Statement by
Thomas A. Scully
Administrator, Centers For Medicare & Medicaid Services
Medicare Payment For Hospital Outliers
before the
Senate Appropriations Subcommittee on Labor, Health and Human Services, and Education

March 11, 2003

Chairman Specter, Senator Harkin, distinguished Subcommittee members, thank you for inviting me to discuss how Medicare pays for inpatient hospital outlier costs. Outlier payments are those Medicare payments that are available to hospitals to help ensure that Medicare's sickest beneficiaries continue to have access to high quality health care. We have an obligation to safeguard this access for Medicare's sickest patients, and to ensure that we are spending taxpayer funds appropriately. Medicare generally pays most hospitals a predetermined amount for a typical patient's stay, based on the average cost of providing care to a patient in a similar condition. The payment is set under a diagnostic related group (DRG) that will allow an efficiently operated hospital to earn a reasonable rate of return overall. However, Medicare law recognizes that there are some beneficiary cases that are more complicated - and therefore more costly to treat - and requires that we pay an additional amount to hospitals for these cases known as outlier payments.

Outlier payments can be viewed as insurance for hospitals against the large losses that could result from extremely expensive cases. In addition to the fixed rate per case, hospitals receive outlier payments when the estimated costs of the case exceed a fixed-loss threshold. This fixed-loss threshold operates like the deductible in a typical insurance policy. When the cost of a case exceeds the fixed-loss threshold, we pay the hospital an additional amount equal to 80 percent of the estimated costs beyond that loss threshold - similar to coinsurance required in most insurance policies. Under the law, CMS must set outlier payments between 5-6 percent of total inpatient payments. In recent years, CMS has set the outlier threshold at a level projected to pay 5.1 percent of total payments for inpatient care for these outliers - which was projected to result in spending of $3.7 billion in FY 2002. However, in reality, we spent $5.3 billion - a difference of $1.6 billion.

We are particularly concerned about excessive claims for outlier payments. Each year, when we update the hospital payment rates, we set a threshold for individual outlier payments designed to keep them at the overall target of 5.1 percent of total payments. As outlier claims increased (and the agency had no idea why) the outlier threshold has skyrocketed - from $14,050 in 2000 to $33,560 in 2003 - as the agency raised the bar to try (very unsuccessfully) to stay within the 5.1 percent target. As a direct result, more hospitals have been forced to absorb the costs of the complex cases they treat, while a relatively small number of hospitals that have been aggressively gaming the current rules benefit by getting a hugely disproportionate share of outlier payments. As you can see, the behavior of a few hundred hospitals - those that took advantage of the outlier program - are the main cause of the sharp increases in the loss threshold.

CMS sets the outlier threshold to reach a target of total outlier payments equal to 5.1 percent of DRG payments. Moreover, we found that these hospitals, in order to maintain that level of outlier reimbursement, significantly exceeded the national average in raising charges between FY99-01, with Hospital A's charges increasing by 111 percent, and Hospital B's charges increasing by 75 percent, compared to the national average of 18 percent. For fiscal year 2002, total outliers exceeded that target and reached 6.9 percent of total inpatient hospital payments. Hospital A in California received outlier payments equaling more than 215 percent of its DRG payments in FY 2002 - that's twice the amount of outlier payments to regular payments. Putting it another way, Hospital A received only approximately $29 million in regular DRG payments, but $62.5 million in outlier payments. However, under the target outlier amount of 5.1 percent, Hospital A would have only received $1.5 million in outlier payments - a difference of $61 million. Hospital B, located in New Jersey, received outlier payments of 129 percent of its DRG payments. In dollars, Hospital B received about $113 million in regular DRG payment and $87.6 million in outlier payments, but under the national target of 5.1 percent, would have only received about $4.7 million in outlier payments. Again, the difference is substantial - about $83 million.

To make sure that outlier payments are used as they were intended, we issued a proposed rule that suggests revising the current outlier policy to prevent further gaming of the system by the few hospitals that get the lion's share of these payments. I wanted to close the loophole immediately, and help the hospitals that need it the most. Instead, we put out changes for comment, partially due to the Committee's interest in evaluating the impact of the changes. I would like to discuss our new proposed rule in greater detail with you today. However, I think it's important that I first discuss the logistics of the outlier system so that I may better explain how the new regulations improve the system, safeguarding access to care for a broader number of Medicare beneficiaries.


