Blog

The 9th Annual Caribbean Health Economists’ Symposium (CHES), held on December 12-14 at Scrub Island Resort in the British Virgin Islands, brought together leading health economists to present cutting-edge research on some of healthcare’s most pressing challenges.

The symposium featured five compelling papers that collectively illuminate critical inefficiencies in the U.S. healthcare system—from technological barriers and administrative burdens to market concentration and organizational restructuring. Researchers from Dartmouth, MIT, Penn, and other top institutions presented evidence-based analyses examining how system design choices, market forces, and policy interventions shape healthcare delivery, pricing, and access. Their findings reveal substantial opportunities for improvement: billions in potential savings from better EHR interoperability, strategic refinements to prior authorization policies, reallocation of hospital inputs to high-value care, greater scrutiny of physician group consolidation, and careful consideration of the trade-offs inherent in hospital privatization. The presentations showcase the power of rigorous empirical analysis to inform policy debates and identify pathways toward a more efficient and effective healthcare system.

The event also featured roundtable discussions on the influence of tariffs on the healthcare economy and the policies needed to sustain biomedical innovation. Stephen Parente, HEAL’s Founding Director & former Chief Economist for Health Policy on the Council of Economic Advisers from 2019 to 2021, and Eric Hargan, former United States Secretary of Health and Human Services led these discussions, providing real time insights on policy developments in Washington, DC on topics such as the impacts of drug pricing reforms (Inflation Reduction Act and Most Favored Nation models) on biopharma innovation, the role of intellectual property (IP) in fueling drug innovation, and reform efforts at the Food and Drug Administration (FDA).

Below are summaries of the five research papers presented at CHES in 2025. For more information about The HEAL Network’s conference in 2026 and opportunities to present, please contact info@thehealnetwork.org.

---

Paper 1: Costs of Technological Frictions: Evidence from EHR (Non-) Interoperability

• Rebekah Dix, Massachusetts Institute of Technology
• Kelsey Moran, Massachusetts Institute of Technology
• Thi Mai Anh Nguyen, New York University

Research Question

• How does the lack of interoperability between Electronic Health Record (EHR) systems affect patients who move between hospitals?

Main Findings

• Patient Outcomes Improve with Better Interoperability
  • When two hospitals switch to the same EHR vendor:
    • Transfer patients: 3.7% lower charges, 2.1% fewer images
    • Referral patients: 2.8% fewer tests, 11% lower readmission rates
    • Benefits largest for high-risk patients and lower-quality hospitals

• Patient Flows Shift Toward Compatible Systems
  • 8% more transfers and 9-10% more referrals between same-vendor hospitals
  • ~100,000 Medicare transfers and 15 million referrals reallocated (2005-2019)
  • $7.3 billion in Medicare spending redirected

• Interoperability Varies Dramatically
  • Within-vendor interoperability (Epic-to-Epic) >> across-vendor (Epic-to-Cerner)
  • Epic has the best within-vendor interoperability among major vendors
  • Average scores (0-1 scale): Across-vendor = 0.23, Within-vendor = 0.41

• Large Welfare Gains Possible
  • Perfect interoperability would increase welfare by 21% (equivalent to 57 km shorter travel)
  • Current progress (2013-2019) achieved only 34% of potential gains
  • 95% of gains from direct benefits, 5% from better hospital allocation

Key Takeaways

• Hidden costs: Fragmented EHR systems cause duplicate testing, higher charges, and worse outcomes
• Network effects: Hospitals strongly prefer sending patients where EHR systems are compatible
• Policy solutions: Government-mandated interoperability standards (like TEFCA) could unlock major gains while preserving competition—better than relying on market consolidation alone
• Uneven progress: Benefits concentrated among hospitals switching to Epic, especially in markets where Epic was already prevalent

Bottom line: Poor EHR interoperability significantly harms patients and wastes resources. Improving across-vendor compatibility could benefit everyone without requiring market monopolization.

---

Paper 2: Informative Ordeals in Healthcare: Prior Authorization of Drugs in Medicaid*

• Samantha Burn, Imperial College, London, UK
• Ljubica Ristovska, Yale University

*This is an in-progress working paper subject to ongoing revisions. Please follow the instructions from the authors’ here to obtain the most current draft.

