Overestimation of test effects in clinical judgment.

Publication Type:

Journal Article


Journal of cancer education : the official journal of the American Association for Cancer Education, Volume 8, Issue 4, p.297-307 (1993)


Adult, Aged, Bayes Theorem, Biopsy, Breast Neoplasms, Decision Making, Decision Support Techniques, Female, Humans, Judgment, Male, Mammography, Nurses, Pharmacology, Clinical, Physicians, Predictive Value of Tests, Probability, Risk Factors, Sensitivity and Specificity


The purpose of this study was to assess the ability of health care professionals to evaluate the effect of test results on disease risk. Fifty health care professionals, including 29 physicians and 21 nonphysicians, associated with a university hospital were studied. Subjects were presented with two hypothetical scenarios involving a common clinical situation to assess the effect of test results on the estimation of disease risk. Estimates of the pretest and posttest probability of breast cancer and mammography sensitivity and specificity were elicited for hypothetical 30- and 70-year-old patients presenting with a breast lump. There was no significant difference between physician and nonphysician probability and sensitivity estimates, although physicians provided higher specificity estimates which were more consistent with literature-derived values. Both physicians and nonphysicians consistently overestimated the risk associated with a positive test result compared to probabilities derived from Bayes' theorem based on the subject's pretest probability and sensitivity and specificity estimates, as well as standard test performance estimates. There was no significant difference between posttest probability estimates for negative test results and those derived from Bayes' theorem utilizing the subject's pretest probability and sensitivity and specificity estimates. Physicians and nonphysicians both estimate test performance characteristics accurately but consistently overestimate the effect of positive test results on the probability of disease. In addition to experience with specific clinical problems, decision making by clinicians could be enhanced by training in the formal methods of decision analysis.