Bayesian modeling of the effect of vaccination and the delta and omicron variants on the COVID-19 epidemic in Burkina Faso using poisson log-linear autoregressive model

Abstract

Introduction: Since its emergence in Burkina Faso in March 2020, the COVID-19 epidemic has undergone several shifts in its trajectory. These fluctuations have been influenced by government control measures, socio-economic dynamics, and biological mutations of the virus. However, the individual impacts of these factors remain insufficiently assessed, and the detailed history of the outbreak is not fully understood. This study aims to objectively evaluate the effects of two major viral variants (Delta and Omicron) as well as the introduction of vaccination on the epidemic's progression. Methods: We used publicly available surveillance data and adopted a Bayesian modeling framework, incorporating a Poisson counting process with log-linear autoregressive component. The events of interest were modeled as dummy time series to estimate their influence. Results: Our findings indicate that the emergence of the Delta variant was associated with an increase in weekly infections, while the arrival of the Omicron variant coincided with a decrease in case numbers. The vaccination was associated with an increase of transmission during the study period. Discussion: This work contributes valuable insights into the epidemic's dynamics in a West African context and offers important lessons for the design of future public health responses.