On March 7 and 8, 2024, the Medicare Payment Advisory Commission (MedPAC) held a virtual public meeting, which included the following sessions:

• Assessing data sources for measuring health care utilization by Medicare Advantage enrollees: Encounter data and other sources;
• Preliminary analysis of Medicare Advantage quality;
• Rural hospital and clinical payment policy: A workplan for 2024-2025; and
• Update on trends and issues in Medicare inpatient psychiatric services.

The full agenda for the meeting and the presentations for the sessions are available here.

MEDPAC EXAMINES STATE OF MEDICARE ADVANTAGE ENCOUNTER DATA

More than half (52%) of eligible Medicare beneficiaries participate in the Medicare Advantage (MA) program. Given the growing importance of MA, CMS has been focused on improving its ability to collect data for the program. In January, the Agency released a Request for Information (RFI) soliciting feedback on how to enhance MA data capabilities, and in February, CMS published a memo on the Health Plan Management System[1] addressing the lack of supplemental benefits data.

Encounter data is the best source of utilization data in MA, and one of the most valuable tools researchers and policy makers must make comparisons between MA and traditional fee-for-service (FFS) Medicare. While encounter submission rates have risen since CMS first began collecting them in 2012, this data set is still incomplete. Unlike FFS, where providers are reimbursed for each claim, MA plans have little incentive to submit encounters that do not impact beneficiary risk scoring, as these encounters do not impact their payment. This incentive mismatch makes it difficult to compare the programs. Accordingly, MedPAC examined the state of encounter data submissions, and discussed strategies to increase data completeness.

MedPAC staffers find that MA encounter submission rates have plateaued

MedPAC staffers reviewed the Commission’s 2019 recommendations on encounter data and presented data on encounter submission rates over time, how these rates varied by program, and how these rates matched with other data sets.

Regarding past recommendations, in 2019 the Commission recommended that Congress direct the Secretary to:

• Establish thresholds for the completeness and accuracy of Medicare Advantage (MA) encounter data.
• Rigorously evaluate MA organizations’ submitted data and provide robust feedback.
• Concurrently apply a payment withhold and provide refunds to MA organizations that meet thresholds.
• Institute a mechanism for direct submission of provider claims to Medicare Administrative Contractors:
  • As a voluntary option for all MA organizations that prefer this method
  • Starting in 2024, for MA organizations that fail to meet thresholds or for all MA organizations if program-wide thresholds are not achieved.

These recommendations still stand and will not be updated as a result of this month’s meetings, though the work presented in this meeting will be made into a chapter in the next report.

When CMS first began collecting encounter data, many plans struggled to set up their systems to submit a properly formatted encounter record. To demonstrate how far plan’s data submission capabilities have come, MedPAC staffers presented the share of contracts that had submitted at least one encounter record across six service categories[2], which rose from 80 percent in 2015 to 96 percent in 2020. To estimate how many encounter records are missing, the staffers then matched encounter data with other data sets that measured MA beneficiary service use across different programs. For example, inpatient services were matched with the Medicare Provider Analysis and Review file (MEDPAR), while skilled nursing services were matched with the minimum data set (MDS). They found that, in 2021, the rate of MA service users that only had a record in the comparison data set ranged from 4 percent for outpatient services, to 15 percent for skilled nursing facilities. To close, the staffers showed how match rates varied across plans, and within plans by service category. High match rates in one category did not guarantee high match rates in other categories, with at least one plan having an over 90 percent match rate with MEDPAR but a 1 percent match rate with the MDS. 28 contracts had a MEDPAR match rate of less than 80 percent, with the average match rate across these contracts being only 21 percent.

The staffers then discussed the data and suggestions for future analysis.

Commissioners highlight importance of understanding patterns in missing encounter data

One flaw that commissioners identified in the staffers research was that matching encounter records with outside data to determine data completeness assumes that performing a similar exercise using FFS claims would yield a 100 percent match rate. This analysis has been tried in the past and the match rate was not 100 percent. To better contextualize the match rates, the commissioners suggested a comparison to FFS match rates.

The majority of the commissioners agreed that the importance of the missing encounters was entirely dependent on whether these encounters are missing at random, of if there were patterns in which encounters are not being submitted. If the encounters are missing at random, then, for example, a 90 percent submission rate is more than enough for accurate data analysis and the missing data is not a major issue. If there are patterns in the missing data, then these patterns need to be understood to avoid drawing incorrect conclusions from data analysis. Given that encounter submission rates were improving steadily before plateauing, it is likely that there is some pattern in the remaining missing encounters that makes them more difficult to collect. Most suggestions for future study were centered around understanding these patterns.

While some commissioners argued in favor of punishing plans that failed to meet encounter submission thresholds, others were concerned that this would punish providers, particularly low-volume providers and those in value-based payment arrangements.

Commissioner Poulsen argued that plans that have value-based payment arrangements with their providers generally have lower costs and better patient outcomes than those that reimburse their providers via traditional FFS. Value-based care will create patterns that are difficult to capture with traditional encounter data. For example, some plans run an anonymous telehealth hotline that is unbilled and cannot be linked to a specific patient. This service adds value for patients but does not facilitate easy data collection and may replace services that would have generated an encounter. Similarly, providers being reimbursed via, for example, a per-beneficiary-per-month capitated payment will have less incentive to submit claims for every service than a FFS provider. Regarding low-volume rural providers, Commissioner Barr pointed out that these providers historically had to deal with 1 to 2 MA plans, but now could be dealing with close to 30 MA plans in their county.

To ease the burden on providers and simplify the submission process, many commissioners supported using the Medicare Administrative Contractors (MACs) to collect MA encounter data and then submit copies to both plans and CMS. A complete description of this process, which the commission recommended in 2019 , can be found on page 28 of chapter 7 of the June 2019 report.

MEDPAC CONDUCTS PRELIMINARY ANALYSIS OF QUALITY IN MEDICARE ADVANTGAGE

As participation in MA grows, it becomes more important that CMS is able to properly evaluate the quality of care provided in the program. Despite the Quality Bonus Program (QBP) distributing over $15 billion annually to plans based on quality metrics, MedPAC does not believe that this program, or any other data collected by CMS, can accurately measure quality in MA. To address this, MedPAC staffers presented preliminary findings from an attempt to measure MA quality via ambulatory care sensitive (ACS) hospitalization rates, the results of a systematic review of studies comparing MA and FFS quality and issues the Commission has identified with the QBP.

MedPAC staffers identify difficulties researchers face when comparing MA and FFS quality

Staffers identified three major issues with the QBP, the level at which the program measures quality, an inability to facilitate comparisons to FFS, and the excessive number of quality measures used in the program. The QBP collects quality measures at the contract level and uses these measures to determine the star rating for all plans under a given contract. These star ratings determine payment. However, contracts cover wide geographies, with the largest MA contract having enrollees in over 46 states. Even if quality measures were perfect, contract level star ratings might not accurately reflect quality at the local level, which is the level at which Medicare beneficiaries choose their plans. The commission has recommended multiple times that quality be measured at the local market-area level, as opposed to the contract level. The commission also recommended replacing the QBP with a different value-payment program in 2020.

To measure MA quality, the staffers calculated risk-adjusted ACS hospitalization rates using 2021 MA encounter data supplemented with data in the Medicare Provider Analysis and Review file (MEDPAR). They found that these rates varied widely across market areas, with market areas in the 90^th percentile having rates of 41.7 hospitalizations per 1,000 enrollees, as compared to rates of only 22.4 per 1,000 enrollees in the 10^thpercentile. Variation was similarly wide between Medicare Advantage Organizations (MAOs) in a given market. At the individual enrollee level, staffers found that Black beneficiaries faced the highest ACS hospitalization rates when compared to White, Hispanic and Asian or Pacific Islander beneficiaries, being the only one of these groups to have a higher hospitalization rate than the national average.

