Adaptive Pair-Matching in the Search Trial and Estimation of the Intervention Effect

• Activity Number: 325
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Laura Balzer*+ and Maya Petersen and Mark J. van der Laan
• Companies: University of California, Berkeley and University of California, Berkeley and University of California, Berkeley
• Keywords: adaptive designs ; causal inference ; cluster randomized trials ; pair-matching ; TMLE

Abstract

In randomized trials pair-matching is an intuitive strategy to protect study validity and to potentially increase study power. In a common design, candidate units are identified, and their baseline characteristics used to create the best n/2 matched pairs. Within the resulting pairs, the intervention is randomized and the outcomes are measured. This design is adaptive, because the construction of the matched pairs depends on the baseline covariates of all candidates. As consequence, the observed data cannot be considered as n/2 independent, identically distributed pairs of units. Instead, the observed data consist of n dependent units. We explore the consequences of adaptive pair-matching in randomized trials for identification and estimation of the sample average treatment effect. We contrast the unadjusted estimator with targeted minimum loss-based estimators (TMLE). We show substantial efficiency gains from matching and further gains by adjustment. This work is motivated by the Sustainable East Africa Research in Community Health study, a community randomized trial to evaluate the impact of immediate and streamlined treatment on HIV incidence in rural East Africa.