Propensity score methods to adjust for confounding in assessing treatment effects: bias and precision

摘要: There is an increasing interest in the use of propensity score (PS) methods for confounding control, with generally three ways estimating adjusted treatment effects pharmacoepidemiological studies: 1) stratification on PS, 2) matching PS and 3) using as a covariate. To assess bias precision different methods, we conducted simulations scenarios: had no effect but crude estimate showed protective effect; was more extreme; increased risk protective. Adjusting confounders all shifted estimates toward true values. Adjusted odds ratios method covariate were biased due to either residual or over-adjustment. Matching produced less average than other lower. --------------------------------------------------------------------------------