A clinical and epidemiological evaluation of tonometry as a screening test

Because of their differences in emphasis, clinical practitioners and clinical epidemiologists often employ different terms and concepts in comparing and rating health screening tests. The apparent inconsistency of a high error rate (96 to 99% over-referrals) yet good test sensitivity and specificity (40% and 95% respectively) for tonometry can be explained by a complete analysis of tonometric screening, encompassing both the clinical and epidemiological view-points. This paper reviews basic terminology, explains the clinical and epidemiological approaches, and demonstrates both of these methods of analysis in a comprehensive evaluation of tonometry using real data from recent clinical investigations.