Ali 2003.

Study characteristics
Methods	Study design: RCT Duration of study: baseline survey January to April 2001 and considering the 8‐week intervention must be provided between May and June‐July 2001
Participants	Country: Pakistan Income classification: low income Geographical scope: semi‐urban: in Qayoomabad, lower middle class semi‐urban community with a population of 80,000 in Karachi Healthcare setting: home Mental health condition: common mental disorders Population: adults 1. Age: 18 to 50 years 2. Gender: female 3. Socioeconomic background: lower‐middle class. Women predominantly aged 26 to 40 years, half with no formal education, not involved in revenue generation, two‐thirds with household income > 3000 PKR, nearly 60% residing longer than 10 years 4. Inclusion criteria a. Women 18 to 50 years old b. Able to communicate in Urdu c. Planning to live in the study area longer than 1 year d. No bereavement in past 6 weeks e. Identified as anxious and/or depressed based on screening with Aga Khan University Anxiety and Depression Scale 5. Exclusion criteria a. Participant women b. Those actively suicidal
Interventions	Stated purpose: to assess effects on levels of anxiety or depression (or both), among women who had attended counselling sessions, provided by briefly trained counsellors in their own community INTERVENTION (n = 70) Name: counselling Delivered by: LHW Title/name of PW and number: minimally trained counsellors ‐ 21 1. Selection: "women were informed by word of mouth and by leaflets; out of 73 women who came for interview, 21 selected based on communications skills, motivation, attitude, ability to read and write Urdu and freedom to move in the community" 2. Educational background: "ability to read and write Urdu" and belonging to local community 3. Training: 11 training sessions held over 4 weeks. Each lesson lasted 3 hours and was led by family practitioner, sociologist, psychiatrist, or 3 clinical psychologists a. Contents: basic information regarding anxiety, depression, stress/anger management, and communication/counselling skills. Communication covered active listening, probing, and feedback, whereas counselling dealt with supportive problem‐solving and cognitive‐behavioural techniques. "Manual incorporating the training material is being published and is planned to train master training who could replicate the study in several urban and rural centres" b. Supervision: "women had ready access to members of the training team throughout the study period" c. Incentives/remuneration: not specified Intervention details 1. Duration/frequency: 8 sessions (?possibly weekly). Supportive, cognitive, and problem‐solving counselling was provided at day and time convenient for the woman 2. Content of intervention: trained counsellors provided supportive, cognitive, and problem‐solving counselling at client residence at convenient time CONTROL (n = 91): usual care, no intervention, just had AKUADS administered at baseline and end of study; however, "as the effectiveness of counselling was proved, for ethical reasons the control group was also counselled" possibly at the end of the study CO‐INTERVENTIONS: nil
Outcomes	Patients Reduction in Aga Khan University Anxiety and Depression Scale scores Carers Not applicable Process/health workers Not specified Economic outcomes None Time points: baseline, end of 8 weeks
Notes	Source of funding: academic body; Aga Khan University Research Council Notes on validation of instruments (screening and outcomes): AKUADS (indigenous screening scale, developed from complaints of patients with anxiety/depression, recorded verbatim in Urdu) previously validated against psychiatrist evaluation as gold standard and compared with SRQ Additional information: declarations of interest ‐ none Handling the data: as per footnotes in data and analysis Prospective trial registration number: not specified
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "every third household was systematically sampled in all of Qayoomabad. [...] One woman was randomly chosen from each selected household and screened for anxiety and/or depression. [...] Using computer‐generated random numbers, 216 [of 1218 women] cases were randomised to the intervention and 150 to the control group". The initial selection was quasi‐random, but then allocation to control or intervention was random
Allocation concealment (selection bias)	Unclear risk	Quote: "computer‐generated random numbers" Comment: even though sequence generation was centrally done, it was unclear how allocation was concealed
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Comment: not able to blind participants or personnel. Unlikely to influence outcomes
Blinding of outcome assessment (detection bias) all outcomes	Low risk	No selective reporting. Independent data collectors blind to allocation and to previous scores
Baseline outcome measurements similar	Low risk	Comment: yes, similar, both across intervention and control, and between dropouts and non‐dropouts
Baseline characteristics similar?	Low risk	Comment: yes, similar. All P values over 0.2 comparing dropouts vs non‐dropouts and intervention vs control groups
Incomplete outcome data (attrition bias) Efficacy data	High risk	Comment:intervention: 68% dropout between baseline and those completing the intervention; control: 33% dropout. Although characteristics are similar between dropouts and non‐dropouts (including baseline scores), scores may have been different at follow‐up
Protection against contamination	Low risk	Quote: "the spontaneous decrease in the score [in the control group] could be attributed to the natural history of depression, which waxes and wanes, but a contaminant effect of counselling cannot be ruled out"; "the effect of summer holidays occurring during the study period was also considered as possibly causing contamination"
Selective reporting (reporting bias)	Unclear risk	Comment: no selective reporting. but no protocol to assess if this is the case
Other bias	Low risk	Comment: none detected