Download as PDF

May 30, 2019 

Adam Boehler
Deputy Administrator
Director of the Center for Medicare & Medicaid Innovation
Centers for Medicare & Medicaid Services
U.S. Department of Health & Human Services
Hubert H. Humphrey Building
200 Independence Avenue, S.W.
Washington, D.C. 20201 

RE: Request for Information on Direct Contracting—Geographic Population-Based Payment Model Option               

Dear Deputy Administrator Boehler: 

The National Association of ACOs (NAACOS) appreciates the opportunity to submit comments in response to the Request for Information on the Geographic Population-Based Payment (PBP) Model Option of the Direct Contracting Model, released by the Center for Medicare and Medicaid Innovation (Innovation Center) on April 22, 2019.[1] As the largest association of accountable care organizations (ACOs), representing more than 6 million beneficiary lives through more than 330 Medicare Shared Savings Program (MSSP), Next Generation Model, and commercial ACOs, NAACOS and its members are deeply committed to the transition to value-based care. NAACOS is an ACO member-led and member-owned non-profit organization that works on behalf of ACOs across the nation to improve the quality of Medicare delivery, population health and outcomes, and healthcare efficiency.  

ACOs, the origins of which date back to the George W. Bush Administration, have been instrumental in the shift to value-based care. ACOs focus on providing high-quality healthcare while controlling costs, and many ACOs are embracing value and preparing to assume greater accountability. Importantly, the ACO model also maintains patient choice of clinicians and other providers. 

NAACOS was pleased to see many ACO principles underpinning the Direct Contracting Model. These include empowering local physicians, hospitals, and other providers to work together and take responsibility for improving quality, enhancing patient experience, and reducing waste. In exchange for taking accountability for a group of patients, Direct Contracting Entities (DCEs) can share in savings generated and are offered tools to improve care for patients. The inclusion of these principles and a call for current ACOs to participate is a testament to the success of ACOs. The Direct Contracting Model, like existing Medicare ACO programs, seeks to improve the quality of care for fee-for-service (FFS) Medicare beneficiaries. 

The Innovation Center’s work to offer more options that improve care coordination for this important segment of patients is admirable and welcomed. We support this being a voluntary model and are very pleased it is designed to qualify as an Advanced Alternative Payment Model (APM).

However, the Innovation Center also needs to be mindful not to undermine the efforts of the now well-established MSSP. NAACOS encouragesthe Innovation Center to use the three options under the Direct Contracting Model to engage in piloting new concepts and policies that might be later incorporated into the MSSP under powers granted to the secretary of Health & Human Services (HHS) by Section 1115A of the Affordable Care Act. 

Now in their seventh year, MSSP ACOs assume responsibility for almost 20 percent of all Medicare beneficiaries and a third of those in traditional or FFS Medicare. The growth and performance of Medicare ACOs are positively impacting care delivery, which is the goal of the Centers for Medicare & Medicaid Services (CMS), the Innovation Center, and their value-based care efforts.  

Data show Medicare ACOs are limiting the growth of healthcare spending at a time when spending continues to grow at a pace that’s faster than inflation of the overall economy. An independent analysis released in December showed the MSSP saved Medicare $2.7 billion between 2013 and 2016.[2] CMS estimated that the overall impact of MSSP ACOs, including “spillover effects” on Medicare spending outside of the ACO program, lowered spending by $1.8 – $4.2 billion (0.5 – 1.2 percent) in 2016 alone.[3] 

The first-year evaluation of the Next Generation ACO Model (Next Gen) showed it reduced Medicare spending in 2016 by $100 million and $62 million after accounting for shared savings and losses.[4] Initial analysis of second-year results show Next Gen ACOs netted at least $165 million to Medicare in 2017.  

Meanwhile, ACOs continue to show they deliver high-quality care. In 2017, MSSP ACOs subject to pay-for-performance measures earned an average quality score of 90.5 percent out of 100 percent. [5] In the Next Gen program, quality in its first year was improved in the form of fewer acute care hospital stays and more annual wellness visits.[6] Other Innovation Center ACOs demonstrated high quality, with the Pioneer ACO Model having and average quality score of 93 percent.[7] 

Our below recommendations on the Geographic PBP Model Option of the Direct Contracting Model reflect our desire to see Medicare achieve long-term sustainability, enhance care coordination for millions of beneficiaries, lower the growth rate of healthcare spending, and improve the quality of care. ACOs play a critical role in achieving these goals.

