Kristin Linn, Assistant Professor of Biostatistics at the University of Pennsylvania, will present:

“Estimation and Evaluation of Individualized Treatment Rules Following Multiple Imputation”

Data-driven optimal treatment strategies promise to benefit patients, care providers, and other stakeholders by improving clinical outcomes and lowering healthcare costs. A treatment decision rule is a function that inputs patient-level information and outputs a recommended treatment. An important focus of precision medicine is to develop optimal treatment decision rules that maximize a population-level distributional summary such as the expected value of a clinical outcome. However, guidance for estimating and evaluating optimal treatment decision rules in the presence of missing data is fairly limited. Our work is motivated by the Social Incentives to Encourage Physical Activity and Understand Predictors (STEP UP) study. In this trial, participants were randomized to a control arm or one of multiple interventions that were designed to increase physical activity. Study participants were given wearable devices which were used to record daily step counts as a measure of physical activity. Many participants were missing at least one daily step count during the study period, and the missingness pattern within individuals was often non-monotone. We propose two frameworks for estimation and evaluation of an optimal treatment decision rule following multiple imputation and compare performance of the frameworks using simulated data. We apply our methods to the STEP UP data to determine whether a personalized intervention strategy might be expected to increase physical activity more than the intervention that had the largest estimated average treatment effect.

A seminar tea will be held at 11:00 a.m. in University Office Plaza, Room 240. All are Welcome.