June 10, 2024

The Honorable Chiquita Brooks-LaSure
Administrator
Centers for Medicare and Medicaid Services
U.S. Department of Health and Human Services
Attention: CMS-1808-P
Submitted electronically to: https://www.regulations.gov

RE: Transforming Episode Accountability Model

Dear Administrator Brooks-LaSure:

The National Association of ACOs (NAACOS) appreciates the opportunity to submit comments in response to Fiscal Year 2025 Hospital Inpatient Prospective Payment System proposed rule which includes a proposed new mandatory model, Transforming Episode Accountability Model (TEAM). NAACOS is a member-led and member-owned nonprofit of more than 470 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 cost. NAACOS represents over 9 million beneficiary lives through Medicare’s population health-focused payment and delivery 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).

NAACOS appreciates the opportunity to provide comments in response to this proposed mandatory model which would hold certain hospitals accountable for costs and quality of care for selected clinical episodes. While NAACOS supports CMS’s efforts to increase participation in alternative payment models, it is critical that models first do no harm. We are concerned with the impact of this model on rural and safety net hospitals and the ability for participants to succeed under the proposed model parameters. Our comments below reflect the concerns of our members and our shared goals to advance value-based payment and participation in APMs.

SUMMARY OF COMMENTS

NAACOS is pleased to see CMS providing additional opportunities for specialists to engage in value-based care arrangements. If structured appropriately, episode models can support driving system changes to advance value-based care and further engage specialists in APMs. However, the proposed TEAM design has inherent challenges including the 3 percent discount factor, 30-day episode length, and target pricing methodology. Should CMS move forward with this model, we strongly recommend revisions to the following model elements, described in more detail in our comments below:

• Incorporate more appropriate levels of risk by allowing for at least two years in upside only before requiring downside financial risk for all model participants.
• Support rural and safety net hospitals by allowing for participation in upside only (non-financial risk bearing tracks) for the duration of the model.
• Delay the start of the model to begin no sooner than 18 months following publication of a final rule.
• Provide more flexible options for participation, including allowing the selection of clinical episodes that align with participants’ patient population needs and clinical focus areas, as well as permitting those not in mandated regions to opt-in to participation in the model and allowing hospitals not located in selected regions to opt-in to the program to allow for additional opportunities for all to participate in value-based care.
• Revise the pricing approach and the three percent discount factor to incorporate an approach that is not penalty-only but allows for opportunities for savings.
• Establish episode lengths specific to each clinical episode, so the duration is more reflective of actual opportunities for savings for participants and clinical needs for patients.
• Exclude low volume hospitals from participation in TEAM to protect against large financial losses due to random variation as a result of assessing a small number of cases and provide reasonable exclusion criteria to protect against random variability in performance driven by low volumes of cases.
• Adjust quality assessments to allow for high quality scores to reduce the discount factor and make other adjustments to reflect more meaningful quality evaluations in the model.
• Finalize overlap policies with an additional voluntary option to carve out patients aligned with full risk models.

PARTICIPATION OPTIONS AND TRANSITION TO RISK

CMS proposes to make TEAM a mandatory, five-year model beginning January 2026 for hospitals in selected geographic regions, defined by Core Based Statistical Areas (CBSAs). CMS proposes to include additional CBSAs with a high number of safety net hospitals and hospitals with less experience in CMS bundled payment models. CMS also proposes three risk tracks, Track one through three, with increasing levels of risk. All participants could participate in the no risk track (Track one) for one year, while only certain participants would be eligible to participate in the moderate risk level, Track two (rural and safety net hospitals, Sole Community Hospitals, Medicare Dependent Hospitals) for performance years (PY) 2 to 5.

