Glossary
Clinical validation
Evidence that a medical tool's output is accurate and meaningful for the patients and setting it is intended for.
Clinical validation shows that a tool, such as an AI model that reads scans, produces accurate and clinically meaningful results in its intended patient population and setting. It is usually distinguished from technical (analytical) validation, which checks that the software works as designed, and from clinical benefit, which asks whether using the tool improves patients' outcomes.
External validation, testing on data from different hospitals or populations than those used to build the tool, is an important part of it, because AI models often perform worse on new sites.
A company saying its tool is "clinically validated" should be able to say in which study, on which patients, and against what comparison. Validation of accuracy is not evidence of patient benefit.
Read more: our explainer