Skilled Nursing Facilities: Medicare Payments Exceed Costs for	 
Most but Not All Facilities (31-DEC-02, GAO-03-183).		 
                                                                 
This report addresses (1) the relationship between Medicare	 
skilled nursing facility (SNF) payments and the costs of treating
Medicare patients in freestanding SNFs, as well as the effect of 
Medicare SNF payments on the financial condition of these	 
facilities, and (2) the relationship between Medicare SNF	 
payments and the costs of treating patients in hospital-based	 
SNFs, as well as the factors that may account for cost		 
differences between hospital-based and freestanding SNFs.	 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-03-183 					        
    ACCNO:   A05789						        
  TITLE:     Skilled Nursing Facilities: Medicare Payments Exceed     
Costs for Most but Not All Facilities				 
     DATE:   12/31/2002 
  SUBJECT:   Cost analysis					 
	     Health care cost control				 
	     Health care programs				 
	     Skilled nursing facilities 			 
	     Comparative analysis				 
	     Medicare Program					 
	     Medicare Prospective Payment System		 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-03-183

Report to Congressional Committees

United States General Accounting Office

GAO

December 2002 SKILLED NURSING FACILITIES

Medicare Payments Exceed Costs for Most but Not All Facilities

GAO- 03- 183

Page i GAO- 03- 183 Medicare Skilled Nursing Facility Payments Letter 1
Results in Brief 2

Background 4 Medicare Payments Considerably Higher than Costs for Most

Freestanding SNFs, Thereby Improving Their Overall Financial Performance
10 Reported Medicare Costs of Hospital- Based SNFs Substantially

Exceeded Medicare Payments 16 Concluding Observations 20 Agency and Other
Comments and Our Evaluation 21

Appendix I Methodology for Calculating and Analyzing SNF Margins 25

Appendix II Medicare Margins of Freestanding SNFs by Selected
Characteristics, 1999 and 2000 29

Appendix III Comments from the Centers for Medicare & Medicaid Services 30

Appendix IV GAO Contacts and Staff Acknowledgments 31 GAO Contacts 31
Acknowledgments 31

Related GAO Products 32

Tables

Table 1: Freestanding and Hospital- Based SNFs by Type of Ownership, 1999
5 Table 2: Freestanding SNFs' Median Medicare Per Diem Payments

and Costs, 1997- 2000 11 Contents

Page ii GAO- 03- 183 Medicare Skilled Nursing Facility Payments

Table 3: Median Medicare Margins for Freestanding SNFs by Type of
Ownership, 1999 and 2000 12 Table 4: Median Medicare Margins for
Freestanding SNFs by

Occupancy Rate, 1999 13 Table 5: Median Margins for Freestanding SNFs,
1999 and 2000 14 Table 6: Median Total Margins for Freestanding SNFs by
Medicaid

Share of Patient Days, 1999 15 Table 7: Median Total Margins for
Freestanding SNFs by Type of

Ownership, 1999 and 2000 15 Table 8: Hospital- Based SNFs' Median Medicare
Per Diem Payments and Costs, 1997- 1999 17

Table 9: Therapies, Special Treatments, and Nursing in HospitalBased and
Freestanding SNFs, 1999 19 Table 10: Distribution of Freestanding SNFs for
Which Margins

Were Calculated, 1999 and 2000 28

Abbreviations

AAHSA American Association of Homes and Services for the Aging AHCA
American Health Care Association AHA American Hospital Association BBA
Balanced Budget Act of 1997 BBRA Medicare, Medicaid, and SCHIP Balanced
Budget

Refinement Act of 1999 BIPA Medicare, Medicaid, and SCHIP Benefits
Improvement and

Protection Act of 2000 CMS Centers for Medicare & Medicaid Services HCFA
Health Care Financing Administration IV intravenous MDS minimum data set

MedPAC Medicare Payment Advisory Commission OACT Office of the Actuary OIG
Office of Inspector General OSCAR Online Survey Certification and
Reporting System PPS prospective payment system RN registered nurse

RUG resource utilization group SNF skilled nursing facility

Page 1 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

December 31, 2002 The Honorable Max Baucus Chairman The Honorable Charles
E. Grassley Ranking Minority Member Committee on Finance United States
Senate

The Honorable William M. Thomas Chairman The Honorable Charles B. Rangel
Ranking Minority Member Committee on Ways and Means House of
Representatives

From the mid- 1980s through 1997, Medicare's spending for skilled nursing
facility (SNF) care rose at an average annual rate of 30 percent, making
it one of the fastest growing components of the Medicare program. During
this period, Medicare paid SNFs based on their reported costs of
delivering care, subject to certain limits that were higher for
hospitalbased SNFs than for freestanding SNFs. 1 Growth in the number of
services provided to an increasing number of patients resulted in Medicare
spending for SNF care reaching $13 billion in 1997. In response, the
Congress established in the Balanced Budget Act of 1997 (BBA) a SNF
prospective payment system (PPS) under which SNFs receive a fixed payment
that covers almost all services provided during each day of a Medicare-
covered stay. 2 With the implementation of the PPS, providers stated that
the payments

were inadequate, threatening their financial viability and their ability
to serve beneficiaries. The Congress subsequently modified the PPS with
several temporary payment increases. Some of these increases expired on
October 1, 2002, and provider representatives have said that they should
be restored due to payment shortfalls from other payers. These

1 A freestanding SNF is a nursing home that provides skilled nursing care
and is not attached to a hospital. A hospital- based SNF is a unit of an
acute care hospital. 2 Pub. L. No. 105- 33, S: 4432( a), 111 Stat. 251,
414 (codified at 42 U. S. C. S: 1395yy( e)).

United States General Accounting Office Washington, DC 20548

Page 2 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

representatives are mainly concerned about Medicaid, a joint federal-
state program for certain low- income individuals. According to this
argument, when Medicaid payments do not cover the costs of Medicaid
patients, higher payments are needed from Medicare to offset current or
anticipated financial difficulties.

The Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act
of 2000 (BIPA) directed us to examine the adequacy of Medicare payments
for SNFs and the extent to which Medicare revenues contribute to SNFs'
financial viability. 3 This report addresses (1) the relationship between
Medicare SNF payments and the costs of treating Medicare patients in
freestanding SNFs, as well as the effect of Medicare SNF payments on the
financial condition of these facilities, and (2) the relationship between
Medicare SNF payments and the costs of treating patients in hospital-
based SNFs, as well as the factors that may account for cost differences
between hospital- based and freestanding SNFs. To address these issues, we
analyzed 1997 through 2000 SNF Medicare cost reports, which are the
financial documents that facilities submit annually to receive payment
from Medicare. 4 We supplemented the cost report data with admissions data
from Medicare claims, data on facility characteristics from the Online
Survey Certification and Reporting System (OSCAR) maintained by the
Centers for Medicare & Medicaid Services (CMS), and data on patient
services from CMS's nursing home minimum data set (MDS). We did not verify
these data, but we excluded a SNF from our analysis if one or more of its
reported data items likely represented data entry or other reporting
errors. We also interviewed representatives of the

SNF industry. (See app. I for a more complete discussion of our methods.)
We performed our work from October 2001 through December 2002 in
accordance with generally accepted government auditing standards.

Under the PPS, most freestanding SNFs' Medicare payments substantially
exceeded the costs of caring for Medicare patients, contributing to
facilities' overall positive financial condition. In 1999, the first full
year under the PPS, the median freestanding SNF Medicare margin* a measure
that compares Medicare payments with Medicare costs* was slightly over 3
Pub. L. No. 106- 554, App. F, S: 311( d), 114 Stat. 2763A- 463, 2763A-
498.