The prospective payment system (PPS) was designed to pay hospitals appropriately while providing an incentive for hospitals to provide care as efficiently as possible. Under the inpatient PPS, Medicare pays hospitals a pre-determined, per-discharge rate for 509 patient categories called diagnosis related groups (DRGs). Each patient discharge is assigned to a DRG based on diagnosis, surgery, patient age, discharge destination and sex. Each DRG has a weight established for it based on charges submitted by hospitals for Medicare patients. Each weight reflects the relative average charge, across all hospitals, of treating cases classified in that DRG. In total, Medicare paid approximately $100 billion in FY 2002 for inpatient hospital services.

Because the PPS payment is based on an adjusted average payment rate, Medicare's payment for some beneficiary cases will be higher than the actual cost of the case while for other cases, payment will be less than the actual cost of the case. The system is designed to give hospitals the incentive to manage their operations more efficiently by evaluating those areas in which increased efficiencies can be instituted without adversely affecting the quality of care. Under inpatient hospital PPS, additional payments are made for those cases that generate extremely high costs when compared to average cases in the same DRG. These outlier payments are intended to protect hospitals from large financial losses due to unusually expensive cases.

The Medicare billing form for inpatient stays provides hospitals with the opportunity to code whether they are requesting an outlier payment. When such a code is not selected, Medicare's fiscal intermediaries (FI), the private-sector contractors that process and pay Medicare hospital claims, identify outlier cases by comparing the estimated costs for a case to a DRG-specific fixed-loss threshold amount. The fixed-loss threshold amount is the sum of the DRG payment for the case, any add-on payments (new technology, indirect medical education, disproportionate share adjustment), and the fixed-loss threshold. We set the fixed-loss threshold each year at an amount that is projected to generate outlier payments equal to 5.1 percent of total payments under the PPS. Medicare then pays 80 percent of hospitals' costs above their fixed-loss threshold amounts. However, in the past few years actual outlier spending has exceeded the projected 5.1 percent. For example, outlier payments totaled 7.6 percent in 1999 ($5.2 billion, or $1.8 billion more than projected); 7.6 percent in 2000 ($5.3 billion, or $1.8 billion more than projected); 7.7 percent in 2001 ($5.5 billion, or $1.9 billion more than projected); and 6.9 percent in 2002 ($5.3 billion, or $1.6 billion more than projected). The national fixed-loss threshold for 2003 is $33,560, up from $21,025 in 2002. As you can see, as hospitals claim more and more outlier cases, we have been forced to raise the fixed-loss threshold to remain close to the 5.1 percent target (and we have not been close). Moreover, taxpayers spent $7.1 billion more than Congress authorized for these payments between 1999 and 2002 because of abusive practices and our inability to track or understand the abuses - until now.


The inpatient PPS outlier policy, prior to our publication of the new proposed rule, had a couple of major problems that allowed hospitals to claim excessive outlier payments. In order to estimate the actual costs incurred by a hospital for a given case, Medicare's FIs use the historical relationship between each hospital's costs and its charges to estimate the "true" cost of individual case. The cost-to-charge ratio is determined by using the most recently settled cost report. So long as hospital costs and hospital charges change at roughly the same rate, this estimate produces a relatively reliable result. However, if a hospital increases its charges dramatically relative to costs in its most recently settled cost report, the use of the historical relationship will yield higher outlier payments than would be appropriate.

In addition, the longer the lag between the historical data and current charges, the more likely it is that the cost-to-charge ratio estimate will be inaccurate. Hospitals must submit their cost reports within 5 months after the end of their fiscal year. CMS makes a decision to accept a cost report within 30 days. Once the cost report is accepted, CMS makes a tentative settlement of the cost report within 60 days. The tentative settlement is a cursory review of the filed cost report to determine the amount of payment to be paid to the hospital if an amount is due on the as-filed cost report. After the cost report is tentatively settled, it can take 12 to 24 months, depending on the type of review or audit, before the cost report is final-settled. So if a hospital quickly starts raising its charges, the cost-to-charge ratio we use to calculate the payment does not immediately reflect that change. The increase in charges will eventually result in a lower cost-to-charge ratio; however, during this lag time, hospitals can receive higher outlier payments than if a more current cost-to-charge ratio was used. While generally there is a lag period of a little under two years, over the past several years, the lag has been slightly longer, providing hospitals with a longer timeframe within which to continue gaming the system.

