Revisiting a Discrepant Result: A Propensity Score Analysis, the Paired Availability Design for Historical Controls, and a Meta-Analysis of Randomized Trials

Journal of Causal Inference

article

Stuart G. Baker¹  Karen S. Lindeman² 

[1] National Cancer Institute;Johns Hopkins Medical Institutions

关键词: causal graph;    causal inference;    Caesarean section;    comparative effectiveness research;    epidural analgesia;   

DOI  :  10.1515/jci-2013-0005

来源: De Gruyter

【 摘 要 】

There is an ongoing controversy over whether epidural analgesia for women in labor increases the probability of Caesarean section. Previous research compared results from three methods for estimating the effect of epidural analgesia on the probability of Caesarean section: a propensity score analysis, the paired availability design for historical controls, and meta-analysis of randomized trials. The propensity score analysis and a paired availability design gave substantially different results with the latter in closer agreement with results of a meta-analysis of randomized trials. We updated this investigation in three ways. First, we discussed the use of causal graphs for variable selection in the propensity score analysis. Second, we introduced new extrapolation estimates to improve generalizability for the paired availability design and the meta-analysis of randomized trials with crossovers. Third, we included the results from more recent studies. This analysis provides a window into various topics in causal inference and comparative effectiveness research.

【 授权许可】

CC BY