Insights^

On September 5, 2025, the Medicare Payment Advisory Commission (MedPAC) convened for the second day of its September meeting. Staff presented findings on the relationship between shifts in Medicare Advantage (MA) enrollment and hospital finances. Commissioners responded positively to the analysis and offered suggestions to strengthen and expand the research.

Association Between Changes in MA Enrollment and Hospital Finances

MedPAC staff examined how rising MA enrollment is affecting hospitals’ finances. Between 2014 and 2025, MA enrollment grew from 31% to 55% of Medicare beneficiaries, driven by beneficiary preferences for added benefits and employer retiree coverage shifts. Hospitals have expressed concern that MA patients often generate lower payment-to-cost ratios than fee-for-service (FFS) patients, and MA plans actively use tools like prior authorization and narrow networks to manage utilization, which can reduce hospital volumes or shift care to lower-paid settings. Additionally, MA plans negotiate rates, downgrade admissions, and deny claims in ways that can reduce hospital revenue.

Using 2013–2023 cost report data, MedPAC found that higher MA penetration is not significantly associated with changes in hospital profit margins, but it is linked to declines in both revenues and costs (about 1.3% and 1.2% for every 10-percentage-point increase in MA penetration). Effects differ by ownership: financially integrated hospitals did not experience significant revenue or cost declines, while non-integrated hospitals did. Critical Access Hospitals saw no statistically significant effects, partly due to cost-based reimbursement and MA per diem structures. Another emerging issue is uncompensated care (UC) payments—because MA plans often mirror FFS add-ons, a decline in FFS discharges raises the UC payment per discharge, potentially increasing hospitals’ UC payments as MA grows.

Overall, the findings suggest that MA growth shifts financial dynamics but does not broadly erode hospital profit margins, with effects moderated by whether hospitals are integrated with MA plans. Policymakers may need to consider integration differences and downstream effects on FFS-related payments when evaluating MA’s hospital impact.

Commissioner Discussion

The Commissioners’ discussion focused on the financial and operational impacts of MA growth on hospitals, with particular attention to integration, utilization management, and methodological considerations. Several Commissioners raised questions about how new services and prior authorization denials affect hospital operations, emphasizing that, while hospitals often maintain profit margins, they must divert resources to manage administrative burdens. Concerns were also raised about retiree health plan enrollees in MA, the role of broker incentives, and whether beneficiaries and taxpayers are truly getting value for money under the current system.

Others highlighted the importance of examining differences between financially integrated and nonintegrated hospitals, noting the leverage and pricing differentials created by integration. Commissioners suggested expanding the analysis to include dollar amounts, stratifications based on MA market characteristics, and thresholds in penetration levels. They also encouraged looking at other sectors, such as post-acute care and skilled nursing facilities, where effects may differ. There was recognition that hospitals owning MA plans tend to be larger and structurally different, and that regional market dynamics and timing of MA penetration may produce varying effects. Overall, while the analysis was well received, Commissioners urged refinement with more current data, stratified analyses, and a clearer picture of how MA growth translates into real-world financial pressures and adaptations by providers.

On September 3, 2025, the House Energy and Commerce Health Subcommittee held a hearing on advancing health care through artificial intelligence (AI). Discussion included the potential applications of AI in prior authorization, rural health care delivery, and the pharmacy sector. Members of both parties agreed that AI holds promise for improving efficiency and access in the health care system. However, they emphasized that human oversight remains essential to address errors and ensure patient safety.

Opening Statements

• Health Subcommittee Chairman Morgan Griffith (R-VA-09)
• Ranking Member Frank Pallone (D-NJ-6)

Witness Testimony

• TJ Parker, Leader Investor, General Medicine – Testimony
• Andrew Toy, Chief Executive Officer, Clover Health – Testimony
• Dr. Andrew Ibrahim, MD, MSc, Chief Clinical Officer, Viz.ai – Testimony
• Dr. Michelle Mello, JD, PhD, MPhil, Professor of Law, Stanford Law School, and Professor of Health Policy, Stanford University School of Medicine – Testimony
• Dr. C. Vaile Wright, PhD, Senior Director, Health Care Innovation, American Psychological Association – Testimony

Member Discussion

Prior Authorization

During the hearing, Ranking Member Frank Pallone (D-NJ) underscored his concern that AI could embed and even magnify existing biases in prior authorization systems if Congress does not establish sufficient guardrails. He recalled that, under the Trump administration there were efforts to incorporate AI into prior authorization processes and warned that, if left unchecked, such initiatives could lead to higher denial rates for Medicare and Medicare Advantage beneficiaries. Rep. Pallone asked how policymakers could ensure AI is deployed in a way that enhances patient care rather than creating new barriers to access. Dr. Mello responded that prior authorization is already a system with high denial rates and significant flaws. She explained that layering AI onto a broken process risks simply “amping up” existing problems rather than fixing them. According to Dr. Mello, the central question is whether AI will serve as a corrective tool that streamlines care or whether it will exacerbate inequities and inefficiencies—something policymakers and regulators cannot yet answer with certainty.

