The American healthcare system desperately needs support. The American population is aging, and life expectancy is increasing, with those 65 or older accounting for one out of every five Americans by 2030—and 80 percent of those having at least one chronic condition. Healthcare costs are increasing; costs have already risen to approximately $4.3 trillion annually, representing at least 17 percent of the U.S. gross domestic product. Finally, and no less troubling, the healthcare workforce is experiencing a growing shortage, with 30 out of the 35 physician specialties projected to see deficits by the 2030s. Rural areas will be facing the brunt. The efficiencies artificial intelligence (AI) offers are vital to overcoming these challenges.

AI uses in healthcare range from back-office support and scheduling help to clinical decision support, encompassing a wide range of risk levels and types. We urge Congress and other policymakers to view the proven capabilities of AI in healthcare to assist patients, providers, technology developers, and others throughout the healthcare ecosystem through the lens of “quadruple aim” framework:

      1. Enhance Patients’ Outcomes and Experiences. AI-supported interventions and treatments offer the ability to improve individual patient outcomes and engagement. Further, AI-supported engagement tools can help patients take steps to prevent disease and to stay engaged in their care after diagnosis.
      2. Improve Overall Population Health. AI can help surface trends and suggest courses of action to address emerging or persistent health issues in population sets. An especially tricky aspect of population health management is that unforeseen factors—which are sometimes aspects of social determinants of health (SDOH) rather than traditional health indicators—often produce significant effects on the health of a given demographic. This is where large datasets and powerful analytical tools can equip providers and public health officials with actionable information and analyses.
      3. Reduce Costs. AI capabilities have already been shown to be critical tools in maximizing efficiencies across the healthcare value chain. AI tools are starkly needed to assist in reducing administrative costs, as well as in leveraging data collected from outside of the doctor’s office between visits to support timely care plan updates and interventions that save the health system significant costs.
      4. Improve Healthcare Professionals’ Experience. The healthcare sector is already experiencing a workforce shortage, with today’s frontline clinicians at high risk of burnout. AI-supported tools offer the opportunity to improve healthcare professionals’ experience by maximizing efficiencies and capabilities, allowing them to reach more patients with better care. Yet, while AI tools are increasing job satisfaction and reducing burnout, they are not intended to replace the provider.

As Congress, federal agencies, states, and the private sector continue to explore how to responsibly bring AI to the healthcare system, the Connected Health Initiative (CHI) recommends that policymakers align with CHI’s cross-sectoral consensus health AI policy principles. These principles recognize the opportunities and challenges AI provides to healthcare and provide baseline recommendations across key areas, including quality assurance/oversight, thoughtful design, access and affordability, bias detection/mitigation, and data privacy and security, among other critical areas.[1] Other discrete opportunities for Congress and the federal government include:

  • Leveraging consensus medical AI terminology[2] and CHI’s cross-sectoral consensus understanding of the unique roles and interdependencies/shared responsibilities amongst the healthcare AI value chain[3] as a baseline for the government’s approach to health AI;
  • Building on the leading efforts of the National Institute of Standards and Technology’s voluntary AI Risk Management Framework[4] and CHI’s health-specific recommendations on its development and implementation[5] to ensure that a coordinated approach is taken to health AI that scales risk mitigation requirements to intended uses and known harms;
  • Helping build trust amongst providers and patients by enhancing transparency (supporting the sharing of information about the AI’s intended use, development, performance, etc.) consistent with CHI’s recommendations in Advancing Transparency for Artificial Intelligence in the Healthcare Ecosystem;[6]
  • Advancing overdue Medicare coverage and payment policy changes that appropriately categorize AI (e.g., recognize that AI software as a medical device is appropriately categorized and paid for as a direct practice expense);
  • Responsibly expanding support for its use in the prevention and treatment of beneficiaries’ acute and chronic conditions;
    • Payment and coverage for healthcare AI systems must be informed by real-world workflow and human-centered design principles; enable physicians to prepare for and transition to new care delivery models; support effective communication and engagement between patients, physicians, and the healthcare team; seamlessly integrate clinical, administrative, and population health management functions into the workflow; and seek end-user feedback to support iterative product improvement.
  • Ensuring that AI can support the transition to value-based care (e.g., eliminating barriers to the responsible use of health AI and other innovative technologies in the Merit-based Incentive Payment System and in Advanced Alternative Payment Models), consistent with recommendations in CHI’s Leveraging Digital Health to Realize Value-Based Care.[7]

We are seeing the positive effects that AI can have in healthcare. As AI technology advances, it is important to promote—and eliminate barriers to—its responsible use while ensuring AI is safe and effective. We need to ensure that patients are getting the care that they need while also making sure that healthcare workers are not overburdened to provide high-quality care.

[1] CHI Health AI Task Force’s deliverables are accessible at https://actonline.org/2019/02/06/why-does-healthcare-need-ai-connected-health-initiative-aims-to-answer-why/.

[2] E.g., https://www.ama-assn.org/practice-management/cpt/cpt-appendix-s-ai-taxonomy-medical-services-procedures.

[3] https://connectedhi.com/wp-content/uploads/2024/02/CHI-Health-AI-Roles.pdf.

[4] https://www.nist.gov/itl/ai-risk-management-framework.

[5] https://actonline.org/wp-content/uploads/Policy-Principles-for-AI.pdf

[6] CHI’s recommendations on necessary policy changes to enhance transparency for healthcare AI are available at https://bit.ly/3Gd6cxs.

[7] https://connectedhi.com/wp-content/uploads/2022/02/LeveragingDigitalHealth.pdf.