TITLE – Detection Dependence in Modeling Binary Response Data: A Bayesian Approach
Diagnostic procedures for some conditions can yield three possible results: positive, negative, or inconclusive. We consider the differential detection scenario, where the probability of obtaining a conclusive test result depends on the presence or absence of the underlying condition. Traditionally, differential detection compromises identifiability and leaves us unable to accurately model the prevalence of the condition. We show how a measure of data quality can be used to correct for detection dependency in the testing data. To illustrate the method, we use an example from a recent reproductive health study.
Contact Name: Gabriel Huerta
Contact Email: firstname.lastname@example.org