Rep. John Joyce (R-PA-13) built on these concerns, noting that he has already heard from physicians in his congressional district who report that AI-based systems are increasing denial rates, particularly within Medicare Advantage. Rep. Joyce argued that AI should only be used as a supportive tool to assist clinical decision-making, and never as the final determinant of coverage. He called for Congress to establish clear guardrails that preserve physicians’ ability to make patient-centered judgments and protect beneficiaries from automated denials. Dr. Toy sought to reassure the subcommittee, testifying that his organization does not use AI in prior authorization decisions and stated unequivocally that AI should never be deployed to deny care. Instead, Toy emphasized that AI’s proper role is to help clinicians deliver services more efficiently, reduce administrative burdens, and ultimately improve patient outcomes.

Finally, Rep. Lizzie Fletcher (D-TX-7) questioned whether the current statutory framework is adequate to govern prior authorization in an era where AI tools are increasingly being integrated into health care. Dr. Mello was direct in her response: the existing legal framework is not sufficient. She argued that without updated oversight and regulation, patients and providers will remain vulnerable to harm, underscoring the need for congressional action to modernize policies around AI and prior authorization.

Rural Health Care

The hearing also explored how AI could support care at rural hospitals and the communities they serve. Chairman Morgan Griffith (R-VA-09) opened the discussion by asking what type of software is needed to bring AI into rural settings and whether such systems are prohibitively expensive. Dr. Toy explained that, while the cost of infrastructure is indeed higher in rural areas, smaller towns can sometimes deploy new systems more quickly, enabling AI solutions to reach patients faster than in large, complex urban health systems.

Rep. John Joyce (R-PA-13) pressed further, questioning whether rural hospitals have a stable environment to invest in new technologies. Dr. Mello acknowledged that most do not, noting that rural facilities often operate with razor-thin margins and require outside assistance to adopt advanced tools. She pointed out that certain technologies, such as radiation tools, have shown promise when paired with AI but emphasized that these opportunities are largely out of reach without federal support. Dr. Ibrahim added that some of the most impactful research on stroke care has been conducted in rural areas, underscoring the potential for innovation outside urban centers. However, he cautioned that reimbursement remains a decisive factor: hospitals are unlikely to invest in AI if payment models do not recognize and support its use.

Rep. Troy Balderson (R-OH-12) asked what steps Congress could take to support wider adoption of AI in rural health care. Witnesses highlighted the need for stronger reimbursement policies, targeted grant funding, and infrastructure investments to give rural hospitals a stable base for innovation. Rep. Kat Cammack (R-FL-03) shifted the focus to physician training, asking how AI could play a role in preparing doctors for practice in rural communities. Dr. Ibrahim responded that rural hospitals are already positioned to serve as training hubs and could leverage AI to expand education and mentorship opportunities for providers who might otherwise face isolation. Dr. Toy stressed the importance of connectivity. He argued that something as simple as linking patients and providers through consumer-friendly devices, like iPads, could make a significant difference in overcoming geographic barriers. In his view, AI is not just about cutting-edge algorithms but also about creating practical tools that connect people and care in underserved areas.

Pharmacy and Drug Development

Rep. Diana Harshbarger (R-TN-01) asked how AI could reshape the pharmacy sector. Mr. Parker answered that AI offers significant opportunities for pharmacists, particularly by applying advanced logic to streamline workflows and build stronger infrastructure for dispensing, monitoring, and counseling. When Rep. Harshbarger pressed further on rural challenges, Dr. Toy observed that limited coordination between pharmacists and physicians remains a barrier in these settings. He argued that AI should be leveraged to strengthen collaboration across the care team to ensure patients in rural areas receive consistent, high-quality services.

Rep. Troy Balderson (R-OH-20) noted that a lack of coordinated care often prevents providers from catching early warning signs of health complications. Dr. Toy responded that a stronger AI-enabled health care ecosystem could close these gaps, with pharmacies playing a central role in connecting patients to the broader system.

Rep. Buddy Carter (R-GA-01) broadened the discussion to the scale of the U.S. pharmacy system, asking how AI could improve services across such a vast network. Mr. Parker explained that AI has the potential to collect and synthesize full historical patient records, giving pharmacists a more comprehensive view of patient needs. Rep. Carter then asked about drug development, and Dr. Ibrahim emphasized that AI excels at detecting data patterns, which can accelerate the identification of promising compounds and support more efficient clinical trials.