Morbidity and mortality due to traffic accidents in the city of São Paulo

• Grant number: 14/17441-2
• Support Opportunities: Regular Research Grants
• Start date: March 01, 2015
• End date: August 31, 2017
• Field of knowledge: Health Sciences - Collective Health - Public Health
• Principal Investigator: Gleice Margarete de Souza Conceição
• Grantee: Gleice Margarete de Souza Conceição
• Host Institution: Faculdade de Saúde Pública (FSP). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated researchers: Gizelton Pereira Alencar ; Maria Do Rosario Dias de Oliveira Latorre

Abstract

Three thousand people die every day in land transport accidents (LTA) worldwide-nearly 1.3 million people a year. In 1997 Brazil's road traffic was considered one of the worst and most dangerous in the world, with extremely high accident rates, one for every 410 vehicles on road. In 2010 the crude mortality rate due to LTA in Brazil was 22.5 deaths per 100.000 inhabitants. São Paulo is the most populous city in the country, with nearly 11.8 million inhabitants and a fleet of more than 7.5 million vehicles. Its population growth and economic development were not followed by the necessary improvements in infrastructure, and as a side effect, the issues related to urban mobility reached a critical level, impairing the quality of life. There are no recent studies describing morbidity and mortality from LTA in São Paulo that identify the subpopulations at higher risks and its characteristics. Objectives: This study aims to describe the current profile of morbidity and mortality due to LTA in São Paulo and its evolution over the past few years, identify subpopulations at risk and to evaluate the spatial distribution of mortality in the city. Subpopulations of interest will be determined by the condition of the victim (all, pedestrian, motorcyclist, vehicle occupant, or unspecified), gender, age, race, education and place of residence. It also aims to estimate the impact of the adoption of security measures in traffic in both morbidity and mortality, as in the spatial distribution of mortality. Methods: This is a longitudinal ecological study. Morbidity data for the period 1998-2012, including age and gender, will be obtained from the Hospital Information System of the Unified Health System (SIH / SUS). Mortality data for the period 1991-2012, including age and gender, will be obtained from the municipal mortality information improvement program (PRO-AIM). Socioeconomic information (race, education level and place of residence of the victim) are available since 1998. The International Classification of Diseases 9th revision, ICD-9 (up to 1995) and ICD-10 (from 1996) will be used In order to classify each death and hospitalization based on the underlying cause in one of five groups namely "condition of victim": 1-All, 2-Pedestrian, 3-Motorcyclist, 4-Vehicle occupant, 5-Unspecified. Statistical analysis: Annual mortality and morbidity rates and proportional distributions of the number of deaths and hospitalizations for each subpopulation of interest will be calculated. Evaluation of mortality and morbidity trends over time will be performed using Generalized Linear Models (GLM). The response variables are the annual number of hospitalizations and the annual number of deaths from LTA. The explanatory variables are year, condition of the victim, sex, age and number of inhabitants, as well as indicator variables for seat belt, traffic restriction of motorcycle, truck traffic restriction and prohibition. The variables race, education and place of residence are also included as explanatory variables in mortality models for the period in which those data are available. Three types of models will be considered: the Classic Linear, the Poisson and the Negative Binomial Response. For spatial analysis, it will be obtained thematic maps and Moran Index; for the variables with cluster pattern, it will be calculated the Local Spatial Association (LISA); regression models are estimated using the rates as dependent variable and the socioeconomic indicators as independents variables. (AU)