Schorling 1997.

Methods	Study design: prospective controlled cohort Sampling frame: Street segments were identified from US Census blocks with at least 50% African American residents and at least 10 African American adults Sampling method: Every household on identified street segments was screened for the presence of 1 or more black adults and was rostered for denominator data on smoking prevalence Collection method: in‐person interview Description of the community coalition: The Buckingham Health Education Board was assembled by a county co‐ordinator hired by the academic researchers. Coalition members were African American and included both volunteer lay persons and clergy. Assistance was offered to the coalition to deal with any health issues of concern, with the provision that smoking cessation must be included
Participants	Communities: 2 rural Virginia counties Country: USA Ages included in assessment: 18+ (n = 452) Reasons provided for selection of intervention community: African Americans have higher smoking‐attributable morbidity and mortality than Caucasians, despite data suggesting higher attempted quit rates. In the rural South, a high percentage of the population is African American, and services for smoking cessation are scarce Intervention community (population size): Buckingham County, VA (11,926) Comparison community (population size): Louisa County, VA (20,325)
Interventions	Name of intervention: Alliance of Black Churches Health Project Theory: Community Empowerment and Transtheoretical Stages of Change Model Aim: to determine if smoking cessation interventions delivered through a coalition of black churches would increase the smoking cessation rate of church members exposed to the intervention and among African Americans community‐wide Description of costs and resources: none stated Components of the intervention: Smoking cessation counselors were trained by participating churches to deliver advice and counseling to individuals interested in quitting.  Smoking cessation devotional booklets were distributed through churches, county‐wide Gospel Quit Nights were held every 6 months, and annual county‐wide smoking cessation contests and in‐school poster and essay contests were held Start date: 1991 Duration: 18 months
Outcomes	Outcomes and measures: smoking cessation. Measure: self reported continuous abstinence (zero cigarettes smoked in the past month) ascertained by in‐person household survey Time points: baseline: n = 648; follow‐up (18 months later): n = 453
Notes	A separate coalition was organized in the control to address hypertension, diet, and exercise. Smoking was not addressed by any coalition activities in this county Funder: US National Heart, Lung, and Blood Institute Funding source: government
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	From 2 pre‐selected counties, selection of a county to receive smoking cessation intervention was “arbitrary”
Allocation concealment (selection bias)	High risk	No allocation concealment
Baseline outcome measurement similar	Unclear risk	After standardization for age distribution, gender‐specific smoking prevalence was similar between intervention and control counties, but no statistical testing was reported
Baseline characteristics similar	High risk	Most participants in intervention and control counties were similar at baseline, but those in the intervention county started to smoke at a significantly younger age on average than participants in the control county
Blinding of outcome assessment (detection bias) All outcomes	High risk	Not blinded. Presumably, study interviewers were aware of intervention allocation, as it was determined by place of residence
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Thirty percent of participants were lost to follow‐up, and these individuals exhibited significant differences from those remaining in the study. However, the attrition rate was similar between study groups, and, according to study authors, no significant between‐county differences were observed in the characteristics of retained study participants. Study author performed secondary analysis with conservative assumption for losses to follow‐up
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants were not necessarily aware of the coalition intervention
Protection against contamination	High risk	Intervention and control counties were in relatively close proximity. 9.4% of control population reported hearing about the intervention program
Selective reporting (reporting bias)	Low risk	Relevant outcome was reported