Download as PDF

August 17, 2023 

The Honorable Chiquita Brooks-LaSure
Centers for Medicare and Medicaid Services
U.S. Department of Health and Human Services
Attention: CMS-5540-NC
Submitted electronically to 

RE: Request for Information, Episode-based Payment Model 

Dear Administrator Brooks-LaSure: 

The National Association of ACOs (NAACOS) appreciates the opportunity to submit comments in response to the request for information on the design of a future episode-based payment model. NAACOS represents more than 400 accountable care organizations (ACOs) in Medicare, Medicaid, and commercial insurance working on behalf of health systems and physician provider organizations across the nation to improve quality of care for patients and reduce health care costs. NAACOS members serve over 8 million beneficiaries in Medicare value-based payment models, including the Medicare Shared Savings Program (MSSP) and the ACO Realizing Equity, Access, and Community Health (REACH) Model, among other alternative payment models (APMs). Additionally, our members have participated in bundled payment models tested by the Innovation Center. Our comments below reflect the concerns of our members and our shared goal to having all Medicare patients and most Medicaid patients in an accountable care relationship responsible for total cost of care and quality by 2030. 

To achieve the 2030 goal, there must be a focus on allowing providers to coordinate care across the continuum of care, working together to achieve optimal patient outcomes. While CMS notes that one avenue for specialists to participate in value-based care initiatives is through episode-based payment models, we must design payment models that bolster coordination and partnership across providers. After more than ten years of payment model design innovation, we have learned that concurrent episode models and total cost of care models results in a complex set of overlap rules, leading to provider and patient confusion and increased burden. Accordingly, episode payment models should be designed to support and align with total cost of care arrangements, creating the proper incentives to encourage coordinated care across the care continuum. CMS should adopt the following principles when considering the design of episode-based payment models.

  1. Bolster total cost of care arrangements, such as ACOs
  2. Develop strategies to engage more specialists and other providers in total cost of care arrangements
  3. Design episodic payment models to compliment total cost of care arrangements
  4. Incent the adoption of total cost of care arrangements 

Bolster total cost of care arrangements, such as ACOs 

As noted above, total cost of care arrangements, such as ACOs, are accountable for the full continuum of care. With the primary care team as the foundation for coordinating ongoing patient care, the ACO is able to support patients with referrals to specialists in the community and transitions between hospitalizations, procedures, post-acute care and back to the home. For example, through care coordination efforts ACOs can better coordinate end of life care and identify health inequities to deploy targeted interventions for improvement. To effectively achieve these goals ACOs currently partner with a broad array of specialists. In 2021, specialists represented 65% of all participants in MSSP ACOs. Additionally, typically one-third of the clinicians receiving the advanced APM incentive payment are specialists. Increasing specialist participation in total cost of care models should be prioritized over development of episode-based payment models. We appreciate CMS’s efforts to increase participation of specialists in the ACO. 

We understand CMS’s aim to build on the lessons learned from prior episode-based payment models as an avenue for engaging more specialists. These efforts should be designed in a manner so that the episode-based payment models are an on ramp to total cost of care and ultimately bolster provider engagement in total cost of care. Accordingly, CMS should give precedence to ACOs in model design by excluding beneficiaries from episode payment models unless the ACO opts-in to participate in the episode-payment model as a convener or in partnership with model participants. 

ACOs and bundled payment participants must coordinate care for the patients they serve. However, when multiple entities are trying to coordinate this care, patient confusion can result. This can also mean physicians and health systems duplicating resources devoted to care coordination activities for multiple programs. Ultimately, ACOs have comprehensive information about the beneficiaries they serve and understand the best opportunities to improve and innovate care for their populations. Accordingly, ACOs should be able to select the episode-payment approaches that are best suited for the populations. Moreover, this approach fosters increased engagement between specialists and the ACO. Specialists are incented to partner with ACOs to increase their patient populations for any particular episode while the ACO is incented to partner with specialists to address areas of care that are ripe for improvement and innovation. 

