Compare Medicare Bundled Payment Models Side-by-Side
Compare bundled payment models by various features, including:
- Key Results
- Details on savings and cost
- Financial Arrangements
- Beneficiary Numbers
What are bundled payment models?
Bundled payment models are a way for Medicare to establish a total budget for all services provided to a beneficiary throughout an episode of care. If the episode’s spending on services is below budget, then the providers may share in Medicare savings; alternatively, if providers’ costs exceed the budget, then the providers may incur losses. In some cases, bundled payment models can span across multiple health care settings. Some of Medicare’s bundled payment models are voluntary and others are mandatory for hospitals in selected areas of the country. The Department of Health and Human Services (HHS) recently proposed reducing the number of mandatory bundled payment models.
CMS is running a range of bundled payments models in Medicare that vary by key design features, including levels of financial risk, up-front payments for infrastructure costs, and beneficiary involvement. To learn more about each bundled payment model and compare key features, see our Bundled Payment Side-by-Side comparison tool.
Q1: What are the different types of bundled payment models in Medicare?
Bundled Payments for Care Improvement (BPCI) Initiative Model 1
BPCI Model 1 was a voluntary model in which CMS bundled payments for all inpatient hospital services for an episode of care and applies a discount to the usual Medicare hospital payment. Participants gained financially if total spending for an episode (determined retrospectively) was below the “target price.” This model ended in December 2016.
Bundled Payments for Care Improvement (BPCI) Initiative Model 2
BPCI Model 2 is a voluntary model in which CMS bundles payment for all inpatient hospital services, physician services, post-acute care services, and other related services and any readmissions throughout an episode of care. Participants gain financially if total spending for an episode (determined retrospectively) is below the “target price.
Bundled Payments for Care Improvement (BPCI) Initiative Model 3
BPCI Model 3 is a voluntary model in which CMS bundles payment for all post-acute care services and any readmissions throughout an episode of care (which begins after discharge from an inpatient hospitalization). Participants gain financially if total spending for an episode (determined retrospectively) is below the “target price.”
Bundled Payments for Care Improvement (BPCI) Initiative Model 4
BPCI Model 4 is a voluntary model in which CMS bundles payment for all inpatient and physician services during the initial hospital stay and any readmissions, but not post-acute care services. In this model, CMS makes a prospective payment to the hospital at the start of the episode of care, rather than reconciling payments after the episode of care. Participants gain financially if total spending for an episode is below the amount CMS paid.
Comprehensive Care for Joint Replacement (CJR)
The CJR model is a mandatory model for acute care hospitals in 67 geographic areas. This model effectively bundles payment for lower extremity joint (hip and/or knee) replacement episodes across all inpatient hospital services, physician services, post-acute care services, and any readmissions or other related services through 90 days after the initial hospital discharge. Participants gain financially if actual expenditures for an episode (determined retrospectively) are below the “target price.” A CMS proposed rule, published August 17, 2017, would allow participation to be voluntary in 33 of the 67 originally designated mandatory areas, as well as for small and/or rural hospitals in the remaining mandatory areas.
Oncology Care Model (OCM)
The OCM is a voluntary model in which oncology practices receive monthly care management fees and are eligible for bonus payments if they lower overall Medicare spending and meet quality goals for episodes of chemotherapy and related care. CMS offers multiple risk/reward options to participating practices. Commercial insurers are also participating.
Episode Payment Models (EPM)
The EPMs include three mandatory bundled payment models—the acute myocardial infarction (AMI) model, the coronary artery bypass graft (CABG) model, and the surgical hip and femur fracture treatment (SHFFT) model. The bundled payments for each of these models include all inpatient hospital services, physician services, post-acute care services, and other related services through 90 days after the initial hospital discharge. Participants can gain financially if total spending for an episode is below a quality-adjusted “target price.” Although a CMS proposed rule, published August 17, 2017, would cancel all EPMs, they are currently scheduled to start on January 1, 2018.
Cardiac Rehabilitation (CR) Incentive Payment Model
The CR Incentive Payment Model is a mandatory, bonus-only model for hospitals in 90 geographic areas. In this model, CMS makes added payments to hospitals based on the number of cardiac rehabilitation sessions that applicable Medicare patients receive after discharge. Although a CMS proposed rule, published August 17, 2017, would cancel the CR model, it is scheduled to begin on January 1, 2018.
Q2: What is the evidence on savings for Medicare bundled payment models?
BPCI Model 1, which focused on Medicare hospital admissions in its bundle, achieved modest net savings ($10 million across two years), attributable mostly to discounted Medicare payments to hospitals. Notably, Medicare spending on care following hospitalizations was higher for the test episodes than the comparison group. Most hospital participants ultimately withdrew and the model has ended.
