Title: User-friendly estimation of optimal adaptive treatment strategies
Authors: Erica Moodie - McGill University (Canada) [presenting]
Abstract: The goal of precision medicine is to tailor treatment strategies on an individual patient level. Although many estimation techniques are available, some are not robust to model mis-specification while others requite a steep learning curve to implement. A previous dynamic weighted ordinary least squares regression model was approachable yet robust, but was limited to binary exposures and continuous outcomes. We will discuss extensions to this approach that allow for multi-valued or continuous treatments, or for censored outcomes and demonstrate their performance in real and simulated data.