The Centers for Medicare & Medicaid Services (CMS) recently issued a memo addressing the use of AI algorithms in determining eligibility for US government healthcare coverage. This action comes in response to legal disputes surrounding Medicare Advantage coverage denials allegedly influenced by flawed AI systems.
The memo outlines that algorithms solely relying on broad datasets, without considering individual patient history, physician recommendations, or clinical notes, would violate Medicare rules. These rules, effective from April 2024, aim to ensure fair and personalized healthcare decision-making.
UnitedHealthcare and Humana are currently facing lawsuits alleging the use of flawed AI systems to deny care to patients. Specifically, the nH Predict AI model has faced criticism for its reliance on generic healthcare data, resulting in shortened hospital stays and high rates of inaccuracies.
The lawsuits highlight concerns regarding the denial of inpatient care, an issue directly addressed in the CMS memo. It emphasizes that AI algorithms cannot independently dictate inpatient admissions or alter coverage criteria without public disclosure and compliance.
Furthermore, CMS raises awareness about the potential for discrimination and bias in healthcare algorithms. It stresses compliance with the Affordable Care Act, which prohibits discriminatory healthcare practices.
In essence, the CMS memo underscores the importance of transparent and patient-centered healthcare decision-making, particularly concerning AI technology.