In April the Centers for Medicare & Medicaid Services published its proposed rule governing how it plans to pay hospitals for Medicare-covered inpatient services in FY 2017. The rulemaking process includes an invitation to stakeholders to submit comments on what CMS has proposed.

NAUH LogoNAUH submitted extensive comments in response to the proposed rule, addressing six aspects of what CMS proposed:

  • Medicare disproportionate share (Medicare DSH) payments
  • the Medicare hospital readmissions reduction program
  • inpatient rates
  • observation status/the two-midnight rule
  • the outlier threshold
  • reporting data for the Medicare area wage index

This week the NAUH blog presents NAUH’s comments. Our schedule is as follows:

  • Monday – objections to the proposed new methodology for calculating Medicare DSH uncompensated care payments
  • Tuesday – NAUH’s own proposed alternative methodology for calculating Medicare DSH uncompensated care payments
  • Wednesday – NAUH’s concern about the size of the proposed Medicare DSH uncompensated care pool
  • Thursday – the Medicare hospital readmissions reduction program
  • Friday – inpatient rates, observation status/the two-midnight rule, and the outlier threshold

Find NAUH’s complete comment letter here.

Medicare DSH: Objections to the Proposed New Methodology for Calculating Medicare DSH Uncompensated Care Payments

In the proposed FY 2017 inpatient prospective payment system regulation, CMS proposes initiating a three-year transition, beginning in FY 2018, from calculating hospitals’ Medicare DSH uncompensated care payments based on its current uninsured proxy of Medicaid and SSI days to calculating those payments based on the uncompensated care hospitals report on their Medicare cost report’s S-10 form. This means using uncompensated care data from line 30 of the S-10 (charity care and non-Medicare bad debt) to calculate Medicare DSH uncompensated care payments rather than the Medicaid/SSI days proxy. To adjust for questionable data, CMS proposes trimming hospitals with cost-to-charge ratios more than three standard deviations above the mean and replace such ratios with average cost-to-charge ratios. In making this proposal, CMS cited a belief that the Medicaid/SSI days proxy does not have adequate correlation with hospital uncompensated care. CMS’s conclusion was based on a report by a consultant that looked at non-profit hospitals’ 990 forms and concluded that the correlation between those reports’ uncompensated care figures and what hospitals report on their S-10 has improved over time and that this suggests hospitals are doing a better job of reporting their uncompensated care data.

NAUH respectfully disagrees with both the CMS proposal and the basis of MedPAC’s recommendation. NAUH believes the data CMS would derive from the current S-10 form, and that it proposes using to calculate hospitals’ Medicare DSH uncompensated care payments is inaccurate, lacks uniformity, and should not be used, under any circumstances, for important public policy-making until steps are taken to improve it and those improvements are verified through rigorous auditing. The study indicates that among hospitals that file a form 990, consistency of reporting uncompensated care has improved. That study does not, however, illustrate or even evaluate whether this data is a reasonable proxy for the costs hospitals incur in providing care to the uninsured. Further, we know that the most notable aberrations in the S-10 data occur among public hospitals, which do not file a form 990 and are therefore missing from the analysis.

NAUH LogoAt the heart of this problem are the instructions for completing the S-10 and the manner in which hospitals interpret those instructions. Those instructions, NAUH believes, are imprecise, leaving them too subject to interpretation – beginning with the question of what, exactly, is “uncompensated care”? Five years ago, a consultant CMS hired to help it implement Affordable Care Act-mandated changes in Medicare DSH policy concluded that “…we found variations in how existing programs and entities define uncompensated care.” Those definitions, the consultant noted, vary among federal programs, the states, rating and research organizations, and provider organizations.”

The result, NAUH has found in its continuing analysis of uncompensated care reporting on the S-10 form over the past five years, is wild inconsistency in how hospitals report their uncompensated care.

Some examples illustrate this conclusion.

  • One hospital with a charity care policy that includes care for Medicaid patients whose stays exceed a day limit may include those days in its charity care line on the S-10 while another hospital may include that shortfall on the Medicaid line of the S-10. In this example, both hospitals are providing the same care for the same reimbursement but only one shows the shortfall as uncompensated care.
  • Some public hospitals report all of the care they provide to their low-income and uninsured patients on the charity care line of the S-10 and payments from local or county funds or indigent care pool resources on line 18 while others report a reduced number in the charity care line to account for payments received; still others do not report their government support at all.
  • Some hospitals report all charges for their uninsured patients on line 23 of the S-10 while others include some of that data on line 29.
  • Some hospitals report full charges for the uncompensated care they provide while others report those charges at a discounted rate.
  • In some states, hospitals draw funds from CMS-approved state uncompensated care pools to help with the cost of caring for their uninsured patients. Some of these hospitals subtract the money they draw down from these pools from their uncompensated care costs while others do not. Then, in some places, uncompensated care payments are reflected in the S-10’s line 19 and not in lines 18 or 20 while other providers report comparable uncompensated care differently.
  • More than 60 hospitals did not include the S-10 in their FY 2014 Medicare cost report and another 23 reported providing negative amounts of uncompensated care.

