Spatio-temporal analysis of violence notifications among the transgender population in Brazil (2015–2023)
摘要
To analyze notifications of violence involving the transgender population in Brazil between 2015 and 2023 from a spatio-temporal perspective, characterizing victim profiles and the types of violence perpetrated. An observational ecological study with a spatio-temporal analysis based on notifications from the Notifiable Diseases Information System (SINAN), aggregated by Federated Units and for Brazil as a whole. Hierarchical Bayesian modeling (Poisson regression) with spatial and temporal random effects was performed via Integrated Nested Laplace Approximations (INLA) to estimate relative risks across Federated Units. There were 45,715 recorded notifications. Victims were predominantly transgender women (65.3%), heterosexual (53.2%), young adults (36.1%), mixed-race/parda (44.6%), and individuals with low educational attainment (23.6%). Physical violence was the primary form (65.9%). While the fully adjusted annual relative risks clustered close to unity, indicating that regional variations are structurally captured by the baseline spatial risk and the national temporal trend, the higher concentration of rates in the South and Southeast macro-regions demonstrates that the distribution of violence is structurally driven rather than random. Violence against transgender people in Brazil exhibits structured spatio-temporal patterns linked to persistent social vulnerability and discrimination. These findings highlight the urgency of strengthening health surveillance and implementing targeted intersectoral policies, particularly across education, employment, and public security, to protect the most vulnerable profiles, notably young transgender women with low educational attainment.