Interpretable Clinical Prediction Models

Some AI models do not need to be explained; evidence of their reliability is enough. But when it comes to many medical applications of AI, the explainability of models is often crucial. This view is shared by the FDA, where regulatory guidelines state ML recommendations must enable a health care provider “… to independently review the basis for such recommendations”.

Although AI systems may be complex, the clinical models produced by them need not be. We investigate state-of-the-art methods (symbolic regression, neurosymbolic AI, and large language models (LLMs)) as tools to generate simple clinical models that clinicians can use to better understand and treat their patients.

An interpretable model of treatment resistant hypertension from La Cava et al. npj Digital Medicine 2023

Code

Feature Engineering Automation Tool

Brush: an Interpretable Machine Learning Library

Funding

PCORI

Selected Papers

A flexible symbolic regression method for constructing interpretable clinical prediction models

William G. La Cava, Paul C. Lee, Imran Ajmal, Xiruo Ding, Priyanka Solanki, Jordana B. Cohen, Jason H. Moore, Daniel S. Herman (2023)

npj Digital Medicine

nature.com  |  medRxiv  |  pdf  |  code  |  HUMIES Silver Medal