GAO-04-772, Medicare: Information Needed to Assess Adequacy of Rate-Setting Methodology for Payments for Hospital Outpatient Services


This is the accessible text file for GAO report number GAO-04-772 
entitled 'Medicare: Information Needed to Assess Adequacy of Rate-
Setting Methodology for Payments for Hospital Outpatient Services' 
which was released on September 17, 2004.

This text file was formatted by the U.S. Government Accountability 
Office (GAO) to be accessible to users with visual impairments, as part 
of a longer term project to improve GAO products' accessibility. Every 
attempt has been made to maintain the structural and data integrity of 
the original printed product. Accessibility features, such as text 
descriptions of tables, consecutively numbered footnotes placed at the 
end of the file, and the text of agency comment letters, are provided 
but may not exactly duplicate the presentation or format of the printed 
version. The portable document format (PDF) file is an exact electronic 
replica of the printed version. We welcome your feedback. Please E-mail 
your comments regarding the contents or accessibility features of this 
document to Webmaster@gao.gov.

This is a work of the U.S. government and is not subject to copyright 
protection in the United States. It may be reproduced and distributed 
in its entirety without further permission from GAO. Because this work 
may contain copyrighted images or other material, permission from the 
copyright holder may be necessary if you wish to reproduce this 
material separately.

Report to the Chairman, Subcommittee on Health, Committee on Ways and 
Means, House of Representatives: 

September 2004: 

MEDICARE: 

Information Needed to Assess Adequacy of Rate-Setting Methodology for 
Payments for Hospital Outpatient Services: 

GAO-04-772: 

GAO Highlights: 

Highlights of GAO-04-772, a report to the Chairman, Subcommittee on 
Health, Committee on Ways and Means, House of Representatives.

Why GAO Did This Study: 

Under the Medicare hospital outpatient prospective payment system 
(OPPS), hospitals receive a temporary additional payment for certain 
new drugs and devices while data on their costs are collected. In 
2003, these payments expired for the first time for many drugs and 
devices. To incorporate these items into OPPS, the Centers for Medicare 
& Medicaid Services (CMS) used its rate-setting methodology that 
calculates costs from charges reported on claims by hospitals. At that 
time, some drug and device industry representatives noted that payment 
rates for many of these items decreased and were concerned that 
hospitals may limit beneficiary access to these items if they could not 
recover their costs. GAO was asked to examine whether the OPPS rate-
setting methodology results in payment rates that uniformly reflect 
hospitals’ costs for providing drugs and devices, and other outpatient 
services, and if it does not, to identify specific factors of the 
methodology that are problematic.

What GAO Found: 

The rate-setting methodology used by CMS may result in OPPS payment 
rates for drugs, devices, and other services that do not uniformly 
reflect hospitals’ costs of providing those services. Two areas of the 
methodology are particularly problematic. The hospital claims for 
outpatient services that CMS uses to calculate hospitals’ costs and set 
payment rates may not be a representative sample of all hospital 
outpatient claims. For Medicare payment purposes, an outpatient service 
consists of a primary service and the additional services or items 
associated with the primary service, referred to as packaged services. 
CMS has excluded over 40 percent of multiple-service claims, claims 
that include more than one primary service along with packaged 
services, when calculating the cost of all OPPS services, including 
those with drugs and devices. It excludes these multiple-service claims 
because, when more than one primary service is reported on a claim, CMS 
cannot associate each packaged service with a specific primary service.
Therefore, the agency cannot calculate a total cost for each primary 
service on that claim, which it would use to set payment rates. The 
data CMS has available do not allow for a determination of whether 
excluding many multiple-service claims has an effect on OPPS payment 
rates. However, if the types or costs of services on excluded claims 
differ from those on included claims, the payment rates of some or all 
services may not uniformly reflect hospitals’ actual costs of providing 
those services. In addition, in calculating hospitals’ costs, CMS 
assumes that, in setting charges within a specific department, a 
hospital marks up the cost of each service by the same percentage. 
However, based on information from 113 hospitals, GAO found that not 
all hospitals use this methodology: charge-setting methodologies for 
drugs, devices, and other outpatient services vary greatly across 
hospitals and across departments within a hospital. CMS’s methodology 
does not recognize hospitals’ variability in setting charges, and 
therefore, the costs of services used to set payment rates may be 
under- or overestimated.

Number and Percentage of Hospitals that Reported Methods to Mark Up 
Drug and Device Charges, 2003: 

[See PDF for image]

[a] Percentage of total hospitals does not total 100 percent due to 
rounding. 

[End of table]

What GAO Recommends: 

GAO recommends that the Administrator of CMS collect data on excluded 
claims and analyze variation in hospital charge setting to determine 
if the OPPS payment rates uniformly reflect hospitals’ costs of 
providing outpatient services, and, if they do not, to make appropriate 
changes to the methodology. CMS stated that it will consider GAO’s 
recommendations.

www.gao.gov/cgi-bin/getrpt?GAO-04-772.

To view the full product, including the scope and methodology, click on 
the link above. For more information, contact A. Bruce Steinwald at 
(202) 512-7119.

[End of section]

Contents: 

Letter: 

Results in Brief: 

Background: 

Payment Rates Were Generally Lower for Separately Paid Drugs, but 
Cannot Be Evaluated for Packaged Drugs and Devices: 

No Type of Hospital Provided a Disproportionate Number of Services 
Associated with Certain Drugs and Devices: 

Payment Rates May Not Uniformly Reflect Hospitals' Costs: 

Conclusions: 

Recommendations for Executive Action: 

Agency and External Reviewer Comments and Our Evaluation: 

Appendixes: 

Appendix I: Scope and Methodology: 

Appendix II: Summary of Hospital Charge-Setting Methodologies: 

Appendix III: Comments from the Centers for Medicare & Medicaid 
Services: 

Appendix IV: GAO Contact and Staff Acknowledgments: 

GAO Contact: 

Acknowledgments: 

Tables Tables : 

Table 1: Payment Rates for Drug Administration APCs, 2002-2003: 

Table 2: Percentage of Medicare Outpatient Services by Type for All 
Hospitals and for Hospitals with Various Characteristics: 

Table 3: Number and Percentage of Hospitals that Reported Methods for 
Setting Base Charges for Clinic Visit Services, 2003: 

Table 4: Number and Percentage of Hospitals that Reported Methods to 
Mark Up Drug and Device Charges, 2003: 

Figures: 

Figure 1: Example of a Single-Service Claim with Packaged Services: 

Figure 2: Example of a Multiple-Service Claim with Packaged Services: 

Letter September 17, 2004: 

The Honorable Nancy L. Johnson: 
Chairman: 
Subcommittee on Health: 
Committee on Ways and Means: 
House of Representatives: 

Dear Chairman Johnson: 

Since 2000, hospitals have been paid fixed, predetermined amounts under 
a prospective payment system (PPS) for outpatient services delivered to 
Medicare beneficiaries. By paying hospitals under a PPS, Medicare seeks 
to encourage them to operate efficiently, as they retain the difference 
if their payments exceed their costs of providing necessary services. 
However, unlike most other Medicare PPSs, where each payment amount is 
designed to cover the combined costs of a large bundle of services, the 
outpatient prospective payment system (OPPS) is more like a fee 
schedule and pays a designated rate for each outpatient service 
provided to a beneficiary.

By law, the initial 2000 OPPS rates were based on hospitals' 1996 
median costs.[Footnote 1] During the development of OPPS, the 
anticipated use of 1996 data prompted concerns that the costs of new 
technology items, such as drugs, biologicals,[Footnote 2] and devices, 
first used after 1996 would not be represented in the 2000 payment 
rates and that hospitals might not provide the newest technology 
because of a perceived shortfall in payment. Accordingly, Congressional 
concerns were raised that beneficiaries might lose access to some of 
these items upon implementation of the payment system.

The Centers for Medicare & Medicaid Services (CMS),[Footnote 3] the 
agency that administers Medicare, sets OPPS payment rates by using 
charges hospitals report to CMS for the outpatient services they 
provide. The agency has used this methodology since setting the 2000 
rates. CMS converts each hospital's charge to that hospital's cost for 
each service using a specific adjustment for each of the hospital's 
departments. Under OPPS, an outpatient service consists of a primary 
service and its packaged services, the additional services or items 
associated with the primary service. For example, the surgical 
insertion of a pacemaker, a primary service, includes packaged services 
such as operating and recovery room services, anesthesia, and surgical 
and medical supplies, including the pacemaker. CMS combines the costs 
of the primary service and packaged services to calculate a total cost 
for that primary service. It assigns primary services to ambulatory 
payment classification (APC) groups and calculates a payment rate from 
the costs of the services in that group. An APC may consist of one 
primary service, but more often consists of two or more primary 
services with clinical and cost similarity. All primary services 
assigned to one APC are paid the same rate.

In response to concerns that the 1996 data that would be used to set 
the 2000 OPPS payment rates did not include cost data for new drugs and 
devices first used after 1996, in 1999, the Congress required that a 
payment be made for a temporary period, in addition to the OPPS amount, 
for certain drugs and devices used in the delivery of outpatient 
services.[Footnote 4] New drugs and devices are eligible to receive 
these temporary additional payments, known as pass-through payments, 
for 2 to 3 years depending on when their eligibility first began and 
when cost data become available to incorporate these items into OPPS as 
either a primary or packaged service. These temporary payments for 
pass-through drugs generally are equal to 95 percent of the average 
wholesale price (AWP),[Footnote 5] and the temporary payments for pass-
through devices are equal to CMS's calculation of the hospital's cost 
for the device.

