Imagine going to a restaurant where, instead of being charged separately for your appetizer, main course, and dessert, you pay a fixed price for the entire meal. That’s the idea behind a bundled payment plan in health care. Instead of billing patients and insurers for every individual test, treatment, or doctor’s visit during a medical episode—like a knee replacement—providers receive a single, comprehensive payment for all services related to a specific treatment or condition over a defined period. While this way of paying for our health care was being tested on a volunteer basis for the past two decades, the imminent requirement of bundled payments across hundreds of hospitals for certain surgical procedures in January 2026 leaves the question: Does this have the positive impact that Medicare is hoping for?
This approach theoretically encourages hospitals and doctors to work together efficiently, aiming for better outcomes at a lower overall cost. The offload of financial risk from health insurance companies to providers would incentivize institutions to deliver coordinated, efficient care by holding them accountable for both the cost and quality of care during the episode and aligning financial incentives across providers.
CMS has launched a number of bundled payment programs over the past two decades, but the most notable programs were Bundled Payments for Care Improvement (BPCI) and its successor, Bundled Payments for Care Improvement Advanced (BPCI-A). Hospitals volunteered to receive bundled payments for a broad range of clinical and surgical episodes, all the way from hip replacements to heart failure hospitalizations. If they reduced costs below the benchmark set by CMS, then CMS rewarded them with a financial bonus. But if their costs exceeded the benchmark, then they owed CMS a penalty payment.
Most current bundles are site-neutral 90-day episodes that start with a hospital admission or outpatient procedure and end after post-acute recovery. They include hospitals, physician groups, skilled nursing facilities, home health agencies, etc., aligning everyone under one target price. Bundled models vary by payment timing.
- Prospective models (e.g., BPCI Model 4) cut a single check upfront; which then distributes funds to clinicians and ancillary providers.
- Retrospective models (e.g., CJR) pay fee-for-service during the episode, reconcile total costs against a benchmark months later, and issue bonuses or clawbacks.
Benchmarks typically incorporate a blend of each hospital’s historical spending and regional averages, with case-mix adjustments to mitigate patient selection bias. To deter repeated admissions, reconciliation payments hinge on complication rates, readmissions, and patient-reported outcomes. CJR, for example, withholds up to 5 percent if a hospital’s composite quality score slips. While BPCI-A is voluntary, CJR and the forthcoming TEAM model are mandatory in selected markets—a signal that CMS is willing to move from pilots to population-level pressure.
Bundled payments function well in theory, but in practice, face major challenges. First, they added to the administrative burden that hospitals and providers were already struggling to manage. Second, bundled payments were meant to encourage care coordination across hospitals, providers, and rehab, but they incentivized providers to shift costs to other specialists and rewarded hospitals for cutting payments to rehab facilities without cutting costs at their own site. Third, bundled payment models penalized hospitals serving medically or socially high-risk patients by not fully accounting for patient complexity.
Despite those challenges, early results from bundled payment programs were promising in certain areas. Bundled payments for hip and knee replacements decreased Medicare payments without compromising quality of care. Encouraged by these findings, CMS expanded bundled payments to a wider range of clinical conditions and surgeries. Hospitals cut costs on these hospitalizations and saved CMS some money, but as a result, CMS had to pay them large bonuses. In the end, CMS saved small amounts in some bundled payment programs but paid more than they saved in others. Bundled payment programs also financially penalized hospitals serving high-risk patients, which had the potential to worsen existing inequities. To date, bundled payment programs have not yet realized the potential savings many hoped they would provide.
The implementation has a profound impact on current and future physicians. If risk adjustment lags reality, safety-net hospitals caring for medically complex patients may face penalties that widen resource gaps. The transition toward more precise demographic risk adjustment in newer models is therefore critical to achieving equitable physician compensation. Bundles reward surgeons who co-manage discharge plans with internists and physical therapists, penalizing siloed decision-making. Additionally, tracking a 90-day episode requires integration across EHRs, post-acute partners, and claims feeds—often with limited interoperability. Practices in early bundled models reported spending more than $12,000 per physician annually on extra staffing and IT in order to comply with increasingly complex reconciliation requirements. However, when bundles center on surgical episodes with clearly defined care pathways, physician groups can reduce costs and participate in substantial shared savings—$421 million across BPCI-A’s first three years alone. In 2026, Medicare is implementing required bundled payment for coronary artery bypass grafting, surgical hip and femur fracture treatment, lower extremity joint replacement, spinal fusion, and major bowel procedures.
Bundled payments are no longer a policy experiment on the outskirts of Medicare. They are fast becoming a competence physicians and health care workers need—both to safeguard their patients’ outcomes and to steer the economics that increasingly shape clinical practice. The implementation of bundled payment models on particular surgical procedures has a higher probability of success than in other sectors based on past research, so the results of this change by Medicare will need to be watched–but may not be replicable if expanded across care types.
The AMA Committee on Economics and Quality in Medicine, Medical Student Section contributors include Daniel Johnson, Kendall Jackson, MD, Ashwin Varma, Davis Giffin, Dhiresh Bandaru, Anuradha Haridhas, Aditya Gunturi, Jordan Pemberton, Aman Singh, and Diana Hla, MD.
