Statistical evaluation of medical diagnostic tests

Vanda Calhau Fernandes Inacio De Carvalho, Maria Xose Rodriguez-Alvarez

Research output: Contribution to specialist publicationArticle

Abstract

This entry reviews the main concepts associated with the statistical evaluation of medical diagnostic tests. Measures of diagnostic test performance for binary tests, such as sensitivity, specicity, and predictive values are reviewed and extensions to the case of continuous-outcome tests are discussed. In particular,
the receiver operating characteristic (ROC) curve, the most popular tool for evaluating the accuracy of continuous tests, and its summary indices are fully detailed. It has been recently acknowledged that several factors (e.g., dierent test settings and subject-specic characteristics, such as age and/or gender) can
aect the test's accuracy beyond disease status. This entry also motivates and discusses the importance of accounting for such external information when evaluating the performance of a test, through information-specific accuracy measures.
Original languageEnglish
Number of pages20
Specialist publicationWiley StatsRef-Statistics Reference Online
DOIs
Publication statusE-pub ahead of print - 13 Nov 2018

Cite this