Medical Decision Making

doi: 10.1177/0272989X9201200101pmid: N/A

Walker, Andrew; Whynes, David K.

doi: 10.1177/0272989X9201200102pmid: 1538628

On the basis of clinical trial data, 13 strategies for initial screening (filtering) for colorectal cancer are modeled for the purpose of economic evaluation. A wide range of detection cost estimates are generated, although ranking options by detection cost ignores the important consideration of undetected cancers. Formulating the problem as one of cost-effectiveness, however, allows the authors to demonstrate that members of a subset of strategies uniquely outperform all others and that the optimum strategy can be identified by the ex ante spec ification of the valuation of cancers missed on screening. Key words cancer; colorectal cancer; mass population screening; fecal occult blood test; colonoscopy; diagnostic inves tigation; cost-effectiveness; economic evaluation. (Med Decis Making 1992;12:2-7)

Alemi, Farrokh

doi: 10.1177/0272989X9201200103pmid: 1538637

This study explores the difference between additive and non-additive indexes in measuring the severity of myocardial infarction. It shows, as an example, the fallacy of adding severity scores in a straightforward manner. An additive severity index was constructed from the judgments of seven experts The experts also identified several exceptions to the additive index. The study used the exceptions to modify the additive index and produce a non-additive severity index. The non-additive seventy index explained 36% more of the variance in the seventy judgments made by five physicians and two nurses on 50 hypothetical cases than the additive index did. In addition, the non-additive index was 3% more accurate in predicting in-hospital mortality of 7,500 patients with myocardial infarction When the study reduced the noise in the data by ignoring 1,200 rare cases in which stable estimates of mortality rate were unavailable, the prediction of the non-additive index was 13% more accurate than that of the additive index. Statistical tests showed that the differences between the additive and the non-additive indexes were significant at an alpha level below 1% The practical impli cations of non-additive seventy indexes are discussed Researchers and physicians who assess the severity of myocardial infarction should systematically explore exceptions that may improve the accuracy of prediction of an additive index. Key words seventy indexes; additive indexes; non-additive indexes; myocardial infarction. (Med Decis Making 1992;12:8-14)

Hooshiari, Alireza; Khorramshahgol, Reza

doi: 10.1177/0272989X9201200104pmid: 1538627

A method for measuring outpatient resource utilization in terms of the amounts of time different categories of patients spend with various providers is described. Patients are cat egorized based on selected attributes, but other attributes could be used. The method is based on two important and measurable variables. 1) frequencies of usage of different resources (e.g., nurse practitioners, physicians, x-ray), and 2) amounts of time used, by each provider type and by ancillary services (x-ray). Using the quantitative measures described, an algorithm is developed for measuring the direct labor costs of delivering primary care to different types of primary care patients. Key words resource utilization measurement, hos pital management. (Med Decis Making 1992;12:15-21)

Diamond, George A.

doi: 10.1177/0272989X9201200105pmid: 1538629

Estimates of sensitivity and specificity can be biased by the preferential referral of patients with positive test responses or ancillary clinical abnormalities (the "concomitant information vector") for diagnostic verification. When these biased estimates are analyzed by Bayes' theorem, the resultant posterior disease probabilities (positive and negative predictive ac curacies) are similarly biased. Accordingly, a series of computer simulations was performed to quantify the effects of various degrees of verification bias on the calculation of predictive accuracy using Bayes' theorem. The magnitudes of the errors in the observed true-positive rate (sensitivity) and false-positive rate (the complement of specificity) ranged from + 11% and +23%, respectively (when the test response and the concomitant information vector were conditionally independent), to + 16% and + 48% (when they were conditionally non- independent) These errors produced absolute underestimations as high as 22% in positive predictive accuracy, and as high as 14% in negative predictive accuracy, when analyzed by Bayes' theorem at a base rate of 50%. Mathematical correction for biased verification based on the test response using a previously published algorithm significantly reduced these errors by as much as 20%. These data indicate 1) that selection bias significantly distorts the determination of predictive accuracies calculated by Bayes' theorem, and 2) that these distortions can be significantly offset by a correction algorithm Key words: diagnosis, pre dictive accuracy; selection bias; sensitivity, specificity, testing. (Med Decis Making 1992;12: 22-31)

Heckerling, Paul S.; Tape, Thomas G.; Wigton, Robert S.

