Resources
Publications and analysis on health AI
QSTOM-IT publishes evidence-focused content on methodology, robustness, data, PRISM, case studies and the European framework. The objective is to make technical reasoning verifiable and reusable.
MethodologyRobustnessPRISMRegulationCase studies
EU AI Act and health: a practical reading for AI projects
Classification, data, documentation, robustness, human oversight and interplay with MDR and IVDR.
View the dossier →PRISM v1: health AI assessment framework
Definitions, evidence levels, decision rules and a first application to an open model.
Explore PRISM →Editorial roadmap
Next topics
- Data leakage and patient overlap
- Temporal leakage in longitudinal data
- Choosing useful health AI metrics
- External validation and multi-site generalisation
- Evaluating hallucinations in generative health AI
QSTOM-IT
Would you like to suggest a topic or contribute to a review?
QSTOM-IT publications are intended to be challenged by technical, scientific, clinical and regulatory experts.