The network epidemiology of an Ebola epidemic
Laurent Hébert-Dufresne,Jean-Gabriel Young,Jamie Bedson,Laura A. Skrip,Danielle Pedi,Mohamed F. Jalloh,Bastian Raulier,Olivier Lapointe-Gagné,Amara Jambai,Antoine Allard,Benjamin M. Althouse
DOI: https://doi.org/10.48550/arXiv.2111.08686
2021-11-17
Abstract:Connecting the different scales of epidemic dynamics, from individuals to communities to nations, remains one of the main challenges of disease modeling. Here, we revisit one of the largest public health efforts deployed against a localized epidemic: the 2014-2016 Ebola Virus Disease (EVD) epidemic in Sierra Leone. We leverage the data collected by the surveillance and contact tracing protocols of the Sierra Leone Ministry of Health and Sanitation, the US Centers for Disease Control and Prevention, and other responding partners to validate a network epidemiology framework connecting the population (incidence), community (local forecasts), and individual (secondary infections) scales of disease transmission. In doing so, we gain a better understanding of what brought the EVD epidemic to an end: Reduction of introduction in new clusters (primary cases), and not reduction in local transmission patterns (secondary infections). We also find that the first 90 days of the epidemic contained enough information to produce probabilistic forecasts of EVD cases; forecasts which we show are confirmed independently by both disease surveillance and contact tracing. Altogether, using data available two months before the start of the international support to the local response, network epidemiology could have inferred heterogeneity in local transmissions, the risk for superspreading events, and probabilistic forecasts of eventual cases per community. We expect that our framework will help connect large data collection efforts with individual behavior, and help reduce uncertainty during health emergencies and emerging epidemics.
Populations and Evolution,Physics and Society