HHS Seeks Sector Input on AI for Fighting Healthcare Fraud

The U.S. Department of Health and Human Services will use “advanced” artificial intelligence tools to more quickly detect and prevent Medicare and Medicaid fraud before scam claims are paid. Regulators are also seeking health sector input on AI as it fleshes out their “major crackdown” on healthcare fraud and consider potential future rulemaking.

See Also: The Power of Peer-to-Peer Communities

HHS’ Centers for Medicare and Medicaid Services on Wednesday announced its AI plans as part of a broader healthcare fraud-busting initiative.

Healthcare legal and privacy experts applauded the effort, but cautioned that the request mentions nothing about how to secure HIPAA-protected information for millions of law-abiding U.S. beneficiaries.

The agency is also calling for a six-month moratorium on new Medicare enrollment for certain durable medical equipment suppliers, and temporarily deferring $259.5 million – and potentially up to $1 billion this year – in federal Medicaid payments to Minnesota for alleged “fraudulent or unsupported” Medicaid claims.

In combination, those measures reflect “a coordinated, data-driven strategy to prevent fraud before it occurs, hold bad actors accountable and protect taxpayer dollars,” HHS said.

“For decades, Medicare fraud has drained billions from American taxpayers – that ends now,” said HHS Secretary Robert F. Kennedy Jr. in a statement. “We are replacing the old ‘pay and chase’ model with a real-time ‘detect and deploy’ strategy, using advanced AI tools to identify fraud instantly and stop improper payments before they go out the door.”

The Centers for Medicare and Medicaid Services “is done trying to catch fraudsters with their hands in the cookie jar – instead, we’re padlocking the jar and letting them starve,” said Dr. Mehmet Oz, CMS administrator in the statement. “This proactive approach will help us crush fraud, protect taxpayer dollars and make sure the vulnerable Americans who depend on our programs get the care they need.”

Seeking Input

In a request for information also issued as part of the effort, the agency is seeking stakeholder input to be considered for potential future rulemaking for “Comprehensive Regulations to Uncover Suspicious Healthcare,” or CRUSH, and “other programmatic changes” to help make the government more effective in “crushing” fraud.

Among other issues, the agency in its RFI is asking for input related to AI being used for Medicare Advantage coding oversight and hospital billing.

That includes requesting stakeholder feedback about “the availability, use, efficacy and cost of using AI, based on machine learning and other methods, to assist with accurately and efficiently abstracting diagnoses from medical record documentation as part of a medical records review.”

More specifically, CMS is seeking comments about the types of AI solutions – including commercial off-the-shelf software – that are most effective and efficient for assisting human coders with large volumes of records; key features and learning capabilities that an AI solution should include to improve accuracy, incorporate coder feedback and prevent hallucinations; how AI-generated coding recommendations should be displayed for human review; compliance risks that should be considered and mitigated; and lessons learned from previous AI implementations.

AI Considerations

CMS and HHS’ Office of Inspector General have been using data analytics and predictive modeling for years to detect potential fraud, but HHS has not publicly discussed whether and the extent to which AI already has been incorporated into those efforts, said attorney Andrew Wirmani of the law firm Reese Marketos LLP, a former U.S. Department of Justice prosecutor (see: Feds Identify $14.6 Billion in Healthcare Fraud Takedown).

“Overall, I see this as a positive development and logical next step to the data analytics that the government is already using to detect potential fraud,” he said. “Healthcare fraud costs taxpayers billions each year so tools that have the potential to enhance the speed and scale of fraud detection can only be seen as a positive,” he said.

“To be effective, however, the government will have to incorporate meaningful human oversight and find ways to avoid false positives, which could unduly burden both government agents and honest healthcare providers. Transparency will also be critical.”

Regulatory attorney Rachel Rose said that if AI is deployed “in a safe, ethical and legal manner” that utilizes accurate data for a limited purpose, then it could speed up the process of detecting and preventing healthcare fraud.

But that said, “It is imperative to appreciate the difference between large language models and generative AI. LLMs have been around for years, so this is a lot safer than a GenAI application,” she added.

Some private sector health insurers, including UnitedHealth Group, also have faced scrutiny – and lawsuits – over their alleged use of AI tools that unfairly deny or reject coverage for necessary medical care (see: Court: UnitedHealth Must Answer for AI Based Claim Denials).

“Because AI output can lead to bias, adverse patient outcomes and upcoding, as well as wrong and incomplete documentation, it poses a risk to patient care,” and other issues, including forming the basis of a False Claims Act case, Rose said.

HHS CMS in its RFI also should have asked for input specifically about the use of AI tools and HIPAA compliance to safeguard the privacy and security of Medicare and Medicaid beneficiaries’ protected health information, Rose suggested.

“Just as the Department of Homeland Security is piloting AI to combat cyberattacks, it is important to use available technology to gain efficiencies but there must be safeguards,” Rose said.

HHS did not immediately respond to Information Security Media Group’s request for additional details about CMS’ healthcare fraud crackdown efforts and its use of AI.

Minnesota Gov. Tim Walz’s office also did not immediately respond to ISMG’s request for comment on HHS’ deferment of Medicaid payments to the state and its allegations of fraud.