BACKGROUND: A correctly specified propensity score (PS) estimated in a cohort ("cohort PS") should, in expectation, remain valid in a subgroup population. OBJECTIVE: We sought to determine whether using a cohort PS can be validly applied to subgroup analyses and, thus, add efficiency to studies with many subgroups or restricted data. METHODS: In each of three cohort studies, we estimated a cohort PS, defined five subgroups, and then estimated subgroup-specific PSs. We compared difference in treatment effect estimates for subgroup analyses adjusted by cohort PSs versus subgroup-specific PSs. Then, over 10 million times, we simulated a population with known characteristics of confounding, subgroup size, treatment interactions, and treatment effect and again assessed difference in point estimates. RESULTS: We observed that point estimates in most subgroups were substantially similar with the two methods of adjustment. In simulations, the effect estimates differed by a median of 3.4% (interquartile (IQ) range 1.3-10.0%). The IQ range exceeded 10% only in cases where the subgroup had < 1000 patients or few outcome events. CONCLUSIONS: Our empirical and simulation results indicated that using a cohort PS in subgroup analyses was a feasible approach, particularly in larger subgroups.
BACKGROUND: A correctly specified propensity score (PS) estimated in a cohort ("cohort PS") should, in expectation, remain valid in a subgroup population. OBJECTIVE: We sought to determine whether using a cohort PS can be validly applied to subgroup analyses and, thus, add efficiency to studies with many subgroups or restricted data. METHODS: In each of three cohort studies, we estimated a cohort PS, defined five subgroups, and then estimated subgroup-specific PSs. We compared difference in treatment effect estimates for subgroup analyses adjusted by cohort PSs versus subgroup-specific PSs. Then, over 10 million times, we simulated a population with known characteristics of confounding, subgroup size, treatment interactions, and treatment effect and again assessed difference in point estimates. RESULTS: We observed that point estimates in most subgroups were substantially similar with the two methods of adjustment. In simulations, the effect estimates differed by a median of 3.4% (interquartile (IQ) range 1.3-10.0%). The IQ range exceeded 10% only in cases where the subgroup had < 1000 patients or few outcome events. CONCLUSIONS: Our empirical and simulation results indicated that using a cohort PS in subgroup analyses was a feasible approach, particularly in larger subgroups.
Authors: Philip S Wang; Sebastian Schneeweiss; Jerry Avorn; Michael A Fischer; Helen Mogun; Daniel H Solomon; M Alan Brookhart Journal: N Engl J Med Date: 2005-12-01 Impact factor: 91.245
Authors: Jeremy A Rassen; M Alan Brookhart; Robert J Glynn; Murray A Mittleman; Sebastian Schneeweiss Journal: J Clin Epidemiol Date: 2009-04-05 Impact factor: 6.437
Authors: Christopher T Rentsch; David A Fiellin; Kendall J Bryant; Amy C Justice; Janet P Tate Journal: Alcohol Clin Exp Res Date: 2019-02-03 Impact factor: 3.455
Authors: Ryan M Carnahan; Joshua J Gagne; Christian Hampp; Charles E Leonard; Sengwee Toh; Candace C Fuller; Sean Hennessy; Laura Hou; Noelle M Cocoros; Genna Panucci; Tiffany Woodworth; Austin Cosgrove; Aarthi Iyer; Elizabeth A Chrischilles Journal: Pharmaceut Med Date: 2019-02
Authors: Christina DeFilippo Mack; Robert J Glynn; M Alan Brookhart; William R Carpenter; Anne Marie Meyer; Robert S Sandler; Til Stürmer Journal: Pharmacoepidemiol Drug Saf Date: 2013-01-07 Impact factor: 2.890
Authors: Joshua J Gagne; Shirley V Wang; Jeremy A Rassen; Sebastian Schneeweiss Journal: Pharmacoepidemiol Drug Saf Date: 2014-04-30 Impact factor: 2.890