These examples clearly illustrate the challenges CMS has not yet overcome in addressing the inconsistencies in how hospitals can and do report their uncompensated care data – if, as the final point illustrates, they report such data at all.

In some places with public hospitals, the local or county government places all of its government-funded FQHCs and section 330 clinics under the public hospital. Because the public hospital does not claim any direct payments for the care these clinics provide, they claim all of those services as uncompensated care, entitling them to a larger share of Medicare DSH uncompensated care pool even though the local government is financing the clinics. While public hospitals in places like Dallas, Los Angeles, and Chicago do this and benefit enormously from this practice through larger Medicare DSH payments, cities without public hospitals cannot do this at all. This is either unfair to federal taxpayers or unfair to the residents of cities that do not have the same opportunity, yet the instructions for completing the S-10 have not been revised to clarify what is and is not permitted nor have the S-10 reports of the hospitals and health systems been audited to identify and correct such situations.

Attempting to attribute different types of data reporting problems to specific types of hospitals is extremely difficult. After spending five years reviewing the data of every general acute-care hospital in the country that participates in Medicare’s inpatient prospective payment system, NAUH has found that same types of hospitals may report their uncompensated care differently; different types of hospitals may report their uncompensated care differently; there can be differences in reporting practices even within distinct hospital types; and there can be differences in reporting practices within regions, within individual states, and even between neighboring hospitals.

NAUH has been monitoring S-10 data reporting ever since CMS began addressing the Affordable Care Act mandate to adjust the uncompensated care portion of hospitals’ Medicare DSH payments to reflect the expected decline in the number of uninsured patients they serve. We have polled our members to learn how they report their uncompensated care; we have polled other private urban safety-net hospitals to learn how they report their uncompensated care; we talk to hospital finance executives continually to learn how they report their uncompensated care; and we talk to hospital association officials almost continually to learn about how their members report their uncompensated care.

The lesson we take from all of these efforts is clear: there is no single, uniform manner in which hospitals report their uncompensated care and very little in the way of any patterns for how different types of hospitals, or hospitals located in different areas, report their uncompensated care. Our experience suggests that if you put 100 hospital finance executives in a room and ask them how they report their uncompensated care on their S-10 you will probably get a dozen different answers – if not more.

NAUH closely monitors the uncompensated care that every general acute-care hospital in the country reports on its S-10 and annually models how the data those hospitals report might alter their Medicare DSH uncompensated care payments if CMS were to shift the basis of those payments from the current Medicaid/SSI days proxy to the S-10. Our latest modeling, based on hospitals’ FY 2014 Medicare cost reports, produced the following examples of questionable practices or outcomes.

  • A county hospital in Texas reported $655 million in uncompensated care on line 30 on its S-10 yet it receives $574 million in county appropriations and $165 million in Medicaid payments in excess of its Medicaid costs, the latter of which included uncompensated care pool funds from the Texas section 1115 waiver. Reason suggests that the purpose of the county appropriation is to support care for the uninsured and that it therefore should be deducted from the line 30 figure to yield a true uncompensated care figure.
  • A public Louisiana hospital reported providing nearly $670,000 worth of uncompensated care per bed. It also reported $2.9 million on line 18 of its S-10 that year. Subtracting this government assistance from its uncompensated care total still leaves $643,000 worth of uncompensated care per bed – 1300 percent of the average amount of uncompensated care per bed reported by hospitals.
  • A 96-bed public hospital in Texas reported $534 million in line 30 uncompensated care. This amounts to more than $5.5 million worth of uncompensated care per bed, or 11,700 percent of the average uncompensated care per bed among all hospitals that report providing charity care. On line 18 of this public hospital’s S-10 form it reports no government financial support at all.
  • A new, 10-bed non-profit hospital in Ohio reported providing nearly $6.5 million in uncompensated care – more than 30 percent of its net patient revenue.
  • A non-profit hospital in Connecticut reported a nearly 2000 percent increase in its uncompensated care from FY 2013 to FY 2014.
  • A hospital in Massachusetts, a state in which 97 percent of the residents are currently insured, reported providing more than $100 million in line 30 charity care – even though Massachusetts has a CMS-approved uncompensated care pool.
  • A for-profit hospital in Florida reported an increase in its uncompensated care from negative $5.5 million in FY 2013 to $33.6 million in FY 2014, leaving uncompensated care to account for more than 16 percent of its net patient revenue.

