When AI models predict how long a patient will survive, who decides whether they can be trusted? Sophie Hanna Langbein was invited to address this very question at an international workshop in the Netherlands.
In February 2026, Sophie Hanna Langbein (BIPS / University of Bremen / Lifespan AI) gave a presentation at the workshop “Methods for Explainable Machine Learning in Health Care” at the invitation of the Dutch Biometric Society (VVSOR). She presented interpretable methods that explain how and why machine learning models make survival time predictions.
Survival time models are indispensable in clinical practice: for cancer prognoses, heart failure risk stratification, or the planning of treatment courses. When complex AI models make these predictions, it often remains unclear on what basis they make their decisions. Explainable AI (XAI) methods can bridge this gap and make AI decisions transparent to clinicians and patients.
The VVSOR (Vereniging voor Statistiek en Operationele Research) is the Dutch professional association for statistics. Its workshops are aimed at statisticians in clinical and epidemiological research.