Any mandatory model should be designed so that participants have adequate opportunity to gain experience in the model before transitioning to downside risk, participants are provided technical or financial support to prepare and adjust to requirements set forth in the mandatory model, and there is opportunity for participants to achieve financial gains in the model. As currently proposed the model does not meet these objectives as the model does not offer any technical or financial support and only provides limited time in tracks without financial risk. In our experience, model participants require 3-4 years in a model before they are financially successful. Additionally, the model does not offer opportunity for participants to achieve financial success. As we describe in more detail below, requiring a 3 percent discount on a 30-day episode limits the opportunities for participants to achieve savings outside of the hospital diagnosis-related group (DRG) payment. The current approach may simply result in a payment cut to the DRG. Moreover, the penalty-only quality approach will not allow for many participants to succeed in the model. In the absence of significant changes to address these concerns, the model should be voluntary.

NAACOS is also concerned that requiring safety net and rural hospitals to participate in downside risk models would have an adverse impact on patients and access to care in these communities. Many rural and safety net hospitals do not have the resources to invest in the infrastructure necessary to do care transformation effectively and already face access issues. CMS runs the risk of placing additional financial burden on hospitals who are not well resourced, perpetuating access issues. CMS should not mandate participation from these hospitals and risk taking away financial resources from hospitals that are already not well resourced. Additionally, rural and safety net hospitals will have fewer cases for episodes, leading to random variability in performance and as a result, financial losses, due to only a very small number of patients. Additionally, hospitals with little or no experience with value-based care will need time to support care transformation and analytic capabilities necessary to be successful before taking on financial risk. At a minimum, CMS should provide a more gradual on-ramp to risk for all participants, allowing for participation for at least two years in Track one (upside only) and allowing all participants to participate in Track two. CMS should support rural and safety net hospitals by not requiring participation in downside risk tracks. Instead, CMS should allow for participation in Track one for the duration of the model for rural and safety net hospitals. This approach will also allow hospitals who are ready to take on higher levels of risk to transition to a more advanced track when they are ready.

MODEL PARAMATERS

Timing for Start of the Model

CMS proposes to start TEAM January 1, 2026, while selected hospitals will not be notified of their required participation until fall of 2024. Given the short timeframe for hospitals to prepare, we urge CMS to consider delaying the start of the model to begin no sooner than 18 months following publication of a final rule. This will give hospitals who have been selected for mandatory participation in the model a more reasonable timeframe to analyze data and prepare for operational changes that will be necessary for success in the model.

Episode Selection

CMS proposes TEAM participants would test five mandatory episodes including Surgical Hip and Femur Fracture Treatment (SHFFT), Coronary Artery Bypass Graft (CABG), Lower Extremity Joint Replacement (LEJR), Major Bowel Procedure, and Spinal Fusion. CMS also notes they intend to test additional episodes in future program years. CMS should allow participants to select the episodes that make the most sense for their patient population and align with their clinical areas of focus. Additionally, CMS should not require participation in any additional episodes that may be added in future model years as participants will need time to develop operational changes necessary for successful participation.

CMS should also allow hospitals not located in selected regions to opt-in to the program to allow for additional opportunities for all to participate in value-based care. CMMI’s strategic refresh has stated a desire to further integrate specialists and other providers in value-based care models, and this could be an opportunity for further inclusion. CMS continues to solicit feedback on ways they can further engage specialists in value-based care and APMs, and allowing for a voluntary opt-in for any willing participant would allow for more participation from hospitals and any additional downstream collaborators such as specialists, to reach more patients to ultimately provide higher quality care at the lowest cost.

Target Prices and Risk Adjustment

CMS proposes to use what it considers to be the most successful policies from Bundled Payments for Care Improvement Advanced (BPCIA) and Comprehensive Care for Joint Replacement (CJR) target pricing methodologies for use in TEAM for establishing target prices and risk adjustments.

Establishing Target Prices
CMS proposes to establish target prices based on three years of baseline data, prospectively trended forward and calculated at the level of MS-DRG/HCPCS episode type and region, with a 3 percent discount factor. CMS proposes to incorporate a prospective normalization factor into preliminary target prices, which would be subject to a limited adjustment at reconciliation. CMS is also seeking comment on including the Accountable Care Prospective Trend (ACPT) as a trending approach, or other potential ways to increase the accuracy of prospective target prices and mitigate the ratchet effect when TEAM target prices are updated.

