Medicare Program; Contract Year 2024 Policy and Technical Changes to the Medicare Advantage Program, Medicare Prescription Drug Benefit Program, Medicare Cost Plan Program, and Programs of All-Inclusive Care for the Elderly

Overview

The Centers for Medicare & Medicaid Services (CMS) published the "Medicare Program; Contract Year 2024 Policy and Technical Changes to the Medicare Advantage Program, Medicare Prescription Drug Benefit Program, Medicare Cost Plan Program, and Programs of All-Inclusive Care for the Elderly" final rule on April 12, 2023. This comprehensive regulation primarily revises policies for Medicare Advantage (MA) and Medicare Prescription Drug Benefit (Part D) programs, with most provisions becoming applicable for coverage beginning January 1, 2024. Although the official title does not explicitly reference Artificial Intelligence (AI), the rule's framework for utilization management, coverage criteria, and patient protections inherently governs the use of advanced technologies, including AI, within MA plans. The rule aims to expand access to affordable healthcare, improve health equity, and enhance consumer protections by addressing various aspects such as marketing and communications, Star Ratings, network adequacy, and prior authorization processes.

Crucially, CMS subsequently issued an FAQ memo on February 6, 2024, specifically clarifying how MA organizations can use AI and algorithms in making coverage determinations under the April 2023 Final Rule. This guidance emphasizes that while AI tools can assist MA plans, they must not be the sole basis for coverage decisions. Instead, all determinations must be individualized, considering the patient's specific medical circumstances, physician recommendations, and clinical notes. The FAQ memo underscored the importance of ensuring that AI systems comply with all existing regulations, particularly those concerning medical necessity and non-discrimination. This approach reflects CMS's intent to foster innovation while safeguarding beneficiaries from potential harms associated with algorithmic decision-making, such as biased outcomes or inappropriate denials of care.

Definitions

While the 2023 Final Rule itself does not introduce specific definitions for 'Artificial Intelligence' or 'algorithms,' the subsequent CMS FAQ memo from February 2024 provides crucial interpretative guidance on their application within the Medicare Advantage context. The memo clarifies that 'an algorithm or software tool can be used to assist MA plans in making coverage determinations.' This implicitly defines these tools as computational systems designed to process data and generate insights or recommendations that inform decisions regarding patient care and coverage. The emphasis is on their assistive role, rather than their capacity for autonomous decision-making in critical areas like medical necessity. The guidance further distinguishes between using AI as a supportive tool and relying on it as the sole determinant, particularly when it comes to individual patient circumstances.

Key terms from the broader 2023 Final Rule that are relevant to AI's application include 'coverage determinations,' which refer to the decisions made by MA organizations regarding whether a particular item or service is covered under a beneficiary's plan. 'Medical necessity' is also a critical concept, defined by CMS as the determination that items or services are reasonable and necessary for the diagnosis or treatment of illness or injury, or to improve the functioning of a malformed body member. The rule reinforces that these determinations must be based on the individual patient's circumstances, including medical history, physician recommendations, and clinical notes, a principle that directly impacts how AI can be integrated. Furthermore, 'utilization management' encompasses the various processes, such as prior authorization, used by MA plans to manage healthcare services, which are now subject to explicit rules regarding AI integration.

Governance and Institutional Framework

The Centers for Medicare & Medicaid Services (CMS), an agency within the U.S. Department of Health and Human Services (HHS), serves as the primary regulatory body overseeing the implementation and enforcement of this final rule. CMS is responsible for administering the Medicare Advantage and Part D programs, establishing the policies and technical requirements that govern MA organizations. Through this rule and subsequent guidance, CMS sets the parameters for how MA plans operate, including their use of advanced technologies like AI. The agency's role extends to monitoring compliance, issuing clarifications, and ensuring that MA organizations adhere to the established standards for patient care, coverage determinations, and consumer protections. This oversight is critical given the increasing adoption of AI in healthcare and the potential impact on beneficiary access to services.

