Abstract:
|
Traditional statistics used in diagnostic study have limited utility for guiding clinical-decision making. Sensitivity, specificity, PPV and NPV focus on subpopulations. Accuracy targets the entire population but is insufficient for clinical decision-making since sensitivity and specificity are rarely equally important in practice. Therefore pragmatic approaches for evaluating the benefit-risk trade-offs of diagnostics are desirable. Benefit-Risk Evaluation of Diagnostics: A Framework (BED-FRAME) is a strategy for pragmatic evaluation of diagnostics designed to supplement traditional approaches. BED-FRAME evaluates diagnostic yield and addresses 2 key issues: (1) that diagnostic yield depends on prevalence, and (2) that different diagnostic errors carry different clinical consequences. We propose the average weighted accuracy (AWA) that targets the entire population, allows differential weighing of sensitivity and specificity, and accounts for the potential heterogeneity of the prevalence within subpopulations, for the pragmatic design and analyses of diagnostic studies. The pragmatic design and analyses of studies using AWA will be illustrated with two examples. We will discuss t
|