4 Hospital- based facilities' 2000 cost reports as well as more recent
cost reports for all SNFs were not available when we did our analysis.
Results in Brief

Page 3 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

8 percent. 5 By 2000, when the temporary payment increases authorized by
the Congress started to take effect, the median Medicare margin had risen
to almost 19 percent. However, nearly one- quarter of SNFs in 2000 had
Medicare margins exceeding 30 percent, while about one- fifth had negative
Medicare margins; that is, the payments they received from Medicare did
not cover their costs of providing care. Medicare margins

were higher for freestanding SNFs affiliated with large, for- profit
nursing home chains and for those with high occupancy. The median SNF
total margin* which reflects total revenues and costs across all patients*
was 1.3 percent in 1999 and 1. 8 percent in 2000. A SNF's total margin
tended to be higher when its Medicare margin was higher despite the fact
that, in most SNFs, Medicare's share of patient days was small. The total
margins for freestanding SNFs tended to be lower when a higher proportion
of a SNF's patients had their care paid for by Medicaid. Unlike
freestanding SNFs, about 90 percent of hospital- based SNFs

reported significantly negative Medicare margins after Medicare's new SNF
payment system was launched. The median hospital- based SNF Medicare
margin was *53 percent in 1999. Under the PPS, per diem payments to
hospital- based SNFs dropped considerably, reflecting the

change from payments based on a facility's own costs to fixed payments
based on average costs for all facilities. At the same time, hospital-
based SNFs' reported per diem costs rose from 1997 through 1999. This is
in contrast to the experience of freestanding SNFs, which had lower per
diem Medicare costs than hospital- based SNFs prior to the PPS and reduced
their costs further after the shift to the PPS. The higher Medicare costs
reported by hospital- based SNFs may stem in part from differences in
services provided to patients. The higher costs may also reflect the
historical allocation of overhead costs to the SNF from the hospital, an
accounting practice that, while consistent with the payment incentives
under the prior cost- based reimbursement system, means that hospitalbased
SNFs' reported costs should be treated cautiously.

We received written comments from CMS stating that our findings are
consistent with a recent analysis conducted by MedPAC and other

5 The Medicare margin is the difference between Medicare payments and
Medicare costs, divided by Medicare payments, expressed as a percentage.
We computed Medicare margins for SNFs using methods similar to those
developed by the Medicare Payment Advisory Commission (MedPAC) and CMS's
Office of the Actuary (OACT). These methods assume that the average
routine costs per day of Medicare patients are equal to the average
routine costs per day of all patients in the SNF. (See app. I.)

Page 4 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

analyses of Medicare margins. CMS noted that this report supports its
position that Medicare SNF payment rates are more than adequate to cover
the cost of services provided to Medicare beneficiaries. CMS's comments
are reprinted in appendix III.

About 15,000 SNFs provide care for patients who are temporarily or
permanently unable to care for themselves, but who do not require the
level of care furnished in an acute care hospital. SNFs provide a variety
of services to patients, including nursing care; physical, occupational,

respiratory, and speech therapy; and medical social services. Medicare
covers these SNF services for Medicare beneficiaries who have recently
been discharged from a stay in an acute care hospital lasting at least 3
days and who need daily skilled care. In addition, many of these
facilities provide long- term care, mostly to Medicaid or private paying
patients. (Over 2,200 nursing homes are not SNFs and treat Medicaid but
not Medicare patients.) A SNF must meet federal standards to participate
in the Medicare or Medicaid program. About 85 percent of SNFs, or roughly

13,000, are freestanding and three- quarters of these are for- profit
entities. Nearly half of freestanding SNFs are owned by for- profit
chains* corporations operating multiple facilities. Hospital- based SNFs,
which number about 1,900, are usually part of not- for- profit acute care
hospitals. (See table 1.)

In 2000, Medicare SNF expenditures were $13 billion for services provided
to 1.4 million Medicare patients. About two- thirds of these patients
received care in freestanding SNFs and the remaining one- third received
care in hospital- based SNFs. On any given day, about 10 percent of
freestanding SNFs' residents were Medicare beneficiaries. 6 Most other
patients cared for in a freestanding SNF were longer- stay patients
receiving nursing or long- term care, which generally is paid for by
Medicaid or by the patients themselves. Medicare patients account for a
larger share of patients in hospital- based SNFs compared to freestanding
SNFs. About 56 percent of patients in hospital- based SNFs are Medicare

patients. 6 The 10 percent refers to SNF patients whose care is covered by
Medicare. It excludes Medicare beneficiaries who are long- stay patients
receiving long- term care or nursing care that Medicare does not cover. In
this report, Medicare patients refers to Medicare

beneficiaries who receive Medicare- covered SNF care. Background

Page 5 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

Table 1: Freestanding and Hospital- Based SNFs by Type of Ownership, 1999
a Numbers in percent Type of ownership Freestanding SNFs Hospital- based
SNFs

For- profit: 10 largest chains b 20 1 Smaller chains 29 9 Independents 26
6 All for- profit 75 16

Not- for- profit 22 65 Government c 3 19

Total 100 100

Source: GAO analysis of OSCAR data and of CMS data based on "Top 50
Nursing Facility Chains,"

Provider, July 1999. a Year refers to each SNF's cost reporting year,
which corresponds to its fiscal year that begins during the federal fiscal
year.

b Chain size is measured by the total number of beds in the chain's SNFs.
This number is selfreported. c Primarily facilities operated by counties
or cities.

During most of the 1990s, Medicare spending for SNF care grew much more
rapidly than spending for most other Medicare services. Under the cost-
based reimbursement system then in effect, Medicare paid SNFs' costs for
routine care (room and board and routine nursing) up to a specified limit,
with higher limits applied to hospital- based SNFs than to freestanding
SNFs. New providers were exempt from the routine- care cost limits for
their first 4 years, and all providers could be granted exemptions to the
limits by demonstrating that their higher costs were due to atypical
patients or patterns of care. Unlike routine- care costs, payments for
ancillary services such as therapy were not subject to cost limits, giving
facilities few incentives to control those costs.

The Congress, in the BBA, directed the Health Care Financing
Administration (HCFA) 7 to replace the cost- based reimbursement system
with a PPS. The PPS is designed to give SNFs incentives to furnish only

necessary services and to deliver those services efficiently by allowing
facilities to retain any excess of Medicare payments over costs, but

7 On July 1, 2001, the Secretary of Health and Human Services changed the
name of HCFA to CMS. In this report, we will continue to refer to HCFA
where our findings apply to the organizational structure and operations
associated with that name. Medicare Payment for SNF

Care

Page 6 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

requiring them to absorb any costs that are greater than payments. Under
the PPS, SNFs receive a per diem payment, adjusted for geographic
differences in labor costs and for differences in the resource needs of
patients. Adjustments for patients' resource needs are based on a patient
classification system, resource utilization group (RUG), version III. This

system assigns patients to 1 of 44 payment groups or RUGs, based on their
clinical condition, functional status, and use or expected use of certain
types of services. With few exceptions, the payment covers all routine,
therapy, and nursing costs incurred in treating patients.

Although we have reported that total SNF PPS payments are likely to be
adequate, we, MedPAC, and others have raised concerns that the Medicare
payments for certain types of patients may be too low because of
inadequacies with the patient classification system. 8 The patient
classification system may not sufficiently reflect the greater resource
needs of those patients who require multiple kinds of health care
services, such as drugs, laboratory services, and imaging. 9 In response
to BIPA's requirement that CMS report on alternatives to the RUG patient
classification system by January 1, 2005, CMS has sponsored research to
determine the feasibility of refinements as well as alternatives to the
RUG system.

