Medicare Hospice Care: Modifications to Payment Methodology May  
Be Warranted (15-OCT-04, GAO-05-42).				 
                                                                 
The Medicare hospice benefit provides care to patients with a	 
terminal illness. For each patient, hospices are paid a per diem 
rate corresponding to one of four payment categories, which are  
based on service intensity and location of care. Since		 
implementation in 1983, the payment methodology and rates have	 
not been evaluated. The Medicare, Medicaid, and SCHIP Balanced	 
Budget Refinement Act of 1999 directed GAO to study the 	 
feasibility and advisability of updating Medicare's payment rates
for hospice care. In this report, GAO (1) compares freestanding  
hospices' costs to Medicare payment rates and (2) evaluates the  
appropriateness of the per diem payment methodology. Because of  
Medicare data limitations, it was not possible to compare actual 
payments to costs or examine the services provided to each	 
patient.							 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-42						        
    ACCNO:   A13002						        
  TITLE:     Medicare Hospice Care: Modifications to Payment	      
Methodology May Be Warranted					 
     DATE:   10/15/2004 
  SUBJECT:   Health care programs				 
	     Home health care services				 
	     Hospital care services				 
	     Medical economic analysis				 
	     Medical services rates				 
	     Patient care services				 
	     Per diem allowances				 
	     Health care costs					 
	     Comparative analysis				 
	     Payments						 
	     Medicare program					 

******************************************************************
** 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-05-42

United States Government Accountability Office

GAO

Report to the Committee on Finance, U.S.

     Senate, and the Committee on Ways and Means, House of Representatives

October 2004

MEDICARE HOSPICE CARE

             Modifications to Payment Methodology May Be Warranted

                                       a

GAO-05-42

[IMG]

October 2004

MEDICARE HOSPICE CARE

Modifications to Payment Methodology May Be Warranted

                                 What GAO Found

Using Medicare cost reports from freestanding hospices, GAO determined
that the per diem payment rate for all hospice care was about 8 percent
higher than the estimated average per diem cost of providing care in 2000,
and over 10 percent higher in 2001. However, the relationship between
payment rates and costs varied across the payment categories and types of
hospices. For all hospice care provided in the home, which accounted for
about 97 percent of care in 2001, GAO estimates that the per diem payment
rate was almost 10 percent higher than average per diem costs in 2000, and
over 12 percent higher in 2001. Small hospices, however, had higher
estimated average per diem costs than medium or large hospices overall and
for each of the four per diem payment categories in 2001.

GAO's analysis indicates that the hospice payment methodology, with rates
based on the historical mix and cost of services, a per diem amount that
varies only by payment category, and a cap on total Medicare payments, may
not reflect current patterns of care. For example, GAO determined that the
relative costs of services, such as nursing care, provided during routine
home care (RHC) have changed considerably since the rates were calculated.
Using limited patient-specific hospice visit data, GAO found that more
visits were provided during the first, and especially last, week of a
hospice stay than during other times in the stay. Finally, few hospices
reached the payment cap, which was intended to limit Medicare hospice
spending.

 Costs of Services Provided During RHC as a Percentage of Total RHC Costs, 1983
                                    and 2001

Percentage

43

20

15

10 5

0 Nursing

             Drugs Social services Home health aide Durable medical

equipment

                          SuppliesOutpatient services

Service

1983

2001 Sources: GAO analysis of CMS data and 1983 hospice final rule (48
Fed. Reg. 56,008 (1983)).

United States Government Accountability Office

Contents

  Letter

Results in Brief
Background
Overall Medicare Hospice Payment Rate Higher Than Freestanding

Hospices' Estimated Average Costs, but Relationship Varied by Payment
Category and Hospice Characteristics Structure of Hospice Payment
Methodology May Not Be Consistent

with Current Service Delivery Conclusions Recommendations for Executive
Action Agency and External Reviewer Comments and Our Evaluation

                                       1

                                      5 6

11

15 19 20 20

Appendix I Scope and Methodology

Appendix II	Comments from the Centers for Medicare & Medicaid Services

  Tables

Table 1: Utilization and Payment Rates for Each Hospice Payment Category 8
Table 2: Estimated Average Per Diem Costs by Payment Category for
Freestanding Hospices, by Size, 2001 14

  Figures

Figure 1: Hospice Provider Types Participating in the 1980-1982
Demonstration and in 2001 (Percentages) 10

Figure 2: Unadjusted Per Diem Payment Rate and Estimated Average Per Diem
Costs for Freestanding Hospices, 2000 and 2001 12

Figure 3: Costs of Services Provided during RHC as a Percentage of Total
RHC Costs, 1983 and 2001 16

Abbreviations

BBRA Medicare, Medicaid, and SCHIP Balanced Budget

Refinement Act of 1999 CHC continuous home care CMS Centers for Medicare &
Medicaid Services DME durable medical equipment GIC general inpatient care
HAA Hospice Association of America HCFA Health Care Financing
Administration IRC inpatient respite care NHPCO National Hospice and
Palliative Care Organization RHC routine home care TEFRA Tax Equity and
Fiscal Responsibility Act of 1982

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. However, 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.