Research Question

• How do prior authorization (PA) requirements—where providers must submit paperwork before insurers cover treatment—affect prescription drug use in Medicaid?

Main Findings

• Overall Impact:
  • PA reduces drug utilization by 58% on average

• Results Vary by Drug Type:
  • Branded drugs (generic available): 70% drop, fully offset by generic substitution ✓
  • Specific formulations (alternatives available): 58% drop, fully offset by substitution ✓
  • Entire drug class (no substitutes): 15% drop, no substitution = patients forgo care ✗

• Poor Targeting:
  • On-label (appropriate) use: ↓ 49%
  • Off-label (questionable) use: ↓ 25%
  • PA doesn’t distinguish between high and low-value care

• Mechanism:
  • Both insurer denials AND provider behavior (ordeal costs) drive reductions

Key Takeaways

• Strategic deployment matters: PA works when cost-effective substitutes exist (brand→generic), but harms patients when restricting drugs without alternatives
• Blunt instrument: Like copays, PA reduces both appropriate and inappropriate care indiscriminately—not well-targeted
• Administrative burden is real: Paperwork requirements significantly deter prescribing, even before insurers review requests

Bottom Line: Prior authorization can reduce Medicaid drug spending without harming outcomes when used strategically to encourage substitution to cheaper, equally effective alternatives. However, broad restrictions without available substitutes lead to care gaps, and PA fails to target low-value use specifically—it just reduces all care.

---

Paper 3: Productivity Variation and Input Misallocation: Evidence from Hospitals

• Amitabh Chandra, Harvard University
• Carrie Colla, Dartmouth College
• Jonathan Skinner, Dartmouth College

Research Question

• Why do hospitals vary so widely in productivity? Specifically, do hospitals misallocate inputs by underusing effective treatments and overusing ineffective ones?

Main Findings

• New Test for Misallocation:
  • Developed general test: if hospitals are efficient, returns to specific inputs should be zero when controlling for total spending
  • Strong rejection of efficiency hypothesis (χ²(3) = 771)

• Impact of Misallocation: Using 1.6M Medicare heart attack patients (2007-2017):
  • Moving from 10th to 90th percentile hospital on misallocation → 3.1 percentage point increase in survival (holding spending constant)
  • Misallocation explains ~25% of variation in hospital productivity

• Patterns by Treatment Category:
  • Category I (effective): Underused treatments like beta-blockers, statins, same-day stenting → positive association with survival when used more
  • Category II (heterogeneous): Mixed results for CT scans, ICU days, number of physicians
  • Category III (low-value): Overused treatments like post-acute care, redundant CT scans → negative or zero association with survival

• Spending ≠ Quality:
  • Weak correlation between expenditures and survival (explains only ~2% of variance)
  • US News “Top 25” hospitals show 5 percentage points higher survival, mostly from better overall productivity (not just less misallocation)

Key Takeaways

• How money is spent matters more than how much: Same spending level produces vastly different outcomes depending on input mix—challenging the focus on total spending levels in healthcare debates
• Misallocation is substantial and measurable: Unlike macro literature testing if returns are equalized across firms, this tests if returns are zero within hospitals, conditional on spending
• Structural barriers drive inefficiency: Information asymmetries, physician beliefs, perverse financial incentives, and poor organizational structure all contribute to systematic underuse/overuse
• Volume helps: Larger hospitals show less misallocation (more incentive to learn optimal practices)

Bottom Line: Hospital productivity differences stem significantly from what they spend money on, not just how much they spend. Misallocation accounts for 25% of productivity variation—hospitals systematically underuse proven treatments (beta-blockers, statins, timely stenting) while overusing low-value care (excessive post-acute care, redundant imaging). This suggests that improving healthcare outcomes requires focusing on input allocation and organizational effectiveness, not just increasing or decreasing overall spending.

---

Paper 4: Provider Concentration and Prices*

• Jean Abraham, University of Minnesota
• Conor Ryan, Pennsylvania State University
• Stephen T. Parente, University of Minnesota

*This paper is not yet published.

Research Question

• How much does physician market power and group size affect commercial healthcare prices, particularly for primary care office visits?