After the implementation of the SNF PPS, some SNF representatives claimed
that Medicare payments were inadequate and contributed to SNFs' poor
financial performance. The Congress responded to provider concerns about
the adequacy of SNF payments by making several temporary modifications to
the PPS payment rates. Two of these changes, which applied to all Medicare
SNF patients and represented about $1.4 billion in annual payments,
expired on October 1, 2002:

8 U. S. General Accounting Office, Skilled Nursing Facilities: Medicare
Payment Changes Require Provider Adjustments but Maintain Access, GAO/
HEHS- 00- 23 (Washington, D. C.: Dec. 14, 1999); Medicare Payment Advisory
Commission, Report to the Congress: Medicare Payment Policy (Washington,
D. C.: March 1999); Korbin Liu, Jennie Harvell, and Barbara Gage, Post-
Acute Care Issues for Medicare: Interviews with Provider and Consumer
Groups, and Researchers and Policy Analysts (Washington, D. C.: May 2000),

http: www. hhs. gov/ aspe. hhs. gov/ search/ daltcp/ Reports/ pacissue.
htm (downloaded August 1, 2002).

9 Medicare Payment Advisory Commission, Report to the Congress: Medicare
Payment Policy (Washington, D. C.: March 2001).

Page 7 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

 an increase provided by the Medicare, Medicaid, and SCHIP Balanced
Budget Refinement Act of 1999 (BBRA) of 4 percent in the payment rate for
all RUGs for fiscal years 2001 and 2002; 10 and  an increase provided by
BIPA of 16.66 percent in the nursing component

of the payment rate for all RUGs for April through September 2001 and
fiscal year 2002. 11 Two additional changes were enacted for selected
types of Medicare

patients. These changes, which affect 26 of the 44 RUGs and total about $1
billion per year, will remain in effect until CMS refines the patient
classification system. CMS has announced that, although it is examining
possible refinements, the system will not be changed for the 2003 payment
year. The two payment changes are:

 an increase provided by BBRA of 20 percent in the payment rate for 15
RUGs, including those for extensive services, special care, clinically
complex care, and certain rehabilitation services; 12 and  an increase
provided by BIPA of 6.7 percent in the payment rate for 14 rehabilitation
RUGs. 13 This redirected the funds from the 3 rehabilitation

RUGs that had received the 20 percent BBRA increase and applied these
funds to all 14 rehabilitation RUGs. As a result of this redirection of
funds, aggregate payments did not increase. Prior to October 1, 2002, when
two of these temporary payment increases

expired, some SNF representatives stated that Medicare payments were
adequate, although they said inadequate Medicaid payments compromised SNF
financial viability. Following the expiration of these two temporary
Medicare payment increases, provider organizations have again expressed
concern that Medicare payments are now not adequate.

Other legislative provisions also affected Medicare payments to SNFs. A
key provision was the 3- year phased transition to the PPS that the BBA
established. Under this transition, which began in 1998, SNFs were paid a

10 Pub. L. No. 106- 113, App. F, S: 101( d), 113 Stat. 1501A- 321, 1501A-
324. 11 BIPA S: 312( a). This increase raised the overall payment rate
from 4 to 12 percent, depending on the RUG. See U. S. General Accounting
Office, Skilled Nursing Facilities: Available Data Show Average Nursing
Staff Time Changed Little after Medicare Payment Increase, GAO- 03- 176
(Washington, D. C.: Nov. 13, 2002).

12 BBRA S: 101. 13 BIPA S: 314.

Page 8 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

blend of facility- specific rates, based on each SNF's 1995 costs, and the
PPS rate. BBRA allowed SNFs to receive the full PPS rate for cost
reporting periods beginning on or after January 1, 2000. 14 This provision
permitted SNFs that were advantaged by the PPS rate to be paid under it,
while SNFs that were disadvantaged by the new rate could transition to it
on the original 3- year schedule.

Medicare- covered SNF use quadrupled from 1985 to 1997, rising from 10 SNF
users per 1,000 Medicare fee- for- service beneficiaries to 41 users. 15 A
variety of factors contributed to this increase:

 In 1983, Medicare began paying hospitals a fixed rate per stay as an
incentive to control costs. Hospitals responded as expected and, to reduce
costs by cutting the length of hospital stays, transferred patients more
quickly to SNFs and other post- acute care settings.  In 1988,
clarification of Medicare coverage guidelines allowed more

beneficiaries to qualify for SNF services.  From 1990 through 1996 the
number of freestanding SNFs increased 49

percent, while hospital- based SNFs increased 82 percent. This growth in
providers was encouraged by Medicare payment policies, which did not
subject new SNFs to payment limits for their first 4 years of operation,
and by the growth in payments. From 1990 through 1996, the average
Medicare payment per SNF day of care climbed from $98 to $292. 16 During
this period prior to the implementation of the SNF PPS, hospitals

that had SNFs were particularly advantaged by transferring acute care
patients sooner to their own SNFs. Transfers enabled these hospitals to
reduce their acute care costs and increase their SNF revenues. To help
ensure that Medicare did not overpay for services at the end of an acute
episode of care, the Congress required HCFA to reduce hospital payments
for patients transferred to post- acute care after a shorter- than-
average

14 BBRA S: 102. 15 Beneficiaries enrolled in managed care plans are
excluded from the calculation. 16 Prospective Payment Assessment
Commission, Medicare and the American Health Care System, Report to the
Congress (Washington, D. C.: June 1997), p. 107. Use of Medicare- Covered

SNF Care

Page 9 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

hospital stay. 17 In fiscal year 1999 HCFA implemented this policy for 10
types 18 of patients with high use of post- acute care. The reduction in
hospital payment for patients transferred to post- acute care lessened the
incentive for hospitals to shorten the stays of these patients. Following
this change, SNF admissions per 1,000 hospital discharges decreased by 4
percent from 1996 to 2000. After adjusting for differences in patients'
clinical conditions, 19 the number of admissions was only 2 percent lower
in 2000 than in 1996, indicating that part of the decline was due to
reduced need for SNF care. However, if the 10 types of patients affected
by the change in hospital payment for transfers are

excluded, SNF admissions were the same in 2000 as in 1996. This suggests
that some of the observed decline in SNF admissions may be due to the
change in payment policy for hospital transfers.

Despite this observed decline in SNF admissions, the evidence does not
suggest major problems with beneficiary access to SNF care. Beginning in
1999, the Department of Health and Human Services' Office of Inspector

General (OIG) has examined SNF access in several surveys of hospital
discharge planners to determine whether they are able to place their
Medicare patients who need care in SNFs. 20 These surveys have found that
planners can place most patients needing care. In the most recent OIG
survey, about three- quarters of discharge planners reported that they
were able to place all patients. However, some planners reported delays in

placing patients with particular medical conditions or service needs,
resulting in these patients continuing to receive care in the hospital
rather

17 BBA S: 4407. Without this provision, Medicare would pay hospitals a
fixed amount per stay* whether the patient had an average or shorter-
than- average hospital stay. In addition, Medicare would pay SNFs for each
day of SNF care, even if* prior to the change in hospital payment* some of
this care would have been provided in the hospital.

18 "Types" refers to diagnosis- related groups (DRG), a classification
scheme that groups acute care hospital patients according to diagnosis,
type of treatment, age, and other criteria. 19 "Patients' clinical
conditions" refers to DRGs. 20 Department of Health and Human Services,
Office of Inspector General, Medicare

Beneficiary Access to Skilled Nursing Facilities: 2001, OEI- 02- 01- 00160
(July 2001); Department of Health and Human Services, Office of Inspector
General, Medicare Beneficiary Access to Skilled Nursing Facilities: 2000,
OEI- 02- 00- 00330 (September 2000); and Department of Health and Human
Services, Office of Inspector General, Early Effects of the Prospective
Payment System on Access to Skilled Nursing Facilities, OEI- 02- 99- 00400
(August 1999).

