A Bayesian Semiparametric Approach to Estimating a Bacterium’s Wild-Type Distribution and Prevalence Estimation: Accounting for Contamination and Measurement Error (BayesACME)

• Activity Number: 1 - Invited E-Poster Session
• Type: Invited
• Date/Time: Sunday, August 2, 2020 : 12:30 PM to 3:30 PM
• Sponsor: Section on Statistical Consulting
• Author(s): Will Eagan* and Bruce A. Craig
• Companies: Purdue University and Purdue University
• Keywords: Antimicrobial Resistance; Measurement Error; Bayesian Analysis; Dirichlet Process Mixture Model

Abstract

Antimicrobial resistance (AMR) is a major challenge to modern medicine and of grave concern to public health. It is primarily driven by the misuse and overuse of antibiotics in humans and food source animals. To monitor AMR, organizations like the FDA and CLSI analyze “drug/bug” collections of clinical assay results with the goal of estimating the prevalence of AMR. This estimation is challenging for several reasons. First, the collection of assay results is a mixture of susceptible and resistant strains. Second, the most commonly used dilution assay produces interval-censored readings. While published methods have addressed these challenges, they do not account for the third challenge, the inherent assay variability that has been shown to encompass a three-fold dilution range. To also account for this, we propose a Bayesian semiparametric method that incorporates measurement error. We also extend our approach to monitor prevalence over time. The feasibility of this approach and its improved precision is demonstrated through a simulation study and an application to a real data set.