Variable Association of Ridesharing Application Implementation with Total and Alcohol-Related Driving Crashes in Florida

Abstract

INTRODUCTION: Alcohol-involved driving crashes are a major public health concern in the US. Ridesharing applications have been highlighted as a commuter service in metropolitan areas that could potentially curb alcohol-related driving, although evidence is mixed. We hypothesize that citywide ridesharing implementation is associated with decreased alcohol-related crashes in Florida. METHODS: Uber and Lyft were implemented in 8 cities in Florida in 2014. Total crash and intoxicated crash data were collected from the 2011 to 2020 Signal4Analytics traffic safety dashboard. An interrupted time series analysis at the introduction of Uber and Lyft was performed using a generalized estimating equation and a Poisson distribution. RESULTS: There were 258,322 total crashes (114,265 before and 144,057 after 2014) and 3,061 intoxicated crashes (1,708 before and 1,353 after 2014) included for analysis (Figure). For all 8 cities combined, the interrupted time series analysis for intoxicated crashes found no significant increase or decrease in crashes; total crashes increased after rideshare implementation (year odds ratio [OR] 1.20, 95% CI 1.12 to 1.27, p < 0.001; rideshare OR 2.38, 95% CI 1.88 to 3.02, p < 0.001; rideshare:year OR 0.82, 95% CI 0.77 to 0.87, p < 0.001). However, when Miami was analyzed separately, rideshare implementation was associated with significantly decreased intoxicated crashes (year OR 0.90, p < 0.001; rideshare OR 0.90, p < 0.001; rideshare:year OR 1.03, p < 0.001), but significantly increased total crashes (year OR 1.16, p < 0.001; rideshare OR 1.99, p < 0.001; rideshare:year OR 0.85, p < 0.001]. CONCLUSION: Citywide rideshare implementation in Florida is associated with variable effects on both total and alcohol-related crashes. Additional research is required to explore the local factors that influence the effect of rideshare implementation on public safety.