In 2003, the first year for which pass-through eligibility expired for 
any drugs or devices, 236 drugs and 95 categories of devices[Footnote 
6] were incorporated into OPPS.[Footnote 7] Of these drugs and devices, 
CMS designated 115 of the drugs primary services and the remaining 121 
drugs and all devices packaged services. While those drugs that became 
primary services have assigned payment rates, the packaged drugs and 
devices do not. At the time CMS made the designations, some drug and 
device and hospital industry representatives noted that in basing the 
payment for these items on hospitals' costs, Medicare payments for 
many had declined significantly. The drug and device industry 
representatives were concerned that if hospitals could not recover 
their costs through OPPS payments, hospitals would not purchase these 
items--in essence, limiting beneficiary access to the products. Some 
hospital association representatives were concerned that certain types 
of hospitals may provide a higher number of services associated with 
drugs and devices, such as cancer center hospitals providing 
chemotherapy services or teaching hospitals performing cardiac 
procedures involving devices, and therefore may be disproportionately 
affected by payment rate decreases for these items. Furthermore, the 
decrease in payment rates for drugs and devices led to broader 
concerns about how CMS ensures that OPPS payment rates for all 
services reflect hospitals' costs.

You asked us to examine these issues. Specifically, we (1) describe how 
payment rates changed for those drugs and devices whose pass-through 
eligibility expired in 2003 and 2004, (2) determine whether a 
particular type or types of hospitals provide a disproportionate number 
of Medicare outpatient services associated with drugs and devices, and 
(3) examine whether the OPPS rate-setting methodology results in 
payment rates that uniformly reflect hospitals' costs for providing 
drugs and devices, as well as all other outpatient services, to 
beneficiaries, and if it does not, to identify specific factors of the 
methodology that are problematic.

To address these objectives, we analyzed 2003 and 2004[Footnote 8] OPPS 
payment rates and the 2003 and 2004 AWPs for former pass-through drugs 
that are primary services, which we refer to as separately paid drugs. 
The remaining drugs and all devices were packaged. Therefore, no 
identifiable 2003 or 2004 payment rate for these items exists and we 
could not analyze any payment rate change. We also analyzed the 
Medicare hospital claims, the bills hospitals submit to CMS for 
payment, that were used to set the 2003 OPPS rates.[Footnote 9] These 
claims were the latest data available at the time of our analysis, and 
we determined they were reliable for our purposes. From the claims 
data, we identified the outpatient services most often associated with 
drugs or devices.[Footnote 10] We determined whether any hospital type 
provided a disproportionate number of these outpatient services, such 
as hospitals with and without an outpatient cancer center or major 
teaching status, with major teaching hospitals defined as those having 
an intern/resident-to-bed ratio of 0.25 or more. We also analyzed 
hospitals by their urban/rural location and by their volume of 
outpatient services. We analyzed information from 113 hospitals on how 
they set their charges for drugs, devices, and other outpatient 
services. Of these hospitals, we interviewed officials from 5, received 
information from another 50 through association and industry 
representatives who gathered the information on our behalf, and 
received information from another 58 who were contacted by 7 state 
hospital associations in geographically diverse areas on our behalf. 
Because these 113 hospitals are not statistically representative of all 
hospitals, we cannot generalize our results to other hospitals. 
Finally, we spoke with officials at CMS, individual hospitals, hospital 
associations, drug and device manufacturers, and trade associations 
representing manufacturers of drugs and devices. We also spoke with 
consultants who advise hospitals on setting charges for their services. 
Our methodology is detailed in appendix I. We conducted our work from 
March 2003 through August 2004 in accordance with generally accepted 
government auditing standards.

Results in Brief: 

The OPPS payment rates of former pass-through, separately paid, drugs 
were generally lower than the pass-through payment rate, but the 
payment rates of former pass-through drugs and devices that were 
packaged cannot be evaluated, as these items are not assigned a 
distinct payment rate. The payment rates for the 115 of 236 former 
pass-through drugs that expired from pass-through eligibility in 2003 
and became separately paid drugs almost universally decreased from the 
pass-through payment rate of 95 percent of AWP. Because the remaining 
121 pass-through drugs and the devices in the 95 pass-through device 
categories were packaged and are not assigned to an APC, we cannot 
evaluate any payment rate changes for these items. In 2003, over 90 
percent, and in 2004, 100 percent, of former pass-through drugs that 
CMS designated as separately paid drugs had payment rates lower than 95 
percent of AWP. In both years, the payment rates were often 
considerably lower than AWP, but decreases varied substantially. For 
example, although the 2003 median drug payment rate was 55 percent of 
AWP, one drug had a 2003 payment rate about 7 percent of AWP, while 
another had a 2003 payment rate about 94 percent of AWP.

No type of hospital provided a disproportionate number of Medicare 
outpatient services associated with certain drugs and devices; in 2001, 
these outpatient services as a percentage of total Medicare outpatient 
services varied little among different hospital types. For example, 
chemotherapy administrations--the outpatient services most frequently 
associated with the use of drugs--accounted for an average of 1.8 
percent of all hospitals' total number of Medicare outpatient services. 
Chemotherapy administration services accounted for an average of 2.0 
percent of cancer center hospitals' and 1.8 percent of noncancer center 
hospitals' total Medicare outpatient services. Cardiac procedures--the 
outpatient services most frequently associated with the use of devices-
-accounted for an average of 0.4 percent of all hospitals' total number 
of Medicare outpatient services. Similarly, these cardiac procedures 
accounted for an average of 0.4 percent of both major teaching and all 
other hospitals' total Medicare outpatient services.

The OPPS rate-setting methodology used by CMS may result in APC payment 
rates for drugs, devices, and other outpatient services that do not 
uniformly reflect hospitals' costs of providing those services. Two 
areas of the methodology are particularly problematic. First, the 
claims that CMS uses to calculate hospitals' costs and set payment 
rates may not be a representative sample of hospital claims, as CMS has 
excluded over 40 percent of multiple-service claims, claims that 
include more than one primary service as well as packaged services, 
when calculating the cost of all OPPS services, including those with 
drugs and devices. It excludes these multiple-service claims because 
outpatient claims list all the services delivered during a visit and do 
not provide a link between primary and packaged services. Because CMS 
cannot associate each packaged service on the claim with one of the 
primary services listed on the claim, the agency cannot calculate a 
total cost for each primary service on that claim. The data CMS has 
available do not allow for the determination of whether excluding a 
sizable percentage of the multiple-service claims has an effect on OPPS 
payment rates. However, if the types or costs of services on excluded 
claims differ from the types or costs of services on included claims, 
the payment rates of some or all APCs will not uniformly reflect 
hospitals' costs of providing those services. Second, in calculating 
hospitals' costs, CMS assumes that, in setting charges within a 
specific department, a hospital marks up the cost of each service by 
the same percentage. However, many hospitals do not use this 
methodology; charge-setting methodologies for drugs, devices, and other 
outpatient services vary greatly both across hospitals and departments 
of a hospital. CMS's methodology does not recognize hospitals' 
variability in setting charges. This may lead to an under or 
overestimation of hospitals' costs for certain services. As these costs 
are used to set payment rates, payment rates may not uniformly reflect 
hospitals' costs.

We recommend that the Administrator of CMS gather the necessary data 
and perform an analysis of the types and costs of services on excluded 
multiple-service claims to determine if they are different from the 
types and costs of services on the claims it includes in setting OPPS 
rates. The Administrator should also analyze the effect that the 
variation in hospital charge-setting practices has on the rate-setting 
methodology. Finally, the Administrator should, in the context of the 
first two recommendations, analyze whether the OPPS rate-setting 
methodology results in payment rates that uniformly reflect hospitals' 
costs of the outpatient services they provide to Medicare 
beneficiaries, and, if it does not, make appropriate changes in that 
methodology. In commenting on a draft of this report, CMS stated that 
it has continued to review and refine its OPPS data collection and 
analysis. CMS stated that it is searching for ways to use more data 
from multiple-service claims, and it has made efforts in recent rate-
setting analyses to include data from more of these claims. We included 
a discussion of these changes in the draft report. In its comments, CMS 
stated that we should recognize that its rate-setting methodology that 
converts hospital charges to costs using a cost-to-charge ratio does so 
at the level of an individual hospital department. The draft report 
noted the fact that cost-to-charge ratios were generally calculated on 
a department-specific basis; however, we have revised the report to 
highlight that information throughout. CMS stated that it will consider 
our recommendations as it continues to assess and refine the rate-
setting methodology. Industry representatives who reviewed a copy of 
this draft generally agreed with the findings, conclusions, and 
recommendations.

