Mandatory Medicare Bundled Payment and the Future of Hospital Reimbursement

Am J Manag Care. 2024;30(Spec. No. 13):SP1050-SP1058

On August 1, 2024, CMS finalized a new mandatory bundled payment initiative, the Transforming Episode Accountability Model (TEAM), which will begin on January 1, 2026, and continue for 5 years.¹ The model will be implemented in 188 metropolitan areas encompassing 25% of Medicare beneficiaries. CMS has identified 741 hospitals that will have to take on financial risk for 5 surgical episodes, including spending for the initial hospital stay and most care provided in the 30 days after discharge.

The TEAM episodes are lower extremity joint replacement (LEJR), coronary artery bypass graft (CABG), spinal fusion, surgical hip/femur fracture treatment (SHFFT), and major bowel procedures. Together, these 5 Medicare episodes accounted for nearly 900,000 inpatient and outpatient surgeries at US hospitals in 2022 and Medicare hospital payments of approximately $18 billion.² The new model puts US hospitals on notice that it’s time to get serious about surgical quality, effective care transitions, and support for vulnerable patients after they leave the hospital. If successful, 30-day episodes could eventually become a new standard for Medicare hospital payment.

TEAM is the latest in a series of models introduced by the Center for Medicare and Medicaid Innovation (CMMI) in support of its stated goal of moving all Medicare beneficiaries into accountable care relationships by 2030.³ CMMI’s prior voluntary bundled payment models have generally not saved money for the government partly because of selective participation.⁴ In 2021, after making major changes to the design of its Bundled Payments for Care Improvement Advanced model, evaluation results showed net Medicare savings.⁵ However, the changes resulted in financial losses for many participants, and two-thirds of them left the model in the subsequent year. The one consistent area of bundled payment savings across programs has been in joint replacement, including in Comprehensive Care for Joint Replacement (CJR), CMMI’s only prior mandatory bundle model.^6,7

TEAM will affect more than twice as many hospitals and approximately 4 times as much Medicare revenue as CJR. It incorporates a regional pricing model, whereas past models all accounted, at least in part, for each hospital’s historical spending so that higher-cost hospitals had higher target prices. Like CJR, hospitals are designated as the model’s risk-bearing entity. Physicians can participate as formal collaborators with TEAM hospitals and are eligible for gain-sharing payments based on criteria that must incorporate quality of care.

There is a logic to focusing on surgery for this phase of mandatory bundles. Episodes for surgical procedures have well-defined triggers and are more clinically homogeneous than acute medical episodes, resulting in more predictable costs. Whereas most prior CMS bundle programs have used 90-day episodes, the TEAM will use 30-day episodes to further reduce spending variation within episode categories.

Because of the surgical focus, a substantial majority of spending for TEAM episodes is concentrated in the initial hospitalization or outpatient procedure. Postacute care makes up only 14% of total episode costs for CABG, 18% for spinal fusion, 21% for major bowel procedures, and 27% for major joint replacement (Figure 1). In contrast, approximately 54% of the costs of SHFFT occur in the postacute period.

The relatively low rates of postacute utilization combined with the 30-day episode window limit hospital savings opportunities. Managing postacute spending—particularly skilled nursing, inpatient rehabilitation, and readmissions—has been a central aspect of provider strategies to manage spending in bundled payment programs. TEAM participants will also need to find savings through improving efficiency inside the hospitals. This will require collaborating with surgeons to standardize purchasing of supplies, limit unnecessary consultations, and avoid complications, which will ultimately improve quality and reduce length of stay.⁸ Hospitals may also invest in improving surgical outcomes through greater use of presurgical care such as physical therapy and better coordination with primary care to help high-risk patients improve diabetic control, lose weight, and stop smoking prior to surgery. Strong partnerships with clinicians and patients will be a key to success in TEAM.

TEAM Features

The 5 TEAM episodes include 29 inpatient diagnostic-related groups (DRGs) and 8 outpatient procedures. The 2025 Inpatient Prospective Payment System Final Rule added 10 new spinal fusion DRGs, and dropped 5 existing DRGs but the implications for TEAM finances are presently unknown. The outpatient procedures are included in LEJR and spinal fusion episodes, whereas the procedures in the other 3 episodes are exclusively done in inpatient settings. CMS will set regional target prices for each DRG episode to account for case-mix differences across hospitals. Outpatient procedure episodes are integrated into a subset of DRG episodes, creating incentives for some hospitals to conduct more surgeries in the outpatient department.

