We read with great interest the recent article by Hemmila et al1 regarding the use of propensity score analyses in the surgical literature for comparative effectiveness research. The authors compare results using standard multivariate modeling vs propensity methods among patients undergoing open vs laparoscopic appendectomy based on data derived from the National Surgical Quality Improvement Program (NSQIP). The results of the multivariate modeling did not differ from the propensity score analysis. Hemmila et al1 conclude that propensity score–adjusted modeling did not add to the analysis, and, in turn, they question the value of propensity score methods. Although the results of the propensity-matched and multivariate analyses are similar in their analysis of the NSQIP appendectomy data, we are concerned that they may have overstated the lack of relevance of propensity score methods as a statistical approach.
Mayo SC, Pawlik TM. Propensity Score Methods: Setting the Score Straight. Arch Surg. 2011;146(7):887–889. doi:10.1001/archsurg.2011.159