Page 10 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

than in a SNF. Patients who took longer to place included those who needed
intravenous (IV) antibiotics or expensive drugs, as well as those who were
ventilator- dependent or who required dialysis or wound care.

In the first 2 full years under the PPS, Medicare payments more than
covered Medicare costs for most freestanding SNFs, although their
experiences varied widely. Many SNFs had very high Medicare margins,
particularly in 2000, although in both years a minority of SNFs had
negative Medicare margins* payments from Medicare did not cover their
costs of serving Medicare patients. The median Medicare margin for SNFs
that were owned by large nursing home chains and for those SNFs with high
occupancy was much higher than the overall median Medicare margin for all
SNFs. SNFs' Medicare margins were sufficiently high that, while Medicare's
share of most SNFs' total patient days was relatively small, SNFs with
higher Medicare margins generally had higher total

margins, which reflect all SNF revenues and costs. For- profit facilities
generally had higher total margins, as did facilities owned by large
chains. SNFs with higher proportions of Medicaid patients generally had
lower total margins.

For their first 2 years under PPS, most freestanding SNFs reported
positive Medicare margins, meaning that their payments more than covered
their costs. 21 In 1999, the median facility had a Medicare margin
exceeding 8 percent, and over one- tenth had margins of 30 percent or
more. By 2000, the median Medicare margin for freestanding SNFs had risen
to nearly 19

percent, 22 and almost one- quarter of SNFs had Medicare margins of 30
percent or more. These positive margins resulted largely from SNFs
reducing their costs. Although Medicare payments per day were 8 percent
lower in 1999 than in 1997, for the median facility these lower payments

21 In aggregate, Medicare payments to freestanding SNFs exceeded Medicare
costs by 11 percent in 1999 and by 24 percent in 2000. 22 We computed
Medicare margins for SNFs using methods similar to those developed by
MedPAC and CMS's OACT. These methods assume that average nursing costs are
the same

for Medicare and other patients. A more refined measure of costs would
reflect the difference in nursing needs between Medicare patients and
other patients in the facility. We tested such a refinement but could not
calculate it for all SNFs due to incomplete data. Using available data, we
estimated that the median Medicare margin in 2000, based on a more refined
measure of costs, would have been between 0.6 and 1. 6 percentage points

lower than that reported here. (See app. I.) Medicare Payments

Considerably Higher than Costs for Most Freestanding SNFs, Thereby
Improving Their Overall Financial Performance

Medicare Margins Generally High, Particularly in 2000 and for SNFs in
Large For- Profit Chains and Those with High Occupancy

Page 11 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

were more than offset by lower costs. (See table 2.) From 1999 through
2000, costs had again declined, although by a smaller amount. At the same
time, payments increased, as the temporary increases authorized by the
Congress began to be implemented. 23 Table 2: Freestanding SNFs' Median
Medicare Per Diem Payments and Costs, 1997-

2000

In dollars Year a 1997 1998 1999 2000

Medicare payments per day 264 270 243 269 Medicare costs per day 273 279
224 220

Source: GAO analysis of Medicare cost report data. a Year refers to each
SNF's cost reporting year, which corresponds to its fiscal year that
begins during the federal fiscal year.

Although most freestanding SNFs had positive Medicare margins, for a
minority of SNFs, Medicare payments did not cover Medicare costs. In 1999,
more than one- third of freestanding facilities reported negative Medicare
margins, with one- tenth reporting margins that were *30 percent or less.
24 By 2000, the number of facilities with negative margins had declined
substantially: about 19 percent had margins that were less than zero, and
4 percent had margins of *30 percent or less.

Freestanding SNFs' Medicare margins differed by type of ownership.
Forprofit SNFs* particularly those associated with the largest chains* had
positive Medicare margins in both 1999 and 2000 that were higher than
those of both not- for- profit and government- operated SNFs. 25 In 1999,
median margins for not- for- profit and government- operated SNFs were

23 The impact of the temporary increases on 2000 payments depended both on
the effective date in the law and on the start date of each SNF's fiscal
year. The 4 percent across- theboard increase was in effect for the entire
year for all facilities. The 20 percent temporary increase in payments for
15 RUGs was in effect for part of 2000 for most facilities. The other two
increases* the increase in the nursing component and the increase for
rehabilitation RUGs* had a smaller effect in the 2000 cost reporting year:
16 percent of SNFs were affected for part of the year.

24 That is, Medicare payments fell short of costs by at least 30 percent.
25 This relationship held true even after accounting for other factors.
(See app. I.)

Page 12 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

negative, while in 2000 the median margins for all types of freestanding
SNFs were positive. (See table 3.)

Table 3: Median Medicare Margins for Freestanding SNFs by Type of
Ownership, 1999 and 2000 Numbers in percent 1999 a 2000 a

For- profit: 10 largest chains b 18.2 25.2 Smaller chains 7.6 16.8
Independents 6.6 17.7 All for- profit 11.7 20.4 Not- for- profit -1.4 11.1
Government c -13.9 8.2 All freestanding SNFs 8.4 18.9

Source: GAO analysis of Medicare cost report data, OSCAR data, and CMS
data based on "Top 50 Nursing Facility Chains," Provider, July 1999. a
Year refers to each SNF's cost reporting year, which corresponds to its
fiscal year that begins during the federal fiscal year.

b Chain size is measured by the total number of beds in the chain's SNFs.
This number is selfreported. c Primarily facilities operated by counties
or cities.

Medicare margins also varied with occupancy. 26 Higher occupancy resulted
in higher margins. For example, in 1999, freestanding SNFs with occupancy
rates 27 of 90 percent or more had a median margin of 10.2 percent, while
SNFs with occupancy rates below 70 percent had a median margin of 0.6
percent. (See table 4.) These results are not surprising, because higher
occupancy reduces per diem costs, as fixed costs are spread across more
patient days.

26 Other factors associated with higher Medicare margins included urban
location, while factors associated with lower margins included having a
small number of Medicare SNF patient days per year.

27 The occupancy rate is based on all beds in the facility, regardless of
whether they were occupied by Medicare patients or other patients.

Page 13 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

Table 4: Median Medicare Margins for Freestanding SNFs by Occupancy Rate,
1999 a Numbers in percent Occupancy rate b Share of SNFs Medicare margin

Less than 70 percent 11 0.6 70 to 79 percent 13 6.1 80 to 89 percent 28
8.3 90 to 100 percent 48 10.2

Source: GAO analysis of Medicare cost report data. a Year refers to each
SNF's cost reporting year, which corresponds to its fiscal year that
begins during the federal fiscal year.

b The occupancy rate is based on all beds in the facility, regardless of
whether they were occupied by Medicare patients or other patients.

