Mosquito and primate ecology predict human risk of yellow fever virus spillover in Brazil

Many (re)emerging infectious diseases in humans arise when pathogens spill over from wildlife or livestock, making accurate prediction of spillover an important public health goal. In the Americas, yellow fever in humans primarily occurs following spillover from non-human primates via mosquitoes. To predict and better understand the conditions limiting yellow fever spillover, we use data from existing studies in combination with remotely-sensed weather and land cover to develop a model based on the mechanisms of spillover. We estimate four risk metrics from hypotheses about the factors limiting spillover including the ecology of mosquito vectors and non-human primate hosts, human susceptibility, and human density. We then test these hypotheses by comparing the risk metrics to the occurrences of human cases of yellow fever in Brazil from 2001 to 2016. We find that risk based only on environmental conditions is the best independent predictor of spillover. Unexpectedly, incorporating information about estimated human vaccine coverage and density did not improve prediction, suggesting that environmental suitability may be more limiting for spillover than availability of susceptible human hosts. More generally, we demonstrate that the interactions between environment, vectors, reservoir hosts and humans can predict spillover of yellow fever, creating the potential to improve preventive action to reduce yellow fever spillover and avoid onward epidemics.