Statement of purpose To estimate the impact of Complete Streets policies (requiring streets to be designed for all users) on the prevalence of bicycling and on bicyclist fatality rates.

Approach We merged Smart Growth America county-level data on Complete Streets policies with Fatality Analysis Reporting System counts of adult bicyclist fatalities due to motor vehicle collision during the period 2000 through 2015. For each county at each year, we estimated the size of the bicycling population using commute data from the American Community Survey and US Census. Based on prior research, we estimated that each bicycle commuter accounted for 4.7 adult bicyclists. To estimate the impact of policies, we used the parametric g-formula, an approach that uses observations from counties without policies to estimate what would have been observed in the absence of policies. To avoid spurious results due to sparse data, we limited analyses to 89 counties in the US where at least 1000 residents reported commuting to work by bicycle in the 2000 Census.

Results There were 4649 adult bicyclist fatalities over an estimated 27.2 million bicyclist-years in these 89 counties, or about 204 fatalities per 1 00 000 bicyclist-years. Complete Streets policies were associated with 5% (95% CI: 1%, 9%) greater adult bicycling prevalence, and about 20 (95% CI: 11, 29) fatalities per 1 00 000 adult bicyclist-years.

Conclusions Adoption of Complete Streets policies may modestly encourage adult bicycling and decrease per-bicyclist fatality risk among adults.

Significance Complete Streets policies appear to make bicycling safer for adults. Further investigation of the impact of Complete Streets policies is warranted, including exploring different Complete Streets policy formulations, acquiring more precise counts of the bicycling population (particularly the child bicyclist population), and determining whether bicyclist fatalities occurred on or near re-engineered streets.