. =
Therapists’ ratings of participant adherence
0.81) and 46 sessions had to be rearranged due to con-
textual circumstances beyond the control of the therap- Promptness (4.3), attendance (3.6) and level of partici-
ist or participant (participant mean = 1.48, S.D. = 0.159), pation (3.1) in therapy sessions were all rated above the
such as legal visits, family visits and security incidents mid-point of the 1–5 rating scale, whilst lower ratings
on wings (‘lockdowns’). Of the participants, 12 (39%) were recorded for mastery of programme (2.8) and
received 12 or more sessions, and 10 (32%) participants the completion of homework (2.8). Ratings for disrup-
completed five or fewer sessions. Four (13%) partici- tive behaviour were very low (1.3). Trial therapist rat-
pants were allocated to receive the CBSP programme ings were not recorded for the four participants who
but were unexpectedly released from the prison did not attend a single therapy session.
 Cognitive–behavioural suicide prevention for male prisoners 3447

(SAPAS; treatment effect = −0.79, S.E. = 0.39, 95% CI

−1.55 to −0.04, p = 0.04).

Clinically significant recovery

Clinically significant recovery for participants was

indicated for total scores of 67 or less on the SPS
(Cull & Gill, 1982). At the end of treatment, over half
(10/18, 56%) of participants in the CBSP group
had achieved a clinically significant recovery com-
pared with a quarter (5/22, 23%) of the TAU group
(χ2 = 4.55, p = 0.03), although this group difference was
not maintained at follow-up (CBSP:53% v. TAU:44%;
Fig. 2. Number of suicidal and self-injurious behaviours. χ2 = 0.25, p = 0.62).
Values are means, with standard deviations represented by
vertical bars. CBSP, Cognitive–behavioural suicide
prevention; TAU, treatment as usual. Discussion

The PROSPeR trial was an exploratory pilot RCT of a

Suicidal and self-injurious behaviours (SIB) novel application of CBSP for individuals at elevated
risk of suicide. The results indicated that delivering
As shown in Fig. 2, the mean number of SIBs for the CBSP within a prison setting is feasible, with the ma-
CBSP group (1.06) was initially lower than for the jority of patients commencing therapy and choosing
TAU group (1.39). However, at the 6-month assess- to complete the programme. Further developments to
ment the mean number of SIBs for the CBSP group the treatment protocol may be required to better sup-
had decreased by almost 50% to 0.58, whereas this port participants’ learning of new coping techniques
figure had changed little (1.48) for the TAU group. and to enhance motivation to complete assignments
As such, the CBSP group engaged in fewer SIBs com- between therapy sessions. For instance, the completion
pared with the TAU group (treatment effect = −0.72, of homework tasks may be improved if additional sup-
S.E. = 0.47, 95% CI −1.71 to 0.09, p = 0.162). The number
port is offered between sessions, perhaps from the pris-
of participants who had recently engaged in suicidal or oner’s personal officer or keyworker. Also, mastery of
self-injurious behaviours at baseline [CBSP: 12 (39%), programme content may be improved by providing
TAU: 13 (42%)] reduced for both groups by the follow- participants with a self-help workbook to be reviewed
up assessment [CBSP: 7 (23%), TAU: 7 (23%)]. At the between sessions.
6-month assessments, no participants within the The participant sample was drawn from a population
CBSP group were found to have increased numbers considered to be at elevated risk of suicide and the inter-
of SIBs relative to baseline, whereas within the TAU vention gave rise to clinically relevant improvements.
group six participants had increased numbers of The importance of this finding is particularly apparent
SIBs. All SIB episodes were determined to be adverse when considered alongside the exceptionally high
events that were not related to the study. rates of suicidal behaviour reported by participants,
with more than half of the prisoners having previously
attempted suicide on two or more occasions. The
Self-report measures
CBSP therapy was found to be associated with improve-
The outcomes for each arm of the trial are compared in ments on measures relating to the primary outcome of
Table 3. Across measures of suicidal ideation, suicide SIB, as well as measures of psychiatric symptomatology,
probability, hopelessness, depression, anxiety and self- but this did not generalize to other established psycho-
esteem there was a consistent pattern that participants logical correlates of suicide. Whilst it would be inappro-
in both the CBSP and TAU groups made improve- priate to emphasize the statistical significance of these
ments between baseline and the follow-up assess- findings within the context of a pilot trial (Lancaster
ments, with greater improvements occurring for the et al. 2004), such results are seen as sufficiently encour-
CBSP group although there were no statistically signifi- aging to warrant further investigation of the efficacy of
cant effects of treatment. The repeated-measures the CBSP intervention.
regression modelling did indicate significant improve- To our knowledge, this is the first study that has
ments in measures of psychiatric symptomatology demonstrated an improvement in the cognitive–
(BPRS-R; treatment effect = −4.60, S.E. = 2.25, 95% CI behavioural prevention of suicidal behaviour delivered
−9.02 to −0.19, p = 0.04) and personality dysfunction within a prison setting. Previous investigations of
3448 D. Pratt et al.

