Constructing Dynamic Treatment Regimes Using Greedy-GQ Algorithm

• Activity Number: 469
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Mental Health Statistics Section
• Author(s): Ashkan Ertefaie*+ and Susan Murphy
• Companies: University of Michigan and University of Michigan
• Keywords: Dynamic treatment regime ; Greedy-GQ ; Electronic Medical Record

Abstract

Dynamic treatment regime is a treatment process that considers individual characteristics and their ongoing performance to decide which treatment option needs to be assigned. We develop a methodology for constructing optimal dynamic treatment regimes based on the individual covariate history in an infinite horizon setting where there is no a priori fixed end of follow up point. We generalize a reinforcement algorithm called Greedy-GQ (GGQ) which can be used in Electronic Medical Record data setting. We discuss the assumptions needed to identify the optimal regime using the GGQ algorithm and derive large sample results necessary for conducting inference. This work is motivated by a multistage intervention study aiming to control the A1c risk factor among diabetes patients.