Introduction:
Research and discovery are an iterative process. This process of testing and refinement is important for understanding how the Centers for Medicare & Medicaid Services’ (CMS) Center for Medicare and Medicaid Innovation (Innovation Center) implements its statutory mandate to test innovative payment and service delivery models to reduce program expenditures while preserving or enhancing the quality of care. Many Innovation Center models have made significant progress through multiple iterations of a concept rather than a one-time test. Lessons learned from an originator model often inform a subsequent, follow-on model, with new innovative features, charting a path to what ultimately might be scaled nationally.
Another important set of tools for the Innovation Center is the ability to leverage and scale lessons learned and promising results into the Medicare and Medicaid programs. These tools are important to understanding the impact Innovation Center models have on the health care system.
Below, we describe in more detail and with recent examples each of these pathways for the Innovation Center.
Part 1. Iterative Change in Payment and Delivery Reform Models
Models operate in a dynamic environment, and most require active management and modification to become successful. This is particularly the case with models aiming at clinical practice transformation at scale. The Innovation Center often tests novel compensation arrangements for groups of professionals caring for patients.
New incentives and supports for practice transformation in a dynamic environment will not always be immediately successful without further adjustment. The Innovation Center can and does learn and change with each model.
Core model design components often require iterative modification to balance beneficiary access and provider viability with strong accountability, quality improvement, and potential savings. CMS balances the need for learning and model changes to support success with the need for stability for model participants.
Some of the major elements that involve evidence-based iteration include:
- Acuity of target populations.
- Benchmarking approach.
- Accountable quality measures.
- Discount percentages.
- Episode length.
- Risk adjustment.
An iterative approach also gives CMS a greater opportunity to interact with private payers and other private sector health system components. A longer trajectory for models enhances the mutual learning and dissemination of payment and delivery innovation.
Two Examples of Successful Iteration: Acute to Post-Acute Episodic Payment Models and Kidney-Focused Accountable Care Organization (ACO) Models
Specialty care-focused models that use an episode-based (bundled) payment design to finance the transition from acute to post-acute care and ACO models focused on kidney care have shown significant progress, learning, and continuity over multiple model iterations. The experience and success of these models provide a foundation for ongoing transformation in specialty care in Medicare.
Example 1: Acute to Post-Acute Episodes
Acute to post-acute episode-based payment models are a pillar of the Innovation Center’s work in transforming care and have strong potential to reduce avoidable rehospitalizations and inappropriate skilled nursing facility (SNF) placements while supporting the successful recovery of Medicare beneficiaries treated for serious conditions. Specifically, evaluation of these models has shown more appropriate hospital discharge decisions regarding whether and for how long people receive different kinds of post-acute care. The Innovation Center episode bundling approach is built on the chassis of Traditional Medicare’s existing payment structure, including a very broad provider network and access to inpatient and post-acute care. To preserve this access, Innovation Center episode models have relied on a retrospective bundled payment arrangement where expenditures are billed fee-for-service but then are reconciled against a target price for an episode of care. Although CMS began operating small episodic payment pilots in the 1990s,[1] the first Innovation Center episode-based Medicare model was the Bundled Payments for Care Improvement (BPCI) Initiative, which began accepting applications in 2011 and operated until 2018. This large and multi-part model tested four different approaches to bundling payments for a wide variety of clinical episodes. Evaluation of BPCI found that fewer patients who were treated by model participants were discharged to an SNF or inpatient rehabilitation facility (IRF) for post-acute services and length of stay was reduced for these patients. Instead, patients were often discharged to heal at home. However, resulting savings were offset by additional incentive payments to participants in the model.[2]
The Innovation Center then followed up BPCI with two currently operating successor models: the BPCI Advanced Model (BPCI Advanced), launched in 2018, and the Comprehensive Care for Joint Replacement (CJR) Model, launched in 2016. Both models are based on 90-day episodes of care. BPCI Advanced participants can select from a range of medical and surgical conditions and are given spending targets for each that are based on their own historical data, along with other factors. Average spending per episode is reconciled against these targets retrospectively, and participants earn bonuses or incur penalties that are based on this reconciliation. The CJR model is a partially mandatory model focused on hip, knee, and ankle replacements that holds participant hospitals financially accountable for the quality and cost of the episode beginning with the procedure and then for up to 90 days following discharge, including post-acute care and recovery at home. Both models have, up to this point, reduced Medicare FFS payments for the majority of episodes in the models while maintaining quality for beneficiaries.[3] Notably, both have achieved higher rates of discharges from hospital to home rather than to IRF or SNF with no increase in readmissions or reduction in functional recovery. There is also a growing body of evidence that these models have spillover effects that benefit enrollees in Medicare Advantage and commercial coverage.[4]
Now, a third generation, the Transforming Episode Accountability Model (TEAM), is being implemented as a mandatory model in a significant portion of the country. TEAM has much in common with the current BPCI Advanced Model and the just-ended CJR Model and includes many new elements building on lessons from those models. These changes include: a shorter (30-day) post-acute episode length to better align episodic and longitudinal, population-based incentives; a revised range of covered clinical episodes based on the BPCI Advanced and CJR experience with episodes that demonstrated success in reducing episode payments and achieving net Medicare savings; and a revised and streamlined geographic benchmarking framework.