For example, let's say that a hospital's cost-to-charge ratio from the settled 2000 cost report is 0.2 (meaning that the average charges were five times higher than average costs), and the patient charges for a particular case were $300,000. The estimated cost for that case would be $60,000 ($300,000 X 0.2), and the outlier payment would be based on that $60,000 cost (the outlier payment is based on 80 percent of the difference between $60,000 and $33,560). Then, assume that hospital starts raising its charges for that same type of patient to $320,000. The cost-to-charge ratio used would not immediately reflect this increased cost, but would rather remain at 0.2, based on data from the 2000 cost report. So, the estimated cost for this same case would rise to $64,000 ($320,000 X 0.2). The hospital would then receive outlier payments based on $64,000, instead of $60,000.

Another problem with the current outlier policy is hospitals' ability in certain cases to have the statewide average number substituted for their own cost-to-charge ratio. If the cost-to-charge ratio for a particular hospital is more than 3 standard deviations away from the national mean, the FIs will substitute the statewide average ratio to calculate costs and determine whether a hospital qualifies for outlier payments. This policy was initially adopted in 1989 in regulation to address a concern that cost-to-charge ratios falling outside such a range were probably because of faulty data reporting. However, using the statewide average instead of a hospital's cost-to-charge ratio clearly increased outlier payments for particular hospitals where there were no errors.

As an example, assume that the cost-to-charge ratio of 0.2, used in the previous example, is more than three standard deviations below the national average. The FI would then substitute the statewide average cost-to-charge ratio (0.4, for example) when calculating outlier payments. So, for the initial $300,000 procedure, the estimated costs would rise to $120,000 ($300,000 X 0.4), and the outlier payment would be based on the $120,000, not $60,000 ($300,000 X 0.2).

Let me return to my previous examples for a moment and look at how these factors affected the cost-to-charge ratios of Hospitals A and B. The current cost-to-charge ratio we used to calculate payments for Hospital A is .339. However, we found by using more current data, Hospital A's actual relationship between cost and charges to be only .093. So basically, we are paying Hospital A at 3 times the actual rate of costs to charges. For our New Jersey hospital, Hospital B, we use a cost-to-charge of .325 to determine payments. However, our analysis found that Hospital B's actual relationship between costs and charges is .291. So clearly under the current outlier policy, we are grossly overpaying these hospital, and likely many more.


Upon discovering the abuse of the outlier policy, we quickly took corrective action. Last December, we instructed FIs to take action to help mitigate any potential vulnerability with outlier payments. First, FIs were instructed to identify hospitals that received outlier payments totaling more than 10 percent of their operating and capital DRG payments for discharges during FY 2002. FIs were also directed to identify other hospitals where outlier payments might be problematic. Later that month, we issued a second program memorandum that initiated a progressive compliance strategy to ensure that Medicare payments for outliers are appropriate. This was designed to ensure that the greatest level of scrutiny is placed on hospitals that appear, through data analysis, to present the greatest risk to the program. We also instructed FIs to identify those hospitals that: (1) have outlier payments of 80 percent or more of their operating and capital DRG payments for discharges during October and November 2002 (excluding outlier, indirect medical education, and disproportionate share payments); or (2) have estimated outlier payments greater than 20 percent of their operating and capital DRG payments for discharges during October and November 2002 (excluding outlier, indirect medical education, and disproportionate share payments) and an increase in average charges per case (calculated including all Medicare discharges) of 20 percent or more from 2000 to 2001 and 2001 to 2002. This comparison may be performed using either hospitals' cost reporting periods or Federal fiscal years.

For hospitals falling into the first category, FIs were instructed to perform comprehensive field audits for indirect medical education, graduate medical education, disproportionate share hospital payments, bad debts, organ acquisition costs, and any other pass through costs. FIs are also required to conduct medical reviews of a random sample of 20 hospital outpatient outlier records, and have the state Quality Improvement Organization (QIO) perform outlier reviews for inpatient stays. For hospitals falling into the second category, FIs were to perform uniform charge reviews, medical review and additional reviews by the state QIO. FIs were instructed to begin the audits by February 1, 2003, and to start all of the audits no later than July 31, 2003. The entire sample should be completed by July 31, 2004.


