Statistical methods for evidence-based medicine: the diagnostic test. Part I

Abstract

According to evidence-based medicine (EBM), physicians must be able to assess and understand scientific evidence generated from biomedical research. Among the many statistical methods involved in the proof of the evidence, an essential area concerns how to select and interpret diagnostic tests to confirm or exclude a diagnosis. Rather surprisingly, these statistical methods are not always correctly reported or satisfactorily explained in some EBM reference books, which leads to misunderstandings and incorrect analyses of published studies in the biomedical literature. In this paper, the authors consider the methodology for performing diagnostic test studies and a correct analysis of the diagnostic test results by explaining it for a clinical audience with a working example.