These aberrant numbers, and the lack of such aberrant numbers for so many other hospitals, illustrate some combination of misinterpretation of the S-10’s instructions, the lack of clarity of those instructions, and the possible lengths to which some providers appear to be going to report their costs and expenses in a manner that will maximize their Medicare DSH money – and that will maximize that Medicare DSH money in a manner that, in the zero-sum environment that Medicare DSH has become, enable them to benefit at the expense of hospitals throughout the country and, in some cases, their neighboring competitors.

Such difficult-to-accept uncompensated care reporting has serious implications for entire states.

California hospitals would collectively see their Medicare DSH uncompensated care payments decline more than $400 million if the S-10 were to become the basis for calculating those payments.

  • Texas hospitals would gain nearly $600 million in additional Medicare DSH uncompensated care payments if the S-10 were to become the basis for calculating those payments. Together, two Texas hospitals alone would see their Medicare DSH uncompensated care payments rise nearly $300 million at a time when the number of uninsured people in that state is declining.
  • While the nation-wide uninsured rate is currently 11.7 percent, Oklahoma has the second-highest uninsured rate in the country: 16.5 percent. Despite this, the proposal to base Medicare DSH uncompensated care payments on S-10 data would result in that state’s hospitals losing $11.7 million in such payments.
  • Meanwhile, New Jersey’s uninsured rate is well below the national average: just 9.7 percent. Despite this, the proposed methodology would result in its state’s hospitals receiving $63 million more in Medicare DSH uncompensated care payments even though that state also has a CMS-approved uncompensated care pool.

These and countless other examples point to the inadequacy of current S-10 data for the purposes of determining how much uncompensated care hospitals provide and calculating eligible hospitals’ Medicare DSH uncompensated care payments. In the current environment, supported by unclear instructions for completing the S-10 and the lack of auditing and enforcement, they have created a system that clearly benefits some states more than others, and at the expense of those others.

Further, while NAUH does not dispute the consultant’s observation that hospitals are doing a better job of reporting their uncompensated care data on the S-10 than they did a few years ago, we disagree about the significance of this observation. Even if is true in the aggregate, the nature of this calculation of Medicare DSH uncompensated care payments is such that the remaining inaccuracy and lack of uniformity in the data reported can have a very large impact on hospitals in the zero-sum environment that Medicare DSH uncompensated care has become: hospitals that, for whatever reason, over-report their uncompensated care benefit financially from doing so while those that do not suffer financial harm. Even a few hospitals with questionable reporting, moreover, can affect a large number of other hospitals. One wonders, for example, how the uncompensated care reported by the two Texas hospitals cited above that would result in them receiving an additional $300 million in Medicare DSH uncompensated care payments under the proposed methodology would affect other DSH-eligible hospitals around the country. For this reason, the possibility that some hospitals are “doing better” is not good enough. They all have to do better, and until they do, this data is not good enough for public policy-making purposes. Today, despite any apparent improvements in data reporting, the data generated through the S-10 remains fundamentally flawed: inaccurate, inconsistent from hospital to hospital, based on questionable and varying interpretations of the instructions for completing the form, and in general unreliable.

NAUH LogoFinally, NAUH supports CMS’s proposal to require hospitals to report charity care data on the S-10 based on the date that charity care is recorded rather than the date on which it is delivered.

In light of all these concerns, considerations, and problems, how can the S-10 instructions be brought up to a standard that reflects CMS’s intent and produces the kind of reliable uncompensated care data needed to constitute the basis of fair Medicare DSH uncompensated care payments?

NAUH believes CMS can improve the accuracy, reliability, and uniformity of S-10 data by taking the following steps:

  • Define all the terms – especially “uncompensated care – more thoroughly and more clearly.
  • Improve the instructions for each line.
  • Be clearer, and much more specific, about what revenue must be reported as offsetting uncompensated care and where on the form that offsetting revenue should be reported.
  • Audit hospitals’ data reporting to verify both the effectiveness of the improvements and the quality of the data hospitals report.

The value of auditing data cannot be overstated. When Medicaid DSH reporting standards were modified some years ago, audits uncovered significant problems in hospitals’ initial filings and many hospitals needed to revise their filings in keeping with the auditors’ findings. This is a natural aspect of the evolution of data reporting: even though the hospitals were being diligent, their understanding of the new requirements was incomplete and required the guidance of auditors. The anomalies in S-10 data reporting presented above, as well as the many more that can be found, illustrate what can happen when unclear instructions are combined with lack of accountability. This auditing, moreover, must be rigorous: audits comparable in thoroughness to those CMS employs to review wage index data reporting, as opposed to the less rigorous HITECH audits. That auditing might even be used as part of the kind of “probe and educate” approach CMS has used in the past, with the new policy to be implemented only when the agency is satisfied that hospitals are now ready to meet the challenge of providing the data CMS needs to implement this important public policy.

