Evaluation of Diagnostic and Predictive Accuracy of Medical Tests and Biomarkers
*Andrew Zhou, HSR&D Center of Excellence, VA Puget South Health Care System, and University of Washington 

Keywords:

Diagnostic tests play a pivotal role in medicine, often determining what additional diagnostic tests, treatments, and interventions are needed and ultimately affecting patients' outcomes. This workshop provides a comprehensive approach to designing and analyzing diagnostic accuracy studies, so as to aid clinicians in understanding these studies and in generalizing study results to their patient populations. The basis for the course is the upcoming second edition of "Statistical Methods in Diagnostic Medicine", by Zhou, Obuchowski and McClish. We define various measures of diagnostic accuracy, describe strategies for designing diagnostic accuracy studies, and present statistical methods for estimating and comparing tests' accuracies, calculating sample size, and synthesizing the literature for meta-analysis. We then present more advanced statistical methods for describing a test's accuracy when accuracy is affected by patient characteristics, for analyzing multi-reader studies, for correcting for verification bias or imperfect gold standard bias, and for performing meta-analyses. The attendees are assumed to have a basic understanding of maximum likelihood and Bayesian methods as well as generalized linear models.