642 Prevalence of behavioural risk factors for road traffic injuries: regional differences in Brazil

Background Road traffic injuries (RTIs) are estimated to be the fourth leading cause of death in Brazil. This burden is not uniform throughout the country with differences between the wealthy developed South and less-developed North. Understanding regional differences in specific behaviours (speeding, helmet use, seatbelt use and drink driving) associated with RTIs is critical in designing effective RTI interventions. This study aimed to assess such differences by comparing observed road safety behaviours in the cities of Sao Paulo (South) and Fortaleza (North). Methods As part of the Bloomberg Initiative for Global Road Safety (BIGRS), observational studies were conducted in Sao Paulo and Fortaleza, Brazil. Six sites were randomly selected in each city and trained observers recorded road safety behaviour (speeding, helmet use, seatbelt use and drink driving) for three full-day periods between August–October 2015. Results 190,730 observations were collected: Sao Paulo (113,539), Fortaleza (77,191). Fortaleza had higher rates of speeding (16.7%) with motorcycles being the most common speeders compared to Sao Paulo (8.1%) where sedans were most responsible. Rates of seatbelt and helmet use among all passengers were higher in Sao Paulo (90.2% and 99.1%, respectively) than Fortaleza (63.4% and 82.8%, respectively). Drivers testing positive for any alcohol was higher in Sao Paulo (14%) than Fortaleza (1.2%). Conclusions Important regional differences in road safety behaviour highlight the need for tailored interventions in Brazil. National-level interventions might not be as effective as regional ones. While both cities had high rates of helmet use, seatbelt use remains low in Fortaleza. Speeding, especially among motorcycles, is a priority for Fortaleza and parts of Northern Brazil. Sao Paulo’s alcohol results illustrate the urgency for drink driving-based interventions for areas of Southern Brazil. These results will be used to develop specific interventions for BIGRS.