Table 3. Primary and secondary outcome measures at baseline, and at 4 and 6 months

Baseline 4 months 6 months

CBSP (n = 31) TAU (n = 31) CBSP (n = 18) TAU (n = 22) CBSP (n = 17) TAU (n = 18)

Number of SIB episodes in 1.06 (2.10) 1.39 (3.28) N.A. N.A. 0.58 (1.52)a 1.48 (3.23)a
previous 6 months
BSSI 13.2 (10.8) 14.5 (11.2) 5.8 (9.9) 6.7 (10.5) 6.6 (10.4) 7.7 (11.4)
SPS 86.9 (19.9) 87.3 (20.8) 67.9 (24.3) 82.6 (23.2) 67.4 (21.8) 76.4 (23.8)
BHS 11.4 (6.1) 10.8 (6.8)b 6.8 (5.8) 8.6 (6.6) 7.9 (7.1) 7.3 (7.1)c
BDI 34.2 (11.7) 35.3 (13.4)b 17.1 (13.0) 26.6 (15.3) 20.2 (19.2) 23.4 (16.6)
BAI 21.7 (13.0) 21.5 (12.3)b 12.7 (13.1) 14.1 (11.1) 9.5 (11.1) 12.1 (13.7)
RSCQ 102.8 (21.9) 100.6 (29.2)b 117.6 (29.3) 106.3 (21.7)d 122.9 (25.5) 113.7 (31.6)
SAPAS 4.8 (1.7) 4.2 (1.7)b 3.8 (1.7) 4.7 (1.7) 3.7 (1.8) 4.3 (1.7)
BPRS 44.6 (11.6) 46.0 (11.0)b 35.1 (9.2) 39.7 (9.6) 34.9 (6.8) 41.1 (11.5)

Data are given as mean (standard deviation).

CBSP, Cognitive–behavioural suicide prevention; TAU, treatment as usual; SIB, suicidal and self-injurious behaviours; N.A.,
not applicable; BSSI, Beck Scale for Suicidal Ideation; SPS, Suicide Probability Scale; BHS, Beck Hopelessness Scale; BDI, Beck
Depression Inventory; BAI, Beck Anxiety Inventory; RSCQ, Robson Self-Concept Questionnaire; SAPAS, Standardised
Assessment of Personality – Abbreviated Scale, BPRS, Brief Psychiatric Rating Scale, expanded version.
a
n = 31.
b
n = 30.
c
n = 17.
d
n = 21.