Despite the expiration of two temporary Medicare payment increases and the
completed transition from payments based on a facility's own costs to PPS
rates, SNFs' positive Medicare margins are likely to continue. MedPAC has
estimated that freestanding SNFs' aggregate Medicare margin for 2002 would
be 9.4 percent, excluding for the entire year the temporary

payment increases that expired on October 1, 2002, and assuming that all
facilities had completed the transition to the PPS. 28 Although most
freestanding SNFs had positive Medicare margins, most had few Medicare
patients and Medicare accounted for a small share of their revenue. In
1999, the median SNF had about six Medicare patients each day and received
about 13 percent of its revenue from Medicare. By contrast, the care for
about two- thirds of patients was paid for by

Medicaid with the remainder generally paid for by the patients themselves.
Despite Medicare's small share of most freestanding SNFs' patients,
Medicare contributed substantially to these facilities' total margins,
because Medicare payments were much higher than costs. In general, 28 In
its estimate, MedPAC updated 1999 costs and payments through 2002 for
inflation and

used payment policies expected to be in effect in 2003: temporary payment
increases due to expire on October 1, 2002, were excluded and all
facilities were assumed to be paid entirely under the PPS. MedPAC included
the temporary payment increases that will be in effect in 2003 and will
not expire until the patient classification system is refined. It
estimated the margin using aggregate Medicare SNF payments and aggregate
Medicare SNF

costs. Medicare Payment Advisory Commission, Report to the Congress:
Medicare Payment Policy (Washington, D. C.: March 2002). Despite
Medicare's Small

Share of Patients, Medicare Margins Significantly Affected Freestanding
SNFs' Total Margins

Page 14 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

facilities with higher Medicare margins had higher total margins.
Moreover, in 1999 and 2000, the median total margin would have been
negative without Medicare; for example, in 2000, it would have been *1.2
percent. With Medicare, the actual median total margin was 1.8 percent.
(See table 5.)

Table 5: Median Margins for Freestanding SNFs, 1999 and 2000

Numbers in percent 1999 a 2000 a

Medicare margin 8.4 18.9 Total margin 1.3 1. 8 Total margin without
Medicare -0.1 -1.2

Source: GAO analysis of Medicare cost report data. a Year refers to each
SNF's cost reporting year, which corresponds to its fiscal year that
begins during the federal fiscal year.

Medicaid's share of freestanding SNFs' residents influenced facilities'
overall profitability. The larger Medicaid's share of a SNF's patient
days, the smaller its total margin. 29 (See table 6.) For- profit status
and ownership by a chain also affected freestanding SNFs' total margins.
For- profit facilities showed higher median total margins than not- for-
profit and government- operated facilities, and large chains displayed the
highest total margins. (See table 7.)

29 Our statistical analysis of SNF total margins showed that the Medicaid
share of patient days remains significant after accounting for SNF
characteristics, including for- profit status. Freestanding SNFs' Total
Margins Varied by

Medicaid Share of Residents and Type of Ownership

Page 15 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

Table 6: Median Total Margins for Freestanding SNFs by Medicaid Share of
Patient Days, 1999 a Numbers in percent Medicaid share of patients b Total
margins

Less than 56 percent 2.2 56 to 69 percent 1.8 70 to 79 percent 1.0 80
percent and above 0.6

Source: GAO analysis of Medicare cost report data. a Year refers to each
SNF's cost reporting year, which corresponds to its fiscal year that
begins during the federal fiscal year.

b Each group contains roughly one- quarter of all freestanding SNFs.

Table 7: Median Total Margins for Freestanding SNFs by Type of Ownership,
1999 and 2000

Numbers in percent 1999 a 2000 a

For- profit: 10 largest chains b 2.7 3. 8 Smaller chains 0.6 1. 5
Independents 1.8 1. 8 All for- profit 1.6 2. 2 Not- for- profit 0.6 0. 3
Government c 0.3 0. 6 All freestanding SNFs 1.3 1. 8

Source: GAO analysis of Medicare cost report data, OSCAR data, and CMS
data based on "Top 50 Nursing Facility Chains," Provider, July 1999. a
Year refers to each SNF's cost reporting year, which corresponds to its
fiscal year that begins during the federal fiscal year.

b Chain size is measured by the total number of beds in the chain's SNFs.
This number is selfreported. c Primarily facilities operated by counties
or cities.

Many other factors were also related to differences in freestanding SNFs'
total margins. Factors contributing to high total margins included high
occupancy and location in a rural area. Factors associated with low total
margins were a high concentration of SNFs in a geographic area, and
location in a state with relatively high average wages for nursing staff.

Page 16 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

In contrast to freestanding SNFs, hospital- based SNFs reported very
negative Medicare margins after the introduction of the PPS. These low
margins reflected a substantial decline in Medicare payments to
hospitalbased SNFs under the PPS as well as hospital- based SNFs' weak
response to PPS incentives to reduce costs. Differences in services
between hospital- based SNFs and freestanding SNFs could have resulted in
higher costs for hospital- based SNFs that may not have been fully
accounted for by the patient classification system in the PPS. The
negative margins reported by hospital- based SNFs were also due in part to
their high costs per day, which may reflect the historical allocation of
hospitals' overhead to their SNF units.

In 1999, about 90 percent of all hospital- based SNFs reported Medicare
costs exceeding Medicare payments, and the median hospital- based SNF
posted Medicare margins of *53 percent. 30 While insufficient data were
available to compute margins for hospital- based SNFs for 2000, their
margins likely improved with the payment increases, but remained
significantly negative. 31 Only a small minority* about 10 percent*
reported positive margins in 1999. These more successful hospital- based
SNFs generally had high occupancy and did not rely heavily on Medicare
payments.

The explanation of these low margins lies partly in the large decline in
Medicare per diem payments that followed the shift to the PPS. Prior to
the PPS, Medicare's payments to SNFs were based on each facility's own
costs. This led to higher payments for hospital- based SNFs: a median of
$378 per day for hospital- based SNFs in 1997 (see table 8), compared to a
median of $264 for freestanding SNFs.

30 In aggregate, Medicare costs of hospital- based SNFs exceeded their
Medicare payments by 55 percent. 31 We were not able to calculate margins
for 2000 because most hospitals' cost reports were not yet available when
we performed our work. Reported Medicare

Costs of HospitalBased SNFs Substantially Exceeded Medicare Payments

Unlike Freestanding SNFs, Most Hospital- Based SNFs Had Very Negative
Medicare Margins

Page 17 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

Table 8: Hospital- Based SNFs' Median Medicare Per Diem Payments and
Costs, 1997- 1999

In dollars Year a 1997 1998 1999

Medicare payments per day 378 347 281 Medicare costs per day 461 484 490

Source: GAO analysis of Medicare cost report data. a Year refers to each
SNF's cost reporting year, which corresponds to its fiscal year that
begins during the federal fiscal year.

In the first year of the PPS, hospital- based SNFs, unlike their
freestanding counterparts, did not respond to the incentives in the PPS by
reducing costs: compared to 1997, hospital- based SNFs' costs in 1999 were
higher by $29 per day. 32 By contrast, freestanding SNFs reduced costs by
$49 per day. As a result, per diem costs continued to be substantially
higher in hospital- based facilities than in freestanding SNFs* more than
twice as high in 1999. Some differences in costs between hospital- based
and freestanding SNFs

may also reflect differences in services in the two settings. Although
patients in hospital- based SNFs had received less therapy as of their
initial Medicare assessment than patients in freestanding SNFs (and
slightly more as of their second assessment), 33 they were more likely to
receive other kinds of services, including IV medications, oxygen therapy,
and

transfusions. Hospital- based SNFs also gave significantly more nursing
care, as measured by the ratio of nurses to patients. However, when
patients' resource needs were measured by RUGs, patients in the two

32 Nearly one- quarter of hospital- based SNFs closed after the
implementation of the PPS. These closures accounted for 4 percent of all
Medicare- covered SNF days in 1997 and 22 percent of Medicare- covered SNF
days provided by hospital- based SNFs. Hospital- based SNFs that opened
after 1995 were more likely to close than those that had been in business
longer.

