Davidson 1994.

Methods	Study design: interrupted time series with control group Sampling frame: Central Harlem, NYC, USA Sampling method: entire population (5‐ through 16‐year‐olds) Collection method: hospital and death records Description of the community coalition: The Safe Kids/Healthy Neighborhoods Coalition was initiated by the Harlem Hospital Injury Prevention Program in response to a request from parents and educators for a program in playground safety. The original coalition consisted of hospital, public health, public school, law enforcement, fire and emergency medical service, municipal government, parents, and community volunteers. Initial funding for the coalition was received from Robert Wood Johnson
Participants	Communities: low‐income, predominantly African American Central Harlem and predominantly Hispanic Washington Heights Districts of New York City Country: USA Ages included in assessment: 5 through 16 Reasons provided for selection of intervention community: high risk of serious injury for youth in Central Harlem Intervention community (population size): 19,254 (5 to 16 years —1990 census) Comparison community (population size): 44,535 (5 to 16 years — 1990 census)
Interventions	Name of intervention: Safe Kids/Healthy Neighborhoods Theory: not reported Aim: to reduce the rate of serious injury for 5‐ to 16‐year‐olds Description of costs and resources: cost unknown, funding from RWJ, CDC, and municipal government Components of the intervention: improvement to parks and playgrounds, playground supervision, traffic and bicycle safety education, bicycle helmets, activities for youth including sports, arts, and dance Start date: 1989 Duration: 36 months
Outcomes	Outcomes and measures: risk ratio for serious injury after intervention compared with before the intervention Time points: pre‐intervention (1983 to 1988), intervention (1989 to 1991)
Notes	Outcomes measured at population level using vital statistics records and hospital records Funding source: Robert Wood Johnson Foundation, Centers for Disease Ccontrol and Prevention, and municipal government
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk
Allocation concealment (selection bias)	Low risk
Baseline outcome measurement similar	High risk	EPOC criteria for ITS study designs Independent of other change: High‐risk injury defines trends at both intervention and control sites
Baseline characteristics similar	Low risk	Point of analysis at intervention: low risk — annual surveillance data collected
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Intervention effect on data collection: low risk — population‐based injury surveillance system
Incomplete outcome data (attrition bias) All outcomes	Low risk	Incomplete outcome data: low risk — population‐based injury surveillance system
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Intervention effect on data collection: low risk — population‐based injury surveillance system
Protection against contamination	High risk	Independent of other change: high risk — injury defines trends at both intervention and control sites
Selective reporting (reporting bias)	Low risk	Selective reporting: low risk — population‐based injury surveillance system