Moskowitz 2007.

Methods	Study design: controlled before‐after study (propensity score adjusted) Sampling frame: Korean surname‐based telephone lists Sampling Method: random; women age > 50 years oversampled during phase 2 Collection method: random telephone surveys in Korean and English administered by trained bilingual Korean American interviewers Description of the community coalition: collaboration between UC Berkeley Center for Family and Community Health, which is a CDC Prevention Research Center, Asian Health Services (AHS), a community clinic that provides primary care to indigent, limited‐English‐proficient Asian‐American immigrants who reside in the county, and the local Korean American Community Advisory Board (KCAB), which comprises members and leaders of the community
Participants	Communities: women affiliated with Korean churches in Alameda County and Santa Clara County, California Country: USA Ages included in assessment: women > 18 years old (n = 876); for mammogram assessment women > 50 years old (n = 419) Reasons provided for selection of intervention community: low rates of Pap screening and mammography use among Asian Pacific Islander women in California Intervention community (population size): Alameda County (Korean American population: 14,200 in 2000) Comparison community (population size): Santa Clara County (Korean American population: 21,600 in 2000)
Interventions	Name of intervention: Health Is Strength Theory: Precede‐Proceed model, Community Sensitive Research Aim: to improve breast and cervical cancer screening among Korean American women Components of the intervention: educational workshops; materials in Korean language, delivered by Korean American social worker and nurse; adaptation of American Cancer Society  “Tell a Friend” program; financial incentives; volunteer Korean lay health advisors/church members used for recruitment and monitoring/reinforcement of health behaviors. Brochure and access‐resource lists distributed, targeted media campaign Start date: 1994 Duration: 48 months — see notes
Outcomes	Outcomes and measures: self reported breast and cervical cancer screening change over time: Pap test, breast self exam, mammogram, clinical breast exam Time points: pre‐intervention (1994) and post intervention (2002)
Notes	Intervention duration calculated as the sum of 3 separate phases, which occurred between March 1996 and January 2002 Funding source: government
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	No randomization
Allocation concealment (selection bias)	High risk	No allocation concealment
Baseline outcome measurement similar	High risk	Mammogram and clinical breast exam rates significantly higher in intervention group at baseline after propensity score adjustment for other characteristics; change scores compared
Baseline characteristics similar	Unclear risk	At pre‐intervention, “women in the two counties differed significantly on 6 of 12 sociodemographic and health care access measures”; adjustments made through propensity score analysis
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Random‐digit telephone survey
Incomplete outcome data (attrition bias) All outcomes	Low risk	Independent samples, response rate similar over time
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding reported, exposure to intervention assessed
Protection against contamination	High risk	Study authors state: “by 2002, 36% of the comparison community had some awareness of, or participation in, our community intervention”
Selective reporting (reporting bias)	Low risk	All relevant outcomes reported