The systemic review found that all studies comparing MA and FFS quality were hampered by difficulties in accounting for coding intensity[3], both when comparing MA and FFS and comparing different MA plans, difficulty accounting for favorable selection[4] in MA, and problems stemming from a lack of MA encounter data completeness and comparability to FFS data.

Commissioners debate how to measure quality in MA

Both the staffers and commissioners acknowledged that the ACS hospitalization rate analysis was insufficient and that further work on this subject was needed, though multiple commissioners felt that comparing FFS and MA using existing data was a losing battle and that the commission should abandon this work. Much of the debate over how to measure MA quality was focused on the level at which we measure quality. Many commissioners felt that contract level quality measures, which could cover a vast geography, would not be useful to beneficiaries who want to understand quality at the local level where they will actually be receiving care. Some called for analysis at the county level, as this is the level at which MA plans bid. One commissioner pointed out that certain measures would better capture quality at different levels (contract, county, etc.) and that there should be some sort of map showing these relationships. Similarly, another commissioner was concerned that some measures would become very “noisy” when applied to smaller levels like the county. Commissioners also highlighted how much of the existing research was driven by the availability of comparator data sets that could be used to validate encounter data (e.g. inpatient services being validated by MEDPAR), and how the lack of such a data set for certain services hindered analysis.

The commissioners were largely in favor of changing the quality metrics, with many arguing for one quality metric system that covered both MA and FFS to facilitate better comparisons. Other issues raised included concerns over measures that allowed plans to “teach to the test,” concerns over measures driving over screening that could lead to over treatment, and a desire for better measures of patient experience. Commissioner Poulsen cautioned against quickly making to0 many quality metric changes, as plans and providers have made significant investments in order to take part in existing quality measurement programs. The commissioner also argued in favor of a more binary comparison system (but not fully binary) between plans, as the difference between, for example, a 4.2 star rating and a 4.3 star rating means nothing to beneficiaries.

Commissioner Cherry pointed out that MA has evolved beyond its original purpose of driving down costs while still providing good quality care as an alternative to fee for service. Now, plans understand that they can drive up reimbursement through coding intensity, and then reinvest this higher reimbursement into benefits not covered by FFS such as dental, vision, and hearing, or into better case management. These benefits could lead to better health outcomes and increased access. The commissioner believes that given this realignment in MA, the commission needs to have a discussion in the future about what the goal of the program is going forward.

MEDPAC REVIEWS ITS RURAL HOSPITAL AND CLINICIAN PAYMENT POLICY WORK PLAN FOR 2024-2025

MedPAC staff presented issues surrounding rural beneficiaries’ access to hospitals and clinicians. The staff first presented on existing rural payment policies and the four principles that target special payments: target payment adjusters to preserve access, target low-volume adjustments on isolated providers, incentives for cost control, and managing the magnitude of these adjustments. Next, MedPAC staff discussed concerns that the Commission had raised surrounding cost-sharing and how this can raise prices in rural hospitals, leading to concern that high prices could drive away some patients. The MedPAC staff then addressed the increase of MA plans in rural areas and how this could affect the access to care because of the administrative burden to providers and the number of options beneficiaries have.

The commissioners first mentioned that it was important to define ‘rural’ properly. The definition that the MedPAC staff uses currently is urban areas are defined as metropolitan plus suburban areas and everything else is rural. Some commissioners had an issue with this and wanted to re-evaluate and/or come up with options on how patient population could be included, not just geography. The Commission then discussed the reform needed for cost-based reimbursement because of their concern that high prices are driving patients away, noting that rural hospitals’ ability to bill fee-for-service Medicare patients and the uninsured on a ‘cost-plus’ basis is not justifiable and should be changed.

