Optimal Matching Approaches in Health Policy Evaluations Under Rolling Enrollment

• Activity Number: 197 - SPEED: Government and Health Policy
• Type: Contributed
• Date/Time: Monday, July 30, 2018 : 10:30 AM to 11:15 AM
• Sponsor: Health Policy Statistics Section
• Author(s): Jonathan Gellar* and Jiaqi Li and Lauren Vollmer
• Companies: Mathematica Policy Research and Mathematica Policy Research and Mathematica Policy Research
• Keywords: Propensity score; Health policy evaluation; Matching; Rolling enrollment; Causal inference

Abstract

In many policy evaluations, propensity score matching is used to construct a matched control group that appears similar to the treatment group to minimize nonrandom selection bias. When subjects enroll in the treatment group on a rolling basis, several complications are introduced. These complications stem from the fact that, while each member of the treatment group is enrolled on a particular date, no analogous date exists for members of the control group. Thus, defining a single baseline period of matching covariates for the control group is more difficult. When treatment eligibility and enrollment is preceded by an acute event, such as a stroke or other type of hospitalization, the importance of properly aligning control subjects is compounded. We discuss several strategies to handle these complications, including a novel approach to optimize matching while disallowing a unique potential control to be matched to more than one treatment subject. This approach is available as the R package groupmatch, which is compatible with other popular matching packages such as optmatch and MatchIt.