Online Program Home
My Program

Abstract Details

Activity Number: 87 - Invited ePoster Session: a Statistical Smörgåsbord
Type: Invited
Date/Time: Sunday, July 29, 2018 : 8:30 PM to 10:30 PM
Sponsor: SSC
Abstract #328823
Title: Analysis of Paired Binary Data Subject to Misclassification Using a Random Effect Model
Author(s): Hua Shen* and Richard John Cook
Companies: University of Calgary and University of Waterloo
Keywords: paired binary data; misclassification ; clustered latent variables; random effect ; diagnostic tests

Clustered data arise commonly in medical research where observations are obtained from subjects comprising distinct clusters thus grouped. The dependence of the data with clusters need to be accounted for and modeling binary data poses additional challenges. Moreover, misclassification is often encountered especially in the absence of definitive diagnostic tests. Ignorance of misclassification and naively taking the obtained diagnostic test result as the underlying truth can lead to severe consequences in many situations. This is investigated and addressed using maximum likelihood approach under a latent variable model. The proposed method is based on maximizing the marginal likelihood for paired binary data with random effects. By assuming a flexible and parametric family for the distribution of random effect, numerical integration and approximation is not demanded. The procedure also allows the evaluation of the accuracy of the medical diagnostic tests. In addition to the numerical studies demonstrating the performance of the proposed method, a stimulating study is used as illustration.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program