BACKGROUND: Torture is widespread, with potentially broad and long-lasting impact across physical, psychological, social and other areas of life. Its complex and diverse effects interact with ethnicity, gender, and refugee experience. Health and welfare agencies offer varied rehabilitation services, from conventional mental health treatment to eclectic or needs-based interventions. This review is needed because relatively little outcome research has been done in this field, and no previous systematic review has been conducted. Resources are scarce, and the challenges of providing services can be considerable. OBJECTIVES: To assess beneficial and adverse effects of psychological, social and welfare interventions for torture survivors, and to compare these effects with those reported by active and inactive controls. SEARCH METHODS: Randomised controlled trials (RCTs) were identified through a search of PsycINFO, MEDLINE, EMBASE, Web of Science, the Cumulative Index to Nursing and Allied Health Literature (CINAHL), the Cochrane Central Register of Controlled Trials (CENTRAL) and the Cochrane Depression, Anxiety and Neurosis Specialised Register (CCDANCTR), the Latin American and Caribbean Health Science Information Database (LILACS), the Open System for Information on Grey Literature in Europe (OpenSIGLE), the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) and Published International Literature On Traumatic Stress (PILOTS) all years to 11 April 2013; searches of Cochrane resources, international trial registries and the main biomedical databases were updated on 20 June 2014. We also searched the Online Library of Dignity (Danish Institute against Torture), reference lists of reviews and included studies and the most frequently cited journals, up to April 2013 but not repeated for 2014. Investigators were contacted to provide updates or details as necessary. SELECTION CRITERIA: Full publications of RCTs or quasi-RCTs of psychological, social or welfare interventions for survivors of torture against any active or inactive comparison condition. DATA COLLECTION AND ANALYSIS: We included all major sources of grey literature in our search and used standard methodological procedures as expected by The Cochrane Collaboration for collecting data, evaluating risk of bias and using GRADE (Grades of Recommendation, Assessment, Development and Evaluation) methods to assess the quality of evidence. MAIN RESULTS: Nine RCTs were included in this review. All were of psychological interventions; none provided social or welfare interventions. The nine trials provided data for 507 adults; none involved children or adolescents. Eight of the nine studies described individual treatment, and one discussed group treatment. Six trials were conducted in Europe, and three in different African countries. Most people were refugees in their thirties and forties; most met the criteria for post-traumatic stress disorder (PTSD) at the outset. Four trials used narrative exposure therapy (NET), one cognitive-behavioural therapy (CBT) and the other four used mixed methods for trauma symptoms, one of which included reconciliation methods. Five interventions were compared with active controls, such as psychoeducation; four used treatment as usual or waiting list/no treatment; we analysed all control conditions together. Duration of therapy varied from one hour to longer than 20 hours with a median of around 12 to 15 hours. All trials reported effects on distress and on PTSD, and two reported on quality of life. Five studies followed up participants for at least six months.No immediate benefits of psychological therapy were noted in comparison with controls in terms of our primary outcome of distress (usually depression), nor for PTSD symptoms, PTSD caseness, or quality of life. At six-month follow-up, three NET and one CBT study (86 participants) showed moderate effect sizes for intervention over control in reduction of distress (standardised mean difference (SMD) -0.63, 95% confidence interval (CI) -1.07 to -0.19) and of PTSD symptoms (SMD -0.52, 95% CI -0.97 to -0.07). However, the quality of evidence was very low, and risk of bias resulted from researcher/therapist allegiance to treatment methods, effects of uncertain asylum status of some people and real-time non-standardised translation of assessment measures. No measures of adverse events were described, nor of participation, social functioning, quantity of social or family relationships, proxy measures by third parties or satisfaction with treatment. Too few studies were identified for review authors to attempt sensitivity analyses. AUTHORS' CONCLUSIONS: Very low-quality evidence suggests no differences between psychological therapies and controls in terms of immediate effects on post-traumatic symptoms, distress or quality of life; however, NET and CBT were found to confer moderate benefits in reducing distress and PTSD symptoms over the medium term (six months after treatment). Evidence was of very low quality, mainly because non-standardised assessment methods using interpreters were applied, and sample sizes were very small. Most eligible trials also revealed medium to high risk of bias. Further, attention to the cultural appropriateness of interventions or to their psychometric qualities was inadequate, and assessment measures used were unsuitable. As such, these findings should be interpreted with caution.No data were available on whether symptom reduction enabled improvements in quality of life, participation in community life, or in social and family relationships in the medium term. Details of adverse events and treatment satisfaction were not available immediately after treatment nor in the medium term. Future research should aim to address these gaps in the evidence and should include larger sample sizes when possible. Problems of torture survivors need to be defined far more broadly than by PTSD symptoms, and recognition given to the contextual influences of being a torture survivor, including as an asylum seeker or refugee, on psychological and social health.