Project
Improving Prediction Accuracy in the Intensive Care Unit

In high-stakes professions, people make critical decisions by collecting information and combining it—through fixed statistical rules, through clinical judgment or through a combination of the two. Research shows that statistical prediction consistently outperforms clinical judgment (Meehl, 1954), yet professionals resist algorithms and trust their intuition (Dawes et al., 1989). This creates a practical problem: How can we improve decision-making when purely algorithmic approaches face resistance?

I will study the accuracy and acceptance of various prediction approaches such as clinical judgment, clinical synthesis, mechanical judgment, mechanical synthesis, and wisdom of the crowd, in the intensive care unit (Sawyer, 1966; Galton, 1907). In this project we compare multiple prediction methods—evaluating their psychometric properties (accuracy, reliability) and practical acceptance. The findings will inform when and how different judgment methods should be deployed, advancing prediction in intensive care.

Supervisors
Prof. Dr. Rob R. Meijer
Prof. Dr. Laura F. Bringmann
Dr. Jacqueline Koeze

Financed by
RUG

Period
1 October 2025 – 1 October 2029