Optimal Dynamic Treatment Regimes for Treatment Initiation with Continuous Random Decision Points

• Activity Number: 416
• Type: Topic Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yebin Tao* and Lu Wang and Haoda Fu
• Companies: University of Michigan and University of Michigan and Eli Lilly and Company
• Keywords: Dynamic treatment regime ; Treatment initiation ; Splines ; Inverse probability weighting ; Causal inference

Abstract

Stepwise intensification of treatment is necessary for chronic diseases with progressive conditions. An important but challenging problem is to find the optimal timing to initiate a treatment for the next stage of disease condition. We consider estimating the optimal dynamic treatment regimes (DTRs) to determine a personalized timing for treatment initiation given a patient's specific characteristics. Instead of considering a discrete set of decision points as in most DTR literature, our study undertakes the task of dealing with continuous random decision points for treatment modification based on patients' biomarker and treatment history. Under each candidate DTR, we employ a flexible survival model with splines for time-varying covariates to estimate the patient-specific probability of adherence to the regime. Then we construct an inverse probability weighted estimator for the counterfactual mean utility (predefined criteria) to assess the DTR. We conduct simulations to demonstrate the performance of our method and further illustrate the application process with the example of type 2 diabetic patients enrolled to initiate insulin therapy.