CRITERIA FOR THE ASSESSMENT OF CLINICAL EFFECTIVENESS OF MEDICAL LABORATORY TESTS

Some basic statistical criteria for the assessment of the medical effectiveness of clinical chemical test results are presented. Population based figures like sensitivity, specificity, and predictive value are usually derived by comparison of test results with a decision limit and are mainly used for the establishment or confirmation of diagnoses. The dependance of the predictive value on the prevalence of the disease is stressed and illustrated by examples. Individual based figures are useful for an objective assessment of serial test results from the same patient. The concepts of intra-individual variation and critical differences and their application for patient monitoring are presented.