Hensel‐Dittmann 2011.

Methods	Study design: randomised controlled trial
Participants	28 clinic outpatients, refugees from various countries—most still seeking asylum Diagnosis: PTSD Method of diagnosis:DSM‐IV Age: not given, but no differences between groups Sex: not given, but no differences between groups Location: Germany
Interventions	Participants were randomly assigned to: 1. Experimental arm Duration: 10 individual sessions of mean 90 minutes Treatment protocol: Narrative exposure (NET) manualised (Schauer) Therapist: trained, with interpreter when necessary (17/28) 2. Comparator arm Duration: 10 individual sessions of mean 90 minutes Treatment protocol: stress inoculation training (SIT), avoiding any element of exposure Therapist: trained, with interpreter when necessary (17/28) Same therapists for both arms
Outcomes	Time points for assessment: pretreatment and at 6‐month and 1‐year follow‐up Assessment language: measures in German; no information on cross‐cultural use Primary outcome PTSD severity score (clinician‐administered scale: CAPS) Secondary outcome PTSD diagnosis: DSM‐IV Depression: Hamilton Depression Scale
baseline characteristics	76% had been tortured; remainder had experienced war No differences between groups in length of time in Germany, area of origin, education or co‐morbid psychiatric disorders, but no baseline data given
adherence and completion	5 dropouts NET, 2 dropouts SIT (1 SIT participant deported)
Notes	Date of study: 2004 to 2007 Funding source: European Refugee Fund and Deutsche Forschungsgemeinschaft Declarations of interest among primary researchers: no conflicting interests
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants matched pairwise by gender, age and region of origin, then allocated by flipping coin
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not possible to render participants nor practitioners blind to allocation. Expectations of benefit not assessed
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Assessors blind (unless accidentally unblinded) to allocation
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data provided different N at each time point; analysis by intention‐to‐treat, so mixed‐effect models with neither imputation nor LOCF
Selective reporting (reporting bias)	Unclear risk	All outcomes reported in trial methods; no protocol available
Therapist allegiance	High risk	NET: active treatment
Treatment fidelity	Low risk	Manual by Schauer
Therapist qualifications	Low risk	Trained therapists
Other bias	Unclear risk	Most refugees still had asylum undecided, so may have had an incentive to underreport improvement