[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Views 6,790
Citations 0
November 15, 2016

Expanding Payment Reform in MedicareThe Cardiology Episode-Based Payment Model

Author Affiliations
  • 1Department of Medicine, Massachusetts General Hospital, Boston
  • 2Harvard Medical School, Boston, Massachusetts
JAMA. 2016;316(19):1973-1974. doi:10.1001/jama.2016.16146

Cardiovascular disease is the leading cause of death and disability in the United States, accounting for 1 in 4 deaths and more than $200 billion in direct and indirect costs annually.1 In July 2016, the Centers for Medicare & Medicaid Services (CMS) proposed a new episode-based payment model for myocardial infarction. This model aims to improve the quality and efficiency of cardiac care and helps move the United States toward CMS’s goal of having 50% of traditional Medicare payments in value-based models by 2018.

Proposed to launch July 1, 2017, the cardiac payment model would provide hospitals with an episode-based payment covering the index hospitalization for acute myocardial infarction (AMI) or coronary artery bypass graft (CABG) surgery and care paid under Medicare Parts A or B for 90 days after discharge.2 Hospitals will bear financial risk or receive rewards based on episode-level spending and quality—similar to “2-sided” population-based global payment contracts that reward accountable care organizations if spending is below budget and penalize them if spending exceeds the budget. Specifically, in the cardiac model, if spending is less than the episode-based amount, hospitals may keep a portion or all of the difference, provided they meet benchmarks for quality measures. If spending exceeds the episode-based payment, hospitals would reimburse Medicare the difference.

The model features several notable similarities to its predecessors, the Oncology Care Model (OCM) and Comprehensive Care for Joint Replacement (CJR) model (eTable in the Supplement). Similar to the CJR, hospitals in 98 randomly selected metropolitan areas will participate as the risk-bearing entity. Incentives for the 90-day episodes will be phased in over 5 years. To protect beneficiaries from incentives for underutilization of care, hospitals may keep no more than 5% of the episode-based payment in the first 2 years and no more than 10%, 20%, and 20%, respectively, in the last 3 years. Analogous caps on financial risk—including an initial period without risk—protect hospitals. Hospitals’ episode-based payments during years 1 and 2 will be based largely on their prior spending but in the final 2 years will depend exclusively on regional spending, as in the CJR. Similar to the OCM and CJR, participation will satisfy the Alternative Payment Model pathway of the Medicare Access and Chip Reauthorization Act (MACRA), exempting physicians from individual-level incentives in MACRA’s Merit-Based Incentive Payment System pathway. All 3 models maintain beneficiary choice to seek care from any physician.

For acute cardiac care, episode-based payments may be well positioned to improve the value of spending. From 2012 to 2015, about 17% of Medicare beneficiaries hospitalized for AMI were readmitted within 30 days.3 Thus, the new model may encourage hospitals to improve care transitions and postdischarge monitoring. Moreover, variations in postacute spending are a major driver of regional differences in Medicare spending,4 and episode-based payments may reduce the use of postacute services that are of low value or unnecessary.

On the other hand, certain features of acute cardiovascular disease may test the viability of this payment model. First, about 60% to 70% of AMIs are classified as non–ST elevation MIs (NSTEMIs).1 The benefits of percutaneous coronary intervention (PCI) have been demonstrated for high-risk patients with NSTEMI (for example, patients who have refractory angina, hemodynamic instability, or new heart failure) but are less certain for lower-risk patients or for those with type 2 AMIs. Because the risk of future adverse events dictates the optimal treatment course, and risk is best assessed using clinical data (eg, medical history, physical examination, laboratory, and electrocardiogram findings) that are often not captured in administrative data, CMS may encounter challenges in determining whether expensive treatments are being used appropriately or avoided for a substantial proportion of patients with AMI. In an episode-based payment model, these areas of uncertainty related to clinical risk assessment and treatment may magnify incentives for patient selection. Indeed, under the proposed rules, patients with AMI who present to an emergency department in one hospital but are transferred to another hospital for admission would be assigned to the second hospital. Selection effects have been linked to public reporting of AMI, which was associated with reductions in PCI use among high-risk patients.5

Second, the relatively short 90-day period covered by the episode-based payment following the index hospitalization may create incentives to avoid expensive care that improves long-term outcomes. Randomized trials of “routine invasive” vs “selectively invasive” management of NSTEMI—which form the basis for the American Heart Association’s Class I recommendations for PCI—show that routine invasive management produces few short-term clinical benefits (and may even cause harm) but substantially reduces rates of repeat AMI, refractory angina, rehospitalization, and possibly death during the period of 6 to 12 months following PCI.6 Thus, by emphasizing short-term outcomes, the 90-day period may dissuade the use of treatments that could improve long-term outcomes in some patients.

Third, in contrast to chemotherapy or joint replacement surgery—which are typically planned in advance—virtually all AMIs are unanticipated. This unpredictability could compel hospitals that do not have the infrastructure or resources in reserve to manage emergent cardiac cases at times of full capacity to bear financial risk without warning. Consequently, smaller hospitals may become even more pressured to transfer patients to higher-volume centers, which could delay definitive treatment. Moreover, the hospital treating the initial AMI, which owns the financial risk of the episode, also bears the financial risk of subsequent care that it cannot easily control (eg, emergency department visits triggered by acute chest pain or shortness of breath managed by other hospitals).

