What is the problem
Artificial intelligence and increased computing power have long held the promise of improving prediction and prognostication in health care. Now, several commercial algorithms have received regulatory approval for broad clinical use. But the barrier for entry of new advanced algorithms has been low. To unlock the potential of advanced analytics while protecting patients, how should regulatory bodies ensure that advanced algorithms meet accepted standards of clinical benefit?
What are we doing
In the journal Science, we proposed a comprehensive regulatory frameworks for AI predictive analytic tools. These including anchoring evaluations to discrete clinical outcomes and benchmarking against realistic standards.
Our framework was cited in regulatory documents published by the European Union and World Health Organization on trustworthy AI for clinical use. Dr. Parikh frequently cites these recommendations as an advisor for Medicare and Nonprofit committees.