Ambulatory surgery centers (ASCs) are not included in TEAM. Although it is conceivable that hospitals could shift cases into ASCs to reduce their exposure to the model, we believe this is unlikely because Medicare payments for hip and knee replacements in ASCs are $3300 to $3600 less than the rates paid in hospital outpatient departments.⁹ Furthermore, the patients shifted into ASCs would tend to be healthier than those remaining in hospital settings, which would increase the average cost of TEAM cases and either reduce the model’s gains or increase its losses.

Hospital target prices are based on a 3-year historical average episode spending for all hospitals in each of 9 US Census divisions. Regional averages are calculated using standardized Medicare prices, which eliminates payment differences due to regional labor costs as well as payment adjustments for disproportionate share hospitals and indirect medical education. Episodes costs are capped at the 99th percentile for each DRG category in the model. Individual hospital target prices are calculated by multiplying the regional average episode price by each hospital’s risk adjustment factor, called the patient case-mix adjustment, and applying a discount factor of between 1.5% and 2.0% to arrive at the preliminary target price. Target prices are updated to the performance year using a prospective trend factor, but with the possibility of a retrospective adjustment should actual spending trends differ from the prospective trend. Importantly, the semiannual settlements that determine whether hospitals receive bonuses or must pay penalties in TEAM are conducted in standardized dollars. This has the effect of reducing both gains and losses in markets with higher Medicare prices such as New York City and San Francisco, California.

The TEAM risk adjustment model includes independent variables for beneficiary age groups (<65, 65-74, 75-84, and ≥ 85 years), number of Hierarchical Condition Categories (HCCs) identified in the 90 days prior to the episode, patient social risk, hospital size category, safety net status, and the presence of 25 individual HCCs considered clinically relevant to specific episodes. The social risk factors include dual eligibility for Medicare and Medicaid, residence in a community with a high Area Deprivation Index ranking, and eligibility for the Medicare Part D low-income subsidy. CMMI will not use the social risk measures in the risk adjustment model if they have negative coefficients, so social risk adjustment should only raise target prices for hospitals that care for underserved populations.

An important aspect of TEAM is that it will not affect shared savings for Medicare accountable care organizations (ACOs). CMMI did not want to exclude ACO beneficiaries from the model, but ACOs were concerned about the financial implications of model overlap, which negatively affected some of them during the initial CMS Bundled Payments for Care Improvement model.¹⁰ The TEAM initiative’s solution is to keep the models financially separate so that pricing for TEAM participants will not affect ACO savings. Although this means that CMMI will potentially pay some duplicative bonuses (eg, when reductions in episode spending result in a bonus for the TEAM hospital while also contributing to an ACO’s shared savings), we believe this is preferable to creating a complex reconciliation process that would add financial uncertainty for ACOs.

TEAM gains or losses will be adjusted for quality based on 5 measures from the Hospital Inpatient Quality Reporting program: all-cause readmissions, fall rates with injury, postoperative respiratory failure, 30-day risk standardized mortality among surgical inpatients with complications, and patient-reported outcome for hip and knee surgery to generate a Knee Injury and Osteoarthritis Outcome Score (KOOS).¹¹ Each hospital’s individual scores will be ranked across all US hospitals and combined into an overall Composite Quality Score (CQS). TEAM performance payments will be increased by up to 10% and losses reduced by up to 15% based on each hospital’s CQS. CMS has signaled that it will expand the measure set over time, but the current structure reflects a limited view of quality with only modest incentives for improvement. However, the use of the KOOS is noteworthy and may signal a shift to greater use of patient-reported functional outcomes.

As with other recent CMMI models, TEAM has features intended to promote health equity. All participants must submit an annual health equity plan that includes goals, intervention strategies, and performance measures. CMMI’s approach to market selection was intended to ensure that safety-net hospitals are well represented in the model, potentially creating more access to value-based care for historically underserved groups. The model defines safety-net hospitals as those whose proportion of Medicare patients who are dually eligible for Medicaid or eligible for Medicare’s Part D low-income subsidy is above the 75th percentile nationally. We estimate that this results in approximately 40% of TEAM hospitals being designated as safety-net providers.

Finally, TEAM has 3 risk tracks, each with different limits on hospitals’ total gains or losses. Track 1 limits aggregate gains to 10% and aggregate losses to zero. All hospitals can elect Track 1 for the first model year in 2026, and safety-net hospitals can stay in Track 1 until 2029. Track 2 limits aggregate gains or losses to 5%. Rural hospitals can remain in Track 2 for all 5 model years. Track 3 limits total gains and losses to 20%. Safety-net hospitals are required to select Track 2 or Track 3 beginning in 2029. Urban hospitals not designated as safety net enter Track 3 in 2027.

