Big Data for Medical Research: A Unified Approach for Individualized Treatment Recommendation and Subgroup Identification Based on Electronic Medical Record Data

• Activity Number: 72
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Haoda Fu*+ and Jin Zhou
• Companies: Eli Lilly and Company and University of Arizona
• Keywords: Subgroup Identification ; Personalized Medicine ; Individualized Treatment Recommendation ; Biomarker ; Casual Inference ; EMR

Abstract

The EMR data are uniquely powerful yet far underutilized data resources. The data contain patient registries, treatments, clinical outcome, patient history, comorbidities, concomitant medication etc. The goal of our research is to utilize EMR data to improve clinical research efficiency. We develop a new framework called individualized treatment recommendation (ITR). The ITR method provides a general framework for subgroup identification which can be broadly applied to both randomized control trials and observational studies. It can also handle more than two treatment subgroup identification problems.