Background: 

Medicare beneficiaries receive a wide range of services in hospital 
outpatient departments, such as emergency room and clinic visits, 
diagnostic services such as x-rays, and surgical procedures. To receive 
Medicare payment, hospitals report the services they provided to a 
beneficiary on a claim form they submit to CMS along with their charge 
for each service. For Medicare payment purposes, an outpatient service 
consists of a primary service and packaged services, the additional 
services or items associated with that primary service. CMS assigns 
each primary service to an APC, which may include other similar primary 
services, and pays the hospital at the designated APC payment rate, 
adjusted for variation in local wages. A hospital can receive multiple 
APC payments for a single outpatient visit if more than one primary 
service is delivered during that visit.

CMS Methodology for Determining APC Payment Rates: 

On outpatient claims, hospitals identify the primary services they 
provided using a Healthcare Common Procedure Coding System 
(HCPCS)[Footnote 11] code, while they identify packaged services by 
either specific HCPCS codes or revenue codes that represent general 
hospital departments or centers, such as "pharmacy," "observation 
room," or "medical social services." In addition to claims, hospitals 
submit annual cost reports to CMS that state their total charges and 
costs for the year and the individual hospital department charges and 
costs.

As a first step in calculating the OPPS payment rate for each APC, CMS 
obtains hospital charge data on each outpatient service from the latest 
available year of outpatient claims. It calculates each hospital's cost 
for each service by multiplying the charge by a cost-to-charge ratio 
that is computed from the hospital's most recent cost report, generally 
on an outpatient department-specific basis. In those instances when a 
cost-to-charge ratio does not exist for an outpatient department in a 
given hospital, CMS uses one from a related outpatient department or 
the hospital's overall cost-to-charge ratio for outpatient department 
services. The cost of each primary service is then combined with the 
costs of the related packaged services to calculate a total cost for 
that primary service. On single-service claims, claims with one primary 
service, CMS can associate packaged services with the primary service 
and calculate a total cost for the service (see fig. 1). However, in 
the case of multiple-service claims, claims with more than one primary 
service, packaged services and their costs listed on the claim cannot 
be associated with particular primary services, as the costs of a 
packaged service may be associated with one or a combination of primary 
services (see fig. 2). For this reason, CMS excluded all multiple-
service claims from rate setting prior to 2003. Beginning with the 2003 
payment rates, CMS identified several methods that allowed it to 
convert some multiple-service claims into single-service claims, and 
therefore include them in its rate-setting calculations.[Footnote 12]

Figure 1: Example of a Single-Service Claim with Packaged Services: 

[See PDF for image]

[End of figure]

Figure 2: Example of a Multiple-Service Claim with Packaged Services: 

[See PDF for image]

[End of figure]

After calculating the cost of each primary service assigned to an APC 
for each hospital claim, CMS arrays the costs for all claims and 
determines the median cost. To calculate the APC's weight relative to 
other APCs, CMS compares the median cost of each APC to the median cost 
of APC 0601, a mid-level clinic visit, which is assigned a relative 
weight of 1.00. For example, if the median cost of APC 0601 is $100 and 
the median cost of "APC A" is $50, CMS assigns APC A a relative weight 
of 0.50.

To obtain a payment rate for each APC, CMS multiplies the relative 
weight by a factor that converts it to a dollar amount. In addition, 
CMS annually reviews and revises the services assigned to a particular 
APC and uses the new APC assignments and the charges from the latest 
available outpatient hospital claims to recalibrate the relative 
weights, and therefore the payment rates.

Expiration of Drug and Device Pass-Through Eligibility: 

New drugs and devices are eligible to receive temporary pass-through 
payments for 2 to 3 years, depending on when each drug and device's 
eligibility began. January 1, 2003 was the first time that pass-through 
eligibility expired for any drugs or devices. Once pass-through 
eligibility for these items expires, CMS determines whether they will 
be considered a primary service and assigned to a separate APC or a 
packaged service and included with the primary services with which they 
are associated on a claim.

On January 1, 2003, 236 drugs and on January 1, 2004, 7 drugs expired 
from pass-through eligibility. For those drugs expiring in 2003, CMS 
designated any drug with a median cost exceeding $150 (115 drugs) as a 
primary service, and each was assigned to its own, separately paid APC. 
The remaining drugs (121 drugs), those with a median cost less than 
$150, were designated as packaged services, that is, their costs were 
included with the costs of the primary service they were associated 
with on the claim. CMS stated that many of these latter drugs were 
likely present on claims with a primary service of drug administration 
and were therefore packaged with the services assigned to the six drug 
administration APCs, that is, the three chemotherapy administration and 
three drug injection and infusion APCs.[Footnote 13] For these packaged 
drugs, although hospitals had previously received two payments, one for 
the administration of the drug or other primary service and an 
additional pass-through payment for the drug itself, when eligibility 
expires, hospitals receive only one payment for both the administration 
or other primary service and the packaged drug. In 2004, all 7 drugs 
for which pass-through eligibility expired were designated as primary 
services and assigned to their own, separately paid APCs.

On January 1, 2003, the devices in 95 device categories, and on January 
1, 2004, the devices in 2 device categories, expired from pass-through 
eligibility; in both years, the devices in all device categories were 
designated as packaged services and their costs were included with the 
costs of the primary service they were associated with on the claim. 
Although hospitals had previously received two payments, one for the 
procedure associated with the device and an additional pass-through 
payment for the device, hospitals then received only one payment for 
both the procedure and its associated device.

Payment Rates Were Generally Lower for Separately Paid Drugs, but 
Cannot Be Evaluated for Packaged Drugs and Devices: 

The OPPS payment rates of former pass-through, separately paid drugs 
were generally lower than the pass-through payment rate, but the 
payment rates of former pass-through drugs and devices that were 
packaged cannot be evaluated, as these items are not assigned a 
distinct payment rate. In 2003, the payment rates for the 115 of 236 
former pass-through drugs that were designated as separately paid drugs 
almost universally decreased from the pass-through payment rates. In 
2004, for all 7 former pass-through drugs were designated as separately 
paid drugs and the payment rates for all 7 decreased. In 2003, for the 
remaining 121 pass-through drugs and the devices in 95 pass-through 
device categories and, in 2004, the devices in 2 device categories, all 
of which were packaged, we cannot evaluate the payment rate changes 
because individual payment rates were not assigned for these items when 
they expired from pass-through eligibility.

Payment Rates Generally Decreased For Separately Paid, Former Pass-
Through Drugs: 

In 2003, about half of all drugs for which pass-through eligibility 
expired (115 of 236) were assigned to their own APC and paid 
separately. For these drugs, we determined that over 90 percent had 
payment rates lower than 95 percent of AWP, the pass-through payment 
rate; the median payment rate was 55 percent of AWP.[Footnote 14] 
Individual payment rates were often considerably lower than AWP, but 
decreases varied substantially. For example, 1 drug had a payment rate 
of about 7 percent of AWP, while another had a payment rate of about 94 
percent of AWP. However, 10 drugs had a payment rate of more than 100 
percent of AWP. In addition, payment as a percentage of AWP varied by 
drug source. The majority of the 113 separately paid drugs that we 
analyzed were sole-source (70 percent), followed by multi-source (19 
percent), and generic (10 percent).[Footnote 15] Generic drugs, which 
were paid the highest percentage of AWP of the three categories, had a 
median payment rate of 74 percent of AWP, multi-source drugs had a 
median of 56 percent of AWP, and sole-source drugs had a median of 53 
percent of AWP.

In 2004, all seven drugs for which pass-through eligibility expired 
were assigned to separate APCs. The individual payment rate of each 
drug was lower than the pass-through rate of 95 percent of AWP, with a 
median payment rate of 69 percent of AWP. All drugs were sole-source.

Although the decreases in payments for these drugs were often 
substantial and varied greatly across individual drugs, some level of 
decrease is expected when pass-through eligibility expires and payments 
become based on hospital costs instead of AWP, which often exceeds 
providers' acquisition costs. In 2001, we reported that certain drugs 
purchased by individual physicians were widely available at costs from 
66 to 87 percent of AWP.[Footnote 16]

Packaged Drugs and Devices Do Not Have Distinct Payment Rates: 

In 2003, the costs of 121 former pass-through drugs and devices in 95 
former pass-through device categories were packaged. Because CMS 
combines the costs of these items with the costs of the primary 
services with which they are associated on each claim, a specific 
payment rate for each of these drugs and devices does not exist. 
However, to indirectly assess the payment rates of packaged drugs and 
devices, we reviewed the payment rates of the APCs with which CMS 
stated they were likely packaged. CMS stated that, in 2003, former 
pass-through drug costs were most likely packaged with the six drug 
administration APCs. The payment rates for five of the six APCs 
decreased in 2003, when the costs of packaged former pass-through drugs 
were included, compared to 2002, when the costs of these drugs were not 
considered in the rate-setting calculations (see table 1). We are 
unable to determine why the costs of these APCs decreased because 
fluctuations in costs for any of the primary or packaged services in 
these APCs, in addition to the costs of the packaged drug, could have 
affected the payment rates. However, we would have expected that 
combining the costs of up to $150 of packaged former pass-through drugs 
with the costs of the primary services in these APCs would have 
increased the 2003 payment rates for more of these APCs as more than 
half of them are less than $150.

Table 1: Payment Rates for Drug Administration APCs, 2002-2003: 

APC: 0116; 
Description: Chemotherapy administration by other technique except 
infusion; 
2002: $46.32; 
2003: $40.43; 
Difference: -$5.89; 
Percent change: -13%.