cognitive therapy for suicidal behaviour have been determinants was found in the current study, there
conducted within community settings and treatment was no differential impact of CBSP therapy.
guidelines now recommend CBT as an important
part of the longer-term management of suicidal and
Implications
self-injurious behaviour (National Institute for Health
and Clinical Excellence, 2011). Results from the current A number of implications arise from this study, al-
study are consistent with previous trials demonstrating though these must be considered within the limited na-
significant associations between cognitive therapy and ture of a pilot trial. The modularized structure of the
decreases in rates of suicidal behaviour, compared CBSP programme into short, ‘digestible’ components
with routine care (Brown et al. 2005; Slee et al. 2008; may have helped to retain participants with poorer
Tarrier et al. 2014). Since previous studies were con- cognitive abilities, who are more common amongst
ducted in community settings with patients presenting prison groups (Social Exclusion Unit, 2002). Whilst
to hospital emergency departments or mental health the current study drew upon the individualized case
centres, it now appears that the efficacy of cognitive formulations for each participant to prioritize the treat-
therapy for suicide could extend into prison settings. ment modules, the ideal ordering of the modules could
Outside of the primary outcome of actual suicidal be investigated. Also, CBSP was delivered on a once-
behaviour, there has been a mixed set of results con- or twice-weekly basis to participants spread across a
cerning the proposed psychological determinants of 4-month treatment window. In other areas of applica-
suicide. Cognitive therapy has been found to be signifi- tion, CBT has been found to be preferable when deliv-
cantly associated with reducing scores on measures of ered in a more intensive format (Oldfield et al. 2011;
depression, hopelessness, anxiety and self-esteem Ehlers et al. 2014). Participants’ tolerance of an inten-
(Brown et al. 2005; Slee et al. 2008; Tarrier et al. 2014). sive CBSP approach, and the speed of recovery and po-
Contrary evidence has also been reported, where a re- tential efficacy, should be investigated since intensive
duction in suicide behaviour associated with cognitive formats may help to minimize the impact of unexpect-
therapy has been observed without concurrent ed transfers and discharges during therapy delivery.
improvements in these psychological correlates of sui- Similarly, since many offender behaviour programmes
cide (Davidson et al. 2006; Morley et al. 2014). Although are delivered within a group format, the familiarity of
a pattern of reducing scores for psychological this format to prisoners and staff should be considered,
 Cognitive–behavioural suicide prevention for male prisoners 3449

especially if this would enable a more cost-effective de- conducted with prisoner participants (Black et al.
livery of CBSP. However, potential drawbacks of a 2011; Olver et al. 2011). As such, when conducting fu-
group therapy format may include the loss of an indi- ture trials, researchers may need to pay even more at-
vidualistic approach to understanding the participant’s tention to assessing and enhancing motivation
motivation for suicide and the reluctance of some par- amongst the target group of participants, and prison
ticipants to share intimate details with others in fear of staff, to facilitate successful recruitment.
potential subsequent victimization outside the therapy Within the pragmatic constraints of a pilot trial,
group. there was no remit to standardize the ‘TAU’ received
A policy implication of the current study concerns the by all prisoners within this study, including any medi-
availability of cognitive–behavioural interventions to cation treatments provided by the Mental Health
individuals living in prison identified to be at risk of sui- In-Reach team. The nature and content of ‘TAU’
cide. The current ACCT system offers a robust risk man- received by individual participants was not registered,
agement process although it remains limited in terms of although all prisoners were entitled to receive mental
proactive interventions. The targeted provision of CBTs healthcare equivalent to that which would be offered
for the most vulnerable may help to contribute to the to all National Health Service (NHS) service users,
complex challenge of prison suicide prevention. such as psychotropic medication and nursing support.

