Analysis of Clinical Susceptibility Bias in Case-Control Studies

Abstract

Epidemiologic studies of causes of disease rarely contain adjustments for inequalities in disease susceptibility caused by baseline differences in clinical phenomena. In the controversial association between estrogens and endometrial cancer, the menopausal syndrome was suspected as an independent risk factor for the development of endometrial cancer, irrespective of estrogen use. To investigate this suspicion, personal interview data from a case-control investigation were collected and analyzed. The odds ratio for the association between menopausal symptoms and endometrial cancer was 1.12 and 0.85 for two different sets of cases and controls assembled at the same institution. When the data were partitioned according to estrogen usage, the odds ratios became consistently less than one. The results suggest that the menopausal syndrome is not a risk factor for endometrial cancer. (Arch Intern Med139:1111-1113, 1979)