The institutional framework also involves the MA organizations themselves, which are private companies contracting with CMS to provide Medicare benefits. These organizations are directly responsible for implementing the provisions of the final rule and integrating AI tools in a compliant manner. They must establish internal governance structures to ensure that AI systems used for coverage determinations meet CMS requirements, including those related to individual patient assessment, transparency, and bias mitigation. The rule also implies a role for clinical review committees and qualified healthcare professionals within MA organizations, as human oversight and individualized clinical judgment are emphasized over sole reliance on algorithmic outputs. This multi-layered governance approach aims to balance technological innovation with robust patient safeguards within the complex Medicare Advantage ecosystem.

Key Focus Areas

A primary focus area of the 2023 Final Rule, as clarified by CMS's AI guidance, is the individualized nature of coverage determinations. CMS explicitly states that while AI and algorithms can assist MA plans, coverage decisions must always be based on the specific circumstances of the individual patient, including their medical history, physician recommendations, and clinical notes. This prevents MA organizations from using algorithms to make blanket denials or to solely determine coverage based on large datasets without considering unique patient needs. This emphasis ensures that AI acts as a supportive tool for clinicians and plans, rather than a replacement for human judgment and personalized care. For instance, an algorithm predicting a length of stay for post-acute care cannot be the sole basis for terminating services.

Another critical focus is transparency and fairness in AI utilization. The rule, interpreted through the AI guidance, requires MA organizations to disclose how AI algorithms inform clinical decisions. This includes providing information on the data sources used, the methodology employed by the AI, and any potential biases identified within the system. Furthermore, MAOs must ensure that AI-driven recommendations align with evidence-based practices and established clinical guidelines. This focus on transparency extends to the public disclosure of coverage criteria, ensuring that beneficiaries and providers understand the basis for coverage decisions, including where AI tools are involved. The rule also underscores the importance of equity and bias mitigation, recognizing that AI algorithms have the potential to perpetuate or exacerbate existing biases in healthcare. MA organizations are mandated to comply with non-discrimination requirements, such as Section 1557 of the Affordable Care Act, ensuring that AI systems do not lead to discriminatory outcomes.

Implementation Framework

The implementation framework for this regulation centers on Medicare Advantage organizations integrating the rule's requirements into their operational processes, particularly concerning utilization management and coverage determinations. For AI and algorithmic tools, this means MAOs must ensure that any such technology used to assist in coverage decisions is fully compliant with all applicable CMS rules. This includes establishing robust internal protocols for how AI outputs are reviewed and incorporated into final decisions, ensuring that human clinicians retain ultimate authority and that decisions are always individualized. MA organizations are expected to review and update their internal coverage criteria to align with traditional Medicare rules and to publicly explain how any additional criteria provide clinical benefits that outweigh potential harms.

Furthermore, the implementation requires MA organizations to enhance their data governance practices to support the ethical and compliant use of AI. This involves adhering to HIPAA requirements for data privacy and security, including obtaining patient consent for Protected Health Information (PHI) use, implementing strong encryption, access controls, and data anonymization techniques. MAOs must also assess on a case-by-case basis whether the use of PHI for training AI models is permissible. The rule also sets forth requirements for prior authorization processes, mandating that approvals for a course of treatment be valid for as long as medically necessary and that plans cannot deny coverage based on a lack of medical necessity if a prior authorization was already approved, except for specific reasons like fraud.

Monitoring and Evaluation

CMS actively monitors and evaluates the compliance of Medicare Advantage organizations with the provisions of this final rule, including the appropriate use of AI and algorithms in coverage determinations. The agency's oversight mechanisms include reviewing MA plan applications, conducting audits, and responding to beneficiary complaints and appeals. CMS expects MA organizations to maintain detailed documentation relevant to medical necessity reviews, especially when AI tools are utilized, to demonstrate adherence to the requirement for individualized patient assessments. The agency has also emphasized that MA organizations should refer to the coverage criteria and summary of evidence presented by Medicare Administrative Contractors (MACs) as a guide for publicly presenting information, which aids in transparency and allows for better public scrutiny.

Evaluation of the rule's effectiveness, particularly concerning AI, will likely involve assessing trends in coverage denials, appeal outcomes, and the overall impact on beneficiary access to care. CMS's continued issuance of guidance, such as the February 2024 FAQ memo, indicates an ongoing evaluative process to clarify and refine regulatory expectations as AI technology evolves and its application in healthcare expands. The agency's focus on transparency, bias mitigation, and non-discrimination suggests that future evaluations will also consider the equitable impact of AI systems and whether they are exacerbating existing disparities or promoting health equity. This continuous monitoring and evaluation framework is essential to ensure that the benefits of AI in healthcare are realized without compromising patient protections or the integrity of the Medicare program.

