Predicting Warfarin Sensitivity after Cardiovascular Surgery

Anticoagulation with Warfarin is indicated and required for post-operative cardiovascular patients. However, it is a high-risk medication with a narrow therapeutic range where sub-optimal dosing can lead to complications and even death. While multiple risk factors have been associated to Warfarin sensitivity, the prediction of optimal Warfarin dosing strategies remains ineffective and requires trial and error and close patient monitoring. This work proposes the use of machine learning and reinforcement learning algorithms to more accurately predict Warfarin requirements in post-operative cardiovascular patients, leading to decreased hospital stay and re-admission rates and increasing cost savings at cardiovascular surgery centers globally.

Faculty Supervisor:

Marzyeh Ghassemi;Anna Goldenberg

Student:

Julyan Keller-Baruch

Partner:

Vector Institute

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

University of Toronto