Comments Related to the Geographic PBP Model Option

Questions Related to General Model Design

  1. How might DCEs in the Geographic PBP Model Option address beneficiary needs related to social determinants of health (such as food, housing, and transportation) with particular attention to whether the geographic scale contemplated under the payment model option creates new opportunities for success in terms of community-based initiatives? What barriers might prevent DCEs from addressing these social determinants of health? Are there additional incentives that CMS could offer to DCEs to motivate these entities to address social determinants of health? 

There is increasing evidence that social determinants of health (SDOH) are a greater driver of cost and outcomes than genetics or clinical care. Currently, there is little financial incentive or infrastructure to pay for SDOH work or to support community agencies that address SDOH. Few ACOs have found ways to effectively close gaps, but those that have, have been able to demonstrate significant cost reductions while improving on quality of care. All three options within the Direct Contracting Model should align financial sustainability not only with coordination of social resources but also with the work of community partners that close SDOH gaps. Innovations then could come not only from the healthcare delivery side, such as building community networks, screening for social determinants gaps, and referral systems to address those gaps, but also from financing of community agencies. NAACOS recommends the Innovation Center provide upfront funding to DCEs to test the effectiveness and value of how those funds can be used to close SDOH gaps. 

Questions Related to Selection of Target Regions

  1. What criteria should be considered for selecting the target regions where the Geographic PBP Model Option would be implemented? For example, are there attributes of target regions, such as low penetration of Advanced APMs or higher healthcare costs than the national average, which CMS should consider in selecting target regions for the Geographic PBP Model Option? What impact would this have on competition in target regions where the Geographic PBP Model Option is ultimately implemented?

We appreciate the Innovation Center’s goal of driving cost savings and improved quality through competition. However, selecting target regions with more than one DCE must be accompanied by additional guardrails to preserve choice and competition for traditional Medicare beneficiaries. For example, Geographic DCEs should not be allowed to use their market power to mandate or require providers in a specific area contract with them or require patients to see providers with whom they have a negotiated relationship. Similarly, Geographic DCEs must not be permitted to unfairly drive out competition from ACOs or Global or Professional DCEs by taking short-term losses on provider contracts and beneficiary incentives. A Geographic demonstration should be closely monitored for unintended consequences and shifting competitive dynamics. 

ACOs have spent years building the necessary infrastructure to support properly coordinated care, including hiring care managers, installing appropriate information technology systems, and driving cultural changes among their providers and within their communities. As outlined above, Medicare ACOs have shown signs of substantial cost savings at a time when health spending continues to grow. This translates into billions of dollars in savings to Medicare, bending the cost curve, extending the life of the Medicare Trust Fund, and improving the quality of care seniors receive. These successes are likely to increase exponentially as ACOs gain more experience and if the Medicare ACO program continues to grow. Because of this success, CMS should be careful not to interfere with existing ACOs (or Global and Professional DCEs), their networks, and patient populations. By selecting regions that already have ACOs, the Innovation Center could damage their work. 

We urge the Innovation Center to consider the implications of program overlap in a particular region. Geographic DCEs should not displace or take precedence over existing APM entities, including ACOs. Participants in existing models have made significant investments to shift to value-based care, which should be recognized by new models and model participants coming into a target region. We encourage CMS to continue to directly contract with MSSP ACOs, Next Gen ACOs, and other Innovation Center initiatives. Any disruption in an existing model inevitably distracts from the important work of creating more value for Medicare beneficiaries. 

Given the savings and quality improvement Medicare ACOs have generated, NAACOS encourages the Innovation Center to start the Geographic PBP Option in areas where there is low, if any, ACO penetration. Furthermore, ACO assigned beneficiaries should be excluded from the populations for which Geographic DCEs would be responsible so not to conflict with ACO participation. NAACOS continues to work with CMS on addressing the overlap of various payment models, which should work to complement – not conflict with – each other. Furthermore, approving regions for the Geographic PBP Model Option that do not overlap with ACOs may allow for more reliable and actionable evaluation of the option. 