Penalties, Liability, and Appeals

The 2023 Final Rule, along with existing Medicare regulations, establishes a framework for penalties and liability for Medicare Advantage organizations that fail to comply with its provisions, including those pertaining to the use of AI in coverage determinations. While the rule does not introduce new specific penalties solely for AI misuse, non-compliance with coverage determination rules, medical necessity requirements, or anti-discrimination mandates can lead to significant consequences for MA plans. CMS has the authority to deny applications for new contracts or service area expansions based on an organization's past performance, including issues related to compliance actions or low Star Ratings. Furthermore, MA organizations are subject to financial remittances if they fail to meet the minimum medical loss ratio requirements, which could be impacted by inappropriate denials of care.

Beneficiaries retain their right to appeal coverage denials made by MA organizations, regardless of whether AI was involved in the initial decision. The rule emphasizes that if an MA coordinated care plan approved a covered service through a prior authorization or pre-service determination, it is generally barred from denying coverage later based on lack of medical necessity, except for good cause or reliable evidence of fraud. This provides a crucial safeguard against arbitrary denials, including those potentially influenced by AI. The appeals process is a vital mechanism for redress, though concerns have been raised about the length of time patients may wait for decisions and the low rates of appeals. MA organizations bear the liability for ensuring that their AI tools comply with all applicable rules, and any denials resulting from non-compliant AI usage could be overturned on appeal, potentially leading to reputational damage, financial penalties, and increased scrutiny from CMS.

Relationship to Other Instruments

This 2023 Final Rule operates within a complex web of existing federal laws and regulations governing healthcare and technology in the United States. Most notably, it interacts significantly with the Health Insurance Portability and Accountability Act (HIPAA), particularly concerning data privacy and security. MA organizations using AI must ensure their data practices, including the collection, storage, and processing of patient data for AI training and deployment, are fully compliant with HIPAA's privacy and security rules. This includes obtaining patient consent for the use of Protected Health Information (PHI) and implementing robust safeguards against unauthorized access or disclosure.

The rule also builds upon and clarifies existing Medicare statutes and regulations related to medical necessity and coverage criteria. It reinforces the principle that MA plans must follow traditional Medicare coverage rules and limit the use of internal or proprietary guidelines that deviate from these established standards. Furthermore, the guidance on AI explicitly references and reinforces Section 1557 of the Affordable Care Act (ACA), which prohibits discrimination on the basis of race, color, national origin, sex, age, or disability in certain health programs and activities. This connection highlights CMS's concern that AI algorithms could inadvertently perpetuate or exacerbate existing biases, thus requiring MA organizations to ensure their AI systems uphold these non-discrimination mandates. The rule also relates to broader efforts by CMS to improve interoperability and streamline prior authorization processes, as seen in other recent CMS final rules and FAQ documents.

International Alignment

As a domestic regulation issued by a U.S. federal agency, the CMS Medicare Advantage Final Rule primarily focuses on the regulatory landscape within the United States. Therefore, direct international alignment is not an explicit objective or component of this specific document. The rule's provisions are tailored to the unique structure and operational requirements of the U.S. Medicare Advantage program and its beneficiaries. However, the principles underpinning the CMS guidance on AI in healthcare, such as the emphasis on patient safety, transparency, fairness, and the mitigation of bias, resonate with broader international discussions and emerging regulatory frameworks for artificial intelligence.

Many countries and international bodies, including the European Union with its AI Act and the OECD with its AI Principles, are grappling with similar ethical and governance challenges posed by AI in critical sectors like healthcare. While the specific legal mechanisms and enforcement approaches may differ, the shared concerns about algorithmic accountability, data protection, non-discrimination, and human oversight in AI-driven decision-making reflect a growing global consensus. The CMS rule, by permitting AI while imposing strict safeguards, contributes to the evolving global dialogue on responsible AI deployment, particularly in highly regulated and sensitive domains such as healthcare coverage and utilization management. Its focus on individualized care and clinical oversight aligns with a human-centric approach to AI that is gaining traction internationally.

Implementation Timeline

Compliance Checklist

Sources and References