Biogeographic regionalization of human infectious diseases in Brazil based on geographically explicit data

Objective: Biogeographic regionalization represents abstractions of the organisation of life on Earth, and can provide a large-scaled framework for health management and planning. We aimed at determining a biogeographic regionalization for human infectious diseases in Brazil, and at investigating non-mutually exclusive hypotheses predicting the observed regions. Methods: Based on the spatial distributions of 12 infectious diseases with mandatory notification (SINAN database, 2007-2020, n = 15,839), we identified regions through a clustering procedure based on beta-diversity turnover. The analysis was repeated 1000 times by randomly shuffling the rows (0.5 degrees cells) in the original matrix. We evaluated the relative importance of variables using multinomial logistic regression models: contemporary climate (temperature and precipitation), human activity (population density and geographic accessibility), land cover (11 classes), and the full model (all variables). We refined the geographic boundaries of each cluster by polygonising their kernel densities to identify clusters' core zones. Results: The two-cluster solution showed the best correspondence between disease ranges and clusters geographic limits. The largest cluster occurred with more density in the central and northeastern regions, while the smaller and complementary cluster occurred in the south and southeastern region. The best model for explaining the regionalization was the full model, supporting the 'complex association hypothesis'. The heatmap showed a NE-S directional display of the cluster's densities, and core zones showed geographic correspondence with tropical + arid (NE) versus temperate (S) climates. Conclusion: Our findings indicate that there is a discernible latitudinal pattern in the turnover of disease in Brazil, and this phenomenon is associated with an intricate interplay between contemporary climate, population activity, and land cover. This generalised biogeographic pattern may offer the earliest insights into the geographic arrangement of diseases in the country. We suggested that the latitudinal pattern could be adopted as a nationwide framework for geographic vaccine allocation.