Speaker:  Tianchen Qian, Postdoctoral Research Fellow, Department of Statistics, Harvard University

Title:  Causal Inference for Time-Varying Treatments and Binary Outcomes, with Application to Mobile Interventions

Abstract:  Advances in wearables and digital technology now make it possible to deliver mobile interventions to individuals in their everyday life. The micro-randomized trial (MRT) is increasingly used to provide data to inform the construction of these interventions. This work is motivated by multiple MRTs that have been conducted or are currently in the field in which the primary outcome is a longitudinal binary outcome. There were no existing methods to appropriately analyze the causal effect of the mobile interventions in such trials. We develop the first estimator that can be used as the basis of primary and secondary analyses for MRTs with binary outcomes. Our estimator is doubly-robust and is able to address statistical challenges including the large number of time points, high-dimensionality of the time-varying covariates, and endogeneity. We illustrate the method using data from the MRT, SARA. In SARA, the goal is to use mobile intervention to increase self report completion rate among adolescents and young adults at risk of substance abuse.