Avaliação do impacto da fiscalização eletrônica de velocidade e de avanço semafórico na segurança viária: um estudo antes e depois utilizando o Método Empírico de Bayes

Abstract

Automated enforcement strategies are commonly employed at signalized intersections to reduce crashes. The real impact of these strategies can vary between jurisdictions and requires observational studies that address methodological challenges such as regression to the mean and temporal limitations. This study examines the effects of electronic speed and red-light camera enforcement on traffic safety at signalized intersections using the Empirical Bayes (EB) Method. Analyzing data from 2010 to 2019, a Safety Performance Function (SPF) for 2011 crashes involving injuries was developed based on traffic volume (VDMA) and the number of lanes, using a sample of 176 intersections in Fortaleza. To estimate missing VDMA values, Universal Kriging and expansion factors were applied. The 2011 SPF was adjusted for subsequent years using the intercept correction method proposed in the Highway Safety Manual. Results indicated a 21% reduction (8% to 33%; 95% CI) in crashes with injury and fatality, similar to findings with a comparison group and slightly better than the international literature. This methodology supports the effectiveness of enforcement devices at signalized intersections and improves observational “before” and “after” studies in the Brazilian context.