Spatial modeling of homicide mortality in the Northeast region of Brazil

OBJECTIVE: To analyze the spatiotemporal distribution of homicide mortality and association with social determinants of health in the Northeast Region of Brazil. METHODS: Ecological study with spatiotemporal modeling of homicide deaths between 2000 and 2019. Temporal trends were analyzed by segmented linear regression. Crude mortality was calculated and adjusted by smoothing the local empirical Bayesian method and analyzed by the Global/Local Moran Index and spatiotemporal scan statistics. The association between social determinants of health and homicide mortality was performed using multiple linear regression and autoregressive spatial models. RESULTS: 353,089 deaths were recorded. Mortality increased from 2000 to 2019, with an annual increase of 4.37 in males and 3.57 in females. High risk spatial and spatiotemporal clusters were identified in the coastal region of the states. The spatial regression model showed an association with socioeconomic inequalities. CONCLUSIONS: High risk areas for homicides associated with socioeconomic inequality, which should be considered as a priority for designing and investing in public health policies were investigated.