Igreja 2004.

Methods	Study design: randomised controlled trial
Participants	137 people, post civil war, mostly rural population seen in their homes Diagnosis: PTSD caseness Method of diagnosis: self‐Inventory for PTSD (Hovens et al, 2001) Age: mean 40 years (SD 14) Sex: 56% men, 44% women Location: Mozambique
Interventions	Participants were randomly assigned to: 1. Experimental arm Duration: 1 occasionally 2 individual sessions, about 60 minutes Treatment protocol: testimony writing: references to 'testimony method' but no mention of protocol Therapist/interviewer: first study author interpreted into Chi‐Gorongese by native speakers (same sex as participant) 2. Comparator arm Duration: none Treatment protocol: no intervention Therapist: none
Outcomes	Time points for assessment: pretreatment and at 11‐month follow‐up Assessment language: all via structured interview, as participants illliterate, interpreted into Chi‐Gorongese Outcomes (not specified as primary or secondary) Post‐traumatic stress symptoms: self‐Inventory for PTSD; only Western data available on performance of scale Psychiatric symptoms: Self‐Report Questionnaire, validated in non‐Western populations Nightmares: Nocturnal Intrusions after Traumatic Experiences Questionnaire; only Western data available on performance of scale
baseline characteristics	58% intervention group and 55% control group tortured; many other relevant events of organised violence on Harvard Trauma Questionnaire, validated in non‐Western populations Mean 15 years in war zone Mean 4 living children and 3 dead
adherence and completion	6 dropouts
Notes	Date of study: not given Funding source: part by Associação Esperança Para Todos, Mozambique Declarations of interest among primary researchers: none Third arm (not included here) of non‐case participants
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Participants given consecutive numbers, divided according to caseness, then allocated to treatment or control according to odd or even number
Allocation concealment (selection bias)	High risk	None
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	Unclear who conducted assessments—all by interview
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Several dropouts (death, moving away)
Selective reporting (reporting bias)	Unclear risk	All outcomes listed in trial reported; no protocol available
Therapist allegiance	Unclear risk	No information provided
Treatment fidelity	Unclear risk	No information provided
Therapist qualifications	Unclear risk	No information provided
Other bias	Unclear risk	Real‐time translation of assessment measures, so not standardised