Develop strategies to engage more specialists and other providers in total cost of care arrangements 

As noted above, ACOs are currently partnering with specialists but range in their ability to effectively engage specialists, with some ACOs currently engaging in gainsharing arrangements or sub-contracting such as shadow bundles, while other ACOs may focus on building referral network partnerships. Any episode-based payment design should incorporate approaches that foster collaboration between specialists and ACOs. Specifically, CMS should share data on cost and quality performance for specialists with ACOs, improve access to data about care transitions, leverage episode-payment models to standardize episode definitions and address ACO design elements that stall specialist participation in ACOs. 

Share data on cost and quality performance for specialists with ACOs. 

ACOs are very interested and actively engaged in finding ways to further engage specialists in total cost of care models. Providing data beyond what ACOs receive today through the Claim and Claim Line Feed (CCLF), specifically episode cost data, as well as quality data and patient reported outcomes data, will help support this work. Sharing this information with ACOs will allow for enhanced referral management that is based on quality, cost, and outcome data for some ACOs, while more sophisticated ACOs may be prepared to engage in sub-contracting through financial arrangements such as gainsharing with the ACO. 

Regardless of their approach, ACOs need more data on specialist cost and quality performance to identify variations in care, partner with specialists to implement evidence-based protocols to help reduce variation, inform referrals to high-value specialists, and create financial incentives that encourage coordination across the care continuum. Data that would be helpful to provide ACOs to further support this work include episode cost data and quality data along with regional and national benchmarks. Currently, ACOs have data on specialists in their CCLFs, but this is not of sufficient volume per specialist to create actionable individual provider data. CMS should share specialist data beyond the ACO, first all Medicare cost and utilization data for specialists that serve beneficiaries aligned to the ACO and then data for all specialists in the ACO service area. Additionally, ACOs do not have access to data on the quality of care provided by specialists. ACOs are eager to obtain quality data on specialty performance, whether it be Merit-Based Incentive Payment System (MIPS) quality data or other sources to support specialist engagement. For example, if CMS has collected or as CMS collects other quality data through the hospital value-based purchasing program or specialty care models such as the Oncology Care Model (OCM) or the Kidney Care Choices model (KCC), that data could also be shared with ACOs, starting with those providers who provide services to the ACO's attributed population. 

While CMS has noted the agency plans to provide data specific to the ACO, it would be more helpful to provide specialist performance data across a broader population. At a minimum, CMS should provide specialist performance data across Medicare. CMS should work to include specialist data across other payers, such as Medicare Advantage, to provide ACOs with a more wholistic and accurate picture of performance in the market. Benchmarks will then allow ACOs to understand how a specialist’s data compares to the region and nation. 

While there is broad interest in gaining access to specialty data across the spectrum, should CMS need to focus on certain specialties to start with, the most logical could include cardiology, gastroenterology, nephrology, oncology, orthopedics, neurology, endocrinology, retina specialists, dermatology, physical therapy, and behavioral health. Data should be timely and actionable so it can be used at the point of care. Historically, claims data lag has created significant overlap challenges between ACOs and bundled payments because the episode initiator was unknown until the episode hand concluded. CMS has sped data through the Beneficiary Claims Data API (BCDA) and should incorporate similar approaches across models. CMS should also consider:

  • Intended use of data. Information being shared to gain a better understanding of quality and performance information to support referral patterns is different than a use case of designing payment approaches within a total cost of care arrangement. Further, if the data sharing is to help inform patients, there will be very different needs (such as to share performance information with beneficiaries in an understandable way that allows them to make better decisions about their care).
  • Ensuring sufficient sample size. ACOs engaging specialists in shadow or nested bundles are often faced with challenges regarding small numbers. Performance data must be based on a sufficient volume of cases so that spending estimates are statistically reliable. 

