Berlin – A growing chorus of healthcare provider groups is urging federal regulators to reconsider proposed changes to health IT certification criteria, particularly those concerning transparency in artificial intelligence (AI) tools. The groups argue that weakening these standards, intended to ensure patient safety and build trust in increasingly complex AI-driven healthcare technologies, could ultimately hinder adoption and potentially compromise care. The debate centers on a proposal by the Assistant Secretary for Technology Policy/Office of the National Coordinator for Health Information Technology (ONC) to streamline certification requirements, aiming to reduce regulatory burdens and costs for developers.
The ONC released its proposed rule in December, outlining plans to remove 34 certification criteria, revise seven and abandon 19 unchanged. The agency estimates these changes would result in $1.53 billion in savings, with $650 million of that benefiting IT developers, providers, and other stakeholders over the next five years, according to Kate Tipping, deputy director of the regulatory and policy affairs division at ASTP, during a meeting of the Health Information Technology Advisory Committee (HITAC) last month. While acknowledging the need to reduce administrative overhead, provider organizations contend that certain criteria, such as those requiring “model cards” for AI tools, are essential for responsible AI implementation in clinical settings.
The Importance of AI Transparency in Healthcare
The core of the disagreement lies in the balance between fostering innovation and safeguarding patient well-being. The proposed changes arrive as AI is rapidly integrating into various aspects of healthcare, from diagnostic imaging and drug discovery to personalized treatment plans and administrative tasks. This accelerated adoption necessitates careful consideration of the potential risks associated with AI, including algorithmic bias, data privacy concerns, and a lack of explainability – often referred to as the “black box” problem. The Health IT Certification Program, established to outline standards and functionality for health IT, plays a crucial role in mitigating these risks.
Currently, the program is voluntary, but certification is often encouraged or required for participation in federal and state programs. The proposed overhaul, stemming from a broader Trump administration initiative to cut back on regulations and spur economic growth, has raised concerns among those who believe that weakening certification standards could undermine the integrity of AI systems used in healthcare. The American Medical Association (AMA), in a formal comment submitted Friday, warned that eliminating transparency requirements would “weaken trust in clinical AI at the very moment its use is accelerating.” Dr. John Whyte, CEO of the AMA, emphasized the need for enforceable transparency baselines to ensure clinicians can confidently utilize AI tools in patient care.
“Model Cards” and the Debate Over Regulatory Burden
The “model cards” concept, finalized by the Biden administration, requires developers to provide detailed documentation about their AI models, including their intended use, limitations, potential biases, and performance metrics. This information is intended to assist clinicians understand how an AI tool works, assess its suitability for a particular patient, and interpret its results accurately. However, some officials argue that these requirements are creating significant burdens for developers without clear evidence of commensurate benefits. Michael Lipinski, director of the regulatory and policy affairs division at ASTP, noted at the HITAC meeting that the requirements add regulatory challenges for certified developers while not applying to those who do not participate in the certification program.
This disparity in regulatory burden is a key point of contention. The Medical Group Management Association (MGMA) has expressed concern that removing certification criteria could shift compliance responsibilities to medical groups, potentially increasing product variability and creating challenges for new market entrants. Anders Gilberg, senior vice president of government affairs at MGMA, argued that this shift could ultimately increase costs for providers. The debate highlights the complex interplay between regulatory oversight, innovation, and the practical realities of implementing AI in diverse healthcare settings.
Privacy and Security Concerns
Beyond AI transparency, provider groups have also voiced concerns about the ASTP’s proposal to remove privacy and security certification criteria. The regulator contends that the costs and burdens associated with these requirements outweigh their benefits, arguing that certification does not guarantee full compliance with the Health Insurance Portability and Accountability Act (HIPAA) security and privacy rules. However, given the increasing frequency of cyberattacks and data breaches targeting the healthcare sector, many believe that maintaining robust security standards is paramount. Ashley Thompson, senior vice president of public policy analysis and development at the American Hospital Association, warned that removing these criteria could pose significant risks to the industry.
Thompson pointed out that cutting these criteria could simply shift costs onto providers, as developers may charge additional fees for security features that were previously included as part of the certification process. This potential cost shift underscores the importance of carefully considering the broader economic implications of regulatory changes. The ongoing debate reflects a broader tension within the healthcare industry: the desire to leverage the transformative potential of AI while simultaneously protecting patient data and ensuring the responsible use of these powerful technologies.
Looking Ahead: Next Steps and Ongoing Dialogue
The ONC is currently reviewing the comments received from stakeholders, including provider groups, industry representatives, and patient advocates. The agency is expected to issue a final rule in the coming months, outlining the specific changes to the Health IT Certification Program. The outcome of this process will have significant implications for the future of AI in healthcare, shaping the regulatory landscape and influencing the development and deployment of AI-powered tools. The Health Information Technology Advisory Committee will continue to advise regulators on health IT policies, playing a crucial role in navigating the complex challenges and opportunities presented by AI.
The debate over health IT certification criteria is not merely a technical discussion; This proves a fundamental question of how we balance innovation with patient safety and trust. As AI continues to evolve and become increasingly integrated into healthcare, ongoing dialogue and collaboration between regulators, providers, developers, and patients will be essential to ensure that these technologies are used responsibly and ethically. The future of healthcare may well depend on our ability to strike this delicate balance.
The next key date to watch is the expected publication of the final rule by the ONC, anticipated within the next quarter. Readers are encouraged to stay informed about these developments and to participate in the ongoing conversation about the role of AI in healthcare. Share your thoughts and experiences in the comments below.