AI Assurance Frameworks for Digital Health Developers

Mon, 07 Oct 2024 02:54:30 -0400

Assurance Methods

Introduction

The ONC HTI-1 final rule introduced Intervention Risk Management requirements for predictive Decision Support Interventions under § 170.315(b)(11). That raised an obvious follow-up for developers: which assurance framework should you actually use to satisfy it?

Three established programs influenced the early conversation. Stanford’s Fair, Useful, and Reliable AI Models (FURM), Duke University Health System’s Algorithm-Based Clinical Decision Support (ABCDS), and MITRE’s repeatable AI assurance process. The Coalition for Health AI later published its Responsible AI Guide drawing on similar foundations.

AI Assurance Frameworks on AI Governance, HITRUST, and ONC Support

AI Assurance Frameworks for Digital Health Developers

Assurance Methods

Introduction