CMStatistics 2018: Start Registration
View Submission - CMStatistics
B1373
Title: Applications of survival analysis towards building a value-driven pre-emptive genotyping program Authors:  Jonathan Schildcrout - Vanderbilt University (United States) [presenting]
Abstract: Currently, there are more than 250 medications containing genetics-based guidance within the USFDA label inserts, and in recent years, Vanderbilt University Medical Center developed a quality improvement program to incorporate genetic information into the electronic health record that, when appropriate, permits genotype guided prescribing. Since a large percentage of patients are prescribed medications with pharmacogenetic (PGx) effects and many patients are prescribed multiple such medications, the program involves both pre-emptive and multiplexed genetic testing. Pre-emptive genotyping allows physicians to use the genetic information seamlessly at the time they decide to prescribe a PGx medication, and multiplexing permits cost efficiencies when genetic data are reused. It is cost-prohibitive to genotype all patients, and we describe our approach to identifying patients for genotyping based on anticipated benefit to the patient. We will detail 1) a survival analysis-based, predictive modeling approach using clinical data to estimate patient-level risk of being prescribed each PGx medication, 2) a discrete event simulation that uses literature-based estimates of adverse event rates, variant allele frequencies, and secular death rates to capture the impact of genotype guided therapy in patients once prescribed, and 3) a decision theoretic approach to combine 1) and 2) in order to develop genotyping rules.