A0309
Title: A Bayesian approach for chronic hepatitis B prevalence estimation to improve the accuracy of economic evaluation
Authors: Julien Smith-Roberge - University of Waterloo (Canada)
William WL Wong - University of Waterloo (Canada) [presenting]
Abstract: Chronic hepatitis B (CHB) is usually a silent disease. The asymptomatic nature means that the disease often remains undiagnosed, leaving its prevalence highly uncertain. This generates significant uncertainty for the associated economic evaluations. The objective is to establish a mathematical framework for the estimation of CHB prevalence and the undiagnosed proportion. A state-transition model describing infection, disease progression and treatment response was mathematically formulated and developed. Model parameters were obtained from the published literature. The historical prevalence of CHB is estimated through a calibration process based on a Bayesian MCMC algorithm. The algorithm constructed posterior distributions of the historical prevalence of CHB by comparing the model-generated predictions of the annual numbers of health events related to CHB against the observed numbers. The prevalence of CHB in Canada in 2018 was estimated to be 0.28\%, and the percentage of undiagnosed among the total infected was 31.5\%. The results are in line with a recently conducted seroprevalence survey. Prevalence estimates impact economic evaluation results on interventions with respect to CHB interventions. Considering the rapid development of interventions for CHB, updated prevalence estimates will become necessary. A platform is provided to estimate this information in a robust and efficient way.
