BACKGROUND: Persistent (chronic) pain is a frequent complaint in survivors of torture, particularly but not exclusively pain in the musculoskeletal system. Torture survivors may have no access to health care; where they do, they may not be recognised when they present, and the care available often falls short of their needs. There is a tendency in state and non-governmental organisations' services to focus on mental health, with poor understanding of persistent pain, while survivors may have many other legal, welfare, and social problems that take precedence over health care. OBJECTIVES: To assess the efficacy of interventions for treating persistent pain and associated problems in survivors of torture. SEARCH METHODS: We searched for randomised controlled trials (RCTs) published in any language in CENTRAL, MEDLINE, Embase, Web of Science, CINAHL, LILACS, and PsycINFO, from database inception to 1 February 2017. We also searched trials registers and grey literature databases. SELECTION CRITERIA: RCTs of interventions of any type (medical, physical, psychological) compared with any alternative intervention or no intervention, and with a pain outcome. Studies needed to have at least 10 participants in each arm for inclusion. DATA COLLECTION AND ANALYSIS: We identified 3578 titles in total after deduplication; we selected 24 full papers to assess for eligibility. We requested data from two completed trials without published results.We used standard methodological procedures expected by Cochrane. We assessed risk of bias and extracted data. We calculated standardised mean difference (SMD) and effect sizes with 95% confidence intervals (CI). We assessed the evidence using GRADE and created a 'Summary of findings' table. MAIN RESULTS: Three small published studies (88 participants) met the inclusion criteria, but one had been retracted from publication because of ethical problems concerned with confidentiality and financial irregularities. Since these did not affect the data, the study was retained in this review. Despite the search including any intervention, only two types were represented in the eligible studies: two trials used cognitive behavioural therapy (CBT) with biofeedback versus waiting list on unspecified persistent pain (58 participants completed treatment), and one examined the effect of complex manual therapy versus self-treatment on low back pain (30 participants completed treatment). Excluded studies were largely either not RCTs or did not report pain as an outcome.There was no difference for the outcome of pain relief at the end of treatment between CBT and waiting list (two trials, 58 participants; SMD -0.05, 95% CI -1.23 to 1.12) (very low quality evidence); one of these reported a three-month follow-up with no difference between intervention and comparison (28 participants; SMD -0.03, 95% CI -0.28 to 0.23) (very low quality evidence). The manual therapy trial also reported no difference between complex manual therapy and self-treatment (30 participants; SMD -0.48, 95% CI -9.95 to 0.35) (very low quality evidence). Two studies reported dropouts, one with partial information on reasons; none of the studies reported adverse effects.There was no information from any study on the outcomes of use of analgesics or quality of life.Reduction in disability showed no difference at the end of treatment between CBT and waiting list (two trials, 57 participants; SMD -0.39, 95% CI -1.17 to 0.39) (very low quality evidence); one of these reported a three-month follow-up with no difference between intervention and comparison (28 participants; SMD 0, 95% CI -0.74 to 0.74) (very low quality evidence). The manual therapy trial reported superiority of complex manual therapy over self-treatment for reducing disability (30 participants; SMD -1.10, 95% CI - 1.88 to -0.33) (very low quality evidence).Reduction in distress showed no difference at the end of treatment between CBT and waiting list (two trials, 58 participants; SMD 0.07, 95% CI -0.46 to 0.60) (very low quality evidence); one of these reported a three-month follow-up with no difference between intervention and comparison (28 participants; SMD -0.24, 95% CI -0.50 to 0.99) (very low quality evidence). The manual therapy trial reported superiority of complex manual therapy over self-treatment for reducing distress (30 participants; SMD -1.26, 95% CI - 2.06 to -0.47) (very low quality evidence).The risk of bias was considered high given the small number of trials, small size of trials, and the likelihood that each was underpowered for the comparisons it reported. We primarily downgraded the quality of the evidence due to small numbers in trials, lack of intention-to-treat analyses, high unaccounted dropout, lack of detail on study methods, and CIs around effect sizes that included no effect, benefit, and harm. AUTHORS' CONCLUSIONS: There is insufficient evidence to support or refute the use of any intervention for persistent pain in survivors of torture.