Design effects of cycle infrastructure changes: An exploratory analysis of cycle levels

Abstract

Introduction: Cycling infrastructure policies yield health and environmental benefits. However, design variations may impact intervention effectiveness. This study explored relationships between infrastructure form (e.g., width, length) and function (e.g., access, safety) and cycling behavior. Methods: We analysed data from a study evaluating 15 streets with new cycling infrastructure improvements in Paris and Lyon. Comparison streets were chosen based on having similar pre-intervention cycling trends and periods of data availability as intervention streets. The outcome was the difference in daily cycle counts between intervention and control streets. We conducted a virtual street audit with Google Street View to assess 14 street features and derived a function score based on 7 components. We used induced smoothed LASSO regression to identify relevant form or function variables associated with cycle count changes. Results: The effects of new cycling infrastructure varied, with significant increases at half (7/15) of the sites and no change at the rest. For every 1 SD increase in cycle lane length, an increase of 83 cycle counts per day (95% CI 32, 134) was observed. Removing car parking and traffic lanes were associated with an increase of 197 counts (108, 285) and 154 counts (58, 249), respectively. Adding a public transport stop showed a negative association with a change of −83 counts (−158, −8). Functions positively associated with cycle count changes were improving safety (75, 95% CI 8, 141) and space (72, 95% CI 10, 135). Conclusions: This study sheds light on cycling infrastructure design influences on cycling behavior. These insights can help guide future policies and infrastructure development to maximize cycling benefits for health and the environment.

Keywords

• Travel behaviour
• Routinely collected data
• Difference-in-differences
• Virtual street audit
• Cycling infrastructure design