Medical Decision Making

The quantitative principles of test selection and interpretation have been reluctantly integrated into clinical practice. This reluctance may reflect an underlying faculty attitude towards the laboratory. To evaluate this attitude, the scoring standards used for 23 patient management problems (PMPs) in the 1980 and 1983 Medical Knowledge Self-Assessment Programs were reviewed. All diagnostic options were categorized by dollar cost, risk, and a determination of "routineness." Attitudes were probed by reviewing the scores obtained by indiscriminate selection of all items in a category. Our analysis indicated that the examiners valued routine, little-ticket, little-risk items. Such selection would be rewarded 85% of the time and in 95%, no penalty would be received. Further, indifference was more frequent for little-ticket items and accounted for 11.7% of acceptable diagnostic expense. An examinee analyzing these scoring keys could reasonably conclude that routine, little-ticket items should be ordered whenever offered. Hence, at a national level, there is an attitude that implicitly encourages the use of these items in clinical practice. Key words: attitudes; diagnostic tests. (Med Decis Making 8:81-86, 1988)

The use of the acute ischemic heart disease predictive instrument reduced coronary care unit (CCU) admissions for patients without acute ischemic heart disease by 30%. One hypothesis holds that it reinforced physicians' correctly low estimates of the probability of acute ischemia, supporting a decision against CCU admission, another that it lowered phy sicians' over-high probability estimates for acute ischemia so that CCU admission was felt to be unnecessary. The authors asked 86 physicians to estimate the probability of acute ischemia for each of three study cases and to decide on CCU admission. For the low- probability case, the mean of physicians' probability estimates for acute ischemia was 46%, vs. the predictive instrument's calculated probability of 19% (p < 0.00001), a 142% over- estimation by the physicians. For the medium-probability case, the mean of physicians' estimates was 54%, vs. the calculated probability of 58% (not significant). For the high- probability case, the mean of physicians' estimates was 82%, vs. the calculated probability of 78% (not significant). All cases for which physicians considered not admitting to the CCU corresponded to their probability estimates of acute ischemia's being in a threshold range of approximately 10 to 30%. These results support the hypothesis that the mechanism by which the predictive instrument reduces unnecessary CCU admissions is by downward correction of physicians' overly-high suspicions of acute cardiac ischemia into a threshold range for which CCU admission is considered unnecessary. Key words: cardiac ischemia; coronary care; probability threshold. (Med Decis Making 8:90-94, 1988)

The authors performed a decision analysis to determine whether a patient suspected to have herpes simplex encephalitis (HSE) should undergo a brain biopsy or be treated empirically with medical therapy. In most cases, empiric treatment with acyclovir would be slightly favored; brain biopsy was not essential in management. However, brain biopsy was found useful for patients who had low CSF glucose at the time of initial lumbar puncture; such patients may have a very high risk to have other treatable conditions such as tuberculosis, brain abscess, toxoplasmosis, or cryptococcosis. The results of the analysis suggest that even with the advent of safe antiviral drug therapy such as acyclovir, brain biopsy is useful in a well-defined subset of patients with possible HSE. The rationale, however, is not to confirm HSE but rather to detect other treatable conditions. Key words: diagnosis; decision making; herpes simplex encephalitis. (Med Decis Making 8:95-101, 1988)

The authors analyzed two invasive procedures used to visualize the biliary tree, endoscopic retrograde cholangiopancreatography (ERCP) and percutaneous transhepatic cholangiog raphy (PTC), and also explored the utility of preinvasive workups for patients with suspected cholestasis. For this analysis they used published ranges for success, fatality, complications, diagnostic accuracies of the procedures, and prognostic information about the underlying diseases. The choice between ERCP and PTC was found to be a "close call," but ERCP was generally favored as the first-choice procedure. The results suggest that noninvasive imaging does not help decide between ERCP and PTC. Although noninvasive imaging may identify those patients with common duct dilation, the higher success rate with PTC in these patients is offset by a slightly higher mortality rate. Consequently, the choice between ERCP and PTC remains close even if ultrasound has shown that biliary ducts are dilated. Fur thermore, it is shown that these noninvasive tests are most useful when they can conclusively determine the presence or absence of biliary obstruction. For many patients, noninvasive imaging will not obviate the need for invasive tests. Key words: cholestasis; jaundice; decision analysis; endoscopic retrograde cholangiopancreatography; percutaneous transhepatic cho langiography; abdominal ultrasonography. (Med Decis Making 8:102-109, 1988)

Voided-urine cytology as a screen for the early detection of urinary bladder cancer is analyzed to determine its potential in an asymptomatic population. Previous cost-effectiveness as sessments predict that screening will extend life-span at a cost-per-detected-case that is comparable to those of other cancer screening efforts. The focus here is on investigating screening from the perspective of the individual contemplating the screening decision. The analysis is based on a computerized model of bladder cancer which integrates a Markov representation of the induction and progression of the disease with algorithms representing diagnostic and therapeutic intervention strategies, intervention effectiveness, and cost. It is shown that the utilities, as well as the probabilities, of true-positive, false-positive, and false- negative screening outcomes are affected by the particular testing regimen employed and the age at which screening takes place. Screening at age 55 or age 65 was analyzed for individuals of normal risk. Analyses predict that the predisposition of cytology screening to identify the high-grade, aggressive form of the disease will result in gains in life expectancy of more than three years for the asymptomatic true-positive case. Results support the decision to screen, and by requiring a repeatedly positive test result, the probability of a false-positive outcome will not exceed that of a true-positive outcome at age 65. Except for the risk of a false-positive outcome, cytology screening compares favorably with what could theoretically be obtained if a 100% accurate screening test were available. Key words: cytology screening; decision analysis; disease modeling; bladder cancer. (Med Decis Making 8:110-119,1988)

A two-parameter exponential equation for modeling a receiver operating characteristic (ROC) curve is presented, where the area under the curve is a simple function of one of the parameters. The model makes no distributional assumptions about the underlying normal and abnormal patient populations or about the shape of the resulting ROC curves. In a computer simulation of 75 ROC curves, the model provides a fit equivalent to the maximum likelihood estimate method commonly used for ROC curve fitting. Similar results are obtained using the model to fit ROC curve data from the literature. The model's equation calculates the true-positive ratio as a function of the false-positive ratio, and has a first derivative that is useful for finding the optimal decision threshold for a diagnostic testing procedure. In particular, the model is useful in a computer program for finding jointly optimal thresholds for multiple sequential tests. Key words: receiver operating characteristic (ROC) curve; decision theory; optimization; computer simulation. (Med Decis Making 8:120-131, 1988)

In 1984, Huang Yan-ting, MD, spent some months in the United States studying surgical techniques. During his stay he met with Dr. Lee Lusted, then Editor-in-Chief of Medical Decision Making, and the two had several discussions about the field of medical decision making. Dr. Huang published a note on the status of the Society for Medical Decision Making in National Medical Journal ofchina (65:502) in 1985 (also published in MDM, 5:475- 476, 1985). This is Professor Huang Yan-ting's second status report, received by Dr. Lusted.— DGF