Key Impacts of HTI-1 for LTPAC

Historically, the Office of the National Coordinator for Health Information Technology (ONC) has been responsible for establishing Health IT (HIT) standards that enable the implementation of executive and legislative healthcare policies. The recent designation of Assistant Secretary of Technology Policy (ASTP) given to the ONC has enhanced their mandate to not only support but develop policy in the realm of HIT. In its new capacity as ASTP, it has recently introduced two significant policies under the Health Data, Technology and Interoperability initiative (HTI): HTI-1 and HTI-2. This paper will focus on the implications of HTI-1 specifically for HIT developers and their customers.

Positioned as a crucial component of the 21st Century Cures Act, HTI-1 aims to drive meaningful progress in HIT by expanding public health reporting capabilities, fostering health equity, and promoting the responsible development and use of artificial intelligence (AI) through transparency. Through a comprehensive exploration of HTI-1, this paper will analyze the transformative impact it will have on the HIT landscape and provide insights into the opportunities and challenges it presents.

Code Standards and Decision Support Intervention (DSI)

The HTI-1 rule is a significant step towards promoting interoperability, extending it to other Health and Human Services (HHS) organizations and ensuring the availability of patient health information across different healthcare systems. This is accomplished by requiring HIT developers to update their electronic health record (EHR) with specific code set versions to ensure all Certified EHR Technology (CEHRT) systems meet the same standards. These code sets including SNOMED, LOINC, CDC Race and Ethnicity, and others include expanded sociodemographic data that can be captured to track and report on health equity measures that may be tied to value-based care incentives.

As the data set continues to expand through greater access and new standards amid a growing number of AI technologies, the need for clear guidance on the use of health information in AI is critical. HTI-1 introduces the first ASTP policy on the use of AI in Clinical Decision Support (CDS), now classified as Decision Support Intervention (DSI) and represents the first substantial revision to the CEHRT certification program requirements for CDS since 2012.

The final rule includes DSI requirements to expose source attributes for both evidence-based DSI and predictive DSI, with the goal of enabling the user to understand how the DSI works and be able to evaluate how specific data such as demographics are used. The aim of the DSI requirements is to bring transparency to these solutions while continuing to advance health equity. 
For evidence-based DSI, the previous requirements for CDS apply, which include a bibliographic citation of interventions, the developer, funding source of implementation, and release date including revision dates if applicable. HTI-1 further expands the source attributes to include use of specific sociodemographic data:

• Race
• Ethnicity
• Language
• Sexual orientation
• Gender identity
• Sex
• Date of birth
• Social determinants of health
• Health status assessment

Transparency in AI related to predictive DSIs is outlined in HTI-1 as a path for healthcare organizations to assess the solutions for fairness, appropriateness, validity, effectiveness, and safety. The rule lists 31 source attributes for predictive DSIs covering the following nine domains:

• Details and output of the intervention
• Purpose of the intervention
• Cautioned, out-of-scope use of the intervention
• Intervention development details
• Process used to ensure fairness in development of the intervention
• External validation process
• Quantitative measure of performance
• Ongoing maintenance of intervention implementation and use
• Updated and continued validation or fairness assessment schedule

Governance and Feedback

Risk management plays a critical role in the design, development, and use of interventions to ensure they are safely, effectively, and ethically used. There are three key components to governance of DSI:

1. Establishing a multi-disciplinary team
   Forming a team with healthcare professionals, clinicians, clinical informaticists, legal experts, data scientists, and other relevant stakeholders with diverse expertise and perspectives can help ensure DSI meets legal and ethical standards.
2. Ensuring transparency
   One of the main intents behind HTI-1 is to foster transparency by requiring clear documentation on the rationale, evidence, and decision-making process of the intervention.
3. Engaging for feedback
   Establishing collaborative partnerships with your end users and stakeholders is essential to ensure their continued involvement throughout the lifecycle of the DSI. HTI-1 requires that user feedback mechanisms are put in place to ensure the DSI performs as expected.

Getting LTPAC Ready for HTI-1

To prepare for the HTI-1 rule, LTPAC operators should focus on the following key steps

1. Reach out to your EHR company to confirm they are ONC-certified and have implemented the code set versions required by the HTI-1 rule, such as SNOMED, LOINC, and CDC Race and Ethnicity. This not only ensures the seamless exchange of data, but it also protects you from claims of information blocking.
2. Familiarize yourself with the expanded sociodemographic data included in the required versions of the code sets. This data is crucial for tracking and reporting health equity measures tied to value-based care incentives. Understanding this data will help you effectively capture and utilize it, demonstrating your commitment to driving greater health equity.
3. Understand the guidelines introduced by the HTI-1 rule and be able to identify which solutions fall under evidence-based DSI and which AI solutions fall under the predictive DSI definition. Ensure that any DSIs you use or plan to implement comply with these requirements and the applicable source attributes specified in the rule.

By taking these steps, LTPAC operators can effectively prepare for the HTI-1 rule, enhance interoperability, promote health equity, and help ensure the responsible use of AI in healthcare.