Analysis of TEAM’s Financial Impact on Hospitals

The most potentially disruptive aspect of TEAM is that target prices are based on average regional spending. Although this reduces administrative complexity, it means that high-cost hospitals will be under intense pressure to improve efficiency to avoid losses. To illustrate the potential impact of regional pricing, we analyzed the financial impact of TEAM for all US hospitals based on the model specifications published in the final rule. We used 100% of Medicare Part A and B claims from calendar years 2021-2023 to calculate target prices for hospitals required to participate in TEAM. We then compared the target prices against actual 2023 spending per episode for all qualifying hospitals in the regions selected for the model. The analysis illustrates the initial impact of TEAM for model year 2, when most participating hospitals must take downside risk, and prior to any new hospital initiatives to manage spending.

CMS identified 741 hospitals that will be required to participate in TEAM, although many do relatively little surgery. Our analysis is based on the 648 hospitals that had at least 11 TEAM episodes in 2023. Table 1 profiles these 648 hospitals based on annual TEAM volume and urban, rural, and safety-net status. It shows that the 20% of hospitals with annual episode volume of 500 or more account for 55% of the $6.0 billion in total TEAM episode spending, whereas the 50% of hospitals with fewer than 200 cases account for 15% of TEAM spending. Similarly, it shows that half of TEAM hospitals are urban non–safety-net hospitals that account for 68% of TEAM spending.

The average financial impact for all eligible US hospitals after applying the model’s stop-loss and stop-gain thresholds will be losses of approximately $560 per episode. This is partly due to the 1.5% to 2% discounts CMS applies to episode target prices. The 62% of hospitals with estimated losses would lose an average of $1350 per episode, whereas the 38% of hospitals with estimated gains would earn approximately $900 per episode. The average impact varies substantially across metropolitan areas, as shown in Table 2. For example, we estimate hospitals in Minneapolis–St Paul, Minnesota, would earn approximately $915 per case annually whereas those in Denver, Colorado, would lose approximately $1350 based on current spending levels. TEAM essentially creates financial transfers from more expensive hospitals in urban centers to lower-cost institutions in smaller communities. However, there is also substantial variation in the model’s impact within markets. Figure 2 shows the initial financial impact across all TEAM hospitals with at least 300 cases in 2023.

The estimated gains and losses shown in Table 2 and Figure 2 incorporate the TEAM’s caps on hospital gains and losses to illustrate the financial impact of all hospitals facing downside risk. However, in model year 2, safety-net hospitals have the option to remain in the model’s Track 1 with no downside risk. We estimate that 164 safety-net hospitals would incur losses of approximately $36 million or $1200 per case if they were required to enter Track 2. If all these safety-net hospitals elected to remain in Track 1 and remain temporarily held harmless for losses, it would reduce aggregate model year 2 losses by approximately 35%.

Policy Considerations

As policy makers assess TEAM, it is important to consider how this model will advance Medicare’s value-based care strategy both immediately and in the long run. We believe the model has important benefits and some limitations. The most important benefit is that TEAM will bring more hospitals—an industry still primarily focused on volume—into a value-based payment model with strong incentives for efficiency. The proposed bundles are high-dollar services for many hospitals. On average, hospitals selected for TEAM will be at risk for their performance on episodes that represent approximately 15% of their total Medicare revenue. To be successful, hospitals will have to ensure their surgical teams get it right the first time, invest in care transition support, and develop effective postacute care management strategies.

Although hospitals may immediately focus on the potential negative impacts of regional pricing, there is one major advantage to the approach. In most prior bundle models, hospitals have had to compete against their historical performance so that success in lowering spending resulted in lower future target prices. In contrast, TEAM hospitals only need to control spending below their region’s average trend to improve their results.

One concern, however, is that by selecting bundles in which most of the spending occurs during the hospital stay, participants have fewer opportunities for gains and therefore weaker incentives to invest in coordinating care across settings. The model does, however, require that hospitals refer patients to a primary care provider prior to discharge. Equally important is that although TEAM creates incentives to improve the efficiency of episodic care, it does nothing to discourage unnecessary or inappropriate surgeries—arguably a larger problem—which is a reason policy makers have shown interest in nesting bundles within total cost of care payment models. Several employer-based episode models have substantially reduced unnecessary surgeries, although these results have occurred at a limited number of center-of-excellence hospitals.¹²

Finally, we believe that leaving acute medical episodes out of the model misses an important opportunity to focus hospitals on managing postacute care. Older patients are extremely vulnerable after being hospitalized, as evidenced by persistently high Medicare readmission rates.¹³ Acute medical episodes have much more spending variation than surgical episodes, with more financial risk but also more opportunity for hospitals to manage spending. CMS could reduce financial risk without incurring additional federal costs through sharing acute medical episode gains and losses with hospitals, reducing stop-loss levels, or both.