Improve Access to Data about Care Transitions 

Foundational to partnerships between ACOs, hospitals, and specialists who focus on acute and procedural conditions is being able to understand when patients are experiencing a transition of care. CMS requires hospitals to report admissions, discharge, and transfer (ADT) alerts to support care transitions; however, this information is not readily available to ACOs. NAACOS surveyed ACOs and found that 38% of ACOs said they don’t have agreements with hospitals or third-party vendors to receive ADT alerts. Those with access cite several problems with the data, including timeliness, accuracy, usability, and costs. NAACOS has called on CMS to modify the HIPAA Eligibility Transaction System (HETS) to allow access to all eligibility inquiries for ACO-assigned beneficiaries and to develop a proactive, real-time notification system for ACOs when beneficiary eligibility is requested. Doctors’ offices, clinics, hospitals, surgery centers, and other providers ping HETS to check patients’ Medicare eligibility at each encounter. If CMS, which operates HETS, allows access to the system’s data, then Medicare providers would have real-time knowledge of beneficiaries’ visits to medical providers, including hospital admissions, emergency department visits, and specialist encounters. This information is critical to ACOs’ care coordination efforts. 

Leverage episode-payment models to create standard definitions for episode-payments, ultimately supporting nesting bundles within ACOs. 

CMS seeks comment on how to achieve multi-payer alignment through an episode-based payment model. Specialists are currently engaged with ACOs, Medicare Advantage and other payers to implement bundled payments. CMS can support this work by developing industry standard definitions for episodes to be used by ACOs and other payers in the way that best suits their organization and regional market. CMS could also support ACOs who wish to voluntarily participate in an episode-based payment model or nest bundles within their ACO by creating and sharing target prices as well as quality performance data for episodes and appropriate risk adjustment for ACOs to use in designing their own nested bundles or specialist payment approaches. These increased data transparency efforts will be critical in helping ACOs to facilitate better communication among primary care clinicians and specialists. Efforts to engage specialists should allow for options from a menu set of more standardized approaches while still allowing for flexibility. 

Address policy and program design elements that currently are prohibitive to the inclusion of specialists in TCOC arrangements. 

As we note above there is currently strong specialist participation in TCOC models; however, there is opportunity to ensure there are strong incentives for collaboration among primary care clinicians and specialists. First, CMS should allow ACOs to negotiate contracts with specialists that would shift payment for certain services. This supports the concept of nested bundles and other negotiated payment approaches. The ACO and specialist could agree to reduce all or a portion of FFS claims payment and receive episodic payment from the ACO or share in the ACO savings. CMS employed a similar approach in the Next Generation ACO Model under its population-based payment mechanisms which allowed ACOs to enter negotiated payment arrangements with downstream providers. The provider received a reduction in FFS claims in exchange for other compensation determined by the ACO, based on quality and other utilization and outcome metrics. 

Second, CMS should design TCOC program policies that support providing care to high cost, high need, seriously ill populations, and populations without a regular source of primary care. These high need populations are a strong opportunity for ACOs and community specialists to partner to improve care. Currently, certain policies such as risk score caps and ratcheting benchmarks can discourage ACOs from actively seeking inclusion of these beneficiaries. CMS has recognized the need to better address these populations through other initiatives. For example, ACO REACH has a high needs track which provides higher benchmarks for seriously ill populations. Similarly, the Oncology Care Model provided better incentives for caring for oncology patients than are provided to oncology patients in an ACO. Adequately addressing the needs of these populations will create incentives for specialists to join or partner with ACOs. Additionally, CMS should consider non-financial incentives that would encourage specialists to engage in TCOC arrangements, such as a set of ACO waivers that provide more flexibility to specialists in an ACO. 

Finally, CMS should consider attribution approaches that would allow a greater portion of a specialists’ patient panel to align to an ACO. Currently, specialists who join total cost of care models have only a small proportion of their patient panel in the ACO. 

Additional current policies that discourage specialist participation in ACOs that should be addressed are:

  • The MSSP quality requirement to move to electronic clinical quality measures (eCQMs)/MIPS CQMs by 2025 inadvertently penalizes ACOs with specialist participants by requiring reporting and assessment of all-payer and all-patient data rather than focusing on ACO assigned patients. As a result, specialists in the ACO are held accountable for primary care measures that are not clinically appropriate. For example, dermatologists in the ACO would be required to assess and do follow-up on depression screenings, which would not be clinically appropriate. Ultimately this would lead to artificially lowering the ACO’s quality score and assessing ACOs based on the case-mix of their population and the proportion of specialists in the ACO.
  • The high/low revenue distinction in MSSP discourages ACOs from including specialists. ACOs with more participating specialists are likely to have a larger percent of the ACO’s revenue for all expenditures of the assigned beneficiaries. Removing the high/low revenue distinction would remove the disincentive to include specialists in the ACO.
  • Qualifying APM Participant (QP) threshold policies discourage specialist participation in ACOs. 