NAACOS urges CMS to lower the 3 percent discount, which is too aggressive given the efficiencies that have already been completed through successor models. Combined with a 30-day episode, the model does not offermeaningful ways for participants to be successful in the model. CMS has conducted many years of episode-based payment models which achieved efficiencies that spread beyond the model participants. This makes it difficult for TEAM participants to achieve cost savings beyond the three percent discount proposed. In a 30-day episode there may be insufficient opportunity to create savings outside of the hospital DRG payment. Accordingly, this model may simply a test a payment cut to the DRG rather than optimizing follow-up care outside of the hospital. Instead, participants must be given an opportunity to create enough savings to earn a positive payment and avoid financial penalties. To achieve this, CMS should consider offering a significantly lower discount (e.g. 0.5-1%), allowing participants to share in savings after first dollar saved, and allowing for high quality scores to lower the discount amount used in reconciliation. We understand that CMS sets proposed discounts and target prices to ensure that models achieve actuarial savings. However, models should achieve savings based on reduction in unnecessary services rather than cutting provider payments.

Moreover, participants must beat their own performance over time as the model includes updates to target spending. CMS proposes to use a three-year rolling baseline trended forward to the performance year to calculate target prices for each episode type and region, re-basing target prices annually. Given that rebasing will occur annually, the ratchet effect is significant, meaning participants will need to beat their own performance over time to succeed in the model.

NAACOS supports CMS’s proposal to establish benchmarks based on regional spending targets. NAACOS conceptually supports use of the ACPT to mitigate the ratchet effect that occurs when benchmarks or target prices are established using historic spending, however we caution that this approach is new, and we have ongoing concerns about the use of the ACPT in MSSP. Specifically, we remain concerned that the administrative trend is a single national rate and mitigates any regional spending that is incorporated into target pricing and benchmarks. This ultimately punishes participants who operate in regions with spending growth above that of national inflation. We ask that CMS continue to explore how to address these concerns before incorporating this policy in other models.

CMS proposes to cap episode spending at the 99th percentile at the episode type and region level to address outliers. Instead, NAACOS urges CMS to cap both baseline and performance year spending at the 5th and 95th percentile to better reduce variability due to outliers.

Risk Adjustment Methodology
CMS proposes to use a risk adjustment methodology that will adjust target prices by using an expanded risk adjustment variable that accounts for multiple potential markers of beneficiary social risk (dual eligibility for Medicare/Medicaid, Area Deprivation Index values, and Part D Low Income Subsidy eligibility). CMS also proposes to apply a prospective normalization factor to the benchmark price to calculate preliminary target prices for each episode. Participants will receive risk adjustment multipliers prior to the start of the performance year to estimate target prices, calculated at the MS-DRG level (CMS will calculate risk adjustment multipliers prospectively for each MS-DRG/HCPCS episode type).  NAACOS supports CMS’s use of HCC count, social risk factors, and age to adjust for risk. However, we urge CMS to cap the normalization factor, so as not to mitigate the impact of risk adjustments. CMS proposes to calculate a prospective normalization factor based on the data used to calculate risk adjustment multipliers to ensure risk adjustment does not inflate target prices. CMS should cap normalization to ensure the normalization factor does not have greater impact than the risk adjustment itself.

Episode Duration

CMS proposes a 30-day episode for use with all episode types across the model. Using a 30-day episode length can be beneficial for certain clinical episode types, however using a 30-day duration across all episodes fails to recognize the clinical differences and needs of each episode type. For example, for some episodes, this approach could limit participants’ ability to generate cost savings as most of the costs would be incurred during the hospital stay or procedure given that DRG payments are fixed. This does not leave sufficient opportunities for savings through opportunities to improve transitions of care and provide additional care coordination post hospitalization/procedure. This is particularly true for the CABG and spinal fusion episodes. NAACOS urges CMS to establish episode lengths specific to each clinical episode, so the duration is more reflective of actual opportunities for savings for participants and clinical needs for patients.

