Skip to main content
Fig. 8 | BMC Medical Research Methodology

Fig. 8

From: How to perform prespecified subgroup analyses when using propensity score methods in the case of imbalanced subgroups

Fig. 8

REFCOR data: Forest plots. Forest plots representing treatment effects in the subsets using the “across subsets” strategy with or without interaction terms and with the “within subsets” strategy. HR of death (A) and event (B) regarding facial nerve resection are represented. The first three lines refer to naive analyses performed on the original samples, ignoring potential confounding-by-indication bias. HR = hazard ratio. DFS = disease-free survival

Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us