AHCP Blog

New CMS Guidance on AI in Medicare Advantage Plans

Written by AHCP | 2/29/24 6:00 PM

The Centers for Medicare & Medicaid Services (CMS) recently issued guidance on its final 2024 Medicare Advantage rule, detailing new regulations on the use of Artificial Intelligence (AI) in Medicare Advantage plans. This rule reflects CMS's efforts to balance innovation in AI with the need for ethical and patient-centric healthcare delivery.

As Jakob Emerson explains in a February 8, 2024 article in Becker's, algorithms and AI-powered software tools can be used to support Medicare Advantage plans in making coverage decisions. However, these tools must comply with CMS' internal benefits requirements and nondiscrimination rules under the ACA. CMS emphasizes the need to ensure these tools do not exacerbate discrimination and bias.

This guidance is significant given the increased scrutiny of payers' use of AI, as detailed in a February 12th, 2024 article in Fierce Healthcare. Federal lawmakers, including Senate Committee on Finance Chairman Ron Wyden, have expressed concerns about the biases in big data systems and are advocating for stronger regulation to protect patients from bias in AI. In the other chamber, as the American Hospital Association reports, more than 30 members of the House of Representatives urged CMS in a November 3, 2023 letter “to monitor and evaluate how Medicare Advantage plans use artificial intelligence and algorithms to guide their coverage decisions, and ensure these tools comply with Medicare rules and do not create barriers to care.”

The ongoing discussion in Congress highlights the need for a balance between protecting innovation and ensuring patient and privacy protection, especially in federal programs like Medicare and Medicaid.

This push for regulation aligns with the lawsuits facing major Medicare Advantage insurers, accused of using AI algorithms to wrongfully deny care. For instance, two patients filed a lawsuit against Humana, alleging reliance on AI tools for coverage determinations in long-term care. More information on the suit is available from McKnights Long Term Care News. Court cases like this highlight the importance of CMS's memo to insurers, clarifying that AI or algorithms cannot be used to determine coverage or deny care based on larger data sets instead of individual patient circumstances.

But the recent CMS guidance doesn’t clear up all the confusion. Michelle Mello, Ph.D., a professor at Stanford University, mentioned in the Fierce Healthcare article, emphasizes the need for further clarification on the proper use of algorithms. She also highlights the necessity of meaningful human review in healthcare decisions involving AI.

Mark Sendak, M.D., from the Duke Institute for Health Innovation, and Ziad Obermeyer, M.D., a professor at the University of California, Berkeley, echoed these sentiments in their testimonies to Senate lawmakers. They discussed the potential of AI in healthcare and the need for transparent and accountable AI systems, urging Congress to support healthcare organizations in navigating the uncharted territory of AI tools.

Despite these challenges and concerns, CMS's stance on AI, as evident from its May 13, 2021 blog post on the CMS Artificial Intelligence Health Outcomes Challenge, is not entirely prohibitive. This challenge showed CMS's interest in exploring AI's potential to improve patient care, even before platforms like ChatGPT had hit the market. As further evidence that CMS is leaning into the artificial intelligence trend, the agency has actually set up an AI website: ai.cms.gov. Here is an excerpt from the site:

At CMS, Artificial Intelligence (AI) has the power to reshape the way we use data to make decisions. In fact, because CMS is such a data-rich agency, there is no better place to implement AI technology. To do so responsibly, we must educate our workforce, share knowledge with our partners, follow ethical standards, and experiment with new methodologies.

That is why we have created this website. It offers a starting point for stakeholders interested in any aspect of AI at CMS.

In short, artificial intelligence is something that everyone is still trying to figure out. CMS recognizes its potential but also the potential for abuse. That’s why the agency is passing rules to protect patients and make sure that coverage decisions are made based on their unique situation rather than using AI, algorithms, and aggregated data to make these decisions. But it’s also why the agency is moving forward with initiatives to leverage the power of AI to better serve America's healthcare needs.

At AHCP, we first wrote about artificial intelligence in January of 2023, shortly after ChatGPT hit the scene and started creating so much buzz. And we’ll no doubt write about it more as the technology continues to advance and its impact the insurance industry becomes even more evident. We expect that to be sooner rather than later.