Marais 2011.

Study characteristics
Methods	Study design: cluster‐RCT; allocation by clinic; no blinding of patients, interventionists, or assessors; 4 clinics in each arm (8 total) Duration of study: March 2007 to February 2008
Participants	Country: South Africa   Income classification: upper‐middle income  Geographical scope: rural  Healthcare setting: 8 state health clinics  Mental health condition: alcohol use disorder  Population (mention whether patient, carer, or dyad)  1. Age: 15 and older   2. Gender: women  3. Socioeconomic background: majority were poor seasonal employees  4. Inclusion criteria (including threshold cutoff score of measurement tool)  a. Adult pregnant women attending clinic for antenatal services    5. Exclusion criteria: none
Interventions	Stated purpose: to determine the effects of a series of BIs on alcohol drinking behaviour of pregnant women in a high‐risk rural district in the Western Cape Province of South Africa  INTERVENTION (n = 97)  Name: brief intervention  Delivered by: CP Title/name of PW and number: trained fieldworkers ‐ 2   1. Selection: previously trained in brief intervention  2. Educational background: 1 a social worker, 1 a social scientist; both are enrolled in MPhil in social science research  3. Training (contents, duration, by whom): not specified  4. Supervision: none  Intervention details   1. Duration/frequency: 4 sessions: first lasting an hour and performed at baseline, second and third sessions lasting 20 minutes and performed at 41 and 47 days, and fourth and last sessions on 58th day, lasting an unspecified duration of time  2. Content of intervention: first session – questionnaire, AUDIT (Alcohol Use Disorders Identification Test), explaining the meaning of AUDIT results, BI with setting drinking goals, and making notes in a take‐home alcohol booklet. Second and third sessions ‐ BIs consisted of feedback on drinking behaviour, negotiations, goal‐setting, and reinforcement, followed by a questionnaire. Final session ‐ BI and feedback on drinking behaviour, questionnaire, AUDIT  CONTROL (n = 82) Initial interview: questionnaire, AUDIT, take‐home alcohol booklet; second interview: AUDIT, questionnaire on changes in drinking behaviour  CO‐INTERVENTIONS: usual antenatal care
Outcomes	Patients  Post‐intervention AUDIT scores (overall*, no drinking group, unconfirmed drinking group, and confirmed drinking group)  Carers  None  Process/health workers Nnone  Economic outcomes  None  (asterisk for study's primary outcomes; star: outcomes that we have not reported in this review)  Time point post intervention: 0 days
Notes	Source of funding: Western Cape Department of Social Development  Notes on validation of instruments (screening and outcomes): validated  Additional information (e.g. provided by authors, existence of a published study protocol): personal communication regarding background of fieldworkers. Declaration of interests ‐ none Handling the data: as per footnotes in data and analysis Prospective trial registration number: none  Study also included in prevention review, as unsure about the population (roughly half the intervention group has mental distress or a mental disorder at baseline; thus the intervention may be a treatment, whereas for the other half, it could be a prevention strategy)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "randomisation of clinics was done by a statistician as independent researcher in the study" Judgement comment: 8 clinics in the area studied (Western Cape Province, South Africa) were randomised by a statistician independent from the rest of the study team
Allocation concealment (selection bias)	Low risk	Quote: "four clinics were randomised to each arm of the intervention" Judgement comment: allocation of clinics to intervention or control was performed at the beginning of the study by an independent statistician
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding was not performed
Blinding of outcome assessment (detection bias) all outcomes	High risk	Blinding was not performed
Baseline outcome measurements similar	Low risk	Page 469: "the baseline AUDIT score was not considered different for the analysed and total group (7.3 vs 6.9, Table 1)"
Baseline characteristics similar?	Unclear risk	Groups differed in proportion close to anybody with an alcohol problem, language group Afrikaans, coloured, employed, heard about FAS, and in relationship/not living with partner
Incomplete outcome data (attrition bias) Efficacy data	Low risk	Follow‐up rate for intervention group 99%. Follow‐up rate for control group 85%
Protection against contamination	Low risk	Judgement comment: randomisation by clinics
Selective reporting (reporting bias)	Unclear risk	No available protocol for comparison of intended and reported outcomes
Other bias	Low risk	No other risk of bias identified