Index for Rating Predictive Accuracy of Screening Tests

This paper proposes a new pair of indices, the »predictive powers«, for measuring the predictive accuracy (predictivity) of screening tests.

Sensitivity and specificity are indices for measuring the validity of a test. They give the probability of a certain test result given a substance of known condition (a carcinogen or a non-carcinogen). They do not describe the predictive accuracy of a test, which is the probability of a certain condition (a carcinogen or a non-carcinogen) given a known test result. Predictive values are unsuitable measures for characterizing a test since they are seriously affected by the prevalence of carcinogens. However, the predictive powers do not have this limitation and are shown to be useful indices for the purpose of rating the predictive accuracies of various screening tests.

In dieser Arbeit werden zwei neue Indices vorgeschlagen, die »Vorhersageschärfen«, zur Messung der Vorhersagegenauigkeit (predictivity) von Vorsorgeuntersuchungen.

Sensitivität und Spezifizität sind Indices zur Messung der Validität eines Tests. Sie geben die Wahrscheinlichkeit eines bestimmten Untersuchungsergebnisses beim Vorliegen eines bekannten Stoffes (einer karzinogenen oder nicht karzinogenen Substanz) an. Sie beschreiben nicht die Vorhersagegenauigkeit eines Tests; das ist die Wahrscheinlichkeit einer bestimmten Eigenschaft (karzinogen oder nicht karzinogen) beim Vorliegen eines bekannten Untersuchungsergebnisses. Vorhersagewerte sind nicht dazu geeignet, einen Test zu charakterisieren, da sie durch die Prävalenz von Karzinogenen ernsthaft beeinflußt werden. Die Vorhersageschärfen jedoch sind frei von solcher Einschränkung und erweisen sich als nützliche Indices zum Zwecke der Einstufung der Vorhersagegenauigkeiten verschiedener Vorsorgeuntersuchungen.

Key-Words:

Screening Tests - Predictive Accuracy - Predictive Power

Schlüssel-Wörter:

Screening-Tests - prädiktiver Wert - Vorhersageschärfe

^* Research student of the National Cancer Institute of Canada.

• References

• 1 Ames B. N. Identifying environmental chemicals causing mutations and cancer. Science 204 1979; 587-592.
• 2 Armitage P. Statistical Methods in Medical Research. Oxford: Blackwell Scientific Publications; 1971
• 3 Barnoon S, Wolfe H. Measuring the Effectiveness of Medical Decisions — An Operations Research Approach. Illinois: C. C. Thomas; 1972
• 4 Bayes T. An essay towards solving a problem in the doctrine of chances. Philos. Trans 53 1763; 370-418. (reproduced in: Biometrika 45 (1958) 293-315)
• 5 Bennett B. M. On comparisons of sensitivity, specificity and predictive value of a number of diagnostic procedures. Biometrics 28 1972; 793-800.
• 6 Breslow N. E, Day N. E. Statistical Methods in Cancer Research. Vol. 1: The Analysis of Case-Control Studies. Lyon: International Agency for Research on Cancer; 1980
• 7 Bridges B. A. Short term screening tests for carcinogenicity. Nature 261 1976; 195-200.
• 8 Brusick D. J. The role of short-term testing in carcinogen detection. Chemosphere 7 1978; 403-417.
• 9 Calkins D. R, Dixon R. L, Gerber C. R, Zarin D, Omenn G. S. Identification, characterization, and control of potential human carcinogens: A framework for federal decision-making. J. nat. Cancer Inst 64 1980; 169-176.
• 10 Choi B. C. K. Evaluation of the Salmonella typhimurium/ microsome test system (the Ames test) as a short-term screen- ing method for the detection of chemical carcinogens. MSc Thesis. Department of Safety and Hygiene; University of Aston in Birmingham, UK: 1978
• 11 Cole P, Morrison A. S. Basic issues in population screening for cancer. J. nat. Cancer Inst 64 1980; 1263-1272.
• 12 Cooper J. A, Saracci R, Cole P. Describing the validity of carcinogen screening tests. Brit. J. Cancer 39 1979; 87-89.
• 13 Cox D. R. The Analysis of Binary Data. London: Methuen & Co. Ltd.; 1970
• 14 Feinstein A. R. Clinical biostatistics. XXXI. On the sensitivity, specificity, and discrimination of diagnostic tests. Clin. Pharmacol. Ther 17 1975; 104-116.
• 15 Greenhouse S. W, Cornfield J. The Youden index: Letters to the editor. Cancer 3 1950; 1097-1101.
• 16 Heddle J. A, Bruce W. R. Comparison of tests for mutagenicity or carcinogenicity using assays for sperm abnormalities, formation of micronuclei, and mutations in Salmonella. In Hiatt H. H, Watson J. D, Winsten J. A. (Eds): Origins of Human Cancer, Book C Cold Spring Harbor: Cold Spring Harbor Laboratory; 1977: 1549-1557.
• 17 Lusted L. B. Introduction to Medical Decision Making. Illinois: C. C. Thomas; 1968
• 18 Magee P. N. Testing for carcinogens and mutagens. Nature 249 1974; 795-796.
• 19 McCann J. Ames, B. N.: Detection of carcinogens as mutagens in the Salmonella/microsome test: Assay of 300 chemicals: Discussion. Proc. nat. Acad. Sci. USA 73 1976; 950-954.
• 20 McCann J, Choi E, Yamasaki E, Ames B. N. Detec- tion of carcinogens as mutagens in the Salmonella/microsome test: Assay of 300 chemicals. Proc. nat. Acad. Sci. USA 72 1975; 5135-5139.
• 21 McCann J, Spingarn N. E, Kobori J, Ames B. N. Detection of carcinogens as mutagens: Bacterial tester strains with R factor plasmids. Proc. nat. Acad. Sci. USA 72 1975; 983
• 22 Morgan R. W. Prospects for Preventive Medicine. Toronto: Ontario Economic Council; 1977
• 23 MUIC V, Petres J. J, Telisman Z. Validity of a diagnostic test designated by a single function. Meth. Inform. Med 12 1973; 244-248.
• 24 Purchase I. F. H, Longstaff E, Ashby J, Styles J. A, Anderson D, Lefevre P. A, Westwood F. R. Evalua-tion of six short term tests for detecting organic chemical carcinogens and recommendations for their use. Nature 264 1976; 624-627.
• 25 Purchase I. F. H, Longstaff E, Ashby J, Styles J. A, Anderson D, Lefevre P. A, Westwood F. R. An evaluation of six short-term tests for detecting organic chemical carcinogens. Brit. J. Cancer 37 1978; 873-959.
• 26 Spicer C. C. Test reduction: II-Bayes's theorem and the evaluation of tests. Brit. med. J 281 1980; 592-594.
• 27 Stoltz D. R, Poirer L. A, Irving C. L, Stich H. F, Weisburger J. H, Grice H. C. Evaluation of short term tests for carcinogenicity. Toxicol, appl. Pharmacol 29 1974; 157-180.
• 28 Vecchio T. J. Predictive value of a single diagnostic test in unselected populations. New Engl. J. Med 274 1966; 1171-1173.
• 29 Youden W. J. Index for rating diagnostic tests. Cancer 3 1950; 32-35.
• 30 Gahlen R. S, Gambino S. R. Beyond Normality: The Predictive Value and Efficiency of Medical Diagnosis. New York: J. Wiley & Sons; 1975