Date of Award

January 2018

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)



First Advisor

Nihar R. Desai


The Precision Medicine Initiative aims to advance Medicine from “one-size-fits-all” treatments to more individualized approaches. Clinical trials evaluate treatments by analyzing average outcomes, and thus risk overlooking potential differences in treatment effect among different subsets of the study population. The use of multivariate models has been proposed as a way to identify heterogeneity of treatment effect and to determine patients’ individualized treatment risks and benefits.

We analyzed the Randomized Evaluation of Long-Term Anticoagulation Therapy (RELY) trial of dabigatran versus warfarin in patients with atrial fibrillation, to determine if the application of multivariate predictive models could demonstrate heterogeneity of treatment effect among the study population. We developed two models to predict patients’ risk of stroke or systemic embolism and risk of major bleeding if treated with dabigatran or warfarin. We then applied these models to the individual patients in the RE-LY trial, and determined patients difference in risk if treated with dabigatran versus warfarin. Individual difference in stroke risk for dabigatran 110mg and 150mg versus warfarin was -0.78%  0.95% and -1.32%  1.31% and the difference in major bleeding risk was -1.12%  1.44% and -0.41%  2.39%, respectively.

These findings demonstrate heterogeneity of treatment effect in the RE-LY trial, and the ability of multivariate risk models to identify distinct treatment risks for individual patients. Such models could be used in clinical practice provide patients and clinicians with individualized treatment risk information and improve treatment decisions.


This is an Open Access Thesis.

Open Access

This Article is Open Access