Predicting Ebola virus disease risk and the role of African bat birthing.

Abstract

Ebola virus disease (EVD) presents a threat to public health throughout equatorial Africa. Despite numerous spillover events into humans and apes, the maintenance reservoirs and mechanism of spillover are poorly understood. Evidence suggests fruit bats play a role in both instances, yet data remain sparse and bats exhibit a wide range of life history traits. Here we pool sparse data and use a mechanistic approach to examine how birthing cycles of African fruit bats, molossid bats, and non-molossid microbats inform the spatio-temporal occurrence of EVD spillover. We create ensemble niche models to predict spatio-temporally varying bat birthing and model outbreaks as spatio-temporal Poisson point processes. We predict three distinct annual birthing patterns among African bats along a latitudinal gradient. Of the EVD spillover models tested, the best by quasi-Akaike information criterion (qAIC) and by out of sample prediction included significant African bat birth-related terms. Temporal bat birthing terms fit in the best models for both human and animal outbreaks were consistent with hypothesized viral dynamics in bat populations, but purely spatial models also performed well. Our best model predicted risk of EVD spillover at locations of the two 2018 EVD outbreaks in the Democratic Republic of the Congo was within the top 12-35% and 0.1% of all 25\u202f×\u202f25\u202fkm spatial cells analyzed in sub-Saharan Africa. Results suggest that sparse data can be leveraged to help understand complex systems.