Low Volume and Exclusion Criteria

CMS proposes to apply a low volume threshold of 31 episodes across all episode categories. Participants would be subject to Track 1 stop loss/gain limits if they do not meet the low volume threshold (31 episodes in the baseline period) in PY 1 and Track 2 stop loss/gain limits in PY 2 through 5. CMS should instead establish a low volume threshold of at least 40 episodes and evaluate each episode category separately. That is, each episode should be evaluated for low volume and if a participant has less than 40 episodes, the participant should not be included in that specific episode. This criterion is more aligned with the policy used in BPCIA and is more appropriate at reducing impact of variability due to low volume of episodes. It is critical that CMS exclude low volume hospitals from participation in TEAM to protect against large financial losses due to random variation as a result of assessing a small number of cases.

Additionally, CMS proposes using exclusion criteria similar to criteria used in the BPCIA. We are concerned that these excluded costs are too limited in scope and will not appropriately account for items and services that may be unrelated to the episode. Additionally, CMS proposes to apply the excluded costs criteria across all five episodes.

Appropriate exclusion criteria must be targeted to also protect against random variation driven by low volume of cases. Exclusions should be based on clinical appropriateness for each episode, and CMS should not use one set of criteria across all five episode types. CMS should work with stakeholders to identify appropriate exclusion criteria for each episode type and to ensure participants are only held accountable for care that is clinically relevant to the episode.

Quality Requirements

CMS proposes to use a quality approach that requires participants to score 100/100 on quality to earn a complete positive reconciliation. This penalty-only approach is overly punitive. This policy in combination with the steep 3 percent discount makes success in the model difficult. Instead, we urge CMS to consider a policy where high-quality performance would reduce the discount applied at reconciliation. This is consistent with the approach used in the CJR model.

CMS also proposes to evaluate TEAM participants on three quality measures starting in the first performance year: the Hybrid Hospital-Wide All-Cause Readmission Measure (CMS Measure Inventory [CMIT] ID 356); CMS Patient Safety and Adverse Events Composite Measure (CMIT ID 135); and Total Hip and/or Total Knee Arthroplasty Patient Reported Outcome Based Performance Measure (CMIT ID 1618), which is only applicable to LEJR episodes. CMS proposes to use Hospital Inpatient Quality Reporting (IQR) program data to evaluate these measures. CMS also seeks comment on potentially adding additional measures in future model years including Hospital Harm Falls with Injury, Hospital Harm Postoperative Respiratory Failure and 30-day Risk Standardized Death Rate among Surgical Inpatients with Complications.

We appreciate that CMS is using measures already in use in the IQR program to reduce burden associated with the model. However, the proposed measures do not address value-based care principles or reflect opportunities to improve quality related to the specific episodes selected. CMS should work with stakeholders to begin developing measures that are focused on population health and value-based care rather than relying on existing measures that are used in fee-for-service environments. Finally, CMS should also allow for points for improvement in quality scores over time to encourage continuous improvement. We urge CMS to work with stakeholders to identify more appropriate quality measures for the specific episodes selected.

Hospital-Wide All-Cause Readmission Measure
Because this measure primarily leverages administrative claims data and as a result, performance scores are not received until well after the close of the performance period, hospitals do not have the ability to identify potential interventions to improve outcomes quickly. This lack of timely feedback on performance would make it difficult for a hospital to be responsive to gaps in quality and take action in a meaningful way.

Patient Safety and Adverse Events Composite Measure
Because this measure leverages administrative claims data and as a result, performance scores are not received until well after the close of the performance period, hospitals do not have the ability to identify potential interventions to improve outcomes quickly. This lack of timely feedback on performance would make it difficult for a hospital to be responsive to gaps in quality and take action in a meaningful way. Additionally, we oppose use of the Patient Safety measure as it is a duplicative penalty. The measure is currently included in the Hospital Acquired Condition Reduction Program (HACRP), which reduces payments to hospitals based on performance in measures. Both HACRP and TEAM are penalty-only approaches, resulting in the hospital being penalized twice for performance on the same measure.

