A trial in seven to eight UK women’s prisons will test whether four to eight sessions of a talking therapy called Psychodynamic Interpersonal Therapy (PIT) can cut self-harm incidents among female prisoners by a third. Self-harm is common in women’s prisons—the Ministry of Justice records an average of 6.65 episodes per prisoner over eight weeks—yet no standardised, evidence-based care pathway exists. Current interventions, such as cognitive behavioural therapy, are rarely designed for self-harm and have never been formally evaluated in this setting. PIT, which focuses on managing emotions and relationships, is targeted and has shown promise in earlier work. If PIT proves effective, it could give prisons a manualised, low-cost treatment that reduces both the human toll of self-harm and the associated costs to the public sector—currently unknown because no data exist on the expense of self-harm in women’s prisons. The trial will also estimate the cost per self-harm incident averted. Qualitative interviews with 40 prisoners and staff focus groups in all prisons will explore what works and why, helping to shape future implementation.
View original technical description
Design: An RCT of Psychodynamic Interpersonal Therapy vs. treatment as usual for self-harming women in prison. Setting: 7-8 UK women’s prisons. Population: Self-harming women in prison. Inclusion: Experiencing self-harm thoughts and on an ACCT in the last month; self-harm in the last month; 18 years or older. Exclusion: Lacking capacity to consent; too distressed or unwell to participate; likely to remain less than 8 weeks in prison; risk of harm to others. Technology: 4-8, 50-minute sessions of manualised PIT adapted for self-harming women prisoners; delivered by supervised psychology/psychiatry trainees. PIT focuses on managing emotions and relationships. Outcomes: Primary: self-harm incidents at 8-weeks (prison records/self-report). Secondary: level of harm caused by self-harm incidents (e.g. requires hospitalisation); self-harm thoughts; BSSI; BDI; BHS; WEMWS; RSES; EQ-5D-5L or SF12. Scales will be completed at baseline/8 weeks from baseline/12 weeks from baseline. There will be an internal stop/go pilot lasting 12 months to assess recruitment, fidelity and treatment completion. Analyses: intention-to-treat following CONSORT principles. Generalised linear mixed models will be fitted to the primary and secondary outcomes, including a random intercept for therapist, and baseline outcome measure and site as fixed effects. There is currently no information on the costs of self-harm in women’s prisons. The incremental cost-effectiveness of PIT vs. TAU will be estimated from a public sector perspective. Secondary analysis will estimate the cost per self-harm-incident averted. Resource use will be collected using SFSUS/bespoke measure from prison/medical notes and self-report. Modelling of potential future costs/benefits of PIT/TAU will be performed. Qualitative: interviews will be completed with prisoners at the end of the pilot (N=12-20) and intervention (N=40) to explore which aspects of PIT/TAU work. Staff focus groups in all prisons (N=7) will explore perceptions of PIT/TAU. Data will be analysed using framework analysis. Size: For comparison of PIT/TAU, calculations are powered to detect a minimal important clinical difference of 33% reduction in self-harm incidents. From MOJ, the average number of self-harm episodes over 8 weeks in the control group was 6.65. We used Monte Carlo simulations with 1000 simulations to estimate the power for a range of sample sizes, with power defined as the proportion of estimated p-values less than 0.05 for null hypothesis testing (treatment effect is 0), against an alternative hypothesis (33% reduction in the outcome log count). With 5% significance, to obtain 90% power, we will require 105 per group for the primary analyses. Based on WORSHIP II we estimate 20% attrition so will recruit 132 per group at baseline. Difference from current care: No standardised, evidence-based self-harm care pathway exists in prisons. Interventions such as CBT are used, but are rarely targeted for self-harm; with none formally evaluated. PIT is targeted and evidence-based. Timetable: Set up (mths 1-9); Pilot (10-21): N=54 based on WORSHIP II [25]; Stop/go (22); Recruit/follow up: N=210 (23-46); quantitative analysis (47-50) qualitative analysis (47-48); write up (49-54); health economics (1; 22; 47-54). Expertise: Gendered mental health (KA), prison research (KA/KG/LR/JS/TW/EP/AB), service user (FE), forensic practice (LR/AB/JS), statistics (RE), health economics (RM) implementation (KL/SA); HMPPS Deputy Director (LS); Prison Governor (MM).
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know