Future research
Limitations
Further research addressing the limitations highlighted
The study has a number of limitations, which would
above is needed to assess the effectiveness of CBSP for
need to be overcome to conduct a more definitive
suicidal prisoners. Researchers should consider recruit-
trial. The sample size for the pilot trial was sufficient
ing participants from multiple sites to investigate if
to enable a preliminary investigation of the potential
CBSP has differential effects across different types of
of the CBSP therapy, although a larger-scale trial
prisons. The active components of CBSP therapy
would be required for more conclusive results.
should be examined by administering a more detailed
Similarly, the pilot trial was conducted within one
assessment battery specific to the proposed psycho-
site, thus limiting the heterogeneity of participants,
logical mechanisms targeted by the intervention,
and so further investigations should be conducted
as well as measures of the treatment process.
across multiple sites. Generalizability concerns are
Additionally, such a trial should consider the need to
also raised; for instance, whilst the proportion of parti-
compare CBSP therapy with an active comparison
cipants describing themselves as white British in the
intervention, e.g. supportive counselling or befriend-
current study (85%) was in keeping with previous
ing, in order to control for potential non-specific fac-
investigations of suicidal prisoners (e.g. 82%, Hawton
tors. Future investigations should include an
et al. 2014), these proportions are notably higher than
economic evaluation in order to estimate the costs of
that observed amongst the general prison population
use of health and social care within the custodial set-
(74%; Ministry of Justice, 2014).
tings, and beyond for those released back into the com-
Recruiting and retaining suicidal participants into a
munity. Additional metrics on the broader impact of
clinical trial has a tendency to be problematic and chal-
the intervention should also be considered, including
lenging. The high proportion (64%) of eligible prison-
violent incidents, prison infractions, adjudications, etc.
ers who chose not to participate in the current trial
presents a serious threat to the feasibility of the inter-
vention. Indeed, treatment refusal and attrition Conclusions
among prisoners is higher than for most other clinical The CBSP therapy offers a novel approach that has
groups with non-completion of treatment endemic to shown some potential for providing clinical benefit to
all interventions delivered with prison settings prisoners in terms of reduced SIB, decreased psychi-
(Wormith & Olver, 2002). Within prisoner groups, typ- atric symptomatology and personality dysfunction,
ical rates of refusal to enter treatment are up to 70% and some improvement on the psychological determi-
(Dalton et al. 1998; Black et al. 2011) and treatment nants of suicide. This small-scale pilot study now
drop-outs can be as high as 93% (Gondolf & Foster, needs to be replicated within a larger-scale multi-site
1991). Although our previous trial of CBSP for suicidal randomized controlled trial.
people experiencing psychosis (Tarrier et al. 2014)
achieved an attrition rate of less than 30%, attrition
Acknowledgements
for the current study (44%) was more in keeping
with other trials of CBT for suicide prevention This paper presents independent research funded by the
(Morley et al. 2014) and intervention evaluations National Institute for Health Research (NIHR) under
3450 D. Pratt et al.