APC: 0117; 
Description: Chemotherapy administration by infusion only; 
2002: $205.14; 
2003: $187.98; 
Difference: -$17.16; 
Percent change: -8%. 

APC: 0118; 
Description: Chemotherapy administration by both infusion and other 
technique; 
2002: $214.81; 
2003: $286.02; 
Difference: $71.21; 
Percent change: 33%. 

APC: 0120; 
Description: Infusion therapy except chemotherapy; 
2002: $157.80; 
2003: $113.70; 
Difference: -$44.10; 
Percent change: -28%. 

APC: 0352; 
Description: Level I injections; 
2002: $20.87; 
2003: $11.62; 
Difference: -$9.25; 
Percent change: -44%. 

APC: 0359; 
Description: Level II injections; 
2002: $91.63; 
2003: $59.12; 
Difference: -$32.51; 
Percent change: -35%. 

Source: GAO analysis of APC payment rates (67 Fed. Reg. 9,556, 9,569, 
9,572 (2002); 67 Fed. Reg. 66,815, 66,818 (2002)).

[End of table]

To indirectly assess the payment rates of the devices in the 95 device 
categories expiring from pass-through eligibility in 2003, we reviewed 
APCs for which CMS determined that device costs made up at least 1 
percent of the APC's total cost.[Footnote 17] We found that the payment 
rates of these APCs varied substantially between 2002 and 2003, when 
the former pass-through device costs likely were included. For example, 
the payment rate of APC 0688 (Revision/Removal of Neurostimulator Pulse 
Generator Receiver) decreased by 48 percent, while the payment rate of 
APC 0226 (Implantation of Drug Infusion Reservoir) increased by 94 
percent. However, we cannot attribute these fluctuations solely to the 
packaging of pass-through devices, because changes between 2002 and 
2003 in the costs of the primary services and other packaged services 
assigned to the APCs also could have affected the payment rates.

In 2004, the devices in two device categories expired from pass-through 
eligibility. The devices in one category were associated with services 
in one APC--APC 0674 (Prostate Cryoablation). The payment rate for this 
APC almost doubled. We were unable to examine the change in payment for 
the APC or APCs associated with the devices in the other expired pass-
through device category because CMS did not identify the APC or APCs 
into which the costs of the devices in this device category were 
packaged.

No Type of Hospital Provided a Disproportionate Number of Services 
Associated with Certain Drugs and Devices: 

No type of hospital provided a disproportionate number of Medicare 
outpatient services associated with certain drugs and devices, as these 
services, as a percentage of total Medicare outpatient services, varied 
little among hospitals with differences in characteristics such as the 
presence of an outpatient cancer center, teaching status, urban or 
rural location, or outpatient service volume.[Footnote 18]

In 2001, outpatient drugs were most often associated with APCs for 
chemotherapy administration services, and devices in pass-through 
device categories were most often associated with APCs for cardiac 
services.[Footnote 19] We found that chemotherapy administration and 
cardiac services composed only a small proportion of total Medicare 
outpatient services for all hospitals (see table 2). In addition, these 
proportions varied little among different types of hospitals.

Table 2: Percentage of Medicare Outpatient Services by Type for All 
Hospitals and for Hospitals with Various Characteristics: 

All hospitals; 
Number of hospitals: 4,034; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 1.8%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.4%. 

Cancer center hospitals; 
Number of hospitals: 555; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 2.0%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.5%. 

Noncancer center hospitals; 
Number of hospitals: 3,479; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 1.8%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.3%. 

Major teaching hospitals; 
Number of hospitals: 288; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 2.4%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.4%. 

Hospitals without major teaching hospital status; 
Number of hospitals: 3,746; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 1.7%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.4%. 

Urban hospitals; 
Number of hospitals: 2,493; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 1.7%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.5%. 

Rural hospitals; 
Number of hospitals: 1,541; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 2.3%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.2%. 

Small volume hospitals; 
Number of hospitals: 1,258; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 1.0%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.1%. 

Medium volume hospitals; 
Number of hospitals: 1,840; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 1.3%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.3%. 

Large volume hospitals; 
Number of hospitals: 936; 
Chemotherapy administration services as a percent of total Medicare 
outpatient services: 2.2%; 
Cardiac services as a percent of total Medicare outpatient services: 
0.5%. 

Source: GAO analysis of CMS data.

Notes: We used hospital outpatient claims from April 1, 2001 through 
March 31, 2002, the claims CMS used to set the 2003 OPPS rates, applied 
to hospital categories defined in 2003. We defined "cancer center 
hospitals" as those hospitals that were members of the Association of 
Community Cancer Centers as of February 28, 2003, the latest data 
available when we performed this analysis. We defined a major teaching 
hospital as a hospital with an intern/resident-to-bed ratio of 0.25 or 
more and a hospital without major teaching hospital status as a ratio 
of less than 0.25. We defined the urban or rural location of a hospital 
using Medicare's classification of that hospital under OPPS. We defined 
volume based on the number of outpatient services a hospital provided. 
Small volume hospitals were those with fewer than 11,000 services, 
medium volume hospitals were those with at least 11,000 services but 
fewer than 43,000 services, and large volume hospitals were those with 
at least 43,000 services.

[End of table]

Payment Rates May Not Uniformly Reflect Hospitals' Costs: 

The OPPS rate-setting methodology used by CMS may result in APC payment 
rates for drugs, devices, and other outpatient services that do not 
uniformly reflect hospitals' costs. Two areas of CMS's methodology are 
particularly problematic. First, the claims that CMS uses to calculate 
hospitals' costs and set payment rates may not be a representative 
sample of hospital claims, as CMS excluded many multiple-service claims 
when calculating the cost of OPPS services, including those with drugs 
and devices. The data CMS has available do not allow for the 
determination of whether excluding many multiple-service claims has an 
effect on OPPS payment rates. However, if the types or costs of 
services on excluded claims differ from the types or costs of services 
on included claims, the payment rates of some or all APCs may not 
uniformly reflect hospitals' costs of providing those services. Second, 
when calculating hospitals' costs, CMS assumes that, in setting charges 
within a specific department, a hospital marks up the cost of each 
service by the same percentage. However, not all hospitals use this 
methodology, and charge-setting methodologies for drugs, devices, and 
other outpatient services vary greatly across hospitals and across 
departments within a hospital. CMS's methodology does not recognize 
hospitals' variability in setting charges, and, therefore, the costs of 
services used to set payment rates may be under or overestimated.

CMS May Not Be Using a Representative Sample of Claims to Set Payment 
Rates: 

The claims CMS uses to calculate hospitals' costs and set payment rates 
may not be a representative sample of hospital claims. When calculating 
the cost of all OPPS services, including drugs and devices, to set 
payment rates, CMS excluded over 40 percent of all multiple-service 
claims because CMS could not associate particular packaged services 
with a specific primary service on these claims.[Footnote 20] Drug and 
device industry representatives we spoke with raised concerns that 
certain drugs and devices are often billed on multiple-service claims 
that are largely excluded from rate setting. For example, they stated 
that chemotherapy administration and the drugs themselves are typically 
billed on a 30-day cycle; therefore, one claim likely includes 
chemotherapy administration and other primary and packaged services and 
is likely excluded from CMS's rate-setting calculations.[Footnote 21] 
Device industry representatives we spoke with also asserted that 
multiple-service claims represent more complex, and therefore, 
potentially costlier, outpatient visits and excluding them from the 
rate-setting calculations underestimates the actual cost of a service. 
Because of the structure of the outpatient claim, the data CMS has 
available do not allow for the comparison of single-service claims and 
multiple-service claims to determine whether excluding many multiple-
service claims has an effect on OPPS payment rates. It is possible that 
excluding many multiple-service claims has little or no effect on OPPS 
payment rates. However, if the types or costs of services on excluded 
claims differ from the types or costs of services on included claims, 
the payment rates of some or all APCs may not uniformly reflect 
hospitals' costs of performing these services.

Rate-Setting Methodology Does Not Account for Variation in Hospital 
Charge-Setting Practices: 

The costs of drugs, devices, and other outpatient services that CMS 
calculates from hospital charges and uses to set payment rates may not 
uniformly approximate hospitals' costs. CMS multiplies charges by 
hospital-specific cost-to-charge ratios to calculate hospitals' costs, 
which decreases the charges by a constant percentage. This methodology 
is based on the assumption that each hospital marks up its costs by a 
uniform percentage within each department to set each service's charge. 
However, we found that not all hospitals use this methodology to 
establish their charges, and that drug, device, and general charge-
setting methodologies vary greatly among hospitals and even among 
departments within the same hospitals.

We received information from 113 hospitals, although not all hospitals 
responded to each question. Of the 92 hospitals responding, 40 reported 
that they mark up all drug costs by a uniform percentage to establish 
charges, but 33 reported that they mark up low-cost drugs by a higher 
percentage and high-cost drugs by a lower percentage. Of 85 hospitals 
responding, 39 reported that they mark up all device costs using a 
uniform percentage, but 39 reported that they mark up low-cost devices 
using a higher percentage and high-cost devices using a lower 
percentage. In addition, 19 hospitals reported using other methods to 
set drug charges and 7 reported doing so for devices, such as a lower 
percentage markup for low-cost drugs and devices than for high-cost 
drugs and devices. (See appendix II for a more detailed description of 
hospital charge-setting methodologies.)

