About

Andrew Michael Ryan is the Provost's Professor of Health Services, Policy, and Practice at Brown University and serves as the Director of the Center for Advancing Health Policy through Research. His research primarily focuses on understanding and evaluating the effects of health care payment reform. He has performed extensive work evaluating national programs aimed at improving the quality and value of health care by altering incentives faced by hospitals, physician practices, Accountable Care Organizations, and Medicare Advantage plans. Much of his work involves detailed econometric analysis of longitudinal data. Dr. Ryan has also contributed to policy development through participation on national technical expert panels and steering committees, gaining a deep understanding of the policy issues related to the use of financial incentives to improve health care spending value in the United States.

Research topics

• Medicine
• Finance
• Business
• Actuarial science
• Surgery
• Economic growth
• Economics
• Computer Science
• Internal medicine
• General surgery
• Family medicine
• Demographic economics
• Medical emergency
• Public economics
• Emergency medicine
• Geography
• Statistics

Selected publications

• Substantial Variation In Administrative Spending And Profit Across State Insurance Markets, 2023

  Health Affairs · 2026-03-01

  articleSenior author

  The US spends hundreds of billions of dollars annually on health insurance administration, but variation in administrative costs and profits remains inadequately understood. We used regulatory filings to estimate state-level, all-plan administrative spending and profit in 2023 for each of three market segments: fully insured commercial coverage, Medicaid, and self-funded commercial coverage. Across state markets for fully insured coverage, median all-plan administrative spending was $599 per person per year, with median profits of $63 per person per year. Administrative costs for self-funded coverage were substantially lower (median, $285 per person per year), as were profits ($22 per person per year). Administrative spending was highest in Medicaid. In states with Medicaid managed care, administration and plan profit accounted for $772 per person per year. Total spending was about $250 per person per year lower in states without Medicaid managed care. Per person per year spending was highly variable across states for all segments. Conservatively, reducing administrative spending and profits in states with above-median levels to the cross-state median could save approximately $6 billion, $9 billion, and $4 billion in the fully insured, Medicaid, and self-funded segments, respectively.

  PublisherDOI

• Timely Antibiotics and Fluid Resuscitation Are Associated With Increased Discharge to Home After Sepsis

  CHEST Journal · 2026-03-01

  articleOpen accessSenior author

  BACKGROUND: Sepsis is a devastating condition with frequent discharge to nonhome settings such as skilled nursing facilities. Bundled payment incentive programs targeting sepsis have tried to encourage lower spending by avoiding discharge to institutional postacute care. RESEARCH QUESTION: What is the impact of timely antibiotic delivery and fluid resuscitation on discharge to home after sepsis? STUDY DESIGN AND METHODS: This was an observational cohort study of adults hospitalized for confirmed community-onset sepsis at 67 hospitals participating in the Michigan Hospital Medicine Safety Consortium Sepsis Initiative (HMS-Sepsis) from 2022 through 2025. Timely antibiotic delivery and fluid resuscitation were assessed via performance measures used for statewide benchmarking. Antibiotic delivery was measured in patients without positive viral testing results. Target administration was ≤ 3 hours of emergency department arrival among patients with hypotension, or else ≤ 5 hours. Fluid resuscitation (≥ 30 mL/kg body weight) was measured in patients with hypotension or elevated lactate. The primary outcome was discharge to home. RESULTS: Among 38,568 patients with community-onset sepsis (18,941 male patients [49.1%]; median age, 71 years [interquartile range, 61-80 years]), 7,942 patients (20.6%) died in hospital or were discharged to hospice, 9,941 patients (25.8%) were discharged to a postacute care facility, and 20,685 patients (53.6%) were discharged to home. Among 35,025 and 27,393 eligible patients, timely antibiotic delivery and fluid resuscitation occurred in 26,357 patients (75.3%) and 13,561 patients (49.5%), respectively. In multivariable models adjusted for patient characteristics, timely antibiotic administration and fluid resuscitation were associated with a 3.0-absolute percentage point (95% CI, 2.0-4.0 absolute percentage point) and 1.1-absolute percentage point (95% CI, 0.2-2.1) increase in discharge to home, respectively. Findings were robust across sensitivity and subgroup analyses. INTERPRETATION: Our results show that in this multihospital cohort, timely antibiotic delivery and fluid resuscitation were associated with increased discharge to home after sepsis. This finding suggests that timely treatment of sepsis may reduce downstream morbidity and health care expenditures.

  PublisherDOI

• Financial Effects of an Out-of-pocket Cap in Traditional Medicare: a Microsimulation Study

  Open MIND · 2026-02-13

  articleOpen access1st authorCorresponding
  OA PDFDOI

• Rapid Disenrollment Rates Tripled For Medicare Advantage Beneficiaries, 2017–22

  Health Affairs · 2026-03-01

  article

  Rapid disenrollment, when a beneficiary disenrolls from a new Medicare Advantage (MA) plan within the first three months of the calendar year, could signal enrollees' immediate dissatisfaction with their plan. The proportion of MA enrollees who rapidly disenrolled tripled from 3.5 percent in 2017 to 12.2 percent in 2022. We report on trends in rapid disenrollment to guide future actions.

  PublisherDOI

• Bundled Payments For Care Improvement Advanced: Effects On Hospital And CMS Spending, 2018–21

  Health Affairs · 2026-02-01

  article1st authorCorresponding

  The Bundled Payments for Care Improvement Advanced Model (BPCI-A) aims to reduce hospital spending and generate savings for the Centers for Medicare and Medicaid Services (CMS). The design of BPCI-A evolved across four model years; however, previous analyses have not examined its impact on hospital and CMS spending throughout this period. We conducted a synthetic difference-in-differences analysis with a 100 percent sample of Medicare fee-for-service beneficiary data from the period April 2014 through December 2021 to evaluate spending changes across 883 participating hospitals and 1,772 nonparticipating hospitals. BPCI-A led to an average $324 reduction in hospitals' ninety-day episode spending, with larger reductions in model years 3 and 4. The largest spending reductions were for orthopedics and neurological care. Yet large incentive payments to hospitals led to net CMS losses of $171 million during the study period, despite net savings in model year 4. BPCI-A reduced payments to skilled nursing facilities during the study period. Thus, BPCI-A had minimal impact on the CMS budget while shifting payments to hospitals and away from skilled nursing facilities. Our results suggest that voluntary bundled payment is unlikely to generate meaningful savings for CMS.

  PublisherDOI

• Exposure to the new Medicare Advantage risk adjustment model varies across insurers

  Health Affairs Scholar · 2026-04-15

  articleOpen access

  Introduction: Medicare Advantage plan payment depends on the health of enrolled patients. As a result, the extent to which beneficiary clinical severity is documented administratively-known as coding intensity-is greater in Medicare Advantage (MA) than in traditional Medicare, which inflates payment to plans. In 2024, the Centers for Medicare and Medicaid Services began phasing in a new risk adjustment model intended to reduce the susceptibility of MA payments to higher coding intensity. Methods: Using 2021 data, we compared average MA contract risk scores under the new model and the old model. Results: Risk scores were 5.8% lower under the new risk adjustment model. Differences between average risks scores under the new and old model varied substantially across contracts and insurers. For example, 1 large insurer's risk scores were essentially unchanged across models while another large insurer's risk score was 18% lower under the new model. Contracts with higher estimated coding intensity had greater exposure to the new risk adjustment model. Conclusion: Our results suggest that the new risk adjustment model will likely reduce MA payments due to enhanced coding intensity, with these reductions appropriately targeting insurers that code more intensely.

  PublisherOA PDFDOI

• Changes In Medical Debt And Bankruptcy After Acute Traumatic Injuries, 2019–21

  Health Affairs · 2026-02-01

  articleOpen access

  Despite expanded insurance coverage after the Affordable Care Act, medical debt remains a significant burden for millions of Americans, particularly after acute medical events such as traumatic injuries. We evaluated the financial impact of hospitalization for acute traumatic injury, using data from a statewide trauma registry linked to consumer credit reports from the period 2019-21. Using a stacked difference-in-differences event study design, we compared financial outcomes for 12,823 injured patients versus 25,195 not-yet-injured matched controls. At eighteen months post-injury, the proportion of patients with medical debt in collections increased by 5.2 percentage points (a 24 percent relative increase compared with the pre-injury baseline), and the mean medical debt in collections (including patients with no debt) rose by $290 (a 76 percent relative increase). Post-injury changes in bankruptcy filings peaked at 3.2 per 1,000 patients (a 6 percent relative increase) at fifteen months post-injury. Financial hardship disproportionately affected uninsured, younger, lower-income, and privately insured patients, whereas those with Medicare and Medicaid experienced minimal change. These findings highlight persistent financial vulnerabilities, even among privately insured patients, and they underscore the need for policy enhancements that strengthen protections against the financial consequences of unanticipated acute medical events.

  PublisherOA PDFDOI

• Double Bonuses Increased MA Spending In Puerto Rico By $865 Million But Did Not Achieve Plan Improvement Goals

  Health Affairs · 2026-04-01

  articleOpen access

  In 2024, more than 90 percent of Medicare beneficiaries in Puerto Rico were enrolled in Medicare Advantage (MA) plans. MA plans receive capitated payments as well as quality-based bonuses, with MA plans operating in so-called double-bonus counties earning twice the usual bonus payments. Puerto Rico was excluded from the double-bonus payment program until 2018, when the double-bonus policy was extended to the territory. Applying a difference-in-differences approach to Centers for Medicare and Medicaid (CMS) data from the period 2012-22, we found that implementation of MA double bonuses in Puerto Rico was not associated with improvements in plan quality or changes in premiums or cost sharing, although it was associated with an increase in the number of plans offered. The findings imply that the additional payments from double bonuses primarily benefited MA plans rather than enrollees. We estimated that the policy resulted in at least $865 million in excess Medicare spending during its first five years. Together with prior evidence questioning the effectiveness of the double-bonus program, these results underscore the need for CMS to reconsider or eliminate MA double bonuses.

  PublisherDOI

• Accountable Care Organization Savings—Hard to Measure, Hard to Find

  JAMA Health Forum · 2026-02-20

  articleOpen accessSenior author
  PublisherDOI

• Trends in Plan Offerings, Enrollment, and Premiums in Medicare Advantage and Medigap

  Health Services Research · 2025-03-15 · 1 citations

  articleOpen access1st authorCorresponding

  OBJECTIVE: Examine trends in Medicare Advantage (MA) and Medigap plan offerings, enrollment, and premiums across state regulatory regimes. STUDY SETTING AND DESIGN: We used national data between 2014 and 2021 on MA and Medigap plan offerings, enrollment, and premiums. Data on Medigap plan offerings and premiums were acquired from Weiss Ratings and matched with county-level data on the Medicare population from 2014 to 2021 Medicare Regional Variation and MA Landscape files. States were classified into three groups based on Medigap regulations: community rating and guaranteed issue states (Connecticut and New York); community rating-only states (Arkansas, Maine, Vermont, and Washington); and no additional Medigap regulation states (remaining states). DATA COLLECTION/EXTRACTION METHODS: We considered only MA plans offering prescription drug coverage. Premiums for Traditional Medicare beneficiaries included Medigap and prescription drug premiums and were calculated using an inflation-adjusted Paasche price index to account for variation across plan types and market segments. PRINCIPAL FINDINGS: Between 2014 and 2021, Medigap offerings and enrollment were relatively constant, while MA enrollment increased substantially. Medigap offerings were lower and MA offerings were higher in states with community rating and guaranteed issue. Between 2014 and 2021, Medigap premiums increased modestly from $4462 to $4745 in states with no additional Medigap regulations and from $6099 to $6612 in states with community rating and guaranteed issue. MA premiums (increased slightly from $2055 to $2121) in states with no additional Medigap regulations and were similar for other states. CONCLUSIONS: Despite modest changes in recent years, Medigap premiums were substantially higher than those of MA. Medigap offerings and enrollment are lower, and premiums are higher, in states with guaranteed issue and community ratings. Nuanced reforms are needed to reduce supplemental insurance costs in Traditional Medicare while preventing adverse selection in Medigap markets.

  PublisherOA PDFDOI

Recent grants

• Evaluating the Effects of HACRP

  NIH · $2.3M · 2018–2024

• Estimating Hospital Outcome Quality: Applications to Pay-for-Performance

  NIH · $742k · 2010–2015

• The Effects of Bundled Payment on Acute Cardiovascular Outcomes Among Older Adults

  NIH · $4.4M · 2015–2027

Frequent coauthors

• Justin B. Dimick

  University of Michigan–Ann Arbor

  84 shared

• Brahmajee K. Nallamothu

  42 shared

• Ushapoorna Nuliyalu

  VA Ann Arbor Healthcare System

  31 shared

• John M. Hollingsworth

  NorthShore University HealthSystem

  30 shared

• Jyothi R. Thumma

  University of Michigan–Ann Arbor

  29 shared

• David J. Meyers

  Brown University

  28 shared

• Karan R. Chhabra

  Beaumont Hospital, Royal Oak

  27 shared

• Adam A. Markovitz

  University of Michigan–Ann Arbor

  20 shared

Education

• Ph.D.

  University of Michigan