doi: 10.1177/0272989X9201200106pmid: 1538630

To investigate the relation between physicians' predicted probabilities of pneumonia and their utilities for ordering chest x-rays to detect pneumonia, the authors studied 52 physicians who ordered chest x-rays of 886 patients presenting to an emergency department with fever or respiratory complaints Physicians estimated the probability of pneumonia prior to obtaining the results of the chest x-ray. Utilities were assessed by asking physicians to consider a hypothetical patient presenting with acute respiratory symptoms, with unknown chest x-ray status, and to rank on a scale from +50 ("best thing I could do") to - 50 ("worst thing I could do") their rating scale utilities for not diagnosing pneumonia and not ordering a chest x-ray when the patient had pneumonia (i.e, missing a pneumonia), and for diagnosing pneumonia and ordering a chest x-ray when the patient did not have pneumonia (i.e, ordering an unnecessary x-ray) The utility for ordering an unnecessary x-ray was negatively correlated with average predicted probability (r = -0 1495, p = 0 29), whereas the utility for missing a pneumonia was positively correlated with average predicted probability (r = 0 2254, p = 0.11), although the correlations were not statistically significant. Relative chagrin, defined as the difference in these utilities, was significantly inversely correlated with average predicted probability (r = -0 2992, p < 0.035), even after adjusting for the prevalence of pneumonia seen by each physician (partial r = - 0.42, p < 0 0027). It is concluded that physicians who experienced greater regret over missing a pneumonia than over ordering an unnecessary x-ray estimated lower probabilities of pneumonia for patients for whom they ordered x-rays. Thus, these physicians may have had lower thresholds for ordering chest x-rays for patients with acute respiratory illness. Key words: prediction, probability; utility, pneumonia. (Med Decis Making 1992;12:32-38)

Hilden, Jørgen; Glasziou, Paul P.; Habbema, J. Dik F.

doi: 10.1177/0272989X9201200107pmid: 1538631

Among those decisions that may be made by a patient in response to an illness, the authors single out a certain class. contingent investment decisions. They are characterized by the patient's committing him- or herself, on the basis of prognostic counseling, to a certain action or non-action that he or she may regret in retrospect Examples show that, when assessing utilities, the decision analyst runs a risk of handling such investment decisions incorrectly, unless they are made explicit and incorporated into the medical decision process. The anomaly is explained as a violation of the structural rules for decision trees and is also interpreted in terms of "the price of prognostic ignorance," a quantity closely related to the expected utility value of perfect information. Key words: decision theory; physician-patient relations, patient compliance; prognosis; risk-taking, quality of life; utility theory; patients' decisions, decision trees. (Med Decis Making 1992;12:39-43)

Eben-Chaime, Moshe; Pliskin, Joseph S.

doi: 10.1177/0272989X9201200108pmid: 1538632

This report demonstrates the power and usefulness of mathematical optimization as a de cision support tool in the medical services industry by presenting an application to dialysis service planning Models to predict the number of dialysis beds in a given region are usually population-based. Dialysis planners and providers have found a need to accommodate sparsely populated regions by making some allowance for patient travel times. A formal approach to incorporating travel times into dialysis planning, based on the formulation and solution of a mixed-integer programming model, is presented The development of a method for dialysis planning serves as a platform to demonstrate the use of integer programming to support decision making Major modeling principles are presented, output interpretation and sensitivity analysis are illustrated by examples; and computational requirements are dis cussed Key words. dialysis need forecasting, population-based model, travel time; math ematical optimization, mixed-integer programming; location, allocation (Med Decis Making 1992;12:44-51)

Singal, Bonita M.; Hedges, Jerris R.; Succop, Paul A.

doi: 10.1177/0272989X9201200109pmid: 1538633

The serum electrolyte panel (SEP) is commonly ordered in the workup of the emergency department (ED) patient This study was done. 1) to evaluate the efficacy of the SEP in terms of the identification of clinically significant abnormals (yield) and the impact on ther apeutic plan (impact) ; 2) to evaluate the reasons that the test was ordered; and 3) to compare the expected and realized contributions of the test to patient care Pretest and posttest questionnaires were administered to physicians managing 800 ED patients ≥ 55 years old for whom SEPs were ordered. The yield of significant abnormals was 16%. Fluid and electrolyte treatment plans were modified after the SEP results became known in 35% of cases This modification was associated with a normal SEP 48% of the time. Both the yield and the impact of the SEP were related to the reason that the test was ordered. The most common reason given was "to look for an unexpected abnormality" (50%) Physicians' expectations for the contribution of the SEP to patient care decisions were greater than the contributions realized after the results were known. However, in 115 cases, the test con tributed more than expected. Physicians predicted that 13% of the tests would contribute nothing to patient care. After the results were known, they felt that 38% had made no contribution. Physicians tend to overestimate the potential impact of the SEP but are oc casionally surprised by a result that contributes more than expected. Thus, there is consid erable pretest uncertainty about treatment decisions and normal results appear to have a substantial impact. Key words serum electrolytes; efficacy; cost containment, decision mak ing; utilization; utility. (Med Decis Making 1992;12:52-59)

Metz, Charles E.; Shen, Jong-Her

doi: 10.1177/0272989X9201200110pmid: 1538634

From the perspective of signal detection theory, human variation in image reading degrades diagnostic accuracy by broadening the statistical distributions of perceived evidence upon which decisions were based. A new multivariate "random-effects" model formulates the total variation in a diagnostic decision variable as a sum of three uncorrelated components that represent differences among cases, readers, and repeated readings by a given reader. This model provides a basis for quantitative predictions concerning the amount by which diagnostic accuracy, as specified by ROC analysis, can be enhanced by the replication of image readings. Although these predictions apply exactly only to the hypothetical situation in which normally distributed decision variables from equivalent readers are averaged, computer- simulation studies and an analysis of mammographic image-reading data from five radiol ogists show that similar gains in accuracy can be achieved by averaging discrete confidence ratings. Key words diagnostic accuracy, random-effects model; image interpretation; ROC analysis. (Med Decis Making 1992;12:60-75)