In conclusion, we think TEAM is taking a step in the right direction by bringing more hospitals into value-based payment in a model with strong incentives to reduce spending variation.Given the wide variations in per-episode spending across hospitals, the TEAM’s regional pricing approach will be disruptive for some facilities, but the model limits the risk for rural and safety-net providers. If TEAM is successful, a longer-term movement to expand the number of hospital services paid as 30-day episodes could be considered. This would strengthen hospital accountability for patient outcomes that extends beyond hospital walls. It would also be a powerful way to align hospital incentives with ACOs and enhance collaboration between primary care and specialist clinicians. Ultimately, achieving this alignment is essential to the future success of value-based care.

Author Affiliations: Brandeis University (REM, JP), Waltham, MA; Institute for Accountable Care (REM, JP, DK), Washington, DC.

Source of Funding: Arnold Ventures.

Author Disclosures: The authors are employed by the Institute for Accountable Care, which offers data analysis services to hospitals and health systems participating in the CMS Transforming Episode Accountability Model. Mr Koppel also reports attendance of the National Association of ACOs Conference.

Authorship Information: Concept and design (REM, JP, DK); acquisition of data (JP, DK); analysis and interpretation of data (REM, JP, DK); drafting of the manuscript (REM, JP); critical revision of the manuscript for important intellectual content (REM, JP); statistical analysis (DK); and obtaining funding (REM).

Send Correspondence to: Robert E. Mechanic, MBA, Brandeis University, 415 South St, MS 035, Waltham, MA 02453. Email: mechanic@brandeis.edu.

REFERENCES

1. CMS, HHS. Medicare and Medicaid programs and the Children’s Health Insurance Program; hospital inpatient prospective payment systems for acute care hospitals and the long-term care hospital prospective payment system and policy changes and fiscal year 2025 rates; quality programs requirements; and other policy changes. Fed Regist. 2024;89(167):68986-70046.
2. CMS 2022 Medicare inpatient hospitals by geography
   and service and Medicare outpatient hospitals by
   geography and service datasets. CMS. Accessed August 15, 2024. https://data.cms.gov/
3. Fowler E, Fogler S, Schreiber C, et al. The CMS Innovation Center’s strategy to support person-centered, value-based specialty care: 2024 update. Health Affairs Forefront. April 2, 2024. doi:10.1377/forefront.20240328.868596
4. Synthesis of Evaluation Results Across 21 Medicare Models, 2012-2020. CMS. Accessed July 14, 2024. https://www.cms.gov/
   priorities/innovation/data-and-reports/2022/wp-eval-synthesis-21models
5. Henke R, Bradley A, Bacon K, et al. CMS Bundled Payments for Care Improvement Advanced Model: Fifth Annual Evaluation Report. CMS. May 2024. Accessed July 14, 2024. https://www.cms.gov/priorities/innovation/data-and-reports/2024/bpci-adv-ar5
6. Agarwal R, Liao JM, Gupta A, Navathe AS. The impact of bundled payment on health care spending, utilization, and quality: a systematic review. Health Aff (Millwood). 2020;39(1):50-57. doi:10.1377/hlthaff.2019.00784
7. The Lewin Group. CMS Comprehensive Care for Joint Replacement Model: Performance Year 5 Evaluation Report—Executive Summary. CMS. April 2023. Accessed November 14, 2024. https://www.cms.gov/priorities/innovation/data-and-reports/2023/cjr-py5-ar-exec-sum
8. Scally CP, Thumma JR, Birkmeyer JD, Dimick JB. Impact of surgical quality improvement on payments in Medicare patients. Ann Surg. 2015;262(2):249-252. doi:10.1097/SLA.0000000000001069
9. Arthroplasty, knee, condyle and plateau; medial or lateral compartment. Medicare.gov. Accessed October 10, 2024. https://www.medicare.gov/procedure-price-lookup/cost/27446/
10. Mechanic RE. When new Medicare payment systems collide. N Engl J Med. 2016;374(18):1706-1709. doi:10.1056/NEJMp1601464
11. Collins NJ, Prinsen CAC, Christensen R, Bartels EM, Terwee CB, Roos EM. Knee Injury and Osteoarthritis Outcome Score (KOOS): systematic review and meta-analysis of measurement properties. Osteoarthritis Cartilage. 2016;24(8):1317-1329. doi:10.1016/j.joca.2016.03.010
12. Woods L, Slotkin JR, Coleman MR. How employers are fixing health care. Harvard Business Review. March 13, 2019. Accessed November 14, 2024. https://hbr.org/2019/03/how-employers-are-fixing-health-care
13. Krumholz HK. Post-hospital syndrome—an acquired, transient condition of generalized risk. N Engl J Med. 2013;368(2):100-102. doi:10.1056/NEJMp1212324