Characteristics and incidence rates of self-reported bicycle crashes in Delhi: a cohort study.

BACKGROUND: High levels of underreporting in non-fatal crashes limit our understanding of bicycling safety. We examined self-reported bicycle crashes and quantified their risks in a cohort of cyclists in Delhi. METHODS: Using on-road intercept surveys of bicyclists (n=844) at locations selected through spatially stratified sampling of the road network, we recorded self-reported bicycle crashes using both retrospective (14-month recall) and prospective (monthly follow-up) cohort designs and compared those with police-reported fatal and non-fatal cases. Using prospectively collected data, we estimated overall and factor-specific exposure-adjusted incidence rates. RESULTS: Of the 218 self-reported crashes, 3% were reported to the police, half involved collisions with motorcycles and <2% were single-bicycle crashes. Police-reported crashes were more likely to involve heavy vehicles and severe injuries. The rate of crashes resulting in a medically attended injury was 16.8 per million km (95% CI 11.1, 23.2) and its ratio to those resulting in any injury and to those including no injury was 1:1.7:2.6. Crash rates declined with cyclists' age though differences lacked statistical significance. DISCUSSION AND CONCLUSIONS: The near-absence of single-bicycle crashes in our study, compared with their large share (40-70%) in high-income contexts, may be attributable to low cycling speeds resulting from the widespread use of heavy-duty roadster bicycles. The injury rate (28.6 per million km cycled) was comparable to those reported in Australia, Norway and the USA. In contrast, it was 60-200 times higher than in Denmark, Germany and the Netherlands, countries with extensive cycling infrastructure, supporting the critical role of infrastructure improvements for reducing bicyclist injuries.