Limitations of sensitivity, specificity, likelihood ratio, and bayes' theorem in assessing diagnostic probabilities: a clinical example.

摘要: We evaluated the extent to which sensitivity, specificity, and likelihood ratio of exercise test diagnose coronary artery disease vary across subgroups a certain patient population. Among 295 patients suspected disease, as independently determined by angiography, we assessed variation in sensitivity specificity according history, physical examination, results, severity 207 with 88 without respectively. The varied substantially sex (women 30% men 64%), systolic blood pressure at baseline (53% 65%), expected workload (50% peak 67%), relative (33% 68%), number diseased vessels (39% 77%). (men 89% women 97%) (85% 98%). (3.8 17.0) same subgroups, did sensitivity. As each population tends be heterogeneous respect characteristics, no single level these parameters can given that is adequate for all subgroups. Use basis calculating diagnostic probabilities individual using Bayes' theorem has serious limitations.