33 SNFs document the amount of therapy a patient has received in the last
7 days at an initial assessment (often called the 5- day assessment) and
again at a second assessment (often called the 14- day assessment). The
amount of therapy documented includes only therapies provided in the SNF;
if the initial assessment occurs on the fifth or sixth day, then only 5 or
6 days of therapy are documented. Therapy includes physical, occupational,
and speech

therapy. Differences in Services and Accounting Practices May

Contribute to Cost Differences between Hospital- Based and Freestanding
SNFs

Page 18 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

settings appeared identical, suggesting that their service needs should be
comparable. Consequently, the observed differences in the treatments that
patients received may suggest that the RUGs do not fully measure
differences in patients' conditions and could account for part of the cost
difference between hospital- based and freestanding SNFs. 34 (See table
9.)

34 A study by Abt Associates found that hospital- based SNFs have
significantly higher perpatient costs than freestanding SNFs after
controlling for various factors, but could not explain why. Abt
Associates, Inc., Why Are Hospital- Based Nursing Homes So Expensive? The
Relative Importance of Acuity and Treatment Setting, Health Services and

Evaluation (HSRE) Working Paper No. 3 (Cambridge, Massachusetts: February
2001).

Page 19 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

Table 9: Therapies, Special Treatments, and Nursing in Hospital- Based and
Freestanding SNFs, 1999

Hospitalbased Freestanding Average therapy per patient, minutes a Measured
at initial Medicare assessment 337 369

Measured at second Medicare assessment 446 440

Patients receiving special treatments, percentage b IV medication 57 37
Intake/ output 70 53 Monitoring acute medical condition 62 57 Ostomy care
3 6 Oxygen therapy 38 28 Suctioning 4 3 Tracheotomy care 1 1 Transfusions
8 4

Average nurse staffing ratio c Average registered nurse (RN) hours per
patient day 1.81 0.37 Average total nurse hours (RN, licensed practical
nurse, aide) per patient day 5.57 3.12

Source: GAO analysis of MDS data and OSCAR data. Note: Entries are not
adjusted for differences in patients' clinical conditions between
hospital- based and freestanding SNFs. Their RUG scores are similar. a
SNFs document the amount of therapy a patient has received in the last 7
days at an initial

assessment (often called the 5- day assessment) and again at a second
assessment (often called the 14- day assessment). The amount of therapy
documented includes only therapies provided in the SNF; if the initial
assessment occurs on the fifth or sixth day, then only 5 or 6 days of
therapy are documented. Therapy includes physical, occupational, and
speech therapy. Patients who were receiving chemotherapy, dialysis, IV
medications, oxygen therapy, radiation, or ventilator or respiratory care
were excluded from the analysis of therapy. b Based on the initial
Medicare assessment. Similar results were obtained using the second

assessment. c These staffing ratios are based on SNFs' overall direct
nursing care staff and the total number of

patients. These ratios are facility- wide rather than ratios specific to
Medicare patients. Part of the cost differential between hospital- based
and freestanding SNFs may reflect accounting practices that increase
reported costs for individual units of the hospital, such as SNFs, that
had been paid on the basis of these reported costs. MedPAC "believe[ s] a
significant portion of the negative [hospital- based] SNF margin reflects
the allocation of hospital

Page 20 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

overhead costs to cost- reimbursed units." 35 Prior to the SNF PPS, but
after Medicare had implemented its per case PPS for acute inpatient
hospital care, hospitals had an incentive to allocate administrative and
capital costs to cost- reimbursed units, including SNFs, potentially
raising reported costs for these units. (Capital as a share of Medicare
per diem costs for hospital- based SNFs was about 96 percent higher than
it was for freestanding SNFs in 1999.) Now that SNFs are paid a fixed
rate, this incentive no longer exists* but neither is there an incentive
to change historical cost allocations. In fact, capital as a share of
Medicare costs for hospital- based SNFs has changed little since before
PPS. In light of these accounting issues, reported costs of hospital-
based SNFs, as well as margins calculated from these costs, should be
treated cautiously. 36 Our analysis shows that the Medicare PPS generally
pays SNFs adequately for the services that beneficiaries receive.
Freestanding SNFs, which treat most Medicare SNF patients, generally
received Medicare payments that

exceeded their costs, often by considerable amounts. Most hospital- based
SNFs reported costs that were greatly in excess of Medicare payments, but
these hospital- based SNFs did not respond to the incentives in the PPS by
reducing costs. Some of their high costs may also be due to differences in
patients that lead to higher resource use and that are not captured by the
payment system. This problem could be addressed through refinements to the
patient classification system, which CMS is currently studying.

Concerns about the financial conditions of some nursing homes have led to
interest in using Medicare payment policy to offset current or anticipated
financial difficulties. Whatever the merits of the case for aiding these
facilities, an across- the- board increase in Medicare payments, such as
the restoration of the expired temporary increases, would be particularly
inefficient. An across- the- board increase would go to all providers of
Medicare SNF care, even those for which Medicare's current

35 See Medicare Payment Advisory Commission, Report to the Congress:
Medicare Payment Policy (Washington, D. C.: March 2001), p. 67. 36 In
addition to the differences in per diem cost, the average length of stay
in hospital- based SNFs was less than half that in freestanding
facilities. It is uncertain whether the difference in length of stay
resulted from differences in patients, effectiveness of treatment, or
other factors. For analysis of selected differences in outcomes, see
DataPRO Team, Skilled Nursing Facilities Prospective Payment System,
Findings from Rehospitalization Transfer within 30 Days Analysis: Summary
and Selected Results, Ad Hoc Report E (Washington, D. C.: June 2002).
Concluding

Observations

Page 21 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

payments already greatly exceed costs and which are not experiencing any
financial difficulty. Such an increase could also not take account of
differences in the adequacy of revenues from other payers, especially the
state Medicaid programs. Moreover, over 2000 nursing homes would not get
any increase, because they do not participate in Medicare. We received
written comments on a draft of this report from CMS (see

app. III) and oral comments from the American Association of Homes and
Services for the Aging (AAHSA), which represents not- for- profit nursing
facilities; the American Health Care Association (AHCA), which represents
for- profit and not- for- profit nursing facilities; and the American
Hospital Association (AHA), which represents hospitals. We incorporated
their

comments as appropriate. Industry representatives agreed with our basic
findings concerning SNF margins and several stated that it was a good
report.

CMS noted that our findings are consistent with a recent analysis
conducted by MedPAC and other analyses of Medicare margins. CMS stated
that the report supports its position that Medicare SNF payment rates are
more than adequate to cover the cost of services provided to Medicare
beneficiaries.

Representatives of AHCA and AAHSA who reviewed the draft report were
concerned that the report does not address the issue of Medicaid payments
being too low and the role they play in SNFs' financial viability. AHCA
also objected to the prominence given to differences in payments by type
of ownership, which they believe is a less important factor than occupancy
and Medicare percentage of total SNF days in explaining SNF total margins.
They characterized as misleading the 30 percent annual growth rate in
Medicare SNF expenditures that we reported for 1985 through 1997, stating
that spending growth was driven by growth in utilization.

Both the AHCA and AHA representatives commented on our findings concerning
the differences between freestanding and hospital- based SNFs. AHCA stated
that any differences in services between hospital- based and freestanding
SNFs are due to differences in patients' clinical conditions.

The AHA representatives objected to our statement that the higher per diem
costs of hospital- based SNFs could be partly due to the historical
patterns of allocating overhead and other costs to the SNF. They stated
Agency and Other

Comments and Our Evaluation

CMS Industry Associations

Page 22 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

that hospital cost accounting systems are constantly changing as hospitals
add and drop services, and that the cost allocation issue in general is an
artifact of the 15 years of operation of Medicare's inpatient PPS.
According to AHA, hospitals were already operating at an efficient level
when the

SNF PPS was implemented and therefore had fewer excess costs to trim. The
AHA representatives also noted the shorter average length of stay of
hospital- based SNFs and suggested that reporting costs on a per case
basis, which reflects this shorter length of stay, rather than on a per
diem basis, would show that hospital- based SNFs are less costly than
freestanding SNFs.