Commissioners also discussed the growth in MA and how this presents challenges. Specifically, some commissioners wanted to learn why more beneficiaries were switching into MA plans from FFS. Commissioners noted how there is a relatively even playing field between providers and MA plans in urban areas, but not in rural areas. The staff mentioned that providers have expressed concerns about MA plans during site visits because of the administrative burden of having many MA plans available. The Commission then discussed their concern with access to post-acute care and reform options that could be considered. Commissioners would like more transparency on what hospitals pay for this care, why swing bed prices are high, and the quality of care that patients receive.

The Commission asked staff to include other stakeholders in the conversation such as physicians’ assistants, nurse practitioners, and pharmacists because of their important role in rural healthcare. For the next MedPAC cycle, Commissioners are interested in research on the role that the safety-net index could have in the context of these payment policies. They would also like to include more research on how technology and artificial intelligence (AI) could assist rural healthcare providers. Although no recommendations were made for this topic by the MedPAC staff or the Commission, the Commission will continue to discuss suggested approaches to evaluating rural hospital and clinician payment policies over the next cycle.

COMMISSIONERS DISCUSS TRENDS AND ISSUES IN MEDICARE INPATIENT PSYCHIATRIC SERVICES

During the March 2024 meeting, MedPAC staff gave an update on beneficiaries who reach the 190-day lifetime limit in freestanding inpatient psychiatric facilities (IPFs) and provided information on scatter-bed stays. In 2022, there were about 50,000 beneficiaries who reached the lifetime limit of 190 days. IPFs, free-standing or hospital-based, must meet criteria related to staffing and the provision of psychiatric services; and scatter-bed stays have a psychiatric principal diagnosis but must take place in acute care hospitals. There are two Medicare FFS payment systems for inpatient psychiatric systems – IPF stays are paid per-diem while Inpatient Prospective Payment System (IPPS) scatter-bed stays are paid per stay. Overall, Medicare inpatient psychiatric stays are declining but scatter-bed stays are a growing share of those stays.

The Commission first focused on the arbitrary 190-day limit for in-patient stays and why this number has not changed since 1965 when Medicare was created. Commissioners asked what the costs would be if the lifetime limit was removed, what would happen if the limit was increased by 50%, and beneficiary impacts if either of these new policies were enacted. Most beneficiaries who reached the limit qualified for Medicare under age 65, but the commission requested further research on what beneficiaries are using this program. Related to the lifetime limit, one commissioner noted that the length of stay in a psychiatric facility should not be viewed negatively as it is for other medical diagnoses, because often times, a longer length of stay might be clinically appropriate. Commissioners were also concerned with the quality-of-care beneficiaries receive for their psychiatric care needs, and the wait times for an in-patient bed. The Commission noted that swing beds do not always have consistent access to a psychiatrist. They also were concerned with the follow-up visit percentage being so low after 30 days and noted that this was a failure of care delivery. The Commission requested more research with possible interviews of beneficiaries to further understand how long patients are waiting and more information on these quality-of-care issues.

Overall, Commissioners focused on issues surrounding the 190-day lifetime limit and the steps that could be taken to address these issues, especially since the limit has not changed since the creation of Medicare in 1965. The Commission did not make any recommendations on this topic during this session.

********

This Applied Policy® Summary was prepared by Emma Hammer with support from the Applied Policy team of health policy experts. If you have any questions or need more information, please contact her at ehammer@appliedpolicy.com or at 202-558-5272.

[1] https://www.cms.gov/about-cms/information-systems/hpms/hpms-memos-archive-weekly

[2] Physician, Inpatient, Outpatient, Home Health, Skilled Nursing Facility, and Durable Medical Equipment

[3] MA plans are paid based on the expected spend associated with each beneficiary (referred to as a risk score) and therefore have a financial incentive to code diagnosis that is not present in FFS, making it difficult to compare the health of beneficiaries across programs.

[4] Because beneficiaries can choose between MA and FFS there may be certain types of beneficiaries more drawn to the MA program, and unless the characteristics of these beneficiaries are accounted for in a study, this favorable selection can skew study results.