In addition to auditing hospitals with potentially problematic outlier payments, CMS recently issued a rule in the March 5 2003, Federal Register proposing revisions to the outlier payments regulations. In this proposed rule, we suggest changes to the methodology for determining payments for extraordinarily high-cost cases (cost outliers) made to Medicare-participating hospitals under the inpatient hospital prospective payment system. These proposed changes would be effective for discharges occurring on or after the date that we issue a final rule following this proposed rule and comment period.

Currently, we use the most recent settled cost report when determining cost-to-charge ratios for hospitals. The covered charges on bills submitted for payment during FY 2003 are converted to costs by applying a cost-to-charge ratio from cost reports that began in FY 2000 or, in some cases, FY 1999. These covered charges reflect all of a hospital's charge increases to date, in particular those that have occurred since FY 2000 and are not reflected in the FY 2000 cost?to?charge ratios. If the rate-of-charge increases since FY 2000 exceeds the rate of the hospital's cost increases during that time, the hospital's cost?to?charge ratio based on its FY 2000 cost report will be too high, and applying it to current charges will overestimate the hospital's costs per case during FY 2003. Overestimating costs may result in some cases qualifying for outlier payments that, in actuality, are not high cost cases. Overestimating costs will also result in higher outlier payments if a case does qualify for outliers.

Using our Medicare Provider Analysis and Review (MedPAR) file data from FY 1999 to FY 2001, we found 123 hospitals whose percentage of outlier payments relative to total DRG payments increased by at least 5 percentage points over that period, and whose case-mix adjusted charges increased at or above the 95th percentile rate of charge increase for all hospitals (46.63 percent) over the same period. We adjusted for case?mix because a hospital's average charges per case would be expected to change from one year to the next if the hospital were treating new or different types of cases. Because we use settled cost reports to compute hospitals' cost-to-charge ratios, the recent dramatic increases in charges for these hospitals are not reflected in their cost-to-charge ratios. For example, among these 123 hospitals, the mean rate of increase in charges was 70 percent. Meanwhile, cost?to-charge ratios for these hospitals, which were based on cost reports from prior periods, declined by only 2 percent.

Because a hospital has the ability to increase its outlier payments during this time lag through dramatic charge increases, in this proposed rule we are proposing to allow fiscal intermediaries to use more up-to-date data when determining the cost-to-charge ratio for each hospital. We are proposing to specify that fiscal intermediaries will use either the most recent settled or the most recent tentatively settled cost report, whichever is from the latest cost reporting period. Using cost-to-charge ratios from tentative settled cost reports would reduce the time lag for updating cost?to-charge ratios by a year or more.

However, even the more recent data calculated from the tentative settled cost reports would overestimate costs for hospitals that have continued to increase charges much faster than costs during the time between the tentative settled cost report period and the time when the claim is processed. In fact, there would still be a lag of one to two years during which a hospital's charges may still increase faster than costs. Therefore, we are proposing to add a new provision to the regulations that would allow CMS the authority to direct the fiscal intermediary to change the hospital's operating and capital cost-to-charge ratios to reflect the high charge increases evidenced by the later data. In addition, we are proposing to allow a hospital to contact its fiscal intermediary to request that its cost?to?charge ratios be changed if it presents substantial evidence that the ratios are inaccurate. Any such requests would have to be approved by the CMS Regional Office with jurisdiction over that fiscal intermediary.

Because of hospitals' ability to increase their charges to lower their cost-to-charge ratios in order to be assigned the statewide average, we are proposing to remove the current requirement in our regulations that specify that a fiscal intermediary will assign a hospital the statewide average cost?to?charge ratio when the hospital has a cost?to?charge ratio that falls below the floor.