Episodic payment in Medicare has not been a simple or linear success, but it is an increasingly successful approach through iterative evolution and refinement, generating both gross savings and increased discharges to home from the hospital. Across two generations of both mandatory and voluntary successor models, the Innovation Center has refined episode length, selection of diagnoses, and cost and quality benchmarking to zero in on the transition from acute to post-acute care and to drive better cost and discharge outcomes with growing effectiveness.
Example 2: Kidney-Focused Accountable Care Organization Models
The Innovation Center currently operates the second iteration of specialized ACO models for people with kidney disease: the Kidney Care Choices Model (KCC). The first kidney-specific ACO model, Comprehensive End-Stage Renal Disease Care (CEC) Model, began operation in 2015. It showed considerable promise in improving health but challenges in generating net savings for CMS.[5] CEC was CMS’ first effort to apply an ACO approach to a specialized population and specialized providers. Although the model had a clinically specialized focus, it included accountability for all of a beneficiary’s Medicare Part A and Part B payments. The evaluation found a 3% decrease in the number of hospitalizations in the model relative to a comparison group and 5% fewer hospitalizations from ESRD complications, contributing to modest gross savings and better patient survival rates. Patients treated by participants in the CEC model saw an increased number of dialysis sessions, a reflection of participants adding dialysis chairs and offering late shifts for otherwise skipped or missed dialysis appointments. Access to primary care improved significantly, a development plausibly tied to the model’s support of care coordinators in dialysis facilities.[6]However, gross savings were offset by shared savings payments to participants, resulting in narrow net losses.[7]
The Innovation Center initiated the ESRD Treatment Choices (ETC) and KCC Models as two successor models beginning in 2021 and 2022. The ETC Model tests a mandatory payment approach to increase use of kidney transplantation and home dialysis. The KCC model applies a second iteration of the ACO approach to kidney care and expands the scope of the model test to add beneficiaries with late-stage chronic kidney disease and those who have already received transplants. KCC moves upstream from ESRD aiming to delay and improve dialysis starts for those with late-stage chronic kidney disease stage four and stage five, a significant evolution from the first-generation Kidney ACO Model. This population presents a major opportunity for savings and better population health. Another major evolutionary step involves risk contracting: KCC goes beyond its predecessor, CEC, in offering an array of tracks with different levels of upside and downside risk for accountable participants, including the possibility of 100% risk.[8] KCC’s first annual evaluation report shows significant early promise of quality improvement while maintaining Medicare spending.[9]
Part 2. Examples of Scaling Promising Aspects of Models
CMS has multiple tools to apply learnings from models by integrating promising elements of the tests into the Medicare and Medicaid programs. These tools include:
- Model elements incorporated into Medicare through rulemaking in other programs, such as the advance payment option of the ACO Investment Model, which was incorporated into the Medicare Shared Savings Program (SSP).
- Model features adopted through legislation, for example, the inclusion of the expansion of the Repetitive, Scheduled Non-Emergent Ambulance Transport (RSNAT) Model test nationwide through the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA).
- Successor models building on lessons from previous models, as described above in Section 1, that further innovate and advance on earlier bundled payment and kidney models, respectively.
- Broader adoption of a model’s innovative features, such as the screening tool for health-related social needs developed by CMS in the Accountable Health Communities Model that is currently in wide use in the private sector.[10]
These scaling approaches are a large and growing part of the Innovation Center’s strategy to impact health care delivery.
A Different Approach in State-Focused Models
The Innovation Center is taking a slightly different approach to model refinement over time and scaling in state-focused models, emphasizing state flexibility and long pre-implementation and implementation periods rather than utilizing shorter test periods and successor models. CMS is now starting on three new state-federal models: States Advancing All-Payer Health Equity Approaches and Development (AHEAD) Model, Innovation in Behavioral Health (IBH) Model, and Transforming Maternal Health (TMaH) Model, which leverage state flexibility to allow participants iterative change and “management to success” within eight- to 10-year performance periods. The AHEAD Model is influenced by previous individual state models, making it both a successor model and a multi-state scaling strategy.
Conclusion:
The Innovation Center rigorously analyzes success or failure of its models to determine how to effectively advance health care transformation that supports higher quality, lower costs, and better care experience. Achieving structural and behavioral change in health care, which is a huge part of American society and the economy, is complicated, and there are many ways that the Innovation Center can have an impact on the health care system. Success in changing provider organization and financing requires thoughtful iteration and the flexibility to scale parts of Innovation Center models using more than one legal and operational scaling pathway.
###
[2] Center for Medicare and Medicaid Innovation. Bundled Payments for Care Improvement (BPCI) Initiative, Model 2-4. https://innovation.cms.gov/data-and-reports/2020/bpci-models2-4-fg-yr6
[8] https://www.healthaffairs.org/conent/forefront/medicare-cec-model-using-lessons-learned-improve-value-based-kidney-care
[10] See, for example, “A New Screening Tool Identifies Unmet Health-Related Social Needs (HRSN)”, Cancer Support Community, November 7, 2023, at https://www.cancersupportcommunity.org/blog/new-screening-tool-identifies-unmet-health-related-social-needs-hrsn.