Because CMS uses the same rate-setting methodology to determine drug 
and device payment rates as it uses for all other OPPS services, we 
also asked hospitals about more general charge-setting practices and 
found that they varied as well. To set base charges for clinic visits, 
hospitals reported using a wide variety of prices and methods, 
including cost, market comparisons, and the rates Medicare pays for 
outpatient services as well as payment rates for other benefit 
categories. To mark up clinic visits, 29 of the 45 hospitals responding 
used a uniform percentage increase; the remaining 16 hospitals reported 
using a variety of other methods, including using a higher percentage 
markup for low-cost visits than for high-cost visits.

In addition to variation in charge-setting methodologies among 
hospitals, variation also can exist within an individual hospital. 
Hospital consultants told us that a single item can be assigned 
different charges if it is provided through more than one department 
within the same hospital.

All 58 hospitals responding reported that they update their charges for 
inflation; 40 reported they did so annually, 12 did so at other times, 
and 6 did so both annually and at other times. Of the 58 hospitals that 
reported updating their charges for inflation, 25 reported that they 
apply a uniform, across-the-board percentage increase to all their 
charges, and 4 hospitals reported using both a uniform percentage and 
another type of increase. The remaining 29 hospitals reported using 
another method, such as applying an increase only to selected 
departments within the hospital. In addition, 33 of the 57 hospitals 
reported that they excluded some charges from these updates. The type 
of charges they excluded varied widely, but included drug and 
laboratory charges. The variation in methods hospitals use to update 
their charges reduces the likelihood that charges will uniformly 
reflect costs.

Conclusions: 

CMS's rate-setting methodology may result in OPPS payment rates that do 
not uniformly reflect hospitals' costs of providing services. We 
identified two areas of this methodology that are of particular concern 
because not enough data are currently available to assess their impact. 
First, CMS excludes many multiple-service claims from its rate-setting 
calculations. To the extent that the types and costs of services on 
these claims are different from services on the claims included in the 
analysis, OPPS payment rates may not reflect hospitals' costs. The 
current structure of the outpatient claims does not allow for an 
analysis to determine the effect of these exclusions. Second, in its 
rate-setting calculations, CMS assumes that each hospital uses a 
uniform markup percentage to set its charges within each department, 
although we found that hospitals use a variety of markup methodologies. 
Therefore, CMS's application of a constant cost-to-charge ratio may not 
result in an accurate calculation.

Recommendations for Executive Action: 

We recommend that the Administrator of CMS take the following three 
actions. First, the Administrator should gather the necessary data and 
perform an analysis that compares the types and costs of services on 
single-service claims to those on multiple-service claims. Second, the 
Administrator should analyze the effect that the variation in hospital 
charge-setting practices has on the OPPS rate-setting methodology. 
Third, the Administrator should, in the context of the first two 
recommendations, analyze whether the OPPS rate-setting methodology 
results in payment rates that uniformly reflect hospitals' costs of the 
outpatient services they provide to Medicare beneficiaries, and, if it 
does not, make appropriate changes in that methodology.

Agency and External Reviewer Comments and Our Evaluation: 

We received written comments on a draft of this report from CMS (see 
app. III). We also received oral comments from external reviewers 
representing seven industry organizations. They included the Advanced 
Medical Technology Association (AdvaMed), which represents 
manufacturers of medical devices, diagnostic products, and medical 
information systems; the American Hospital Association (AHA); the 
Association of American Medical Colleges (AAMC), which represents 
medical schools and teaching hospitals; the Association of Community 
Cancer Centers (ACCC); the Biotechnology Industry Organization (BIO), 
which represents biotechnology companies and academic institutions 
conducting biotechnology research; the Federation of American Hospitals 
(FAH), which represents for-profit hospitals; and the Pharmaceutical 
Research and Manufacturers of America (PhRMA).

CMS Comments and Our Evaluation: 

In commenting on a draft of this report, CMS stated that it has 
continued to review and refine its OPPS data collection and analysis. 
In responding to our recommendation that CMS gather the necessary data 
and perform an analysis comparing the types and costs of services on 
single-service claims to those on multiple-service claims, CMS stated 
that it is searching for ways to use more data from multiple-service 
claims, and it has made efforts in recent rate-setting analyses to 
include data from more of these claims. We noted these efforts in the 
draft report. CMS noted that there are continuing challenges and costs, 
to both the federal government and hospitals, to expanding its efforts 
in this area. In its comments, CMS suggested that an analysis could be 
done using an algorithm to allocate charges among multiple-service 
claims, but noted that such an approach could create further 
distortions in the relative weights. Our recommendation to CMS, 
however, is that the agency should gather additional data on the 
relative costs of services on single and multiple-service claims, 
rather than continuing to analyze existing data.

In response to our recommendation that CMS analyze the effect of 
hospital charge-setting practices on the OPPS rate-setting methodology, 
CMS stated that we should recognize that its rate-setting methodology 
that converts hospital charges to costs using a cost-to-charge ratio 
does so at the level of an individual hospital department. The draft 
report noted the fact that CMS generally calculates cost-to-charge 
ratios on a department-specific basis; however, we have revised the 
report to highlight that information throughout. CMS also said that the 
application of cost-to-charge ratios to charges of a hospital has long 
been the recognized method of establishing reasonable costs for 
hospital services and was an important component of the cost-based 
reimbursement system that was used by Medicare to pay for hospital 
outpatient services before OPPS was implemented. While we agree that it 
was an important component of the prior payment system, we believe the 
implementation of the current payment system has changed the relevance 
of applying cost-to-charge ratios to determine hospitals' costs. OPPS, 
rather than reimbursing individual hospitals on the basis of their 
costs of providing outpatient services, uses costs from individual 
hospitals to construct a prospective payment system that sets rates for 
individual services that apply to all hospitals. Finally, CMS stated 
that the Medicare Prescription Drug, Improvement, and Modernization Act 
of 2003 specified that cost-to-charge ratios would be used to set 
payment amounts for brachytherapy sources; however, a discussion of 
brachytherapy payment is outside of the scope of this report.

In response to our recommendation that CMS analyze whether the OPPS 
rate-setting methodology results in payment rates that uniformly 
reflect hospitals' costs of the services they provide to Medicare 
beneficiaries and make any appropriate changes in the methodology, CMS 
stated that it will consider our recommendations as it continues to 
assess and refine the rate-setting methodology. CMS said that it 
believes it has made great strides on this issue and is continuing to 
pursue the analyses necessary to create means by which all claims can 
be used to set the OPPS relative payment weights and rates.

CMS also made technical comments, which we incorporated where 
appropriate.

Industry Comments and Our Evaluation: 

Industry representatives generally agreed with the findings, 
conclusions, and recommendations in the draft report. Comments on 
specific portions of the draft report centered on three areas: payment 
rates of former pass-through drugs and devices, provision of services 
associated with drugs and devices, and CMS's rate-setting methodology.

Several industry representatives commented on our analysis of Medicare 
payment for former pass-through drugs and devices. AHA stated that 
although when drugs have expired from pass-through status their payment 
rates may have decreased, they are now more consistent, relative to 
costs, with the payment rates for other OPPS services. PhRMA agreed 
with our finding that the payment rates for former pass-through drugs 
and devices that are packaged cannot be evaluated and suggested that we 
recommend that CMS specifically address this problem.

Industry representatives commented on our analysis of the provision of 
services associated with drugs and devices among different types of 
hospitals. ACCC agreed with the percentages of Medicare outpatient 
services related to chemotherapy administration and cardiac services in 
the draft report; however, it stated that it believed that these 
percentages demonstrated that large hospitals provided a 
disproportionate share of chemotherapy administration. ACCC and AAMC 
stated that these percentages also demonstrated that major teaching 
hospitals provided a disproportionate share of chemotherapy 
administration services. In addition, both groups suggested that we 
perform other analyses by type of hospital, such as the proportion of 
total payments, proportion of total services excluding clinic services, 
or absolute number of services for which chemotherapy administration 
and cardiac services accounted.

Many of the reviewers addressed our finding that CMS's rate-setting 
methodology may result in OPPS payment rates that do not uniformly 
reflect hospitals' costs. Representatives from AAMC, ACCC, AdvaMed, 
BIO, and PhRMA agreed with our conclusion that CMS may not be using a 
representative sample of claims to set payment rates and that CMS's 
rate-setting methodology does not account for variation in hospital 
charge-setting practices. Several of these representatives suggested we 
analyze and discuss other factors that could further skew CMS's 
calculation of hospital costs, such as its use of incorrect or 
incomplete claims in rate setting.

Regarding the suggestion that we specifically recommend that CMS 
address the issue that the payment rates for former pass-through drugs 
that are packaged and former pass-through devices cannot be evaluated, 
we believe that our more general recommendation allows the agency the 
flexibility to determine the most appropriate analyses for examining 
the rate-setting methodology.