Both AAHSA and AHA addressed possible changes to the PPS. The AAHSA
representative stressed that Medicare payments are inadequate for patients
who need medically complex nonrehabilitation ancillary services. She
stated that the report should include language suggesting that the

patient classification system should be changed to better reflect patient
characteristics. Regarding our concluding observations, AHA inferred from
our discussion of across- the- board increases in payment rates that we
would favor targeted increases.

Regarding payments, we accounted for Medicaid payments in our analysis of
total margins but were unable to conduct a separate analysis of Medicaid
payment adequacy for nursing homes because of the lack of suitable data.
Isolating the impact of Medicaid payments was not possible because the
Medicare cost reports do not report Medicaid payments or costs separately
and because there is no source of Medicaid financial data collected on a
consistent and ongoing basis across all facilities and states. Although
occupancy and Medicare percentage of total SNF days were important factors
in explaining differences in total margins, we found that after accounting
for these factors, type of ownership remained a significant factor. We
agree that utilization growth was a key factor in the rise in Medicare SNF
spending, as this report states. Nonetheless, the rapid

growth in spending, which we characterized correctly, provided the impetus
for enactment of the SNF PPS.

With regard to differences between freestanding and hospital- based SNFs,
our analyses reported in the draft show that the average RUG score, a
measure of patients' clinical conditions, is nearly the same for both. We
acknowledged, moreover, that the RUG system may not adequately account for
differences across patients. As we stated in the draft, the higher costs
of hospital- based SNFs are consistent with historical patterns of
allocating overhead costs to SNF units. Whatever changes are occurring Our
Response

Page 23 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

in hospital cost accounting, hospitals have had no incentive to change
their historical cost allocations since the implementation of the SNF PPS.
Moreover, we found no evidence to suggest that they had done so. For
example, as we stated in the draft, capital costs expressed as a share of
Medicare per diem costs have not changed. Although AHA representatives
said that hospitals were more efficient and consequently had less
flexibility to reduce costs after the implementation of the SNF PPS, they
did not offer evidence to reconcile this view with hospital- based SNFs'
higher costs. Hospitals' efficiency may have improved on the inpatient
side as a result of the hospital PPS, but this would not necessarily
improve the

efficiency of hospital- based SNFs. Although we agree that hospital- based
SNFs are less costly than freestanding SNFs on a per case basis, we did
not present this measure because payments under the SNF PPS are not made
on a per case basis. Regarding possible changes in the PPS, we have
previously acknowledged

that the current patient classification system may not adequately
recognize the greater resource needs of some patients. 37 We support CMS's
sponsorship of research to investigate improvements in the system. Our
analysis does not support an increase in Medicare payment rates. Instead,
it would be preferable to refine the patient classification system
underlying the SNF PPS, if necessary redistributing money to ensure that
payments vary appropriately to reflect patient resource needs. We are
sending copies of this report to the Administrator of CMS,

appropriate congressional committees, and other interested parties. We
will also provide copies to others upon request. In addition, the report
is available at no charge on the GAO Web site at http:// www. gao. gov.

37 See U. S. General Accounting Office, Skilled Nursing Facilities:
Medicare Payment Changes Require Provider Adjustments but Maintain Access,
GAO/ HEHS- 00- 23 (Washington, D. C.: Dec. 14, 1999).

Page 24 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

If you or your staffs have any questions, please call me at (202) 512-
7114. Other GAO contacts and staff acknowledgements are listed in appendix
IV.

Laura A. Dummit Director, Health Care* Medicare Payment Issues

Appendix I: Methodology for Calculating and Analyzing SNF Margins Page 25
GAO- 03- 183 Medicare Skilled Nursing Facility Payments

This appendix describes the data and methods used to calculate margins for
SNFs as well the analyses of factors affecting margins.

In general, a SNF margin is the difference between its payments and its
costs, divided by payments; this ratio is expressed as a percentage. Using
this definition, a total margin for a SNF is based on the difference
between its total payments* derived from all payers* and its total costs.
A

Medicare margin for a SNF is based on the difference between its Medicare
payments and its reported costs of serving Medicare patients. 1 We report
the median margins for freestanding and hospital- based SNFs,

as well as for subgroups (for example, not- for- profit freestanding
SNFs). 2 We computed Medicare and total margins for freestanding and
hospitalbased SNFs using methods similar to those developed by MedPAC and
CMS's Office of the Actuary (OACT). Our primary data sources for SNF
payments and costs used in calculating Medicare and total margins were the
1997 through 2000 Medicare SNF cost report files maintained by CMS. 3 Our
methods differed slightly from those used by MedPAC and OACT with

respect to the definition of outliers and the application of an adjustment
to Medicare costs.

Definition of outliers. The data for some SNFs must be excluded from the
analysis because they result in outliers* implausibly high or low margins
that suggest data error. To identify outliers, MedPAC uses a method based
on percentiles. We used a standard statistical distribution (lognormal)
and removed SNFs where margins were plus or minus 3 standard deviations or
more from the mean. We used this method because it improved our ability

to detect and eliminate extreme values. Application of cost adjustment.
MedPAC adjusted 1999 cost data for freestanding SNFs that, after the
implementation of the SNF PPS, changed 1 The definition of Medicare costs
excludes nonallowable costs such as patients' telephones and personal
laundry services. 2 MedPAC primarily reports revenue- weighted, or
aggregate, margins. These are calculated

by summing the revenues of all providers in a group (for example,
freestanding SNFs) and separately summing these providers' costs. Using
these sums, a margin is calculated for the group. Reporting revenue-
weighted margins is consistent with MedPAC's focus on the adequacy of
Medicare's aggregate payments for SNF care. 3 Each annual cost report file
contains data for SNFs' cost reporting periods beginning on or after the
start date of the relevant federal fiscal year. Appendix I: Methodology
for Calculating and

Analyzing SNF Margins Data and Methods Used in Calculating Margins

Appendix I: Methodology for Calculating and Analyzing SNF Margins Page 26
GAO- 03- 183 Medicare Skilled Nursing Facility Payments

from certifying a portion of their beds for use by Medicare patients to
certifying all or most their beds for Medicare patients. This increase in
certified beds resulted in the average cost per day reflecting the cost
experience of a broader range of patients, many of whom may not have
received skilled care. The adjustment made data more comparable over time
by making costs for 1999 more similar to costs in 1997 and 1998, which
were based on a larger share of patients needing a SNF level of care.

To better ensure comparability of cost data across time, we made an
adjustment similar to MedPAC's. Following MedPAC's approach, we identified
freestanding SNFs for which this adjustment should be made by examining
the change between years in the number of Medicare- certified beds. If the
number of certified beds increased 50 percent or more, over 90

percent of the SNF's beds were Medicare- certified, and certain other
conditions were met, MedPAC adjusted the SNF's routine costs. We used
similar criteria: If a SNF had over 90 percent of its beds certified as
Medicare and if, in addition, this percentage had changed by more than 30
percentage points from 1998, we adjusted the SNF's routine costs. The
adjustment raised or lowered routine costs based on the pre- PPS ratio of
Medicare SNF routine costs per day to the entire facility's routine costs
per day. We applied this adjustment in both 1999 and 2000. We found that
our criteria identified about 10 percent of SNFs in 1999 and 18 percent in
2000 for which the adjustment was appropriate. Without the adjustment, the
median Medicare margin for freestanding SNFs in 1999 would have been 10. 2
percent rather than 8.4 percent, and the median Medicare margin in 2000
would have been 21.7 percent rather than 18. 9 percent.

