Authors: Barnieh L, James MT, Zhang J, Hemmelgarn BR
Propensity score methods are used to control for treatment-selection bias in observational studies. A propensity score reduces a collection of covariates into a single composite score. This score is the probability, or propensity, of receiving a specific treatment conditional on the observed covariates. A propensity score can be applied by either matching subjects on the score, stratification by the propensity score or including the propensity score as a predictor in a multivariable model. This paper focuses on propensity score-matched methods. There are 4 steps in a propensity score-matched analysis. The propensity score is derived, the propensity score-matched sample is constructed, the degree to which matching has balanced observe...