Design episodic payment models to compliment total cost of care arrangements 

We reiterate that any episode-based payment model should exclude ACO-aligned beneficiaries unless the ACO opts-into participation as a convener or in partnership with specialists or hospitals. This will create the proper incentives for bundles done within a total cost of care model to ensure there is no incentive for overutilization. 

We appreciate that CMS noted that it is focused on shorter episodes triggered from procedures or hospitalizations. We recommend CMS avoid including chronic condition episodes because attribution is less clear. It is difficult to assign accountability, and more testing and work needs to be done in this area. As noted above, CMS should work with ACOs to understand opportunities to shift payment for specialists providing chronic care. For example, providing up front population-based payments to specialists is one approach ACOs may use to engage specialists. 

Procedural or acute episodes have been successful in programs like the Bundled Payments for Care Improvement-Advanced (BPCI-A) and the Comprehensive Care for Joint Replacement (CJR) models because the episode can be accurately attributed to a facility and provider. Any episode should be defined so that it is clinically relevant.

  • 30-day post episode trigger bundles are positive, but the episode length should be defined by when it is clinically appropriate to transition care from the acute/procedural provider back to the community provider.
  • Combined episodes should be limited to clinically relevant clinicians. CMS should avoid the approach employed in BPCI-A that created a medical and critical care episode service line group that combined conditions treated by disparate clinicians and departments within the hospital.
  • Include low volume thresholds to avoid penalizing providers for minor variations in care.
  • Include quality measures that assess and reward transitions back to the community provider.
  • Address health equity through stratification of quality measures. Given the abbreviated nature of the episodes it may be difficult for providers to address social risk factors, CMS should consider how addressing social risk can be incorporated as part of the transition back to community providers, since this is better addressed in an ACO.
  • Incorporate sharing of social risk factor information with episode initiators. We caution requiring collection of social risk data as CMS is requiring collection of this information through other setting- and program-specific requirements, such as ACO models. To avoid de-duplication of data collected, CMS should consider how it can best facilitate sharing information among providers and incorporate social risk factors into payment approaches. 

Incent the Adoption of TCOC Arrangements 

Given CMS’s 2030 goal, all aspects of payment models should be designed so that providers are incented to engage in total cost of care arrangements. While we do not support mandatory bundles, if CMS considered a mandatory episode-payment model it should exempt providers who are meaningfully engaged in total cost of care initiatives. We note that CMS should work with stakeholders to define meaningful engagement in total cost of care. There are limitations to the current approaches (i.e., participation lists and QP designation) as TCOC entities may not include all hospital and specialist engagement on a participant list. CMS could consider other factors such as ownership structure or preferred provider lists to help identify meaningful participation in a TCOC. Finally, CMS should explore approaches for incenting specialty providers in other payment system programs. Nearly all settings of care have a value-based purchasing program but only the clinician Quality Payment Program includes explicit incentives for adopting value. At a minimum, CMS should assess TCOC participation in other VBP programs. 

Finally, CMS should consider the broader payment incentives that encourage specialists to join APMs. For example, setting QP determinations at the individual level rather than the APM entity may have an unintended consequence of excluding specialists as QPs. Additionally, the discontinuation of advanced APM incentive payments, which are set to expire this year, results in stronger financial incentives for remaining in FFS. Moreover, nonfinancial incentives in the form of FFS regulatory relief and waivers in APMs have been limited. CMS should design all APMs so that the financial and nonfinancial incentives encourage APM adoption. 


Thank you for the opportunity to provide feedback on the request for information on the design of a future episode-based payment model. NAACOS and its members are committed to providing the highest quality care for patients while advancing population health goals for the communities they serve. We look forward to our continued engagement on episode-based payment models. If you have any questions, please contact Aisha Pittman, senior vice president, government affairs at [email protected]


Clif Gaus, Sc.D
President and CEO