Bayesian inference in an extended SEIR model with nonparametric disease transmission rate: an application to the Ebola epidemic in Sierra Leone

The 2014 Ebola outbreak in Sierra Leone is analyzed using an extension of the SEIR compartmental model. The unknown parameters of the system of differential equations are estimated by combining data on the number of new (laboratory confirmed) Ebola cases reported by the Ministry of Health and prior distributions for the transition rates elicited using information collected by the WHO Response Team (2014) during the follow-up of specific Ebola cases. The evolution over time of the disease transmission rate is modeled nonparametrically using penalized B-splines. Our framework represents a valuable and robust stochastic tool for the study of an epidemic dynamic from irregular and possibly aggregated case data. Simulations and the analysis of the 2014 Sierra Leone Ebola data highlight the merits of the proposed methodology.