Anthropogenic landscapes and vector-borne disease dynamics: Unveiling the complex interplay between Human Footprint and disease transmission in Colombia.

Anthropogenic landscape transformations are fundamentally reshaping the epidemiology of vector-borne diseases (VBDs), yet their causal impacts remain poorly quantified across diverse ecological and socio-economic contexts. This study evaluates the causal effect of the Human Footprint (HFP) on the transmission dynamics of malaria, dengue, and visceral leishmaniasis in Colombia. We conducted an ecological analysis using municipal-level retrospective data from Colombia's National Public Health Surveillance System (SIVIGILA, 2007-2019). Epidemiological records were integrated with environmental and socio-economic indicators, and a Double Machine Learning (DML) framework was applied to estimate the Average Treatment Effect (ATE) and Conditional Average Treatment Effect (CATE) of HFP on excess disease cases. Model robustness was assessed through refutation tests and non-parametric sensitivity analyses for unmeasured confounding. A one-standard-deviation increase in HFP significantly reduced the probability of excess malaria cases by 7.6 percentage points (ATE = -0.076, 95% CI: -0.094 - -0.058), an effect that was more pronounced in socio-economically deprived municipalities and modulated by temperature and precipitation gradients. Conversely, the ATE for dengue and visceral leishmaniasis was not different from zero. Robustness tests suggest the presence of residual bias, but the sensitivity test points to a low plausibility of an unobserved confounder. These findings underscore the necessity of integrating HFP monitoring into public health planning to design context-specific, multi-sectoral interventions that address the evolving landscape of VBD risk in rapidly transforming regions.