2. What are the benefits and/or risks to access, quality, or cost associated with the implementation of the Geographic PBP model option in a target region that includes a rural area? What safeguards might CMS consider to preserve access and quality for beneficiaries in rural areas in a Geographic PBP target region? How would rural market forces (for example, out-migration, hospital closures, and mergers/acquisitions) affect the DCE’s ability to lower cost and improve quality under the payment model option? 

CMS’s questions related to rural providers are appreciated since uptake of Medicare value-based payment programs are lower in rural areas. NAACOS encourages the Innovation Center to start the Geographic PBP Model Option in areas where there is low ACO penetration. For that request to be heeded, CMS must address how to increase participation of Medicare value-based payment programs in rural areas. 

Perhaps the biggest challenge to implementing the Geographic PBP option in rural areas is the Innovation Center’s desire to select two entities in a region to foster competition. In rural areas, there may not be interest from more than one entity to participate, nor may there be a sufficient number of beneficiaries to support two DCEs. CMS would need to consider additional incentives to attract multiple potential applicants for a single region or consider an exception to its demand for two entities in rural areas. 

The Innovation Center should also be mindful of the differences in hospital systems that are reimbursed under both a prospective payment system and cost-based reimbursement. Under the cost-based reimbursement system that governs critical access hospitals, rural providers could face a payment system that puts them at a disadvantage over counterparts being paid under a prospective payment system. 

Questions Related to DCE Eligibility

  • What are the benefits and/or disadvantages of the DCE selection criteria under consideration for the Geographic PBP model option, described above? What other selection criteria and core competencies should CMS consider requiring applicants to address? Please describe the benefits of including such additional selection criteria. What criteria are of the greatest importance and therefore should receive the greatest weight in our selection decisions?

To this end, CMS should give particular weight to the following criteria:

  • Whether the applicant has a historical presence in the target region as demonstrated by (1) overlap of its health insurance coverage service area with the target region (in the case where the applicant is a health plan); (2) overlap of its direct or affiliated provider network with the target region; and/or (3) contractual relationships with organizations described in (1) and (2).
  • Whether the applicant has the capacity to provide strategic and operational direction and technical assistance to healthcare providers to support delivery transformation.
  • The strength of the applicant’s strategy for leveraging current CMS models or programs and existing and/or planned delivery system transformation efforts within the target region to support success.
  • Relevant experience of the applicant, taking into account formal partnerships or other contractual relationships with healthcare providers or other organizations that have experience in risk-sharing arrangements. 
  • What types of entities might participate in the Geographic PBP Model Option that have not participated in CMS Innovation Center Models or other Advanced APMs offered by CMS, such as the MSSP, to date? What conflicts of interest issues might arise and how should CMS and/or the DCE address them?

Given the amount of risk that a Geographic DCE will take on, the most likely participants are large payers and very large integrated health systems. Smaller entities, such as those that provide significant savings to the Medicare Trust Fund through their participation in ACO programs, would not be able to participate. Not only can they not take on risk of that size, they are unlikely to be able to provide a financial guarantee of the size required by the Innovation Center. 

While this option may bring new players into the Direct Contracting Model, the Innovation Center must ensure that these entities do not participate to the detriment of current participants that have demonstrated success in ACO initiatives. As stated above, Geographic DCEs should not be allowed to use their market power to mandate or require providers to contract with them, or to require patients to see providers with whom they have a negotiated relationship. Similarly, Geographic DCEs must not be permitted to unfairly drive out competition of ACOs or Global or Professional DCEs by taking short-term losses on provider contracts and beneficiary incentives. A Geographic demonstration should be closely monitored for unintended consequences and shifting competitive dynamics. 

Questions Related to Beneficiary Alignment

1. CMS currently plans to select target regions with at least two DCEs to encourage competition. In the event that there are two or more DCEs in a given target region, we are considering either randomly aligning beneficiaries in the target region to one of the DCEs or allowing beneficiaries in the target region to voluntarily align themselves to a specific DCE. One potential benefit of a random alignment approach is that it could help to reduce reliance on risk adjustment, which is intended to account for differences in health risk in a given population. Where risk is taken on a large population basis, such as in the Geographic PBP Model Option, we would expect risk to be evenly distributed, making risk adjustment less necessary to account for differences, particularly if beneficiaries are aligned on a randomized basis as between DCEs operating in the same target region. Notwithstanding this interest, we seek information on what alternative alignment methodologies CMS might consider and the relative pros and cons of alternative approaches for beneficiaries and for DCEs operating in the same target region. Are there hybrid approaches to consider? For example, would stratified randomization of the beneficiary’s residence be a preferable approach to complete randomization? What implications would either stratified randomization or allowing for voluntary alignment have on risk adjustment considerations? 