With respect to the comment that the percentages of Medicare outpatient 
services accounted for by chemotherapy administration demonstrate that 
certain types of hospitals provide a disproportionate share of these 
services, we disagree. As noted in the draft report, we found that 
these percentages differ by type of hospital, but the differences are 
not substantial, as all types of hospitals provided a relatively small 
proportion of these services. No type of hospital provided a 
disproportionately large number of these services. We analyzed the 
proportion of services, rather than payments as industry 
representatives suggested, because we believe that is the better 
analysis for determining whether a certain type of hospital provides a 
disproportionate share of these services. We did not analyze the 
proportion of total services except for clinic services or the absolute 
number these services made up, as we do not believe such an analysis 
would accurately and comparably reflect potential differences between 
hospitals for all outpatient services they perform.

The industry representatives also made technical comments, which we 
incorporated where appropriate.

We are sending a copy of this report to the Administrator of CMS. The 
report is available at no charge on GAO's Web site at [Hyperlink, 
http://www.gao.gov]. We will also make copies available to others on 
request.

If you or your staff have any questions, please call me at (202) 512-
7119. Another contact and key contributors to this report appear in 
appendix IV.

Sincerely yours,

Signed by: 

A. Bruce Steinwald, 
Director, Health Care--Economic and Payment Issues: 

[End of section]

Appendixes: 

Appendix I: Scope and Methodology: 

We analyzed Medicare claims data used by the Centers for Medicare & 
Medicaid Services (CMS) to set the 2003 outpatient prospective payment 
system (OPPS) payment rates. In addition, we analyzed drug average 
wholesale prices (AWPs), drug sources (sole-source, multi-source, or 
generic), and OPPS payment rates obtained from CMS. We interviewed 
officials at CMS and representatives from the American Hospital 
Association, Association of American Medical Colleges, Association of 
Community Cancer Centers (ACCC), Federation of American Hospitals, 
Greater New York Hospital Association, as well as from one large 
hospital system, one large hospital alliance, and five individual 
hospitals. In addition, we spoke with representatives from the Advanced 
Medical Technology Association, Biotechnology Industry Organization, 
California Healthcare Institute, Pharmaceutical Research and 
Manufacturers of America, as well as from seven drug manufacturers and 
three device manufacturers. We also spoke with consultants that advise 
hospitals on setting their charges.

To compare payment for drugs to previous pass-through payments, we 
relied on information provided by CMS on drug sources and 2003 and 2004 
drug payment rates, and on CMS's calculations of the AWPs for these 
drugs, which we supplemented with our own calculations. From CMS, we 
obtained the drug source and the payment rate for the 115 drugs and the 
7 drugs whose pass-through eligibility expired as of January 1, 2003 
and January 1, 2004, respectively, that were assigned to separate 
ambulatory payment classification (APC) groups. We used Medicare's 
January 2003 and January 2004 Single Drug Pricer files to determine the 
2003 and 2004 AWPs, respectively, for most of the drugs. For the 37 
drugs that were not included in the 2003 Single Drug Pricer file, we 
used the 2002 Drug Topics Red Book, published by Thomson Medical 
Economics, to calculate their AWPs. For the 2 drugs that were not in 
the 2004 Single Drug Pricer file, we used the 2003 Drug Topics Red 
Book, published by Thomson PDR, to calculate their AWPs. We calculated 
payment rates as a percentage of AWP for all drugs in 2003 and 2004. 
From our 2003 analysis, we excluded 1 multi-source drug for which we 
calculated an AWP from the 2002 Drug Topics Red Book that was 
inconsistent with the 2002 AWP CMS provided to us and another multi-
source drug with an AWP of $0.34, but a payment rate of almost 29,000 
percent of that amount.

To determine whether a particular type or types of hospitals provide a 
disproportionate number of outpatient services associated with drugs 
and devices, we used the outpatient claims file that CMS used to 
calculate the 2003 OPPS payment rates.[Footnote 22] To perform our own 
data reliability check of this file, we examined selected services to 
determine the reasonableness of their frequency in the data set, given 
the population of the beneficiaries receiving services and the setting 
in which they are delivered. We determined the data were reasonable for 
our purposes.

Using the claims, we determined which outpatient services were most 
often associated with drugs and devices and found that drugs were most 
often associated with chemotherapy administration services and devices 
were most often associated with cardiac services. Then, also using the 
claims, we compared proportions of chemotherapy administration and 
cardiac services for all hospitals, as well as for cancer center and 
noncancer center hospitals, major teaching and other hospitals, urban 
and rural hospitals, and hospitals with different outpatient service 
volumes.[Footnote 23] We included only those hospitals identified in 
CMS's 2003 OPPS impact file, a data file CMS constructs to analyze 
projected effects of policy changes on various hospital groups, such as 
urban and rural hospitals. We excluded hospitals with fewer than 1,100 
total outpatient services, or approximately 3 outpatient services per 
day, as we believe such hospitals are not representative of most 
hospitals with outpatient departments. We defined cancer center 
hospitals as those hospitals that were members of ACCC as of February 
28, 2003, the latest data available when we performed this analysis. We 
obtained the membership list from the ACCC. Using the September 2002 
Medicare Provider of Services file and information obtained directly 
from the ACCC, we determined the Medicare provider numbers of ACCC 
members to identify claims billed by these hospitals. We defined major 
teaching hospitals as those hospitals having an intern/resident-to-bed 
ratio of 0.25 or more. We defined the urban or rural location of a 
hospital based on the urban/rural location indicator in the Medicare 
hospital OPPS impact file from calendar year 2003. We defined volume 
based on the number of services a hospital provided, also as indicated 
in the impact file. Small volume hospitals were those with fewer than 
11,000 services, medium volume hospitals were those with at least 
11,000 services but fewer than 43,000 services, and large volume 
hospitals were those with at least 43,000 services.

We interviewed representatives from hospitals, hospital associations, 
and drug and device manufacturers and the associations that represent 
them to obtain information about hospital charging practices. We 
received information on charge-setting practices from 5 hospitals whose 
officials we interviewed. We indirectly received information from 50 
other hospitals through association and industry representatives with 
whom we spoke. Finally, we contacted seven state hospital associations 
in geographically diverse areas not well represented in our previous 
sample to identify their members' charging practices. Some hospitals 
responded directly to us and others responded to their state 
association, which forwarded the responses to us. We received responses 
from 58 hospitals. The 113 hospitals from which we received information 
are not a statistically representative sample of all hospitals.

We conducted our work from March 2003 through August 2004 in accordance 
with generally accepted government auditing standards.

[End of section]

Appendix II: Summary of Hospital Charge-Setting Methodologies: 

We received information from 113 hospitals, although not all hospitals 
responded to each question. Hospitals reported using a variety of 
methods to set the base charges for their clinic visit services (see 
table 3). To set the base charges for drugs, 25 of 57 hospitals 
responding reported that they used acquisition cost, 30 used the drug's 
average wholesale price (AWP), and 2 used a combination of acquisition 
cost and AWP. To set the base charges for devices, 55 of 57 hospitals 
responding reported that they used acquisition cost. After setting base 
charges, 29 of 45 hospitals responding reported that they marked up all 
of their clinic visit services by the same percentage increase, 
although they reported using a variety of other methods as well. To 
mark up base charges for drugs and devices, most hospitals responding 
used either the same percentage for all drugs and for all devices, or 
used a graduated percentage markup, marking up low-cost items by a 
higher percentage (see table 4).

Table 3: Number and Percentage of Hospitals that Reported Methods for 
Setting Base Charges for Clinic Visit Services, 2003: 

Cost of visit; 
Number: 13; 
Percentage: 29.

Comparable charges in market; 
Number: 10; 
Percentage: 22.

Medicare physician fee schedule; 
Number: 8; 
Percentage: 18.

Cost of visit and comparable charges in market; 
Number: 4; 
Percentage: 9.

Unspecified Medicare payment; 
Number: 3; 
Percentage: 7.

Outpatient prospective payment system amount with an adjustment; 
Number: 1; 
Percentage: 2.

Cost of visit and Medicare physician fee schedule; 
Number: 1; 
Percentage: 2.

Unspecified Medicare payment and comparable charges; 
Number: 1; 
Percentage: 2.

Other; 
Number: 4; 
Percentage: 9.

Source: GAO.

[End of table] 

Table 4: Number and Percentage of Hospitals that Reported Methods to 
Mark Up Drug and Device Charges, 2003: 

Same percentage for all items; 
Drugs: Number: 40; 
Drugs: Percentage: 43; 
Devices: Number: 39; 
Devices: Percentage[A]: 46.

Graduated percentage, higher for low-cost items; 
Drugs: Number: 33; 
Drugs: Percentage: 36; 
Devices: Number: 39; 
Devices: Percentage[A]: 46.

Graduated percentage, lower for low-cost items; 
Drugs: Number: 6; 
Drugs: Percentage: 7; 
Devices: Number: 4; 
Devices: Percentage[A]: 5.

Other; 
Drugs: Number: 13; 
Drugs: Percentage: 14; 
Devices: Number: 3; 
Devices: Percentage[A]: 4.

Source: GAO.

[A] Percentage of total hospitals responding does not total 100 percent 
due to rounding.

