Propensity Score Matching in Randomized Clinical Trials

• Activity Number: 296
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Zhenzhen Xu*+ and John David Kalbfleisch
• Companies: University of Michigan and University of Michigan
• Keywords: Cluster randomized trials ; Experimental design ; Optimal full matching ; Propensity score matching ; Randomization test ; Optimal tripartite matching

Abstract

Cluster randomization trials with few clusters have been widely used in recent years for evaluation of health care strategies. On average, randomization achieves balance in both known and unknown confounding factors. In practice, however, investigators can only introduce a small amount of stratification and cannot balance on all the variables simultaneously. A new design, the balance match weighted(BMW) design, which applies the optimal matching with constraints technique to a prospective randomized design, was introduced (Biometrics 2009). In this paper, we generalize the BMW design to other major areas of applications. One example is clinical trials with three arms where the generalized design would apply an optimal tripartite matching technique to randomized samples. BMW Design-based randomization tests are also developed and their performance investigated under various setting.