Predictive Model for Estimating Annual Ebolavirus Spillover Potential

Abstract

Forest changes, human population dynamics, and meteorologic conditions have been associated with zoonotic Ebolavirus spillover into humans. High-resolution spatial data for those variables can be used to produce estimates of spillover potential and assess possible annual changes. We developed a model of Ebolavirus spillover during 2001–2021, accounting for variables measured across multiple spatial and temporal scales. We estimated the annual relative odds of Ebolavirus spillover during 2021 and 2022. The highest relative spillover odds estimates occurred in patches that closely followed spatial distribution of forest loss and fragmentation. Regions throughout equatorial Africa had increased spillover estimates related to changes in forests and human populations. Spillover events in 2022 occurred in locations in the top 0.1% of overall spillover odds estimates or where estimates increased from 2021 to 2022. This model can be used to preemptively target surveillance to identify outbreaks and mitigate disease spread and educate the public on risk factors for infection.