Adds definitions for "Predictive Decision Support Intervention" (Predictive DSI) involving AI algorithms for healthcare decision-making. Requires certification and configuration of AI-based decision support systems, ensuring fairness, transparency, risk management, and external validation processes. Establishes standards for data interoperability and API access.
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is a binding federal regulation from the Department of Health and Human Services with mandatory compliance requirements, enforcement mechanisms, and legal authority under 42 U.S.C. 300jj-11 and 300jj-14. The document uses mandatory language throughout ('must', 'shall') and establishes certification requirements for health IT systems.
The document has good coverage of approximately 6-8 subdomains, with strong focus on discrimination and fairness (1.1, 1.3), privacy (2.1), AI system transparency and robustness (7.3, 7.4), and governance structures. Coverage is concentrated in healthcare AI safety, fairness in clinical decision support, and technical standards for interoperability.
This regulation primarily governs the Health Care and Social Assistance sector, with specific focus on health information technology systems used in clinical settings. It also has implications for the Information sector (health IT developers) and Professional and Technical Services sector (IT consultants implementing these systems).
The document primarily focuses on the Deploy and Operate and Monitor stages of the AI lifecycle, with significant coverage of Verify and Validate. It establishes requirements for how Predictive Decision Support Interventions must be validated, deployed with appropriate documentation, and monitored for ongoing performance. There is some coverage of Build and Use Model through requirements for training data documentation and model development practices.
The document explicitly defines and covers Predictive Decision Support Interventions (Predictive DSI) as a specific type of AI system for healthcare. It also addresses evidence-based decision support interventions. The document focuses on task-specific AI for clinical decision support rather than general-purpose AI. There is no mention of compute thresholds, foundation models, or open-weight models.
Department of Health and Human Services; Office of the National Coordinator for Health Information Technology (ONC)
The document is a federal regulation issued by the Department of Health and Human Services, with the Secretary adopting certification criteria. The ONC is explicitly identified as the responsible office.
Department of Health and Human Services; Office of the National Coordinator for Health Information Technology (ONC); The Secretary
The Department of Health and Human Services, through the Secretary and ONC, has enforcement authority over the certification program and compliance with the established criteria.
Office of the National Coordinator for Health Information Technology (ONC); limited set of identified users
The ONC oversees the certification program. The regulation also requires health IT systems to enable designated users to monitor intervention performance, validity, and fairness over time.
health IT developers; Health IT Module developers; certified Health IT Module providers
The regulation explicitly targets health IT developers who create and provide certified Health IT Modules, including those with Predictive Decision Support Interventions. These entities must comply with certification criteria and update requirements.
7 subdomains (4 Good, 3 Minimal)