Background The City of Edinburgh implemented 20mph speed limits on most of the streets (an increase from 50% to 80%) in the city between April 2016 and March 2018. This paper is part of the research undertaken by the ‘Is Twenty Plenty for Health?’ project team which examines the impact of the 20mph speed limit policy in the City of Edinburgh. This paper assesses, specifically, the impact of the speed limits on the average vehicular speed and volume in the City, and the road traffic collision and casualty rates. Police recorded data (road traffic collisions and casualties), and the City of Edinburgh’s street survey data are used in this study.

Methods In this paper we calculate vehicular speed and volume summaries, and the average annual rate for collisions and casualties for selected subgroups. In addition, we calculate the percentage difference in annual rates before and after the 20mph speed limits (un-adjusted and adjusted for secular trend) for various street categories. The observed reduction in collisions (overall) is compared with that predicted from Elvik’s power model.

Finally, we construct a regression model with key explanatory variables treatment (implementation zone vs control zone), time (adhering to the stepped wedge nature of the implementation), and an interaction term constructed from the two latter variables. The dependent variable is the number of collisions (and separately, casualties) and the model coefficient of the interaction term is the ‘difference in differences’ estimator.

Results Our results show that there was a statistically significant reduction of in average speed in the City (1.34mph) since the speed limit implementation and that there was no significant change in vehicular volume. We identified a reduction in road traffic collisions post speed limit implementation and a reduction in road traffic casualties in key vulnerable groups such as children (a provisional, un-adjusted 39% reduction in casualty rate).

Conclusion In this paper, we reveal the changes in vehicular speed and volume post 20mph speed limits and also present changes in road traffic collisions and casualties. In addition we apply statistical methods to evaluate the impact of the speed limits in a natural experiment context. This study is important not only to public health and city planning officials, but to academics involved in natural experiment evaluations.