Saxe 2006.

Methods	Study design: controlled cross‐sectional before‐after study Sampling frame: adults age 16 to 44 years and living in experimental and control communities Sampling method: random‐digit dial population survey Collection method: telephone interview Description of the community coalition: Robert Wood Johnson Foundation’s Fighting Back program to prevent and control drug and alcohol abuse. Community coalitions were established at each site and involved local political, business, and grassroots leaders. Community leaders chose, developed, and implemented strategies with autonomy, thus sites differed in their approaches
Participants	Communities: 41 communities in sections of the following US cities: Washington, DC, Santa Barbara, Vallejo, Little Rock, New Haven, San Antonio, Kansas City, Milwaukee, Columbia, Charlotte, Newark, and Worcester. On average, sites were more urban, more African American, and poorer than the USA at large Country: USA Ages included in assessment: 16 to 44 years old (n = 2804) Reasons provided for selection of intervention community: Targeted neighborhoods were economically disadvantaged with higher rates of drug and alcohol problems Intervention community (population size): Area represented by each coalition ranged from 100,000 to 250,000 residents, typically portions of cities Comparison community (population size): communities similar in size and demographics within the same city as the intervention community
Interventions	Name of intervention: Fighting Back Theory: not reported Aim: to develop a comprehensive system of prevention, treatment, and aftercare for substance abuse Description of costs and resources: Communities received, on average, US $3 million over a 5‐year implementation period to bring coalition groups together, hire staff, and develop a plan for interventions Components of the intervention: broad‐based community initiatives, including public awareness, prevention targeted at youth, early identification of substance abuse problems, and treatment and relapse prevention  Start date: 1995 Duration: 4 years
Outcomes	Outcomes and measures: use of any illicit drug, heroin use within 12 months, daily marijuana use Time points: baseline (1995) and follow‐up (4 years later)
Notes	Funding source: government
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Intervention was not randomly assigned
Allocation concealment (selection bias)	High risk	No allocation concealment
Baseline outcome measurement similar	Low risk	Baseline measures were similar in control and intervention groups
Baseline characteristics similar	Low risk	Reported baseline characteristics were similar in control and intervention groups
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Population‐based telephone survey
Incomplete outcome data (attrition bias) All outcomes	Low risk	Random‐digit dial population survey with large sample size
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding
Protection against contamination	Unclear risk	No statement regarding possibility for contamination
Selective reporting (reporting bias)	Low risk	Relevant outcomes reported