Dengue is a serious public health problem in Minas Gerais, with incidence influenced by multiple factors. The Brumadinho dam collapse (2019) raised concerns about its health impacts. This study analyzes the dynamics of dengue incidence (2014–2023) in Minas Gerais, comparing “Case” municipalities (Rio Paraopeba, impacted), “Control” municipalities (same health region, non-riparian), and “Other Municipalities”. Data on dengue cases, hospitalizations, and deaths (SINAN, SIH, SIM, 2014–2023) were used. Incidence and hospitalizations were calculated per 1,000 inhabitants and standardized as Z-scores by municipality. Contextual variables (Health Region, Quadrilátero Ferrífero, riparian status, and Study Group) were incorporated. Hierarchical cluster analysis was performed to identify 10 groups with similar epidemiological dynamics (correlation distance). Visualizations included boxplots, heatmaps, and t-SNE. PERMANOVA assessed the influence of these factors on Z-score dispersion. The same analyses were applied to hospitalizations, in which the health region explained 39% of the observed variability. Boxplots revealed annual distributions of dengue Z-scores, differentiating pre- and post-collapse periods and study groups by region and river. Cluster analysis identified 10 distinct epidemiological patterns, visualized by Z-score heatmaps and t-SNE, confirming intra-cluster similarity. PERMANOVA of incidence showed that Health Region (R²=0.453, p<0.001) and Study Group (R²=0.0095, p<0.001) were significant factors. Other variables showed no impact. The hospitalization profile was similar to dengue case records, indicating that the health region is the most important factor, accounting for 39% of the observed variability. Z-score and cluster analyses identified distinct epidemiological patterns of dengue incidence in Minas Gerais. PERMANOVA results indicate that the Health Region is the main determinant of dengue epidemiological dynamics. Although several variables were considered, the impact of the dam collapse on this dynamic, at the municipal and temporal level evaluated, proved irrelevant compared with regional influence. This reinforces the need for dengue control strategies tailored to the realities and heterogeneities of each health region.
Fernandes et al. (Sun,) studied this question.