CMS Weighs in on the Use of Algorithms and Artificial Intelligence in Coverage Determinations

Manatt, Phelps & Phillips, LLP
Contact

Manatt, Phelps & Phillips, LLP

Apparently in light of recent litigation against Medicare Advantage (MA) plans, on February 6, 2024, CMS published responses to frequently asked questions (FAQs) regarding the use of algorithms and artificial intelligence in coverage determinations.  The FAQs clarify that MA plans can use algorithms and software tools, including AI, to assist in decisions as long as all regulatory requirements are met, including, in the case of adverse medical necessity determinations, individualized determinations based upon a patient’s circumstances.1

As set forth in the April 5, 2023 final rule concerning, among other things, MA coverage criteria for basic benefits and the use of prior authorization, in making coverage determinations, MA plans must take into account individual circumstances, including medical history and physician recommendations. 42 C.F.R. § 422.101(c)(i)(C). Additionally, for benefits covered under Parts A and B, plans must apply criteria stated in Medicare statutes, regulations, national coverage determinations (NCDs) and local coverage determinations (LCDs) for their service areas. 42 C.F.R. § 422.101(b), (c)(i)(A). Plans can impose additional internal coverage criteria only if Medicare coverage criteria are not fully established, criteria are evidence-based, clinical benefits of use outweigh clinical harms, and they are publicly accessible. 42 C.F.R. § 422.101(b)(6). Finally, if an adverse medical necessity decision is made, it must be reviewed by a physician or other appropriate health care professional with expertise in the field of medicine or health care that is appropriate for the service at issue. 42 C.F.R. § 422.566(d).

As long as plans meet these requirements, the FAQs indicate that plans can employ AI and other software tools. (See Questions Nos. 2, 7.) For example, AI can suggest the tests an individual should receive for a given set of diagnoses or predict the length of stay for inpatient or post-acute care. However, coverage decisions cannot be based on larger data sets alone without considering patient-specific circumstances, including a patient’s medical history, physician recommendations, and clinical notes. Likewise, a plan cannot terminate services or discharge a patient without reassessing the patient’s condition and providing a specific and detailed explanation to why services are either no longer reasonable and necessary or are no longer covered, including a description of the applicable coverage criteria and rules. Additionally, algorithms cannot adjust coverage criteria or deviate from the publicly posted criteria under § 422.101(b)(6).

The FAQs also note that CMS is concerned that algorithms and new artificial intelligence technologies can exacerbate discrimination and bias and remind MA organizations of the non-discrimination requirements under Section 1557 of the Affordable Care Act. Accordingly, the FAQs state that MA organizations should ensure that any software tools used are not perpetuating or exacerbating existing bias or introducing new biases, although the FAQs do not indicate how that should be accomplished.2

This guidance comes amidst challenges to plan deployment of AI in ways that allegedly deprive members of fair and thorough claim investigations. For example, lawsuits have challenged algorithms that allegedly automatically denied authorizations for tests that did not match plan-determined diagnosis or that denied payment for post-acute care longer than AI-predicted lengths of stay without individualized determinations.

While emphasizing the need for individualized review if making adverse clinical coverage determinations, the FAQs indicate that plans can use AI and other software tools for a wide range of activities, including determining compliance with coverage criteria, facilitating coverage approvals, identifying alternative treatments and levels of care, and studying the effectiveness of interventions, all of which could ultimately bring significant benefits for enrollees, providers and plans.

The bottom line is plans can leverage AI as a tool to facilitate coverage decisions but cannot fully delegate the process to AI.


1 The FAQs are available at https://www.aha.org/system/files/media/file/2024/02/faqs-related-to-coverage-criteria-and-utilization-management-requirements-in-cms-final-rule-cms-4201-f.pdf. The final rule, titled 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, was published in 88 FR 22120 and is available at https://www.federalregister.gov/documents/2023/04/12/2023-07115/medicare-program-contract-year-2024-policy-and-technical-changes-to-the-medicare-advantage-program. 

2 The Biden Administration’s regulations implementing the Section 1557 prohibitions on discrimination in the context of clinical algorithms remain pending.  https://www.federalregister.gov/d/2022-16217/p-1615

DISCLAIMER: Because of the generality of this update, the information provided herein may not be applicable in all situations and should not be acted upon without specific legal advice based on particular situations.

© Manatt, Phelps & Phillips, LLP | Attorney Advertising

Written by:

Manatt, Phelps & Phillips, LLP
Contact
more
less

Manatt, Phelps & Phillips, LLP on:

Reporters on Deadline

"My best business intelligence, in one easy email…"

Your first step to building a free, personalized, morning email brief covering pertinent authors and topics on JD Supra:
*By using the service, you signify your acceptance of JD Supra's Privacy Policy.
Custom Email Digest
- hide
- hide