Treatment Outcomes: Conditioning on Electronic/Personal Health Records for congestive heart failure

We will build a prototype system that can be used to predict the outcome of treatments for congestive heart failure given a patient’s personal health record or electronic medical record. This will be based on building a relational probabilistic model that is based on standard medical ontologies, initially on expert knowledge, and can make predictions from electronic/personal health records. The system will be evaluated against clinical practice guidelines. This will be used as the basis for a future system, that can learn the models from data, and so provide auditable best-practices recommendations.

Faculty Supervisor:

David Poole

Student:

Partner:

Treatment

Discipline:

Computer science

Sector:

Health and Related Sciences & Technology; Information and Communications Technology

University:

The University of British Columbia

Program: