February 2, L Street, NW, Washington, DC T: F:

Transcription

1 February 2, 2018 Dr. Jeet Guram Special Advisor to the Administrator Centers for Medicare & Medicaid Services U.S. Department of Health & Human Services 200 Independence Avenue, S.W. Washington, D.C Amy Bassano, M.A. Acting Deputy Administrator for Innovation and Quality & Acting Director of the Center for Medicare and Medicaid Innovation 7500 Security Boulevard, Baltimore, MD Dr. Guram and Ms. Bassano, The American Health Care Association (AHCA) represents over 12,000 skilled nursing facilities (SNF) in 50 states and the District of Columbia. The Association is proud of the strides made in quality and our efforts to modernize SNF payment policy intended to support Centers for Medicare & Medicaid Services (CMS) quality improvement efforts as well as payment modernization including SNF participation in bundling. However, we are deeply troubled by the Bundled Care Payment Improvement Advanced (BPCI-Advanced) program. While we recognize the focus on creating advanced alternative payment methods (A-APMs) for physicians, the Association and its members believe SNFs have little meaningful role in BPCI- Advanced. Furthermore, for the Medicare program and beneficiaries, years of BCPI Model 3 expertise and strides in quality, savings strategies, improved care coordination and infrastructure development would be lost without a well defined PAC role in BCPI-Advanced. PAC providers would become a commodity who services are controlled largely by third party conveners many of which have limited PAC expertise. Such a sweeping change likely would negatively impact BPCI-Advanced and bundling in the long term. In brief, AHCA requests a meeting with CMMI to discuss: - Continuation of Model 3 to inform possible development, and inclusion, of a PAC role for the January 2020 BCPI-Advanced application opportunity (see Attachment A for more detail; - Clarification of how CMMI envisions PAC provider participation in BPCI-Advanced, both short-term and long-term options. Specifically, AHCA recommends guidance on gainsharing and use of the 3-Day Waiver; and - Strategies for making a Model 3-like PAC bundle a viable Advanced Alternative Payment Method for physicians. 1

2 Related to item, two above, AHCA also would like to discuss the possibility of specific PAC provider provisions in BCPI-Advanced to facilitate short-term participation. Suggestions include the flexibility to submit a Letter of Intent to Apply (LIA) in March rather than a complete application. The new model will require significant work for SNF and other PAC providers to establish needed relationships with physicians and hospitals. Related to the LIA, we also envision formal applications to follow in April or May 2018 and updated financials in July Additional time greatly would improve PAC provider participation potential as Conveners. AHCA Quality and Payment Improvement Efforts Informative to BPCI-Advanced AHCA has long focused on quality and Medicare modernization which could inform CMMI bundling work. In 1998, AHCA launched its National Quality Award Program, based on the Baldrige Performance Excellence Criteria. Since then, 2,614 member centers have been awarded Gold Level Quality Awards, 593 with Silver and 34 Bronze. The AHCA/NCAL National Quality Award program has grown to be the largest Baldrige-based program in the county, receiving over 1,200 applicants per year. Additionally, a number of payers, including states (Florida, Tennessee and Utah), Medicare Advantage (MA) plans and Accountable Care Organizations (ACOs), include SNFs Quality Award status when making network participation and payment decisions. AHCA also maintains a project called Long-Term Care Trend Tracker (LTC- TT). The platform allows members to monitor and analyze their quality performance across an array of measures. In addition to a member dashboard report which members may view on their own, AHCA also pushes out a LTC-TT Top Line Report offering them highlights on their performance. Additionally, in early 2012, AHCA launched a bold, three-year national effort to further improve the quality of care in America s skilled nursing care centers through its Quality Initiative. The profession s ongoing efforts have improved the lives of the individuals AHCA members serve while also reducing health care costs. The Quality Initiative includes important goals such as reducing the use of off-label antipsychotics, reducing the number of hospital readmissions, as well as improving functional outcomes and customer satisfaction. To date, over half of AHCA members have achieved the antipsychotic goal and the rehospitalization goals, and many more are making significant progress to meeting these goals. In regard to payment, for over four years, AHCA has advanced a SNF payment reform concept. Specifically, the condition category-based concept would move SNFs from an archaic per diem payment system with problematic incentives to a stay or episode-base payment system which addresses many of the challenges noted by CMS, MedPAC, the U.S. Department of Health and Human Services (DHHS) and the U.S. Department of Justice. In its March 2017 payment policy report to Congress, as well as in Congressional testimony, MedPAC notes AHCA proposal would address concerns also addressed by CMS proposal, the Resident Classification System 2

3 Version 1 (RCS-1). The In addition to its own proposal, AHCA is sharing recommendations with CMS aimed at improving its RCS-1 proposal. Additionally, AHCA has completed extensive work on a possible chronic condition episode concept for BPCI as well as possible modifications to BPCI which could be incorporated into BPCI-Advanced (see Attachment A). BPCI Model 3 In terms of bundling, as of October 1, 2017, BPCI Model 3 has 675 participants in BPCI Phase 2 and of that figure, 550 are SNFs. The Lewin Group Evaluation Report 3 indicates evidence of patient selection, challenges with quality and service outcomes as well as deminimis savings. However, the Lewin Report also notes many of its findings are not statistically valid. And, we would respectfully highlight that the Evaluation Report 3 data is derived from 2015, Quarter 4. We do not believe this older data incorporates more mature data and member experiences. Below we offer some highlights from a sample of Model 3 SNF participants performance: 1 Figure 1. SNF Achievements in Reducing Lengths of Stay 1 Source: Avalere Health, Achieving Success in Episodic Bundling, Lessons Learned and Roadmap for Success Client Webinar, April 5,

4 Figure 2. Average Length of State Reductions in Days In Figures 1 and 2, above, SNFs have reduced lengths in several clinical episode groups. Reductions are present in other areas we highlight these as illustrative examples. SNFs also have saved Medicare funds by reducing hospital readmissions. See Figure 3, below. Figure 3. Model 3 SNF Reductions in Hospital Readmissions Additionally, also resulting in saving to Medicare, SNFs have reduced downstream costs via care transition protocols and post-discharge follow along programs. See Figure 4, below. 4

5 Figure 4. Reduced Post-SNF Discharge Service Reductions by key Clinical Conditions Conclusion AHCA has concrete data from members on their achievements under Model 3 as well as ideas for making a PAC bundle a viable A-APM under the MACRA criteria and chronic condition episodes. We would like to schedule a meeting among your bundling team, AHCA members and our researchers to discuss our work and options to aid our members with participating in this critical avenue to risk-bearing for SNFs. Sincerely, Michael W. Cheek Senior Vice President Reimbursement Policy 5

6 Attachment A Recommendations to Modify BPCI -Advanced in Short and Long-Term January 2020 Application Cycle 6

7 Recommendation Issue(s) Action(s) 1. Continue Model 3 and make necessary modifications to ease participant challenges SNF Participation: There is strong interest among SNFs to participate in BPCI, and Model 3 offers the only real opportunity to play a meaningful role. CMMI should maintain Model 3 (or other PAC-only bundle models) in any future iteration of BPCI. Episode Definition: Episodes defined by MS-DRG assigned in the hospital create difficulties in the PAC setting. DRGs do not accurately predict cost and care delivered in PAC settings. DRG-based episodes are narrowly defined and result in volumes that are too low for any one facility to be able to effectively manage risk. AHCA suggests substituting procedure or diagnosis codes to allow patient identification at time of admission to the PAC provider. Moving Target Prices: Shifts in target prices were invisible to most participants, and while some data contractors/conveners helped them track changes, most did not know what their target prices were from quarter to quarter or what drove any changes. Patient Identification: Identifying patients who fall within episodes continues to be one of the largest operational challenges in BPCI. This is directly linked to DRG-defined episodes, and the delay that often occurs with final assignment of the DRG. AHCA member interviews indicate identification error rates between 20 and 40%, making the program almost unsustainable. This causes providers to either overinform or under-inform patients, creating confusion. Providers also invest extra time and resources trying to confirm DRG assignment with the hospital, often met with resistance or annoyance. Low-Volume Challenges: DRG-based episodes are narrowly defined, and therefore the average SNF does not see sufficient volumes in any one episode to be able to adequately spread risk. Most Model 3 participants who had low volumes have failed in achieving savings. CMS may want to consider volume thresholds that providers must meet if they want to participate. Maintain Model 3 and/or add other PAC-only models to Advanced BPCI Define non-drg-based episodes by using ICD-10 codes to define episodes for two different types of medical care: recovery from single acute condition hospitalization (such as infections); and severe chronic conditions (may be multiple) Establish static target pricesetting methodology Explore alternative methodologies for identifying BPCI patients, including using non-drg coding at the point of PAC admission Consider setting volume threshold requirements (e.g., 50% or more of available clinical episodes) to protect facilities from large losses due to low volume 7

8 Recommendation Issue(s) Action(s) 2. Using Model 2 as a framework, increase opportunities for SNF participation Model 2 hospitals have very little incentive to share in risk (and gains) with PAC providers, even though the episodes that lead to financial savings are usually a result of efficiencies gained in the PAC portion of the episode (due to higher cost variations that exist in PAC than in acute care hospitals). CMMI should explore ways to incentivize Model 2 hospitals to partner more meaningfully with PAC providers. Model 2 Patients Discharged to Model 3 PAC: For Model 2 patients discharged to a provider who is also a Model 3 episode initiator for that same episode, CMMI should develop a sharing methodology that would split the loss or the gain between the Model 2 and Model 3 participant. Model 2 Patients Discharged to non-bpci PAC: CMMI should create incentives for Model 2 providers to gainshare with PAC providers that are not in Model 3. CMMI may specify requirements for PAC providers to become gainsharing partners with Model 2 providers (e.g., they must perform certain care management functions for the hospital); PAC providers should have the opportunity to benefit from the savings that they create. Having such an option would allow more small PAC providers to participate in BPCI without bearing full risk and without needing to work under a third-party convener. Require a split of the savings or losses between Model 2 and Model 3 providers where their episodes overlap, rather than allowing one to have complete precedence over another Establish requirements that non-bpci PAC providers would have to meet, that would benefit Model 2 hospitals, to incentivize more shared risk and gainsharing between hospitals and PAC in Model 2 3. Replace Winsorization with appropriate risk adjustment policies Most Model 3 participants are in risk tracks where the top and bottom 1% or 5% of nationally distributed cases are excluded from reconciliation. This approach does not adequately protect the average PAC provider from the typical risk of abnormally high-cost episodes. A more appropriate risk adjustment model for post-acute care must be developed if BPCI is to be scaled nationally, and such a model does not yet exist. Implement a pilot to collect data that can be used to develop a PAC-specific risk adjustment model for EPMs. 8

9 Recommendation Issue(s) Action(s) CMMI could pilot a data collection approach and develop an appropriate risk adjustment model for future iterations of bundling models. AHCA and other industry stakeholders could assist CMMI in this effort. 4. Develop and include uniform, consistent quality measures that are appropriate for post-acute care Advanced APMs require the use of quality measures that are consistent with those provided in the physician Merit-Based Incentive Payment System (MIPS). Most of those measures are specific to physician specialty, patient satisfaction, or physician decision-making. There are not adequate PAC quality measures in MIPS. Examples could include: maintenance or improvement in functional status (risk adjusted); reduce health care system acquired infections; patient satisfaction surveys applicable to PAC (e.g., coreq, which is NQF-endorsed); medication review at end of episode; and selected clinical episode-related measures. Develop a consistent and limited set of quality measures that aligns to those to which physicians and hospitals are already subject 5. Limit reconciliation to 2 rounds Currently a single episode may require more than 12 months to close out completely, as DRG assignments shift with resubmissions and corrections of hospital claims and delays in claims processing. AHCA believes that CMS should bear some risk for the correct identification of BPCI patients to episodes and for claims processing delays by limiting the number of rounds of reconciliation for each episode. This could also help to curb any potential action by hospitals to deliberately adjust claims to impact episode assignments (e.g., deliberately shifting highcost episodes outside of BPCI). Limiting episode reconciliation to two rounds would pose little risk to CMS, as relatively few episodes undergo multiple reconciliation rounds, but it would provide significant protection to any individual PAC provider. Limit the number of reconciliation rounds for any single episode to two. 9

10 Recommendation Issue(s) Action(s) 6. Provide more timely data to BPCI participants, and expand audience Model 3 Awardees: Model 3 participants have reported delays of 3+ months in providing claims data on the reconciliation cycle. Participants actively managing their programs are trying to interface different sources of data on their own. CMMI should work to provide better access to real-time data on service utilization based on what they have learned across all demonstrations. Model 3 Awardee Conveners: Model 3 PAC participants who have a relationship with a third-party convener have reported challenges with obtaining their data from the convener. CMMI should impose stricter data sharing requirements on third-party conveners. Model 2 Non-Participant PAC Providers: PAC providers engaged with Model 2 hospitals do not have access to any claims data or reports from CMMI, making it difficult for them to negotiate with hospitals. CMMI should explore allowing downstream Model 2 PAC providers to have access to rolled up (de-identified) claims data reports on their patients so they see the same data/numbers as hospitals and can better prepare for negotiations and/or process improvements. Share claims data directly with participants in a timely manner Require third-party conveners to share claims data in a timely manner with Model 3 participants. Consider allowing nonparticipant Model 2 downstream PAC providers to access rolled up claims data reports. 7. Modify SNF 3-day rule waiver policies and waive SNF 30- day benefit period for BPCI patients SNF 3-day Waiver in Model 2: AHCA member Model 2 downstream providers have reported very limited use of the SNF 3-day waiver in BPCI. AHCA believes that 3-day waivers can be very helpful in operationalizing care redesign and cost containment strategies, but Model 2 hospitals are cautious to experiment. CMMI could begin to identify successful models of care under a waiver and educate Model 2 hospitals to help assuage concern over its usage. CMMI could also allow Model 2 hospitals/physicians to execute agreements with SNFs individually, if they feel that the SNF should be able to participate in the waiver, since they are at risk for the spending and quality outcomes. SNF 3-day Waiver in Model 3: CMMI currently does not allow Model 3 participant SNFs to access the 3-day waiver. AHCA feels that CMMI is missing a savings opportunity and should consider allowing Model 3 Explore alternate qualitybased criteria for SNFs to access the SNF 3-day waiver under Model 2 Consider policies to allow the SNF 3-day waiver in Model 3 Disconnect waiver eligibility from NH Compare ratings and use alternate quality indicators as criteria Allow waivers of the SNF 30-day benefit period 10

11 Recommendation Issue(s) Action(s) SNFs to access the waiver. CMMI could explore policies to discourage inappropriate admissions, such as disallowing certain populations (e.g., long-term care residents) from being admitted under the waiver. SNF 3-day Waivers Relationship to Nursing Home Compare Rating: Current program policies put requirements on SNFs who are eligible to admit patients under a 3-day waiver based on the Five Star rating on NH Compare. Although referrals are not denied to 1- and 2- star facilities under BPCI, hospitals are effectively limiting their network to waiver-eligible SNFs thereby excluding large numbers of available providers in some MSAs. CMMI should consider more thoughtful criteria to determine if a SNF should be eligible to use a 3-day waiver, such as focusing on the QM portion of the Five Star score, or some other measure-based performance to guarantee quality. SNF 30-day Benefit Period: If a patient is discharged from a hospital, under current Medicare rules they are eligible for coverage of SNF care only within the first 30 days post-discharge. With 60- and 90-day episodes it may be appropriate and cost-effective for a patient to go into a SNF later in the episode for any number of reasons and avoid a costly hospital readmission. CMMI should consider waiving this regulation for both Model 2 and Model 3 participants to allow hospitals and SNFs to partner on new care models for late SNF admission within an episode. regulation in both Model 2 and Model 3 8. Review all overlaps across models and demonstrations for fairness Overarching Overlap Policies: There is currently no overarching policy governing how different CMMI models and demonstration models should interact. Rather, policies have been implemented piecemeal as issues are identified. CMMI should establish one overarching overlap policy for all APMs that does not disproportionately favor one provider type over another. Model 2/Model 3 Precedence Rules: Under current policy, Model 2 episodes take precedence over Model 3 episodes depending on Establish one overarching overlap policy for all APMs Develop a fair Model 2/Model 3 precedence policy that equally rewards or penalizes both Models participants if their episodes overlap 11

12 Recommendation Issue(s) Action(s) relative start dates of participants. As a result, many Model 3 participants are losing episodes to Model 2 while being excluded from gainsharing opportunities. This challenge for Model 3 participants will likely be exacerbated with a new iteration of BPCI, as all new Model 3 organizations will lose precedence to any Model 2 if they are in the same market. CMMI should explore policies that would allow Model 2s and Model 3s to split the difference if their episodes overlap. It could involve determining what the savings or loss would have been under each Model, and then splitting the gain or loss between the two providers. 9. Provide adequate oversight to thirdparty conveners and establish performance requirements Convener Ombudsman : AHCA member Model 3 participants report mostly challenging experiences with third-party conveners, primarily Remedy Partners. Common issues reported are related to fee structure and amount, the quality of services provided, and the inability to end the relationship while remaining a participant. Participants felt that fees were too high for services offered (or lack thereof). AHCA feels that inappropriate fee structures unnecessarily draw resources out of the system that should be used to coordinate care, do care redesign, and invest in quality improvement efforts (e.g., staff, technology). CMMI should set up a process (e.g., ombudsman) to identify participants issues and field complaints about conveners. Fee Structures: AHCA members report high fees to work with thirdparty conveners. For example, one convener charges a percentage of projected episodes (based on CMS-provided historical claims data) plus a 50/50 split of gains and losses with the episode initiator. Many participants feel that the services received do not justify the high fees. CMMI should consider capping fees. Services: Conveners typically offer services such as data analytics, reports, and shared learning events. Often, they do not provide timely access to data reports, leaving the provider with no visibility into their data and no way to access it from another source. Furthermore, if a Establish a process for participants to elevate issues, questions and complaints about conveners to CMMI Place a limit or cap on fees conveners can collect from episode initiators Require that any convener claiming >25% of gains provide services that have been associated with successful programs: timely and comprehensive data feedback, care navigation services that are fully documented, claims review and reconciliation with adequate time for appeals, regular meetings with EIs, and demonstrated accountability in delivery of required services 12

13 Recommendation Issue(s) Action(s) convener is taking some threshold of gain/risk (e.g., 25%), they should be required to take a more active role in partnering with the provider to manage risk. There are certain activities that have been associated with successful bundling programs, such as care navigation and patient identification services, claims review, and regular meetings. CMMI should consider establishing more stringent requirements on third-party conveners who are set to make large gains on behalf of their episode initiating providers. 13

14 14