Background The risk from road infrastructure and traffic is reported as a major barrier to cycling. In low-and-middle-income countries with high levels of traffic injury rates, there is a lack of evidence using study designs with comparison (control) groups to identify interventions to improve the safety of cyclists.

Objective We aimed to identify road infrastructure and traffic characteristics contributing to fatality risk for cyclists using a robust epidemiological, case-control study design.

Methods Our study population is the road locations in the peri-urban areas of New Delhi, India. We used a retrospective case-control study design that compares cases (road sites where a fatal cycle crash occurred) with controls (road sites where no fatal cycle crash occurred). We identified 50 case sites during the three-year period using police reports. To identify control sites, we intercepted and interviewed three cyclists at each case site at the same time of day as the fatal cycle crashes. We recorded the starting point of the respondents’ trips, mapped their routes from the stated origin location to the case site, and randomly selected one control site on each of those routes. We recorded traffic counts, speed, and road infrastructure characteristics at the case and control locations. To estimate effect of road design on the likelihood of fatality risk to cyclists, we developed logistic regression models with cases and controls as binary outcomes.

Results We found that the presence of median U-turn was associated with 3.7 times (95% confidence interval: 1.3, 10.6) greater likelihood of bicycle fatality. This finding is robust to sensitivity analyses. We found that volume of cars and that of bicycles was associated with increased likelihood of bicycle fatality while the volume of motorised two-wheelers had a protective effect.

Conclusions Our results indicate that median U-turns are a major risk factor for bicycle fatalities. As a replacement for signalised mid-block crossings and intersections, median U-turns are being constructed on a large scale in Delhi. The strong evidence of their impact on the safety of bicyclists suggests that their installations should be discontinued, and those already in existence should be reverted to crossings or junctions.