[Diagnostic accuracy and linked evidence--testing the chain]

Abstract

In this article we describe the methodological underpinnings in assessing the confidence in estimates of effects (quality of evidence) for diagnostic questions and resulting recommendations in guidelines. In particular, we will describe current considerations and research regarding the link between data relating to diagnostic accuracy and patient important outcomes and how to move from evidence about diagnostic tests and strategies to recommendations. Our deliberations and suggestions, a result of an invited presentation at the annual fall conference of the German Institute for Quality and Efficiency in Healthcare (IQWiG), are based on our work with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group.