Systematic review -- systematic and replicable methods of searching the literature and selecting studies -- and meta-analysis -- combining the weighted results of these studies for a well-powered answer -- have become the gold standard for clinical guidelines. Their strengths are transparency, replicability, and can ease of updating, but since they were not designed for trials of psychological treatment, it can be difficult to apply the principles and methods. Despite this, we argue that it is the best way to synthesise evidence. We explore and demonstrate these methods for randomised controlled trials of psychological treatment for survivors of torture.
We carried out the review within the Cochrane Collaboration, which involves rigorous attention to methodology. We searched up to June 2014 and aim to update the review and meta-analysis. We searched both peer-reviewed and non-peer-reviewed literature databases: over 2000 titles eventually yielded 9 eligible studies. We were sorry to exclude a few excellent non-randomised trials. We extracted data on participants, interventions, control conditions and outcomes from all 9 trials and combined it in analyses of outcome: immediate and follow-up results (where available) for PTSD symptoms, depression/distress, and quality of life. We also used standard and other relevant tools to identify possible sources of bias in trials, to aid interpretation.
Six trials were carried out in Europe, three in African countries. They included 507 adults; most were refugees diagnosed with PTSD to enter treatment. Four studies used narrative exposure therapy, one used cognitive behavioural therapy, and four were mixed. Meta-analysis showed no improvement in PTSD, distress, or quality of life at the end of treatment, but some improvement at six-month follow-up in distress and PTSD symptoms, although levels were still high. There were several high risks of bias particularly from small numbers and therapist allegiance, and barely any attention to issues of cultural and language differences in using outcome instruments. The impression overall is of disappointingly small change from trials whose methods -- although we recognise the difficulty of conducting such trials -- risk undermining their results. This is a contrast to many narrative reviews of the field but, by virtue of the review methods, reliable and robust.
Funding & No Conflicts Declaration
We are grateful to the Oak Foundation for financial support. We have no conflicts of interest to declare.