A more refined measure of the routine costs attributable to Medicare
patients in all SNFs would reflect the difference in nursing needs between
Medicare patients and other patients in the facility. To test the impact
of such an adjustment, we used patient- specific data on services from
CMS's nursing home minimum data set (MDS) to approximate the difference in
nursing needs between Medicare patients and other patients for each SNF.

Using this estimate, we adjusted the portion of total facility routine
costs attributable to employee wages and benefits, which we used as a
proxy for nursing costs. Based on this analysis, we estimate that using a
more refined measure of Medicare costs would likely have reduced the
median Medicare margin we reported for freestanding SNFs by between 0.6
and 1.6 percentage points. This adjustment would not affect SNF total
costs or

total margins. Refining the Measurement

of Medicare Costs

Appendix I: Methodology for Calculating and Analyzing SNF Margins Page 27
GAO- 03- 183 Medicare Skilled Nursing Facility Payments

The SNF cost report files we used to calculate 1999 and 2000 Medicare
margins were the most current files as of May 2, 2002. The 1999 and 2000
files differed with respect to their completeness. The 1999 file was over
97 percent complete, while the 2000 file was 80 percent complete. 4 After

excluding freestanding SNFs that had outliers or lacked key data,
including data necessary to adjust routine costs, 7,805 facilities were
available for analysis in 1999 and 6,975 facilities in 2000. After
exclusions, 1,506 hospital- based SNFs were available for analysis in
1999. The 2000 file contained very few records for hospital- based SNFs;
as a result, we could

not reliably calculate and report 2000 margins for these providers. Table
10 shows that the distribution of freestanding SNFs is similar in both
years for type of ownership, location (urban versus rural), and census
region. Compared to the 1999 file, the 2000 file has more SNFs that
provide 4,000 or more days of care to Medicare beneficiaries and
correspondingly fewer that provide less than 1,500 days of care.

4 A large proportion of these cost reports* 36 percent in 1999 and 63
percent in 2000* are "as submitted," meaning that, although they have
passed initial automated checks and edits, final payment has not been
made. Characteristics of the Files

Used to Calculate Margins

Appendix I: Methodology for Calculating and Analyzing SNF Margins Page 28
GAO- 03- 183 Medicare Skilled Nursing Facility Payments

Table 10: Distribution of Freestanding SNFs for Which Margins Were
Calculated, 1999 and 2000 Numbers in percent Year a 1999 b 2000 c

Type of ownership For- profit 79 80 Not- for- profit 18 17 Government 3 3
Location Urban 63 63

Rural 37 37 Census region Northeast 15 14

South 34 35 Midwest 33 33 West 18 18 Medicare days Under 1,500 32 29

1,500 * 2,499 23 22 2,500 * 3,999 22 23 4,000 or more 23 26

Source: GAO analysis of Medicare cost report data. a Year refers to each
SNF's cost reporting year, which corresponds to its fiscal year that
begins during the federal fiscal year.

b Based on 7,805 facilities, after excluding SNFs that lacked key data or
were outliers. c Based on 6,975 facilities, after excluding SNFs that
lacked key data or were outliers. To account quantitatively for factors
that potentially influence SNF margins, we analyzed SNF margins using
multiple regression. This statistical technique accounts for variation in
margins by estimating the separate contribution of each of several
explanatory factors included in the analysis, while controlling for the
effect of all other included factors. For freestanding SNFs, we estimated
separate regressions for Medicare margins and for total margins. Each
regression included contextual factors, such as the number of SNFs in a
geographic area and the state in which the SNF was located, and individual
factors, such as each SNF's

proportion of Medicaid patients, its occupancy rate, and whether it was
for- profit. We report only results that are statistically significant.
Methods for Analysis of

Factors Affecting SNF Margins

Appendix II: Medicare Margins of Freestanding SNFs by Selected
Characteristics, 1999 and 2000

Page 29 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

Numbers in percent Year a 1999 b 2000 c

Type of ownership For- profit 11.7 20.4 Not- for- profit -1.4 11.1
Government -13.9 8. 2 Location Urban 10.7 19.4

Rural 5.2 18.0 Census region Northeast 4.3 15.7

South 11.6 19.8 Midwest 4.6 17.7 West 15.3 21.0 Medicare days Under 1,500
3.2 16.1

1,500 * 2,499 8.0 18.9 2,500 * 3,999 10.0 19.4 4,000 or more 12.8 20.2

Source: GAO analysis of Medicare cost report data. a Year refers to each
SNF's cost reporting year, which corresponds to its fiscal year that
begins during the federal

fiscal year. b Based on 7,805 facilities, after excluding SNFs that lacked
key data or were outliers. c Based on 6,975 facilities, after excluding
SNFs that lacked key data or were outliers. Appendix II: Medicare Margins
of Freestanding SNFs by Selected

Characteristics, 1999 and 2000

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

Page 30 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

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

Appendix IV: GAO Contacts and Staff Acknowledgments

Page 31 GAO- 03- 183 Medicare Skilled Nursing Facility Payments

Jonathan Ratner, (202) 512- 7107 Phyllis Thorburn, (202) 512- 7012

Contributors to this report were Dae Park and Eric Wedum. Appendix IV: GAO
Contacts and Staff

Acknowledgments GAO Contacts Acknowledgments

Related GAO Products Page 32 GAO- 03- 183 Medicare Skilled Nursing
Facility Payments

Skilled Nursing Facilities: Available Data Show Average Nursing Staff Time
Changed Little after Medicare Payment Increase. GAO- 03- 176. Washington,
D. C.: November 13, 2002.

Skilled Nursing Facilities: Providers Have Responded to Medicare Payment
System by Changing Practices. GAO- 02- 841. Washington, D. C.: August 23,
2002.

Nursing Homes: Quality of Care More Related to Staffing than Spending.

GAO- 02- 431R. Washington, D. C.: June 13, 2002.

Skilled Nursing Facilities: Services Excluded from Medicare's Daily Rate
Need to be Reevaluated. GAO- 01- 816. Washington, D. C.: August 22, 2001.

Nursing Homes: Aggregate Medicare Payments Are Adequate Despite
Bankruptcies. GAO/ T- HEHS- 00- 192. Washington, D. C.: September 5, 2000.

Skilled Nursing Facilities: Medicare Payment Changes Require Provider
Adjustments but Maintain Access. GAO/ HEHS- 00- 23. Washington, D. C.:
December 14, 1999.

Skilled Nursing Facilities: Medicare Payments Need to Better Account for
Nontherapy Ancillary Cost Variation. GAO/ HEHS- 99- 185. Washington, D.
C.: September 30, 1999. Related GAO Products

(290145)

The General Accounting 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.

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 fulltext 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 daily E- mail alert for newly released products* under the
GAO Reports heading.

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. General Accounting 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

Contact: Web site: www. gao. gov/ fraudnet/ fraudnet. htm E- mail:
fraudnet@ gao. gov Automated answering system: (800) 424- 5454 or (202)
512- 7470 Jeff Nelligan, managing director, NelliganJ@ gao. gov (202) 512-
4800

U. S. General Accounting Office, 441 G Street NW, Room 7149 Washington, D.
C. 20548 GAO*s Mission Obtaining Copies of

GAO Reports and Testimony

Order by Mail or Phone To Report Fraud, Waste, and Abuse in Federal
Programs Public Affairs
*** End of document. ***