Optimizing Digital Interventions through Micro-Randomized Trials and Causal Modeling

Optimizing Digital Interventions through Micro-Randomized Trials and Causal Modeling Contact Person:

Tianchen Qian

Project Summary:

The development in smartphone and wearable technology now makes it possible to deliver digital health interventions to individuals in real time. These interventions include notifications and reminders for physical activity, stress management exercises, etc. To optimize such digital health interventions and to reduce user burden, it is crucial to understand when, under what circumstances, and what intervention is more effective. Micro-randomized trials (MRT) is an experimental design to answer these questions. In a MRT, each user is repeatedly randomized to various versions of the digital intervention for often hundreds or thousands of times. In this project we develop causal inference methods for analyzing MRT data, in order to understand the time-varying causal effect of the interventions and how such effect interacts with user’s contextual information. The results can be used to optimize the delivery and content of the interventions.

Publications:

– Tianchen Qian, Hyesun Yoo, Predrag Klasnja, Daniel Almirall, and Susan A. Murphy. Estimating time-varying causal excursion effect in mobile health with binary outcomes. Biometrika (2020)

– Tianchen Qian, Michael Russell, Linda Collins, Predrag Klasnja, Stephanie Lanza, Hyesun Yoo, and Susan A. Murphy. The Micro-Randomized Trial for Developing Digital Interventions: Data Analysis Methods. https://arxiv.org/abs/2004.10241