A Computer Program for Rapid Generation of Receiver Operating Characteristic Curves and Likelihood Ratios in the Evaluation of Diagnostic Tests

Abstract

We describe a MUMPS program to facilitate the evaluation of the diagnostic effectiveness and efficiency of laboratory tests using receiver operating characteristic (ROC) curves and likelihood ratios (LR). Test results from a population with and without a disease are stored into retrievable, editable data bases. Either single or a combination (≤9) of different test data files can be accessed to generate tables of calculated frequency distribution, sensitivity, 1-specificity and LR at variable decision thresholds. The program also calculates the area under the ROC curve, confidence intervals and SE. The selection of the optimum cut-off value is assisted by graphical plots of frequency distribution, ROC and decision level curves as illustrated by using some ‘cardiac’ enzyme tests.

References

Yerushalmy

. Statistical problems in assessing methods of medical diagnosis, with special reference to X-ray techniques. Pub Health Rep 1947; 62: 1432–49.

Galen

Gambino

. Beyond Normality: the Predictive Value and Efficiency of Medical Diagnosis. New York: J Wiley and Sons, 1975.

Robertson

Zweig

. Use of receiver operating characteristic curves to evaluate the clinical performance of analytical systems. Clin Chem 1981; 27: 1569–74.

Beck

Shullz

. The use of relative operating characteristic (ROC) curves in test performance evaluation. Arch Pathol Lab Med 1986; 110: 13–20.

Albert

. On the use and computation of likelihood ratios in clinical chemistry. Clin Chem 1982; 28: 1113–9.

Radack

Rouan

Hedges

. The likelihood ratio: an improved measure for reporting and evaluating diagnostic test results. Arch Pathol Lab Med 1986; 110: 689–93.

Gerhardt

Keller

Evaluation of test data for clinical studies. Scand J Clin Lab Invest 1985; 46 (Suppl. 181): 1–74.

Feinstein

. Clinical Epidemiology: the Architecture of Clinical Research. Philadelphia: WB Saunders Co., 1985.

Ward

. The differential positive rate, a derivative of receiver operating characteristic curves useful in comparing tests and determining decision levels. Clin Chem 1986; 32: 1428–9.

10.

Bulpitt

. Confidence intervals. Lancet 1987; i: 494–7.

11.

Jablonsky

Leung

Henderson

. Changes in the ratio of lactate dehydrogenase isoenzymes 1 and 2 during the first day after acute myocardial infarction. Clin Chent 1985; 31: 1621–4.

12.

Leung

Henderson

. Thin-layer agarose electrophoresis of lactate dehydrogenase isoenzymes in serum: A note on the method of reporting and on the lactate dehydrogenase isoenzyme-1/isoenzyme-2 ratio in acute myocardial infarction. Clin Chem 1979; 25: 209–11.

13.

Sox

. Probability theory in the use of diagnostic tests. Ann Int Med 1986; 104: 60–6.

14.

Kreig

Abendroth

Bongiovanni

. When is a diagnostic test result positive? Arch Pathol Lab Med 1986; 110: 787–91.

15.

Linnet

Brandt

. Assessing diagnostic tests once a optimum cut-off has been selected. Clin Chem 1986; 32: 1341–6.

16.

Youden

. Index for rating diagnostic tests. Cancer 1950; 3: 32–5.

17.

Suchman

Dolan

. Odds and likelihood ratios. In: Clinical Diagnosis and the laboratory: Logical Strategies for Common Medical Problems. Griner

Panzer

Greenland

, eds. Chicago: Year Book Medical Publishers, Inc., 1986; 36–43.

18.

Abendroth

Bongiovanni

Kreig

. A microcomputer program for critical evaluation of diagnostic tests. Arch Pathol Lab Med 1986; 110: 952–8.

19.

Galen

Weiss

. Predicitive Value Calculator. Helena Laboratories, Beaumont, Tx 77704-0752. 1985.