Of the other BPCI models, the most prevalent is BPCI Model 2, which spans across hospitals, physicians, and post-acute providers. In its first two years, this model generated statistically significant Medicare savings per episode for one clinical category—hip/knee replacement (technically, major joint replacement of the lower extremity)—among the 23 categories that had enough episodes for statistical analysis. Medicare savings in the hip/knee replacement category was $1,273 per episode, relative to a comparison group. While the other 22 clinical categories did not have statistically different spending from their comparison groups, among the 10 clinical categories with the highest number or episodes, spending generally trended lower than their comparison groups, per episode.
Further analysis of the hip/knee replacement episodes shows that BPCI participants had lower Medicare spending for facility-based post-acute care, such as skilled nursing facility (SNF) care, but not for home health care, which is typically less expensive.
For the newer model, the CJR model, which was mandatory for 800 hospitals and focused on hip and knee replacements, Medicare spending results are limited. Analysis of CMS data shows that in the first year of the program, about half of the hospitals received “reconciliation payments” indicating that they received added payments from Medicare because their spending was below their target benchmark in 2016. Reconciliation payments totaled $37.6 million, averaging $1,134 per episode. No spending information is available for hospitals that did not receive reconciliation payments.
Notably, episode-based spending results, including those reported for the CJR model and BPCI, do not account for changes in the number of episodes per participating provider. Therefore, any changes in the overall volume of episodes could have additional spending effects on Medicare. For specific results and more details about each model, see the Bundled Payment Side-by-Side comparison tool.
Q3: What is the evidence on quality for Medicare bundled payment models?
Quality results varied by clinical episode and model, but generally did not reveal major differences between the BPCI models and comparison groups. However, in BPCI Model 3, evaluation results showed increased rates of unplanned readmissions, emergency department use, and mortality (though greater patient complexity may have biased mortality results somewhat) for SNF-based BPCI participants relative to comparison groups. In contrast, for other measures, including functional improvement and patient-reported experience, these BPCI participants showed either greater improvement or no statistical differences.
For the CJR model, quality results are limited. CMS data show that in the first year of the program, among the hospitals that received any reconciliation payments for lower episode spending, 39 percent were classified as “excellent”; 53 percent as “good”; 8 percent as “acceptable”; and none as “below acceptable.” In general, these classifications were based on CMS analysis of medical complications and patient-reported assessments for each hospital. No quality information is available for hospitals that did not receive reconciliation payments.
For further details on quality results by model, see the Bundled Payment Side-by-Side comparison tool.
Q4: Are doctors in Medicare bundled Payment models eligible for automatic bonuses under MACRA?
In 2015, Congress enacted the Medicare Access and CHIP Reauthorization Act (MACRA), which changed the Medicare payment system for physician services. In this law, physicians and other health professionals who are affiliated with “advanced alternative payment models” (APMs) are eligible for automatic 5-percent bonuses on their Medicare payments, starting in 2019. Several bundled payment models—namely, the Comprehensive Care for Joint Replacement (CJR) model, the Oncology Care Model (OCM), the acute myocardial infarction (AMI) model, the coronary artery bypass graft (CABG) model, and the surgical hip and femur fracture treatment (SHFFT) model—all qualify as advanced APMs, either in full or in certain “tracks” of each model. However, HHS has recently proposed canceling several of these models (specifically, the AMI, CABG, and SHFFT models), which have not yet started.
Q5: Where are Medicare bundled payment models located?
For 2017, CMS reports that 987 hospitals/practices and another 1,019 “episode initiator” organizations are participating in Medicare bundled payment models in almost all states and the District of Columbia. Bundled payment models are more prevalent in the eastern half of the country, but are also concentrated along the west coast, with others located in Alaska and throughout the Midwest.
Q6: How many Medicare beneficiaries have received care in bundled payment models?
BPCI Model 1 accounted for over 240,000 beneficiary episodes among participating hospitals during its first two years. Models 2-4, for which only first year results are available, accounted for about 61,000 episodes, with the majority (74%) initiated under Model 2. Although beneficiary episodes are not an exact match to the number of beneficiaries receiving care in bundled payment models, it is likely a close estimate. To learn more about the number of beneficiaries in other models and how bundled payment models compare, see the Bundled Payment Side-by-Side comparison tool.
Q7: When did each Medicare bundled payment model start and end?
BPCI started in 2013. Model 1 concluded in December 2016, but the other three BPCI models were extended into 2018. The other ongoing models—the CJR model and the OCM—will be active at least through 2020. Future models are slated to begin in 2018, but CMS has recently proposed canceling the Episode Payment Models (the AMI, CABG, and SHFFT models) and the Cardiac Rehabilitation Incentive Payment model.