Total Hip/Knee Patient Reported Outcome Measure
Hospitals are still in their initial round of data collection for this measure and as a result, do not have complete experience in gathering all of the data required for the measure. Even with this limited experience, the measure is extremely burdensome and many are concerned with potential survey fatigue resulting in reduced response rates. Hospitals must collect responses from three surveys for risk adjustment in addition to the hip and knee disability, osteoarthritis and joint replacement outcome surveys and achieve a minimum 50% response rate for the post-operative survey, which will be challenging given the length of time required (up to 425 days after the surgery).

Hospitals need additional time to gain experience collecting and reporting the data required for the measure. In addition, they should be able to see how successful they were in meeting the data collection requirements and minimum survey responses as well as their performance scores prior to use of the measure in this mandatory model.

Overlap Policies

In an effort to promote a cooperative relationship between TEAM participants and other accountable care models, CMS proposes that a beneficiary can be included in a TEAM episode and be attributed to a provider participating in a total cost of care model or program. This policy proposal would allow savings generated in each respective model to be retained by the respective participant. CMS also proposes that TEAM reconciliation payments/repayments would not be included in total cost of care models’ expenditures. CMS also seeks comment on an alternative approach that would prohibit ACO aligned beneficiaries from being attributed to TEAM episodes.

NAACOS supports CMS’s proposed overlap policy that eliminates confusion and allows for coordination between TEAM episode participants and other total cost of care models by allowing beneficiaries to be assigned to multiple models and allowing each respective participant to retain savings generated under their respective model. Avoiding complex attribution and financial reconciliation overlap polies will allow for collaboration among APM participants and reduce confusion created for patients.

We also encourage CMS to allow ACOs have the option to exclude their beneficiaries from being attributed to TEAM episodes. In managing the total cost of care, ACOs often identify areas of care, including episodes, that can benefit from care redesign and achieve savings. It is feasible that ACOs are currently engaged in interventions, with gainsharing agreements in place for downstream providers, for some of the areas of care addressed in the model. Allowing ACOs to exclude their aligned beneficiaries will ensure that we do not introduce additional areas of complexity that could lead to beneficiary confusion.

Beneficiary Incentives and Program Rule Waivers

CMS proposes waivers for both telehealth geographic site and originating site requirements, as well as the Skilled Nursing Facility 3-day rule (allows discharge of TEAM beneficiaries to a SNF with a quality rating of three or more stars). NAACOS supports the inclusion of telehealth and the 3-day SNF rule waivers to support participants in care coordination activities and expanding access to care for beneficiaries. In addition to these waivers, we urge CMS to also include the post discharge home visit waiver for TEAM participants to better support transitions of care from the hospital to home. The post-discharge home visit waiver allows patients to receive visits in their home from a physician or other licensed clinician after an inpatient hospital discharge to help support the transition from the hospital to the patient’s home. This can support not only the beneficiary but also their caregivers and can be helpful in preventing readmissions to the hospital. This waiver was successfully used in the Next Generation ACO model and would be a useful tool for TEAM participants.

We also recommend that CMS and the Innovation Center establish a standard set of waivers across all APMs to reduce administrative burdens and increase use of the waivers across programs and models for entities participating in multiple APMs.

Data Sharing

CMS proposes to share certain beneficiary identifiable claims data for episodes, as well as certain aggregate regional total expenditure data, specifically total expenditures during an anchor hospitalization or procedure and the 30-day post-discharge period for all beneficiaries who have initiated an episode during baseline performance years. This data will be shared for all Parts A and B claims associated with episodes in TEAM for United States Census Divisions which the TEAM participant is located in. NAACOS supports the sharing of regional expenditure data, and we urge CMS to share data in advance of the model start date to allow for the complex evaluation of performance data and to make operational changes that are necessary to succeed in such a model.

