What Does an Observed Relative Risk Convey about Possible Misclassification?

Summary

Misclassification of exposure in epidemiological studies has the effect of distorting the results in various directions. We give quantitative information about the possible order of magnitude of misclassification and of “true” effects given the observations of an epidemiological study. It will be demonstrated that misclassification can also distort trends in relative risks. Real examples from the epidemiological literature are used for illustration.

Mißklassifikationen der Exposition können in epidemiologischen Studien die Ergebnisse in verschiedener Hinsicht verzerren. Für gegebene beobachtete Daten einer epidemiologischen Studie werden quantitative Aussagen über die Größenordnung der Mißklassifikationen und der zugrundeliegenden »wahren« Effekte gemacht. Es wird gezeigt, daß durch Mißklassifikation ein Trend in den relativen Risiken verzerrt werden kann. Die Ergebnisse werden durch Beispiele aus der epidemiologischen Literatur illustriert.

Key-Words:

Case-Control Study - Dose-Response - Epidemiology - Misclassification - Odds Ratio - Relative Risk

Schlüssel-Wörter:

Fall-Kontroll-Studie - Dosis-Wirkung - Epidemiologie - Mißklassifikation - Odds Ratio - relatives Risiko

• References

• 1 Barron B. A.. The effects of misclassifica-tion on the estimation of relative risk. Biometrics 1977; 33: 414-418.
• 2 Breslow N. E., Day N. E.. Statistical Methods in Cancer Research, Vol.1. The Analysis of Case-Control Studies. IARC Scientific Publications No. 32 (International Agency for Research on Cancer, Lyon 1980
• 3 Bross I.. Misclassification in 2 x 2 tables. Biometrics 1954; 10: 478-486.
• 4 Byers T. E., Rosenthal R. I., Marshall J. R., Rzepka T. F., Cummings K. M., Graham S.. Dietary history from the distant past: a methodological study. Nutr. Cancer J 1983; 69-77.
• 5 Copeland K. T., Checkoway H., McMichael A. J., Holbrook R. H.. Bias due to misclassification in the estimation of relative risk. Amer. J. Epidemiol 1977; 105: 488-495.
• 6 Elton R. A., Duffy S. W.. Correcting for the effect of misclassification in a case-control study using data from two different questionnaires. Biometrics 1983; 39: 659-665.
• 7 Goldberg J. D.. The effects of misclassification on the bias in the difference between two proportions and the relative odds in the fourfold table. J. Amer. Statist. Ass 1975; 70: 561-567.
• 8 Greenland S., Kleinbaum D. G.. Correction for misclassification in two-way tables and matched pair studies. Int. J. Epidemiol 1983; 12: 93-97.
• 9 Korn E. L.. Hierarchical log-linear models not preserved by classification error. J. Amer. Statist. Ass 1981; 76: 110-113.
• 10 Marshall J. R., Prior R., Graham S., Brasure J.. On the distortion of risk estimates in multiple exposure level case-control studies. Amer. J. Epidemiol 1981; 113: 464-473.
• 11 McDevitt H. O., Bodmer W. F.. HL-A, immune-response genes and diseases. Lancet 1974; 1269-1275.
• 12 Møller Jensen O., Wahrendorf J., Rosenqvist A., Geser A.. The reliability of questionnaire-derived historic dietary information and temporal stability of food habits in individuals. Amer. J. Epidemiol. (in press)
• 13 Newell D. J.. Errors in the interpretation of errors in epidemiology. Amer. J. Publ. Hlth 1962; 52: 1925-1928.
• 14 Quade D., Lachenbruch P. A., Whaley F. S., McClish D. K., Haley R. W.. Effects of misclassifications on statistical inferences in epidemiology. Amer. J. Epidemiol 1980; 111: 503-515.
• 15 Schlesselman J. J.. Case-Control Studies. New York: Oxford University Press; 1982
• 16 Simon D., Yen S., Cole P.. Coffee drinking and cancer of the lower urinary tract. J. nat. Cancer Inst 1975; 54: 587-593.
• 17 Teta M. J., Lewinsohn H. C, Meigs J. W., Vidone R. A., Mowad L. Z., Flan-nery J. T.. Mesothelioma in Connecticut. 1955-1977. J. occup. Med 1983; 25: 749-756.