Flash Points

  • OIG’s recently issued guidance puts MA organizations on notice.  Automated tools don’t eliminate compliance responsibility under Medicare requirements.
  • OIG is concerned about AI-enabled EMR prompts that encourage unsupported diagnoses.
  • AI tools are not exempt from traditional oversight. Governance and auditing frameworks must extend to all processes affecting payment and documentation.

The OIG’s newly issued Medicare Advantage Compliance Program Guidance is not an AI regulation, but it includes explicit references to AI and algorithmic tools that offer important lessons for healthcare compliance programs broadly.

Several points in the guidance are especially instructive for compliance professionals across healthcare sectors:

  • AI and algorithms in utilization management:
    The guidance cautions against relying on algorithms, including those based on artificial intelligence, to drive coverage and medical necessity determinations without appropriate individualized, patient-specific review and oversight. OIG emphasizes that MA organizations remain responsible for ensuring decisions comply with Medicare requirements and are not effectively made by automated tools alone.
  • Risk adjustment and documentation tools:
    In its discussion of risk adjustment, OIG raises concerns about the use of electronic medical record tools and prompts—including algorithmic and AI-enabled features—that encourage diagnosis capture when diagnoses are unsupported or do not affect patient care. OIG links these practices to fraud and abuse risk.
  • Algorithms for monitoring and analysis:
    The guidance acknowledges the use of algorithms and analytics to identify outliers, anomalies, and coding patterns in claims and utilization. These tools are framed as appropriate for monitoring and auditing purposes, while accountability for decisions and outcomes remains with the organization.

Implications for healthcare compliance programs beyond Medicare Advantage:
While this guidance is MA-specific, OIG’s treatment of AI is instructive. AI and algorithmic tools are not discussed as experimental or exempt; instead, they are addressed within existing compliance expectations when they influence decisions, documentation, or payment.

For compliance professionals, this reinforces the importance of applying established compliance frameworks—governance, monitoring, auditing, and oversight—to AI-enabled processes, rather than treating AI as outside traditional compliance structures.

Also for you …