Details
Date: Jun 28, 2026
Location: Pamplona - Spain
Abstract
Tuberculosis (TB) remains a critical public health challenge in Brazil, particularly in the state of Rio Grande do Sul (RS), where incidence rates (46.2 cases per 100,000 inhabitants in 2021) exceed the national average (32.0). Designing effective intervention strategies requires reliable local risk estimates. However, analyzing municipal data is complicated by zero- inflation, small population sizes, and complex spatiotemporal dependencies. Standard disease mapping models typically rely on adjacency-based spatial structures. These approaches overlook the specific geometry of regions, often failing to account for the fact that large, sparsely populated municipalities exert a different spatial influence than small, densely populated ones. To address these limitations, we propose a Generalized Linear Mixed Model to analyze TB incidence across 497 municipalities in RS from 2011 to 2021. We introduce a novel separable spatiotemporal extension of isotropic Gaussian Processes (GP) tailored for areal data. A key innovation is using the spatial covariance through the BallHausdorff distance, which explicitly accounts for the shape and size of the regions. We explicitly establish the validity (positive-definiteness) of these covariance functions. To ensure computational scalability, we implement a nearest-neighbor approximation for the spatial component. This framework yields robust incidence estimates despite zero-heavy data and provides accurate one-year-ahead forecasts, offering policymakers critical tools to target social determinants and allocate resources effectively.