A1013
Title: Subgroup analysis in the observational studies
Authors: Hoeun Lee - Konkuk University (Korea, South) [presenting]
Youngjoo Cho - Konkuk University (Korea, South)
Abstract: Matching is one of the widely used methods in observational studies to estimate causal effects. In the past decade, the virtual twin's method, the combination of imputation and machine learning method on potential outcomes, has provided many useful results in the estimation of treatment effect and subgroup analysis in the randomized study. Virtual twins are proposed with matching in observational studies for subgroup analysis. The relevance of the proposed method is evaluated by the consistency and accuracy of estimating treatment effects in simulation studies.
