Improving comprehensibility and performance of risk prediction models for decision support in clinical environments

Visual analytics (VA) combines automated analysis with visualisation techniques to gain insight into complex datasets. The overall objective is to create a feedback loop and to enable end-users to steer the analysis process. Whereas much advancement has been made in the field, there are still several open challenges associated with this “human-in-the-loop” approach to data analysis. A first key challenge is to enable end-users to understand the reasoning of a prediction model. Whereas several interesting examples of VA applications have been elaborated, user trust and confidence in the prediction is often low. Particularly in high risk application domains, such as healthcare, the challenge has prevented adoption of VA techniques in real-life settings. In this project, we will research how we can design, develop and evaluate interactive VA techniques on top of different prediction models for chronic diseases (T2D and CVD) risk estimation. The overall objective is twofold: 1) to explain the rationale of predictions as well as the uncertainty of these predictions to healthcare practitioners as a basis to support decision making and increase user trust, and 2) incorporating input and feedback from end-users in the prediction process, such as missing lab values  and parameter refinement. The approach will be evaluated in two elaborate user studies in clinical environments.