Liao 2010a.

Methods	Study design: controlled before‐after Sampling frame: combination of banks of telephone numbers and listed telephones (target group surnames) Sampling method: random digit Collection method: telephone survey (in‐person survey in Lowell, MA) (n = 98,206) Description of the community coalition: Study authors state that each coalition comprised a community‐based organization and at least 3 other organizations, of which 1 was a local or state health department, university, or research organization
Participants	Communities: Los Angeles County and Orange County, CA; Santa Clara County, CA; Lowell, MA; and King County, WA Country: USA Ages included in assessment: 18 and older Reasons provided for selection of intervention community: communities with large populations of Asian Americans Intervention community (population size): unknown Comparison community (population size): unknown
Interventions	Name of intervention: Racial and Ethnic Approaches to Community Health (REACH) Theory: Social‐Ecologic Model Aim: to reduce health disparities among targeted groups (reduction of prevalence of current smoking) Description of costs and resources: unknown Components of the intervention: Interventions included health communications campaigns and health education and promotion programs and varied among communities Start date: 2002 Duration: 60 months
Outcomes	Outcomes and measures: prevalence of current smoking from survey data (logistic regression used to examine trends) Time points: annual risk factor surveys (2002 to 2006)
Notes	Funding source: government
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Not randomized
Allocation concealment (selection bias)	High risk	Not randomized
Baseline outcome measurement similar	Low risk	Same survey
Baseline characteristics similar	High risk	Baseline differences noted
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Population‐based survey
Incomplete outcome data (attrition bias) All outcomes	Low risk	Adequately reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Telephone survey respondents were not necessarily aware of intervention, but data collectors were aware
Protection against contamination	Low risk	Intervention sites in different cities
Selective reporting (reporting bias)	Low risk	Relevant outcomes reported