its Research for Patient Benefit (RfPB) Programme Department of Health (2002). National Suicide Prevention
(Grant Reference Number PB-PG-0609–19126). The Strategy for England. Department of Health: London.
views expressed are those of the author(s) and not Department of Health (2007). Improving Health,
necessarily those of the NHS, the NIHR or the Supporting Justice: A Consultation Document. A Strategy for
Improving Health and Social Care Services for People Subject to
Department of Health. We would like to acknowledge
the Criminal Justice System. Department of Health: London.
Chris Wall, Aisha Mirza and Kieran Lord for assistance
Department of Health (2012). Preventing Suicide in England: A
with data collection and data entry; Chris Wall for Cross-government Outcomes Strategy to Save Lives.
administrative support; the Mental Health Research Department of Health: London.
Network (especially Heather Morrison) for their sup- Efron B, Tibshirani RJ (1993). An Introduction to the Bootstrap.
port and assistance; and Peer Bhatti, Claire Daniels, Chapman and Hall: London.
Angel Delight, David O’Brien and Natasha Peniston Ehlers A, Hackmann A, Grey N, Wild J, Liness S, Albert I,
for their contribution throughout the project as mem- Deale A, Stott R, Clark DM (2014). A randomized controlled
bers of the Service User Reference Group (SURG). trial of 7-day intensive and standard weekly cognitive
therapy for PTSD and emotion-focused supportive therapy.
American Journal of Psychiatry 171, 294–304.
Fazel S, Benning R, Danesh J (2005). Suicides in male
Declaration of Interest
prisoners in England and Wales, 1978–2003. Lancet 366,
None. 1301–1302.
Fazel S, Grann M, Kling B, Hawton K (2011). Prison suicide
in 12 countries: an ecological study of 861 suicides during
2003–2007. Social Psychiatry and Psychiatric Epidemiology 46,
References
191–195.
Beck AT, Epstein N, Brown G, Steer RA (1988). An inventory Gendreau P (1996). Offender rehabilitation what we know
for measuring clinical anxiety: psychometric properties. and what needs to be done. Criminal Justice and Behavior 23,
Journal of Consulting and Clinical Psychology 56, 893–897. 144–161.
Beck AT, Steer RA (1988). Manual for the Beck Hopelessness Gondolf EW, Foster RA (1991). Pre-program attrition in
Scale. The Psychological Corporation: San Antonio, TX. batterer programs. Journal of Family Violence 6, 337–349.
Beck AT, Steer RA (1991). Manual for the Beck Scale for Suicide Hawton K, Linsell L, Adeniji T, Sariaslan A, Fazel S (2014).
Ideation. The Psychological Corporation: San Antonio, TX. Self-harm in prisons in England and Wales: an
Beck AT, Steer RA, Brown GK (1996). Manual for the Beck epidemiological study of prevalence, risk factors, clustering,
Depression Inventory-II. The Psychological Corporation: San and subsequent suicide. Lancet 383, 1147–1154.
Antonio, TX. Hollin CR, Palmer EJ (2006). Offending Behaviour Programmes:
Birmingham L (2003). The mental health of prisoners. Development, Application and Controversies. Wiley: Chichester.
Advances in Psychiatric Treatment 9, 191–199. Humber N, Hayes A, Senior J, Fahy T, Shaw J (2010).
Black G, Forrester A, Wilks M, Riaz M, Maguire H, Carlin P Identifying, monitoring and managing prisoners at risk of
(2011). Using initiative to provide clinical intervention self-harm/suicide in England and Wales. Journal of Forensic
groups in prison: a process evaluation. International Review Psychiatry and Psychology 22, 22–51.
of Psychiatry 23, 70–76. Jacobson NS, Roberts LJ, Berns SB, McGlinchey JB (1999).
Bowcott O, Taylor M, Laville S (2014). England and Wales Methods for defining and determining the clinical significance
prisons chief warns of ‘terrible toll’ of inmate suicides. In of treatment effects: description, application, and alternatives.
The Guardian. http://www.theguardian.com/society/2014/ Journal of Consulting and Clinical Psychology 67, 300–307.
oct/21/prisons-uk-chief-hardwick-suicide-prisoners-safety Johnson J, Gooding P, Tarrier N (2008). Suicide risk in
(accessed April 2015). schizophrenia: explanatory models and clinical
Brown GK, Ten Have T, Henriques GR, Xie SX, Hollander implications, The Schematic Appraisal Model of Suicide
JE, Beck AT (2005). Cognitive therapy for the prevention of (SAMS). Psychology and Psychotherapy: Theory, Research and
suicide attempts: a randomized controlled trial. Journal of Practice 81, 55–77.
the American Medical Association 294, 563–570. Konrad N, Daigle MS, Daniel AE, Dear GE, Frottier P,
Cull J, Gill W (1982). Manual for the Suicide Probability Scale. Hayes LM, Kerkhof A, Liebling A, Sarchiapone M (2007).
Western Psychological Services: Los Angeles, CA. Preventing suicide in prisons, Part I: recommendations
Dalton R, Majoy S, Sharkey M (1998) Nonattenders and from the International Association for Suicide Prevention
attrition from a forensic psychology outpatient service. Task Force on Suicide in Prisons. Crisis: The Journal of Crisis
International Journal of Offender Therapy and Comparative Intervention and Suicide Prevention 28, 113–121.
Criminology 42, 174–180. Lancaster GA, Dodd S, Williamson PR (2004). Design and
Davidson K, Norrie J, Tyrer P, Gumley A, Tata P, Murray H, analysis of pilot studies: recommendations for good
Palmer S (2006). The effectiveness of cognitive behavior practice. Journal of Evaluation in Clinical Practice 10, 307–312.
therapy for borderline personality disorder: results from the Landenberger NA, Lipsey MW (2005). The positive effects of
Borderline Personality Disorder Study of Cognitive cognitive–behavioral programs for offenders: a
Therapy (BOSCOT) trial. Journal of Personality Disorders 20, meta-analysis of factors associated with effective treatment.
450–465. Journal of Experimental Criminology 1, 451–476.
 Cognitive–behavioural suicide prevention for male prisoners 3451