[End of table]

In addition, 24 of the 57 hospitals responding reported that they 
include nonproduct costs as a portion of their drug charges, and 25 of 
57 responding reported that they include nonproduct costs as a portion 
of their device charges. The most common nonproduct costs included were 
administrative and overhead costs. Of the 24 including nonproduct costs 
in drug charges, 12 reported that they do so by adding an additional 
percentage of the drug acquisition cost to the drug charge. Of the 25 
including nonproduct costs in device charges, 16 reported that they do 
so by adding an additional percentage of the device acquisition cost to 
the device charge. However, the amount of the nonproduct costs as a 
percentage of the charges varied widely among hospitals. Of the 24 
hospitals including nonproduct costs in drug charges, 16 reported that 
the amount varied by the route of administration for the drug, such as 
intravenous or intramuscular administration.

Of the 58 hospitals responding, all reported that they update their 
charges for inflation; 40 reported they did so annually, 12 did so at 
other times, and 6 did so both annually and at other times. While many 
used a standard across-the-board percentage increase to update their 
charges, the majority used other methods. In addition, 33 of the 57 
hospitals responding reported that they exclude certain charges from 
these updates. The types of services whose charges they excluded, such 
as drug, laboratory, and room charges, varied widely. Finally, 49 of 58 
hospitals responding reported that they periodically review all their 
charges.

[End of section]

Appendix III: Comments from the Centers for Medicare & Medicaid 
Services: 

DEPARTMENT OF HEALTH & HUMAN SERVICES: 
Centers for Medicare & Medicaid Services:
Administrator: 
Washington, DC 20201:

DATE: AUG 19 2004:

TO: A. Bruce Steinwald:
Director, Health Care-Economic and Payment Issues:

FROM: Mark B. McClellan, M.D., Ph.D. 
Administrator:

SUBJECT: Government Accountability Office (GAO) Draft Report: MEDICARE: 
Information Needed to Assess Adequacy of Rate-Setting Methodology for 
Payments, for Hospital Outpatient Services (GAO-04-772):

Thank you for the opportunity to review and comment on the draft 
report. We appreciate the interest of the House of Representatives 
Subcommittee and the efforts of GAO in the methodology used to set 
rates under the Medicare Outpatient Prospective Payment System (OPPS).

As required by the Balanced Budget Act of 1997, the Centers for 
Medicare & Medicaid Services (CMS) implemented in August 2000 a new 
prospective payment system to pay for most types of Medicare services 
provided in hospital outpatient departments. The new system is based on 
ambulatory payment classifications (APCs) that are groupings of 
clinically similar services that require similar resources. The 
relative weights and payments assigned to the APCs for the first 
several years of the OPPS were based on an analysis of hospital 
outpatient claims under the prior reasonable cost-based methodology.

Beginning with the update for calendar year 2003, the relative weights 
for most APCs have been based on hospital claims submitted and paid 
under the new system. The CMS establishes a median cost for each APC by 
applying a cost-to-charge ratio derived from hospital cost reports to 
the charges on the claims. The median cost for each APC is then 
compared to the median cost for the mid-level clinic visit, one of the 
most frequently performed OPPS services, in order to establish the 
relative weight for each APC. Excluded from payment under the OPPS, 
and, therefore, excluded from consideration in setting OPPS rates are 
such services as services provided under Medicare's physician fee 
schedule, and clinical laboratory services paid under the clinical 
laboratory fee schedule.

Since the implementation of the OPPS, CMS has continued to review and 
refine its data collection and analysis. For example, in the first 
year, CMS did not have the capability to analyze multiple procedure 
claims. As a result, 40 percent of outpatient claims were excluded 
from the weight-setting process. In the past 2 years, CMS has made 
significant improvements in the use of multiple procedure claims.

Meanwhile, payment rates for certain items, such as drugs (and for 
2003, devices of brachytherapy) that have been priced using claims data 
and paid separately were calculated from 100 percent of the line items 
in usable claims for those products. As a result, there was no 
packaging for those products and use of multiple claims for such items 
is not a problem. Overall, for the 2004 update, CMS used at least some 
portion of the 82 percent of claims for services that fell under the 
OPPS.

It is in this context that we address the recommendations in the GAO 
report.

Attachment:

Centers for Medicare & Medicaid Services' Comments to the GAO 
Draft Report: MEDICARE. Information Needed to Assess Adequacy of Rate-
Setting Methodology for Payments for Hospital Outpatient Services (GAO-
04-772):

GAO Recommendation:

That CMS gather the necessary data and perform an analysis that 
compares the types and costs of services on single-service claims to 
those on multiple-service claims.

CMS Response:

The CMS is searching for ways to use the data from all valid claims for 
OPPS services, including multiple procedure claims. We have moved to 
line item date of service segregation to attempt to associate revenue 
code charges submitted without Healthcare Common Procedure Coding 
System (HCPCS) codes with charges for the related, separately paid 
HCPCS codes. We have also established criteria for bypassing charges 
for separately paid HCPCS codes in determining whether a claim may be 
treated as a single procedure claim. This has allowed us to use 
significantly more claims data for 2004 OPPS than for 2002 or 2003 
OPPS. We expect to further refine the methodology for 2005 OPPS. We are 
also exploring ways of allocating line item charges among multiple 
separately paid HCPCS codes on a claim. However, we note that any 
methodology we use will allocate such charges generally but not 
specifically.

When we have spoken to hospitals about this issue, they describe the 
difficulties in allocating line item charges for packaged services 
among the separately paid services, and that is why they bill charges 
without HCPCS codes. For example, when there is a charge for 
administration of an anesthetic during a surgery in which two 
procedures are performed, the hospital cannot allocate the charge for 
the anesthetic between the two procedures. Similarly, when a patient 
incurs charges in an operating room and recovery room for two 
procedures that were performed in the same operative session, hospitals 
are unable to split the charges between the two separately paid 
procedures as would be necessary to secure an accurate allocation of 
the total charge for the use of those spaces.

An allocation of charges to use in multiple procedure claims would 
require use of assumptions in an algorithm, rather than an actual split 
based on the amount of the drug or the amount of operating room or 
recovery room time that relates to each procedure separately. The 
result of such a process may, in fact, create more distortion in 
relative weights than would occur if we continue to set weights based 
on single procedure claims. Moreover, there would be no way to confirm 
that there is more or less distortion unless hospitals can provide the 
data needed to create a "gold standard" against which to compare both 
methodologies. Again, the cost to hospitals of providing, and to 
Medicare of collecting and analyzing, such data needs to be considered.

As we have explained above, we have made and are continuing to make 
strides in the amount of data from multiple procedure claims we are 
able to use in rate setting. In our forthcoming proposed rule for 2005, 
we expect to discuss additional improvements to our processes for using 
data from multiple procedure claims.

GAO Recommendation:

That CMS analyze the effect that the variation in hospital charge-
setting practices has on the OPPS rate-setting methodology.

CMS Response:

The GAO says that "CMS assumed that to set charges, all hospitals mark 
up their costs by the same percentage." We believe this statement is an 
overgeneralization. Our department-specific cost-to-charge ratios 
(CCRs) capture variability in charging practices by cost center. The 
CMS uses department (cost-center) specific, or absent them, hospital 
overall CCRs applied to the charges from that hospital by related 
revenue code in order to determine the relative costs of services for 
that hospital. The CCRs are determined from the cost and charge data 
provided by the hospital and therefore reflect the differential 
charging practices by department and by hospital. For example, for the 
2004 final rule, median CCRs by department ranged from 0.12 for CAT 
Scan (3230) to 1.12 for Family Practice (4040).

Hence, the CMS methodology does not assume that "all hospitals mark up 
their charges by the same percentage." However the methodology does 
assume that the CCR for the department or for the hospital reflects the 
markup practices of that hospital for all services within the 
department or within the hospital and therefore it determines costs on 
a basis that is broader than each specific item or service in the 
department. The CCRs are not available for specific items and services 
within the hospital.

While we recognize that the application of CCRs to the charges of a 
hospital does not result in precise costing for individual items and 
services, it has long been the recognized method of establishing 
reasonable costs for hospital services. It has been supported by 
hospital organizations in public comments as the best means of 
establishing relative costs for OPPS rate setting. The application of 
CCR to charges is a fundamental principle of cost reimbursement that 
was in effect in Medicare for many years, supported by the hospital 
industry as resulting in an appropriate reflection of the costs of 
services they furnished. Moreover, as recently as the passage of the 
Medicare Prescription Drug, Improvement, and Modernization Act of 2003 
(MMA), Congress specified that this methodology would be used to set 
payment amounts for brachytherapy sources at cost. In a budget neutral 
payment system, the relative costs serve as the means for distributing 
the payments relative to the costs of items and services paid under the 
system and the application of hospital reported CCR to hospital 
reported charges in order to determine relative cost is the best system 
we currently have on which to base the payment distribution.

The issue raised by manufacturers is whether the department-specific 
ratio can adequately capture the variation in mark-up between low and 
high cost services within the same cost center. This variability may 
reflect hospital pricing sensitivity to income elasticity on the part 
of beneficiaries, among other influences, such as competitive pricing. 
The CMS' use of variable OCRs and the specific distinction of "within" 
department differences should be addressed. GAO might demonstrate that 
this was the issue raised in its interview guide. Further, GAO should 
note that its acquisition cost survey required by the MMA will provide 
some evidence on variability in charges for drugs, assuming that 
sampling is by hospital.

GAO Recommendation:

That CMS analyze whether the OPPS rate-setting methodology results in 
payment rates that uniformly reflect hospitals' costs of the services 
they provide to Medicare beneficiaries, and, if it does not, make 
appropriate changes in that methodology.

CMS Response:

We will consider GAO's recommendations as we continue to assess and 
refine our rate-setting methodology. In that regard, we would welcome 
more specific recommendations from GAO based on the findings in this 
report and/or as part of the studies that GAO is undertaking pursuant 
to its mandate under the MMA on hospital costs for drugs, biologicals, 
radiopharmaceuticals, and brachytherapy sources. We also look forward 
to receiving the results of GAO's surveys on hospital acquisition costs 
of the items mandated by the MMA. However, we believe the GAO 
recommendations should take into account the costs to providers and to 
the Medicare program of gathering additional data for analysis and 
possible methodological changes.

We continue to welcome specific recommendations from GAO and others on 
additional ways to increase the percentage of data used from multiple 
procedure claims to update OPPS rates. In spite of the progress we have 
made, we recognize that there are some services that are very 
frequently performed with other services that may be underrepresented 
in our rate-setting data. However, we believe we have made great 
strides on this issue and we are continuing to pursue the analyses 
necessary to create means by which all claims can be used to set the 
OPPS relative payment weights and rates. 

[End of section]

Appendix IV: GAO Contact and Staff Acknowledgments: 

GAO Contact: 

Nancy A. Edwards, (202) 512-3340: 

Acknowledgments: 

Beth Cameron Feldpush, Joanna L. Hiatt, Maria Martino, and Paul M. 
Thomas made major contributions to this report.

(290290): 

FOOTNOTES

[1] The Balanced Budget Act of 1997, Pub. L. No. 105-33, § 4523, 111 
Stat. 251, 445 (1997).

[2] In this report, we use the term "drugs" to refer to both drugs and 
biologicals.

[3] In July 2001, the agency's name was changed from the Health Care 
Financing Administration to CMS.

[4] The Medicare, Medicaid, and SCHIP Balanced Budget Refinement Act of 
1999, Pub. L. No. 106-113, App. F, § 201(b), 113 Stat. 1501A-321, 
1501A-337 (1999).

[5] Often described as a "sticker price" or "list price," AWP is the 
average price that a manufacturer suggests wholesalers charge 
pharmacies.

[6] Devices were initially eligible for pass-through payments based on 
the individual device. Effective April 1, 2001, devices are eligible 
for pass-through payments based on device categories, with an 
individual device eligible if it meets a category description. 

[7] Since this group included all drugs and devices eligible over a 4-
year period (from January 1, 1997 through January 1, 2001), many more 
items expired in 2003 than are expected to expire in any subsequent 
year. 

[8] In our analysis, we used the 2004 OPPS payment rates set by CMS in 
the November 7, 2003 final rule, which were based on hospital costs and 
hospital outpatient claims. 68 Fed. Reg. 63,398 (2003). These rates do 
not reflect provisions that limited the amount of fluctuation between 
the 2003 and 2004 rates that were implemented on January 1, 2004 as a 
result of the Medicare Prescription Drug, Improvement, and 
Modernization Act of 2003. We did not analyze the updated rates because 
they were not based on hospital costs or hospital outpatient claims. 
Pub. L. No. 108-173, § 621, 117 Stat. 2066, 2307 (2003).

[9] These claims are for services performed from April 1, 2001 through 
March 31, 2002.

[10] We analyzed all outpatient drugs that were individually identified 
in the outpatient claims data, not only the drugs classified as pass 
through. We could analyze only pass-through devices because these 
devices were individually identified in the outpatient claims, while 
other devices were not.

[11] The HCPCS is a uniform system of codes used by providers and 
medical suppliers to report professional services, procedures, and 
supplies.

[12] For multiple-service claims that have no packaged services, CMS 
considers each primary service its own single-service claim. Similarly, 
CMS treats each pathology service on a multiple-service claim as its 
own single-service claim. If a multiple-service claim contains one 
primary service together with certain other primary services that CMS 
states do not typically have packaged services associated with them, 
such as a chest X-ray or an electrocardiogram, CMS assigns all packaged 
services to that one primary service and treats it as a single-service 
claim. In addition, if the claim includes the date for each service and 
each primary service has a different date, CMS uses the dates of 
service associated with packaged services listed on the claims to match 
them to primary services with the same dates of service, and makes each 
primary service its own single-service claim. 

[13] CMS expects that most drug charges would be present on claims that 
also include the service for the administration of the drug; however, 
it is possible that drug charges are present on claims with primary 
services other than an administration and are included in the APCs to 
which those primary services are assigned. 

[14] This analysis excludes 2 drugs: 1 for which we were unable to 
determine a reliable AWP, and 1 for which the payment rate was an 
extremely high percentage of AWP. 

[15] Generally, "sole-source" drugs are brand-name drugs produced by 
only one manufacturer, "multi-source" drugs are drugs with generic 
equivalents or drugs for which there are two or more competing 
therapeutically-equivalent brand-name products, and "generic" drugs 
are not patented and can be produced by many manufacturers.

[16] GAO, Medicare: Payments for Covered Outpatient Drugs Exceed 
Providers' Cost, GAO-01-1118 (Washington, D.C.: Sept. 21, 2001).

[17] These APCs are identified in 67 Fed. Reg. 66,801-2 (2002).

[18] We defined "cancer center hospitals" as those hospitals that were 
members of the Association of Community Cancer Centers as of February 
28, 2003, the latest data available when we performed this analysis. We 
defined teaching status by a hospital's intern/resident-to-bed ratio. 
We defined a major teaching hospital as a hospital with an intern/
resident-to-bed ratio of 0.25 or more and a hospital without major 
teaching hospital status having a ratio of less than 0.25.

[19] We analyzed all outpatient drugs identified by a HCPCS code in the 
outpatient claims data, not only the drugs that had pass-through 
eligibility. We could analyze only pass-through devices because these 
devices were specifically identified in the outpatient claims while 
other devices were not.

[20] In 2003 and 2004, CMS used 53 percent of the approximately 20.4 
million and 58 percent of the approximately 16.9 million multiple-
service claims to set its rates, respectively. In the same years, the 
exclusion of the multiple-service claims from the analysis resulted in 
CMS using only 81 and 83 percent of all claims, respectively.

[21] Beginning in 2004, CMS uses the dates of service associated with 
packaged services listed on a claim to match them to primary services 
with the same dates of service to create a single service claim. Thus, 
claims with only chemotherapy administration and packaged services 
including drugs, and no other primary services delivered on the same 
dates, would be included in rate setting, however claims with 
chemotherapy administration, packaged services, and additional primary 
services delivered on the same date or dates would be excluded.

[22] This data file contains claims for services performed from April 
1, 2001 through March 31, 2002.

[23] For both chemotherapy administration and cardiac services, we 
included in our analysis the procedure or administration codes 
associated with those services. We also included any chemotherapy or 
cardiac drugs that were assigned to their own APC for payment in 2003. 
We identified only separately paid drugs, and did not include packaged 
drugs, because the structure of the data file would have counted the 
packaged codes twice in our analysis - once with the procedure code and 
again if they were also listed separately on the claim. For the same 
reason we excluded all of the device codes from our analysis, as all 
devices that lost pass-through eligibility were packaged in 2003.

GAO's Mission: 

The Government Accountability Office, the investigative arm of 
Congress, exists to support Congress in meeting its constitutional 
responsibilities and to help improve the performance and accountability 
of the federal government for the American people. GAO examines the use 
of public funds; evaluates federal programs and policies; and provides 
analyses, recommendations, and other assistance to help Congress make 
informed oversight, policy, and funding decisions. GAO's commitment to 
good government is reflected in its core values of accountability, 
integrity, and reliability.

Obtaining Copies of GAO Reports and Testimony: 

The fastest and easiest way to obtain copies of GAO documents at no 
cost is through the Internet. GAO's Web site ( www.gao.gov ) contains 
abstracts and full-text files of current reports and testimony and an 
expanding archive of older products. The Web site features a search 
engine to help you locate documents using key words and phrases. You 
can print these documents in their entirety, including charts and other 
graphics.

Each day, GAO issues a list of newly released reports, testimony, and 
correspondence. GAO posts this list, known as "Today's Reports," on its 
Web site daily. The list contains links to the full-text document 
files. To have GAO e-mail this list to you every afternoon, go to 
www.gao.gov and select "Subscribe to e-mail alerts" under the "Order 
GAO Products" heading.

Order by Mail or Phone: 

The first copy of each printed report is free. Additional copies are $2 
each. A check or money order should be made out to the Superintendent 
of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or 
more copies mailed to a single address are discounted 25 percent. 
Orders should be sent to: 

U.S. Government Accountability Office

441 G Street NW, Room LM

Washington, D.C. 20548: 

To order by Phone: 



Voice: (202) 512-6000: 

TDD: (202) 512-2537: 

Fax: (202) 512-6061: 

To Report Fraud, Waste, and Abuse in Federal Programs: 

Contact: 

Web site: www.gao.gov/fraudnet/fraudnet.htm

E-mail: fraudnet@gao.gov

Automated answering system: (800) 424-5454 or (202) 512-7470: 

Public Affairs: 

Jeff Nelligan, managing director,

NelliganJ@gao.gov

(202) 512-4800

U.S. Government Accountability Office,

441 G Street NW, Room 7149

Washington, D.C. 20548: