Statistical Analysis of Stratified 2X2 Tables

Abstract

Previous papers in this series have discussed design and analysis issues for epidemiologic data that can be presented in a 2 × 2 table.‘” the primary focus was on methods for determining the extent of association between a dichotomous disease state and a dichotomous risk factor. methods appropriate for cross-sectional, prospective, and retrospective studies using matched or unmatched samples were described.In order to study the association between disease status and a dichotomous risk factor, it is often necessary to take into account other variables that may influence either the disease status or the risk factor. For example, Hemming, Overall, and Britt studied the association between the occurrence of nosocomial infections (risk factor) and mortality (disease status) in a large cohort of infants in a newborn intensive care unit. these authors analyzed the data using methods for summarizing the results from a single 2 × 2 table. However, because very low birth weight is associated strongly with both the acquisition of nosocomial infection and mortality, it may be more appropriate to control or adjust for the effects of birthweight.