NAUH also would like to offer four suggestions for improving the S-10; we shared two of these with CMS in the past.

  • Line 21 reduces the charity care charges reported on line 20 by a cost-to-charge ratio. The charges in column two represent co-pays and deductibles waived under a charity care policy. NAUH does not believe it is appropriate to reduce these amounts by a cost-to-charge ratio because they do not represent full charges. If CMS agrees with this suggestion, NAUH believes it may be necessary to clarify the instructions for line 20 to make it clearer that any charges associated with Medicaid non-covered services or days exceeding length-of-stay limits should be reported in column one.
  • Line 29 reduces the bad debt expense reported on line 28 by a cost-to-charge ratio. Similar to our concern with the charity care cost calculation above, some portion of this bad debt expense is attributable to patient coinsurance and deductibles which are already reported at a net payment amount and should not be further reduced by a cost-to-charge ratio. Unlike charity care, however, this portion is not separately identifiable in bad debt. We believe the S-10 should be revised to avoid reducing amounts that do not represent full charges by a cost-to-charge ratio.
  • Line 30, which is a hospital’s total unreimbursed and uncompensated care cost, should reflect the impact of revenue associated with government grants and appropriations and transfers for support of hospital operations, private donation and grant funds, hospital-funded uncompensated care pools, intergovernmental transfers and grants, and endowment income restricted to funding charity care. To that end, NAUH believes line 30 should be revised to instruct hospitals to subtract lines 17 and 18 when calculating the value of this field. With this revision, line 30 would represent a more appropriate value to use for calculating future Medicare DSH adjustments. Even if the form is not revised, NAUH suggests that for the purposes of future Medicare DSH adjustments, this same methodology should be used to determine hospitals’ uncompensated care costs.
  • All charges on the S-10 are reduced by a single cost-to-charge ratio reported on line 1. In our experience, hospitals often have significantly different cost-to-charge ratios for their inpatient and outpatient services and often have a significantly different mix of inpatient and outpatient service utilization among the charity care population and the overall hospital population. For these reasons, we believe it would be more appropriate for the S-10 to distinguish between inpatient and outpatient charges and adjust them by separate cost-to-charge ratios.

NAUH would be pleased to meet with CMS officials to review these proposals and, if you wish, to convene a task force of our own members to develop recommendations for other aspects of the S-10.

Finally, NAUH is aware that there there is a perception that S-10 data will not improve until it is actually used for payment purposes and that it is therefore better to start using the data now than it would be to spend time trying to improve it. We disagree with such an approach, for several reasons.

  • Using the data in its current form could result in Medicare DSH payment reductions that could cause serious and even potentially irreparable harm to some safety-net hospitals.
  • Proceeding with use of this data now is especially a problem because there is a fixed pool of money for Medicare DSH uncompensated care payments. This means that one hospital’s “mistake” or “misinterpretation” would directly harm others.
  • Hospitals have no appeal or litigation rights if they believe the methodology for making these payments treats them unfairly. This, in our view, makes it inappropriate to begin using the approach with the assumption that it will treat some unfairly and that they will just have to accept that they have no recourse over that unfair treatment.

For these reasons, NAUH opposes using data that is seriously flawed and urges officials to fix the data before using it for payment purposes.

NAUH believes that in its current form, the S-10 is inadequate for public policy-making purposes: inaccurate, unreliable, deeply flawed, and ultimately misleading about the work many hospitals are doing in their communities. Numerous examples illustrate that using the S-10 to calculate hospitals’ Medicare DSH uncompensated care payments would use erroneous data to unjustly reward some hospitals and do so at the expense of others. It also would reward hospitals in states that chose not to expand their Medicaid programs while penalizing those in states that did. NAUH recommends that CMS withdraw its proposal to use the S-10 to calculate Medicare DSH uncompensated care payments beginning in FY 2018 and instead begin a process of examining the S-10 and the data reported on it; clarifying the meaning and intention of each line on the report; improving the instructions for completing the report; and initiating rigorous auditing of hospitals’ data reporting on a revised S-10. NAUH further recommends that CMS only use the S-10 to determine hospitals’ Medicare DSH uncompensated care payments when it can do so with a high degree of confidence that the data reported on the form is accurate and that distributing the payments in this manner will be fair to all eligible hospitals, the residents of the communities they serve, and American taxpayers. 

Find NAUH’s complete comment letter here.