Liebling A, Maruna S (eds.) (2005). The Effects of Perry AE, Marandos R, Coulton S, Johnson M (2010).
Imprisonment. Willan: Cullompton. Screening tools assessing risk of suicide and self-harm in
Little RJ, Rubin DB (2002). Statistical Analysis with Missing adult offenders: a systematic review. International Journal
Data, 2nd edn. Wiley: New York. of Offender Therapy and Comparative Criminology 54, 803–828.
McGuire J (2002). Offender Rehabilitation and Treatment: Rivlin A, Hawton K, Marzano L, Fazel S (2010). Psychiatric
Effective Programmes and Policies to Reduce Re-Offending. disorders in male prisoners who made near-lethal suicide
Wiley: Chichester. attempts: case–control study. British Journal of Psychiatry
Medical Research Council (2008). Developing and Evaluating 197, 313–319.
Complex Interventions: New Guidance. Medical Research Robson P (1989). Development of a new self-report questionnaire
Council: London. to measure self-esteem. Psychological Medicine 19, 513–518.
Ministry of Justice (2013). PSI 64/2011: Management of Senior J, Hayes AJ, Pratt D, Thomas SD, Fahy T, Leese M,
Prisoners at Risk of Harm to Self, to Others and from Others. Bowen A, Taylor G, Lever-Green G, Graham T, Pearson
Ministry of Justice: London. http://www.justice.gov.uk/ A, Ahmed M, Shaw JJ (2007). The identification and
downloads/offenders/psipso/psi-2011/psi-64-2011-safer-cus management of suicide risk in local prisons. Journal of
tody.doc (accessed April 2015). Forensic Psychiatry and Psychology 18, 368–380.
Ministry of Justice (2014). Offender Management Statistics Slee N, Garnefski N, van der Leeden R, Arensman E,
Prison Population 2014. Ministry of Justice: London. Spinhoven P (2008). Cognitive–behavioural intervention
Moran P, Leese M, Lee T, Walters P, Thornicroft G, Mann A for self-harm: randomised controlled trial. British Journal of
(2003). Standardised Assessment of Personality – Psychiatry 192, 202–211.
Abbreviated Scale (SAPAS): preliminary validation of a Social Exclusion Unit (2002). Reducing Re-offending by
brief screen for personality disorder. British Journal of Ex-prisoners. Social Exclusion Unit: London.
Psychiatry 183, 228–232. StataCorp (2009). Stata Release 11. StataCorp: College Station, TX.
Morley KC, Sitharthan G, Haber PS, Tucker P, Sitharthan T Stone A, Wheeler C, Barge A (2007). Improving the design of
(2014). The efficacy of an opportunistic cognitive behavioral phase II trials of cytostatic anticancer agents. Contemporary
intervention package (OCB) on substance use and comorbid Clinical Trials 28, 138–145.
suicide risk: a multisite randomized controlled trial. Journal Tarrier N, Gooding P, Pratt D, Kelly J, Maxwell J, Awenet Y
of Consulting and Clinical Psychology 82, 130–140. (2013). Cognitive Behavioural Prevention of Suicide in
National Institute for Health and Clinical Excellence (2011). Psychosis: A Treatment Manual. Routledge: London.
Self-Harm: Longer-Term Management (Clinical Guideline Tarrier N, Kelly J, Maqsood S, Snelson N, Maxwell J, Law
CG133). National Institute for Health and Clinical H, Dunn G, Gooding P (2014). The cognitive behavioural
Excellence: London. prevention of suicide in psychosis: a clinical trial.
Naud H, Daigle MS (2010). Predictive validity of the suicide Schizophrenia Research 156, 204–210.
probability scale in a male inmate population. Journal of Tarrier N, Taylor K, Gooding P (2008). Cognitive–behavioral
Psychopathology and Behavioral Assessment 32, 333–342. interventions to reduce suicide behaviour: a systematic
Oldfield VB, Salkovskis PM, Taylor T (2011). Time-intensive review and meta-analysis. Behavior Modification 32, 77–108.
cognitive behaviour therapy for obsessive compulsive Ventura J, Lukoff D, Nuechterlein K, Liberman R, Green M,
disorder: a case series and matched comparison group. Shaner A (1993). Manual for the expanded brief psychiatric
British Journal of Clinical Psychology 50, 7–18. rating scale. International Journal of Methods in Psychiatric
Olver ME, Stockdale KC, Wormith JS (2011). A Research 3, 221–224.
meta-analysis of predictors of offender treatment attrition Wormith JS, Olver ME (2002). Offender treatment attrition
and its relationship to recidivism. Journal of Consulting and and its relationship with risk, responsivity, and recidivism.
Clinical Psychology 79, 6–21. Criminal Justice and Behaviour 29, 447–471.