Outcomes of psychological therapies in prisons: a systematic review and meta-analysis

Thesis

Background: The imprisoned population, currently more than 10 million worldwide, is reported to have higher rates of suicide and prevalence of severe mental illness than the general population. With these identified needs, the effectiveness of psychological therapies on mental health outcomes in prisons warrants a systematic evaluation. The purpose of this thesis is to systematically review and analyse randomised controlled trials of psychological therapies in prisons.

Methods: A systematic search was conducted to identify studies that report mental health outcomes of psychological therapies among prisoners. Effect sizes were calculated from extracted data and presented using forest plots. Eligible studies were assessed for quality. Meta-regression analysis was performed to examine sources of heterogeneity between studies.

Results: Thirty-seven randomised controlled trial studies from 31 individual papers were eligible for data synthesis. In total the included studies showed a medium pooled effect size (0.50, 95% CI: 0.34-0.66) with a high level of heterogeneity. Subset of studies that adopted no treatment or waitlist as the control group had significantly larger pooled effect size (0.74, 95% CI: 0.51-0.97) than those that had treatment-as-usual (TAU) or another type of psychological therapy as controls (0.21, 95% CI: 0.01-0.41). Few studies that measured follow-up outcomes did not produce a significant pooled effect size. Effect sizes did not significantly differ between group therapy and individual therapy. Differences among varied treatment types were also not significant. Meta-regression of multiple study characteristics showed that presence of an adequate fidelity measure in a study was associated with lower effect size in treatment outcome. In the subgroup analysis with studies with no treatment/waitlist controls, lower participation rate and lower sample size were associated with higher effect size. Qualitative synthesis reported the difficulties in conducting RCTs of psychological therapy in prisons.

Conclusions: While psychological therapies in prisons tend to show medium treatment effect compared to no treatment or waitlist controls, the effect is reduced when the comparator is treatment-as-usual or other forms of psychological therapy. Studies should adopt an adequate fidelity measure, and caution is needed in interpreting studies with small sample sizes (≤ 100) and high attrition rates (≥ 20%).

Authors: Yoon, IA

• DOI: 10.5287/ora-9gjpwerav
• Type of award: MSc by Research
• Level of award: Masters
• Awarding institution: University of Oxford