Machine Learning for a Heart Failure Electronic Health Care Predictive Analytics System - Faraz Ahmed

Heart failure is a morbid condition affecting a large population of patients. Northwestern Medicine, like other health systems, has taken a reactive approach to the management of heart failure patients by focusing on reducing heart failure readmissions. Proactive strategies have become increasingly important as new payment models are adopted. Health care electronic predictive analytics systems represent a novel approach to heart failure prevention. By utilizing the massive repository of electronic data at Northwestern and machine learning, a computer science method that learns from data, models can be derived that predict outcomes. In this proposal, we aim to develop an electronic health care predictive analytics system to inform a proactive approach to heart failure management. Using data from the unique and robust Northwestern Medicine Enterprise Data Warehouse, we propose to develop, validate, implement, and evaluate an electronic heath care predictive analytics system for incident heart failure hospitalization. To date, there are no systems yet developed that anticipate the initial hospitalization for heart failure. Our work will be among the first of its kind. These results will provide the preliminary data and infrastructure for the development of a larger system of comprehensive care for heart failure patients focused on the promotion of health.