Other Issues

Equity Provisions
CMS proposes three equity related requirements of TEAM participants: submission of a health equity plan starting in PY 2; reporting of demographic data starting in PY 2; and screening attributed beneficiaries for four health related social needs (HSRNs) and screened positive data, starting in PY 1.

Addressing health inequities is a laudable goal and one that APMs can support through their focus on population health with the support of certain waivers to allow for flexibilities to address non-medical health needs. However, lack of standardization has made collection of data to support this work, such as social determinants of health (SDOH) and HRSN data, difficult. CMS should allow for more standardization to occur, which will improve all APMs’ effectiveness in reducing health inequities, reduce administrative burdens and better support sharing of best practices.

For example, the REACH model currently has similar equity requirements and participants have struggled to capture data across disparate electronic health records (EHRs) and instances of EHRs. This is due to the numerous taxonomies used by EHRs and providers (for race and ethnicity as an example) as well as varied reporting of the information and screening tools. For example, even when the same screening tool is used, there is variation in how the data is reported by different providers due to variation in underlying definitions.

Additionally, CMS should rely on reporting of this information that is already being done for the Hospital IQR Program rather than requiring duplicative reporting of the same information by TEAM participants. CMS is currently requiring collection of this data across multiple setting-specific programs, which could result in duplicative efforts and penalties when used to affect payment. NAACOS has provided detailed recommendations on this topic, including more impactful ways CMS could engage with APMs on this issue.

Beneficiary Notifications
CMS proposes to require TEAM participants to require all ACOs, providers, and suppliers who execute a Sharing Arrangement to provide beneficiary notification materials prior to discharge from the anchor hospitalization or procedure and that the TEAM participant must provide the beneficiary notification materials if there is not a Sharing Arrangement. CMS also proposes to require that TEAM participants require every TEAM collaborator to provide written notice to TEAM beneficiaries of the existence of its sharing arrangement with the TEAM participant and the quality and payment incentives under the model, which must be provided no later than when the beneficiary receives the first item or service from the TEAM collaborator.

If implemented as proposed, these policies would require multiple notifications from multiple entities with duplicative information being shared with the same beneficiary. This could result in beneficiary confusion and frustration as well as wasted resources for participants. NAACOS is also concerned with the timing requirements for notifications which CMS notes are infeasible. On page 36472 CMS states, “We propose that the notice must be provided no later than the time at which the beneficiary first receives an item or service from the TEAM collaborator during an episode. We recognize that due to the patient’s condition, it may not be feasible to provide notification at such time, in which case the notification must be provided to the beneficiary or his or her representative as soon as is reasonably practicable.” CMS should not finalize beneficiary notification requirements that participants cannot reasonably comply with.

NAACOS also strongly encourages CMS to work with patient and consumer advocacy organizations and Medicare beneficiaries to ensure notification materials are using language that is simple, easy to understand and meaningful to patients. For example, in the proposed rule CMS notes that any ACOs, providers, or suppliers with a Sharing Arrangement with the TEAM participant must be identified as a “financial partner of the hospital for the purposes of participation in TEAM.” Using the term “financial partner” in that way may have negative connotations for some beneficiaries and does not capture any care coordination or care delivery arrangements between model collaborators, which would be more meaningful to beneficiaries and their care. Instead, NAACOS prefers CMS’s alternative proposal that would require all beneficiary notification materials are provided by participating hospitals at the point of admission for all beneficiaries, as it would streamline the notifications and avoid duplicate notices being provided to the same beneficiaries and reduce administrative burdens. CMS should work with stakeholders to ensure the information being provided is useful for patients and should allow for participants to customize notifications. CMS should allow participants to tailor the timing and information communicated to beneficiaries based on the needs and preferences of their populations.

CONCLUSION

Thank you for the opportunity to provide feedback on the proposed TEAM. 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 designing models that further engage specialists in total cost of care models in ways that all participants can be successful. If you have any questions, please contact Aisha Pittman, senior vice president, government affairs at [email protected].

Sincerely,

Clif Gaus, Sc.D.
President and CEO
NAACOS