Model-based Estimation of Relative Risks and Other Epidemiologic Measures in Studies of Common Outcomes and in Case-Control Studies

摘要: Some recent articles have discussed biased methods for estimating risk ratios from adjusted odds when the outcome is common, and problem of setting confidence limits ratios. These overlooked extensive literature on valid estimation risks, ratios, differences logistic other models, including that remain rate case-control studies. The present article describes how most these can be subsumed under a general formulation also encompasses traditional standardization projecting impact partially successful interventions. Approximate variance formulas resulting estimates allow interval estimation; intervals closely approximated by rapid simulation procedures require only standard software functions.