Sample size to test for interaction between a specific exposure and a second risk factor in a pair-matched case-control study

• Published in: Statistics in medicine 2000-04, Vol.19 (7), p.923-935
• Main Authors: Qiu, Peihua, Moeschberger, Melvin L., Cooke, Glen E., Goldschmidt-Clermont, Pascal J.
• Format: Article
• Language: English
• Subjects: Algorithms; Biological and medical sciences; Case-Control Studies; Computerized, statistical medical data processing and models in biomedicine; Data Interpretation, Statistical; Homocysteine - metabolism; Humans; Medical sciences; Medical statistics; Myocardial Infarction - etiology; Myocardial Infarction - genetics; Polymorphism, Genetic; Risk Factors; Sample Size; Smoking - adverse effects
• ISSN: 0277-6715; 1097-0258
• DOI: 10.1002/(SICI)1097-0258(20000415)19:7<923::AID-SIM341>3.0.CO;2-O

We discuss a sample size calculation for a pair‐matched case‐control study to test for interaction between a specific exposure and a second risk factor. The second risk factor could be either binary or continuous. An algorithm for the calculation of sample size is suggested which is based on a logistic regression model that relates the logarithm of the disease‐exposure odds ratio to the second risk factor. This problem is motivated by a study comparing the prevalence of GP‐IIIa PlA2 polymorphism (the exposure) in individuals with and without myocardial infarction (case‐control). One of the hypotheses in this study is whether or not there is an interaction between the prevalence of GP‐IIIa PlA2 polymorphism and a second risk factor such as smoking status and homocysteine level. We introduce the algorithm in detail with several numerical examples. Copyright © 2000 John Wiley & Sons, Ltd.