MedTech Terms
    The authoritative reference
    All terms

    ONC HTI-1 / HTI-2

    ONC certification rules adding transparency and risk-management requirements to certified health IT, including AI/predictive DSI.

    Reviewed by Christian Espinosa, Founder, Blue Goat CyberLast reviewed May 5, 2026

    Definition

    HTI-1 (final 2024) introduced the Decision Support Intervention criterion requiring source attributes and intervention risk management for predictive models in certified EHRs. HTI-2/HTI-3 extend transparency, interoperability, and information-blocking requirements.
    What the regulation says
    The ONC Health IT Certification Program, specifically the HTI-1 Final Rule, mandates new requirements for certified electronic health records (EHRs) regarding decision support interventions. These regulations aim to enhance patient safety and transparency when predictive algorithms are used in clinical settings. The rule requires that certified health IT modules identify source attributes for decision support interventions and implement intervention risk management, as detailed in 45 CFR Part 170. This directly impacts how AI and machine learning models integrated into EHRs must be developed, validated, and deployed.

    What this means in practice

    Even non-EHR vendors are affected: model 'nutrition labels' are increasingly expected by HDOs procuring AI-enabled tools.

    Examples

    • A developer integrating a diagnostic AI algorithm into a certified EHR must ensure the algorithm's training data, validation metrics, and known biases are clearly documented as source attributes.
    • A hospital procuring an AI-powered clinical decision support system requests explicit documentation of its intervention risk management plan, demonstrating compliance with ONC HTI-1 requirements.
    • A health IT vendor creates a mechanism within their EHR to display the source attributes of a predictive model whenever a clinician receives a decision support alert.
    Common pitfalls
    • •Misconception: These rules only apply to EHR vendors, but any developer whose software integrates with or provides data to certified EHRs may be indirectly affected.
    • •A common pitfall is failing to adequately document the 'source attributes' of predictive models, which can lead to non-compliance during certification.
    • •Overlooking the necessity of a robust intervention risk management plan for all decision support interventions is a frequent error.
    • •Assuming that existing risk management processes are sufficient without specifically addressing the unique risks of AI-driven interventions can lead to gaps.
    • •Failure to keep up with subsequent ONC rules, such as potential HTI-2 or HTI-3 updates, could lead to outdated compliance strategies.

    Frequently asked questions

    Source attributes refer to the information that explains the basis of a decision support intervention, including the algorithms, data sources, and evidence used. This is crucial for transparency and user understanding in clinical decision-making.
    Grouped by theme

    Primary references

    3 sources
    Link health: 3 verified· last checked 2026-06-20
    ONC·1FDA·2
    1. 1
      ONC HTI-1
      Verified
      ONChealthit.gov
    2. 2
      FDA - Software as a Medical Device (SaMD)
      Verified
      FDAfda.gov
    3. 3
      FDA - AI/ML-Enabled Medical Devices
      Verified
      FDAfda.gov

    Inline markers like [1] jump to the matching reference above.