Neglected tropical diseases risk correlates with poverty and early ecosystem destruction

Poverty and deforestation are usually treated as two separate crises — one a failure of economic development and the other an environmental emergency. However, for nine tropical diseases that kill and disable millions of people across Brazil, Ramalho Magalhães and colleagues have shown they are the same problem. Adding poverty data to a disease-risk model improves its accuracy by up to 18 percent. That number is the whole argument in compressed form. These are diseases that the global health system has historically deprioritized. Neglected tropical diseases — including dengue, malaria, Chagas disease, leishmaniasis, Brazilian spotted fever, schistosomiasis, leptospirosis, and hantavirus — affect at least one point seventy-four billion people worldwide. They cause chronic, debilitating illness, carry stigma, and hit low-income households hardest. Several are zoonoses, meaning they jump between animals and people: hantavirus spreads through wild rodent excreta, while Brazilian spotted fever spreads through tick bites from reservoirs like capybaras and cattle. Others are water or vector-borne. All of them share a social footprint, disproportionately burdening communities that cannot afford to escape them. Brazil is a particularly revealing case study because these communities are distributed across an enormous country undergoing rapid, uneven deforestation — which means the social and environmental drivers of disease are visible at a continental scale. To map that risk, Magalhães and colleagues turned to ecological niche modeling. The basic idea is that diseases, like animals, occupy a habitat — a set of conditions that make sustained transmission possible — and that habitat can be mapped. The team assembled seven hundred twenty-three thousand one hundred nine confirmed case records across all nine diseases from 2007 to 2018, assigned them to municipalities, and standardized everything onto an approximately eighteen-kilometer grid. They built two sets of models and compared them head-to-head. The first used only environmental predictors: mean annual temperature, annual precipitation, and crucially, loss of natural vegetation cover between 2008 and 2018. The second added socioeconomic predictors, including mean gross domestic product per municipality, the Gini index for income inequality, and sanitation measures such as the percentage of households without toilets or with piped water. Multiple modeling algorithms — including random forest, Maxent, and support vector machines — were run with spatial cross-validation to produce ensemble consensus maps. Model quality was evaluated with a partial receiver operating characteristic metric, where values above one indicate better-than-random performance. The central question was simple: what do you miss when you leave human poverty out of the equation? The answer, it turns out, is a lot. Across the nine diseases, adding socioeconomic variables improved model accuracy by ten percent on average, and that improvement was statistically significant with a p-value below zero point zero one. The gains were not uniform — dengue showed the largest jump, at eighteen point eight percent, followed by leptospirosis at sixteen percent and malaria at fourteen percent. Brazilian spotted fever was the one exception, where adding socioeconomic predictors slightly reduced predictability. This result is noted by the authors but not fully explained, given the disease's distinct epidemiological profile. The single most important socioeconomic variable was gross domestic product, or GDP. Across composite models, GDP carried a mean relative variable importance of thirty-seven percent — relative variable importance simply means how much the model's predictions depend on that variable compared to all the others. For seven of the nine diseases, GDP was the top socioeconomic predictor. Its relationship with disease was inverse: poorer municipalities showed higher modeled disease probability. That pattern was strongest for Chagas disease and hantavirus. The logic isn't just statistical. Poor household construction is linked to triatomine bug infestations, the insects that transmit Chagas. Lack of piped water and toilets drives schistosomiasis risk. Poverty creates the material conditions that allow transmission to occur. Dengue was the notable exception — in that case, higher-GDP municipalities showed higher risk because dengue sustains itself in dense urban environments regardless of income. That inversion is worth noting. It indicates these diseases have distinct ecologies, even if poverty shapes most of them in the same direction. The environmental aspect of the story is equally counterintuitive. In environment-only models, loss of natural vegetation cover between 2008 and 2018 was the dominant predictor, with a mean relative variable importance of forty-two percent. However, the relationship ran in a direction many listeners might not expect. Disease probability was highest not in untouched wilderness, but in places at early stages of ecosystem loss — municipalities that had recently started losing native cover. Places where destruction was already advanced showed lower modeled risk. This is the edge effect. Active deforestation and landscape fragmentation bring natural transmission cycles into new contact with people. The authors provide a specific example: during deforestation, the malaria vector Anopheles cruzii shifts from the forest canopy down to ground level, increasing malaria exposure for both primates and humans. It is the frontier that poses danger, not the deep jungle. Spillover — when a pathogen jumps from its usual animal host into humans — peaks at the boundary between what is being cleared and what remains. That ecological dynamic and the socioeconomic one do not sit in parallel. They interact. Magalhães and colleagues find that disease probability is highest when both GDP and ecosystem destruction are low simultaneously — meaning poor communities at the frontier of early deforestation are the most exposed. Poverty places people in those areas and strips away the sanitation and housing quality that might buffer transmission once they are there. The authors describe this explicitly as a poverty trap: chronic disease burden reduces economic productivity, which deepens poverty, which sustains the conditions that keep disease risk high across generations. That cycle can persist indefinitely without deliberate intervention. The researchers frame their findings within what's called the One Health approach — the recognition that human health, animal health, and ecosystem health are inseparable. One Health proposes coordinated monitoring across species and environments, integrating agriculture, health, and environmental sectors. What Magalhães and colleagues add to that framework is a quantitative demonstration that socioeconomic conditions belong in the same model as forest cover and rainfall. Not as a soft add-on, but as a variable that explains thirty-seven percent of the predictive signal. The practical stakes of that demonstration are real. A risk map that ignores poverty doesn't just perform worse statistically — it misplaces apparent risk areas, which misdirects surveillance and the allocation of scarce public health resources. The authors note that a dengue hospitalization alone can last between fourteen and eighteen days and cost between five hundred fourteen and one thousand five hundred US dollars per patient. When the map is wrong, the money and personnel go to the wrong places. When the map is correct, interventions — including prevention, treatment, and social support — can be concentrated where they will make a real difference in outcomes. Magalhães and colleagues conclude by naming two continental priorities: deforestation suppression and socioeconomic improvement. Not as background conditions for health policy to operate within, but as active levers for directly reducing disease transmission. Their evidence supports that framing. The model accuracy improvement from including GDP is not merely a statistical curiosity; it implies that reducing poverty would measurably decrease the geographic footprint of these diseases. Furthermore, controlling deforestation would minimize the spillover events that generate new cases at the frontier. Neither solution is sufficient on its own. The diseases that cluster in poor, forest-edge communities require both approaches — simultaneously, coordinated, at scale. Ultimately, this research offers a more honest picture of what drives neglected tropical diseases. It's not just climate or the insects, ticks, or contaminated water in the abstract. Rather, it is the specific, mappable intersection of economic deprivation and environmental destruction. Treating poverty and deforestation as separate policy problems has not worked for the one point seventy-four billion people living with these diseases. The models show they are one problem — and they illustrate exactly where that problem resides. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.