A case study of physicians' use of liver-spleen scans. Are we doing what we think we're doing?

We evaluated physicians' perceptions of the performance capabilities of the liver-spleen scan (LSS) in detecting metastases and the inferences physicians draw from LSS results. Physicians' perceptions of the sensitivity and specificity of the LSS in detection of metastases, as well as their estimates of likelihood ratios for various scan results, varied broadly over a range that could result in clinically important variations in patient treatment. In addition, independent of any variations in estimates of test performance characteristics, physicians had difficulty drawing appropriate probabilistic inferences from LSS results. Our findings suggest that data regarding the performance capabilities of the LSS and other diagnostic tests within a particular hospital should be made available to physicians and that physicians should be given microcomputer assistance in estimating the impact of test results on the probability of disease.