Main Findings

• Massive Price Variation: Using Transparency in Coverage data for 3 national insurers:
  • Median price for moderate complexity office visit: $180
  • Interquartile range: $93 (75th percentile is 67% higher than 25th percentile)
  • 30% of variation occurs within same area, same insurer (physician-level differences)
  • Within-zip code spread (90th-10th percentile): $109

• Geographic Concentration Effects (Small):
  • Standard deviation increase in employment concentration → 0.7-2.3% price increase
  • Geographic market structure matters, but relationship is quantitatively small
  • Conclusion: Local concentration explains little geographic price variation

• Group Size Effects (Large):
  • 90th percentile group size → ~10% higher prices vs. median-sized groups
  • 98th percentile group size → ~20% higher prices vs. median
  • Larger groups gain substantial bargaining power within markets

• Quality Not the Driver:
  • No evidence that physicians in larger groups have higher demand (using Medicare fee-for-service shares as proxy)
  • Some evidence of better quality scores, but modest
  • Interpretation: Price premium from large groups driven by market power, not quality

Key Takeaways

• Within-market variation dominates geographic variation: Provider market power explains ~20% of within-market variation but very little across-market variation
• Group size = pricing power: Largest physician groups command substantial price premiums through collective bargaining, not superior quality or demand
• Same physician, same area, different prices: 30% of price variation is physician-specific within same insurer-geography, suggesting significant contracting inefficiencies
• Antitrust blind spot: Physician groups receive less scrutiny than hospitals despite substantial pricing power, especially for largest groups

Bottom Line: While geographic differences in physician concentration have minimal impact on prices, physician group size within markets has substantial pricing power—the largest groups command 10-20% price premiums over median-sized groups. This market power appears driven by collective bargaining strength rather than quality differences, suggesting current antitrust scrutiny may miss important sources of healthcare price variation by focusing on geographic concentration rather than within-market group consolidation.

---

Paper 5: The Impact of Privatization: Evidence from the Hospital Sector

• Mark Duggan, Stanford University
• Atul Gupta, University of Pennsylvania
• Emilie Jackson, Michigan State University
• Zachary Templeton, University of Pennsylvania

Research Question

• What are the causal effects of privatizing government-owned hospitals on hospital finances, patient access, and quality of care?

Main Findings

• Hospital Finances (Large Gains):
  • Profitability turnaround: From -3% operating margin to modest surplus
  • Revenue increases: 8% increase per bed; Cost reductions: 6% decrease in personnel spending
  • Government savings: $2 million/year per privatization on average

• Patient Access (Significant Reduction for Vulnerable Populations):
  • Medicaid patients: 15.6% decrease in admissions at privatized hospitals
  • Uninsured patients: 37% decrease (state-level data)
  • Medicare/privately insured: Small, insignificant changes
  • Market-level: ~4% net decrease in Medicaid admissions; 11% in concentrated markets
  • Health consequences: Markets with larger Medicaid declines saw higher mortality among 55-64 year-olds

• Quality of Care (Decline):
  • Medicare mortality: 3% increase in 30-day mortality for FFS patients 65+
  • Lives lost: 3.4 additional deaths/year per privatization (elderly Medicare FFS only)
  • Life-years lost: 18.4 per privatization per year (conservative estimate)
  • Persistent: Effect sustained over 5-year follow-up

• Mechanism – how privatized hospitals achieve their financial gains and why patient outcomes worsen:
  • Revenue Enhancement:
    • 54% reduction in obstetric admissions (unprofitable service)
    • 6.5% increase in list prices
    • Payer mix and pricing changes explain ~50% of revenue increase
  • Cost Cutting:
    • 6% decrease in FTE staff; 25% decrease in physicians
    • 1.7% shorter hospital stays
    • No evidence of offsetting quality improvements

Key Takeaways

• Trade-off quantified: $0.6 million in savings per additional death, or $110,000 per life-year lost—below federal cost-effectiveness standards (~$369,000/life-year)
• Vulnerable populations suffer: Medicaid/uninsured patients lose access; elderly Medicare patients face higher mortality
• Market structure amplifies harm: Worst effects in concentrated markets and high-poverty areas
• Active restructuring: Not just selection—hospitals close unprofitable services, raise prices, cut staff

Bottom Line: Hospital privatization eliminates deficits and saves governments money but imposes substantial social costs through reduced access for vulnerable populations and increased mortality for elderly patients. The savings per life-year lost fall below federal healthcare cost-effectiveness thresholds, suggesting privatization may not pass benefit-cost tests despite improved finances.