Improving Medical Guidelines using Reinforcement Learning

In this project we attempt to improve medical guidelines by employing reinforcement learning methods. Computer-interpretable guidelines (CIGs) are a set of formalisms that encode clinical practice guidelines (CPGs). CPGs are sets of recommendations used to help inform the clinical expert with their decisions. We research whether we can use data from actual treatments to improve such guidelines by correcting mistakes or by adding important exceptions to rules.

Partner

Team

Emile van Krieken

PhD Students

Emile van Krieken

Dr. Diederik M. Roijers

Assistant professor

Dr. Diederik M. Roijers