NAACOS supports CMS’s increased emphasis on beneficiary engagement, a trend we have seen across CMS programs. Voluntary alignment is one way to engage beneficiaries and would be important in a model that seeks to attract new entrants. As such, it would make sense to educate beneficiaries about voluntary alignment. This approach also reinforces CMS’s goals of fostering competition. We have significant concerns related to beneficiary confusion that could result from a random alignment approach. Further, as discussed above, we support CMS selecting target regions with low ACO penetration and at the same time excluding ACO assigned beneficiaries from Geographic DCEs. As such, ACO beneficiaries should be excluded from any random alignment, should CMS move forward with that approach. 

CMS uses a variety of risk adjustment methodologies across Medicare APMs. For example, the CMS Hierarchical Condition Category (CMS-HCC) prospective risk adjustment model is used in MSSP to calculate beneficiary risk scores, adjust the benchmark years used for the historical benchmark, and compute the rebased historical benchmark. Accurate risk adjustment is imperative to assess APM participant performance, as risk adjustment should remove or minimize differences in health and other factors that impact performance but are outside the participant’s control. We recommend CMS implement a risk adjustment approach that is consistent with other Medicare APMs to not advantage or disadvantage APM participants solely based on CMS risk adjustment. Using consistent risk adjustment principals across APMs avoids situations of arbitrage across APMs. We urge CMS to identify and implement fair and consistent risk adjustment approaches across Medicare APMs, including the Geographic PBP Option. 

2. Are there transparency/notification requirements, in addition to or in lieu of the requirements described above, that CMS should consider to protect beneficiary freedom of choice of any Medicare provider or supplier for beneficiaries aligned to a DCE participating in the Geographic PBP Model Option?

With CMS’s emphasis on beneficiary engagement and voluntary alignment, it will be essential for DCEs to make it abundantly clear that beneficiaries retain full choice of providers. Having the ability to see any Medicare providers is a hallmark of Medicare FFS and one that DCEs should adhere to and clearly communicate to beneficiaries to avoid confusion or limit choice. Beneficiary communications should be in clear and concise with easy to understand language explaining this freedom of choice.  

Questions Related to Program Integrity and Beneficiary Protections

3. Providing incentives to beneficiaries to positively influence their behavior and healthcare decision-making could implicate the fraud and abuse laws and potentially raise quality of care, program cost, or competition concerns, particularly if the incentives would cause beneficiaries to be aligned to one DCE over another entity participating in Direct Contracting or another CMS initiative. What safeguards should CMS put in place to ensure that any beneficiary incentives provided do not negatively impact quality of care, program costs, or competition?

CMS has traditionally provided a broader set of beneficiary incentives in models where the participants take on risk. For participants in total-cost-of-care models such as current ACO programs, concerns of fraud or gaming are greatly diminished as these participants are liable for increased spending under risk-based models. In keeping in line with this policy, Geographic DCEs should be allowed flexibility in the incentives they provide. At a minimum, they should be allowed to identify groups of beneficiaries to target with an incentive, which could help the DCE direct patients to higher-value care or prevent avoidable healthcare costs. However, CMS should similarly allow participants in other initiatives (such as the Next Gen program and the Global PBP and Professional PBP Options) the same slate of beneficiary incentives. This is necessary for CMS to remain consistent in its policy and ensure the Geographic PBP Option does not negatively impact competition.


NAACOS and the Innovation Center share the goal of wanting to transform healthcare payment and delivery systems to better rewards value and incentivizes quality, well-coordinated care. We appreciate CMS’s willingness to listen to feedback on how to implement the new value-based care option. ACOs are proving to be successful in bending the cost curve and improving quality. As such, we urge the Innovation Center to be mindful of their work and build the Geographic PBP Option in a way that doesn’t disrupt ACOs’ work in creating more value for Medicare beneficiaries. 


Clif Gaus, Sc.D.
President and CEO
National Association of ACOs