In addition, outcomes after AMI depend on cardiac rehabilitation, but its adoption is highly variable.7 While the proposed rules included incentives for cardiac rehabilitation, regional variation in access to cardiac rehabilitation suggests that certain hospitals are better positioned to refer patients to cardiac rehabilitation and reap its associated benefits than others. It remains to be seen whether the proposed incentives will be large enough to induce investments in cardiac rehabilitation.

To encourage buy-in, improve accountability, and protect hospitals and beneficiaries, CMS could potentially strengthen this payment model in several ways: it could publicly report risk-adjusted 1-year mortality for all participating hospitals, include long-term quality-of-care measures for AMI, and test bundles of different durations to assess differential effects on outcomes.8 Moreover, CMS could monitor rates of PCI and CABG surgery for different types of AMI to evaluate for potential underutilization and track patterns of interhospital transfers. In addition, CMS could ask hospitals that transfer AMI patients to share accountability for the 90-day episode with the receiving hospital.

CMS estimates that the model (and associated incentives for cardiac rehabilitation and expansion of CJR) will save Medicare $170 million over 5 years.2 Notably, virtually all of the savings ($169 million) will accrue during the program’s last 3 years, when caps on financial risk are increased to 10% or 20% and episode payments are based on regional spending. This shift from hospital- to regional-based benchmarks is important. Because hospitals initially have a spending target that reflects their own practice patterns, early efforts to reduce readmissions or other improvements will likely generate savings but may not sufficiently entice underperforming hospitals to change practices enough to resemble higher-performing ones. Although regional benchmarks create stronger incentives to reduce such variation, they could also weaken the link between a hospital’s absolute improvements and its performance under the model. For example, higher-cost hospitals with PCI or CABG surgery capabilities may improve efficiencies but remain above their regional benchmarks (which are derived from average spending among higher- and lower-cost profile hospitals in the region). Conversely, hospitals without higher-cost capabilities may spend more but remain below their benchmark. Stratifying benchmarks by hospital capabilities could help address this concern.

More than 200 000 Medicare beneficiaries are hospitalized for AMI and CABG surgery annually.1 Despite challenges, the cardiac payment model will be a meaningful step toward improving the quality and efficiency of cardiac care. Its performance might also serve as a bellwether for the applicability of episode-based payments to other clinical scenarios.

Back to top
Article Information

Corresponding Author: Zirui Song, MD, PhD, Department of Medicine, Massachusetts General Hospital, 55 Fruit St, Bigelow 730, Boston, MA 02114 (zirui_song@post.harvard.edu).

Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Funding/Support: Dr Song acknowledges support from a National Institute on Aging Predoctoral MD/PhD National Research Service Award (F30 AG039175).

Role of the Funder/Sponsor: The National Institute on Aging had no role in the preparation, review, or approval of the manuscript or decision to submit the manuscript for publication.

Mozzafarian  D, Benjamin  EJ, Go  AS,  et al; American Heart Association Statistics Committee and Stroke Statistics Subcommittee.  Heart disease and stroke statistics—2016 update: a report from the American Heart Association. Circulation. 2016;133(4):e38-e360.
Centers for Medicare & Medicaid Services. Medicare Program; Advancing Care Coordination Through Episode Payment Models (EPMs); Cardiac Rehabilitation Incentive Payment Model; and Changes to the Comprehensive Care for Joint Replacement Model (CJR). https://innovation.cms.gov/Files/x/advancing-care-coordination-nprm.pdf. Accessed September 1, 2016.
Centers for Medicare & Medicaid Services. Readmissions and deaths—national. July 29, 2016. https://catalog.data.gov/dataset/readmissions-and-deaths-national. Accessed September 1, 2016.
Newhouse  JP, Garber  AM.  Geographic variation in health care spending in the United States: insights from an Institute of Medicine report. JAMA. 2013;310(12):1227-1228.
Joynt  KE, Blumenthal  DM, Orav  EJ, Resnic  FS, Jha  AK.  Association of public reporting for percutaneous coronary intervention with utilization and outcomes among Medicare beneficiaries with acute myocardial infarction. JAMA. 2012;308(14):1460-1468.
Fanning  JP, Nyong  J, Scott  IA, Aroney  CN, Walters  DL.  Routine invasive strategies vs selective invasive strategies for unstable angina and non-ST elevation myocardial infarction in the stent era. Cochrane Database Syst Rev. 2016;(5):CD004815.
Aragam  KG, Dai  D, Neely  ML,  et al.  Gaps in referral to cardiac rehabilitation of patients undergoing percutaneous coronary intervention in the United States. J Am Coll Cardiol. 2015;65(19):2079-2088.
Jha  AK.  Will episode payment models show how to better pay for hospital care?  JAMA Forum. August 4, 2016. https://newsatjama.jama.com/2016/08/04/jama-forum-will-episode-payment-models-show-how-to-better-pay-for-hospital-care/. Accessed October 14, 2016.