After issuing our December 2002 Program Memorandum instructing fiscal intermediaries to identify all hospitals receiving the statewide average operating or capital cost?to?charge ratio because their cost-to-charge ratios fell below the floor of reasonable parameters, we identified 43 hospitals that were assigned the statewide average operating cost?to?charge ratio and 14 hospitals that were receiving the statewide average capital cost?to?charge ratio. Three hospitals were common to both lists. Prior to application of the statewide average cost-to-charge ratios, the average actual operating cost?to?charge ratio for the 43 hospitals was 0.164, and the average actual capital cost?to?charge ratio for the 14 listed hospitals was 0.008. In contrast, the statewide average operating cost?to?charge ratio for the 43 hospitals was 0.3425 and the statewide average capital cost?to?charge ratio for the 14 hospitals was 0.035. In the proposed rule, we suggest a revision that would give hospitals their actual cost?to?charge ratios, no matter how low their ratios fall.

However, we are proposing that statewide average cost-to-charge ratios would still apply in those instances in which a hospital's operating or capital cost?to?charge ratio exceeds the upper threshold. Cost?to?charge ratios above this range are more likely to be the result of faulty data reporting or entry, and should not be used to identify and pay for outliers. In addition, hospitals that have not yet filed their first Medicare cost reports with their fiscal intermediaries would still receive the statewide average cost?to?charge ratios.

The proposed rule would greatly reduce the opportunity for hospitals to manipulate the system to maximize outlier payments by updating cost-to-charge ratios using the most recent tentative settled cost reports and using actual rather than statewide average ratios for hospitals that have cost-to-charge ratios that are more than 3.0 standard deviations below the geometric mean cost?to?charge ratio. However, these two steps would not completely eliminate all such opportunity for aggressive gaming. A hospital would still be able to dramatically increase its charges by far above the rate of increase in costs during any given year. This possibility is of great concern, given the recent findings that some hospitals that have been able to receive large outlier payments by doing just that. Therefore, we are proposing to add a provision to our regulations to provide that outlier payments will become subject to adjustment when hospitals' cost reports are settled. Payments would be processed throughout the year using operating and capital cost-to-charge ratios based on the best information available at that time. When the cost report is settled, any reconciliation of outlier payments by fiscal intermediaries would be based on operating and capital cost?to?charge ratios calculated based on a ratio of costs to charges computed from the cost report and charge data determined at the time the cost report coinciding with the discharge is settled. The language we propose to add would allow outlier payments to be adjusted to account for the time value of the money during the time period it was inappropriately held by the hospital. This adjustment would also apply in cases where outlier payments were underpaid to the hospital. In those cases, the adjustment would result in additional payments to hospitals. Any adjustment would be based upon a widely available index to be established in advance by the Secretary, and would be applied from the midpoint of the cost reporting period to the date of reconciliation (or when additional payments are issued, in the case of underpayments). This adjustment to reflect the time value of a hospital's outlier payments would ensure that the outlier payment received by the hospital at the time its cost report is settled appropriately reflects the hospital's true costs of providing the care.


Outlier payments are a vital part of the Medicare payment system because they provide hospitals with insurance against extraordinarily high patient costs and ensure access to care for Medicare's beneficiaries. However, Congress intended that outlier payments would be made only in situations where the cost of care is extraordinarily high in relation to the average cost of treating comparable conditions or illnesses. Under the existing outlier methodology some hospitals' recent rates of charge increases greatly exceed their rates of cost increases. This disparity results in an overestimation of their current costs per case. In an effort to ensure that the outlier payments are used as they were intended, we issued a proposed rule that would prevent further gaming of the system by a few hospitals that obtain the majority of these payments. We are anxious to move forward with this effort as quickly as possible and are eager to receive comments on the proposed rule so that we can return the outlier policy to its original legislative purpose, which is to safeguard access to care for a broader number of Medicare beneficiaries and ensure appropriate use of taxpayer funds. I'm embarrassed that as a longtime hospital analyst I was not able to understand this problem and I am embarrassed for the Agency that we did not catch this sooner. Nevertheless, we firmly believe that this situation demands and immediate remedy. Thank you for allowing me to discuss this very important issue with you today. I look forward to answering your questions.

ATTACHMENT: Chart of Hospitals with Large Percentage of Outlier Payments (DOC) (36KB)

Documents in DOC format require Microsoft Word® or Microsoft Word Viewer®. If you experience problems with DOC documents, please download Word Viewer 2003®.

Last Revised: March 14, 2003