United States Government Accountability Office Washington, DC 20548

October 15, 2004

The Honorable Charles E. Grassley
Chairman
The Honorable Max Baucus
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

Since the Medicare hospice benefit was implemented in 1983,1 neither the
payment methodology nor the payment rates for hospices have been
evaluated; however, there are indications that hospice services and
patients have changed. The payment methodology and rates for the benefit
were developed using cost data collected from 26 hospices providing care
to Medicare patients with terminal cancer under a 1980 to 1982 Health
Care Financing Administration (HCFA)2 demonstration. Since then,
advances in medical technology, such as new palliative drugs,
chemotherapy, and radiation, have influenced the mix of hospice services
provided. In addition, the mean length of a hospice stay has declined
since
1983, and the number of patients with noncancer diagnoses has grown
considerably, representing approximately one-half of all patients
receiving
hospice care in 2000.

In addition to changes in the delivery of hospice care, there has been
substantial growth in the hospice program. From 1992 through 2002,
Medicare payments for hospice care increased fivefold, to about

1Tax Equity and Fiscal Responsibility Act of 1982 (TEFRA), Pub. L. No.
97-248, S: 122, 96 Stat. 324, 356.

2In July 2001, the name of the agency that administers Medicare, HCFA, was
changed to the Centers for Medicare & Medicaid Services. In this report,
we refer to the agency as HCFA when discussing actions it took under that
name.

$4.5 billion, and the number of Medicare patients increased fourfold, to
approximately 640,000. During the same period, the number of
Medicareparticipating hospices grew by almost 90 percent to 2,275.

The Medicare hospice benefit was designed to provide patients who have a
terminal illness with comfort and pain relief, as well as emotional and
spiritual support, generally in a home setting. By law, Medicare hospice
services include nursing care, counseling, and home health aide services,
as well as drugs and medical supplies.3 Hospice services are delivered by
providers that operate as freestanding entities or are based in hospitals,
home health agencies, or skilled nursing facilities. Under the law,
Medicare pays hospices a daily rate4 that covers all services provided to
the patient, except for physician services.5 HCFA developed four payment
categories, with corresponding per diem payment rates to reflect variation
in service intensity and the location of service.6 These categories, as
set out in Medicare regulations, are routine home care (RHC), a typical
day of home care; continuous home care (CHC),7 home care that is provided
during periods of patient crisis for at least 8 hours in a 24-hour period;
inpatient respite care (IRC), care in an inpatient facility for a short
period to provide respite for primary caregivers; and general inpatient
care (GIC), inpatient care to treat symptoms that cannot be managed in
other settings. In 1986, annual updates to the rates for the four payment
categories were set in law.8 A hospice is paid for each day that a patient
is under its care, regardless of whether the patient receives a visit or
other service on that day. Medicare pays the RHC payment category rate for
days when a patient does not receive CHC, IRC, or GIC. In 2001, RHC days
accounted for 96 percent of all days provided to hospice patients. By law,
each hospice is subject to an aggregate payment cap,9 which limits its
total

342 U.S.C. S: 1395x(dd)(1) (2000).

442 U.S.C. S: 1395f(i)(1)(B) (2000).

5Medicare pays physicians separately for services provided to hospice
patients.

642 C.F.R. S: 418.302(b)(1) - (4) (2003).

7Although a per diem rate for CHC is determined, Medicare pays hospices an
hourly rate for CHC calculated from the per diem rate.

8Consolidated Omnibus Budget Reconciliation Act of 1985, Pub. L. No.
99-272, S: 9123, 100 Stat. 82, 168 (1986).

942 U.S.C. S: 1395f(i)(2) (2000).

annual Medicare payments to a specified amount based on the number of
Medicare patients it served.

The Medicare, Medicaid, and SCHIP Balanced Budget Refinement Act of 1999
(BBRA) directed us to study the feasibility and advisability of updating
Medicare's payment rates for hospice care.10 Specifically, we (1) compared
hospices' costs to the Medicare payment rates and (2) evaluated the
appropriateness of the per diem payment methodology.

To compare hospices' costs to Medicare payment rates, we used 2000 and
2001 Medicare hospice cost reports, the most recent data available at the
time of our analysis. We confined our analysis to cost reports of
freestanding hospices.11 Hospital-based and home health agency-based
hospices must allocate overhead costs between the hospice and the hospital
or home health agency, which may distort the costs of providing hospice
services.12 For freestanding hospices, the only costs incurred are for
delivering hospice care to patients. We verified the reliability of the
cost report data by comparing descriptive statistics calculated using the
cost reports with those calculated using the Medicare hospice claims for
all hospices from the same years, and we verified the reliability of the
claims data by comparing descriptive statistics we calculated using the
claims with statistics published by the Centers for Medicare & Medicaid
Services (CMS). We excluded certain cost reports; for example, those with
fewer than 11 total patients or an average of less than 1 patient per day.
We determined the remaining data were suitable for our purposes. For
freestanding hospices, we estimated average per diem costs by summing the
total costs for all hospices and dividing by the total number of days of
care for all hospices. We compared the estimated average per diem cost for
freestanding hospices to an unadjusted payment rate based on national
payment rates for each category.13 We also determined average per diem
costs for each payment category using the same methodology and compared
each to the unadjusted payment rate calculated for that

10Pub. L. No. 106-113, App. F, S: 132, 113 Stat. 1501A-321, 1501A-333.

11In 2001, over half of all Medicare hospice patients were served by
freestanding hospices.

12Estimated per diem costs for these hospices were generally much lower
than those of freestanding hospices.

13To calculate a single unadjusted hospice payment rate, we weighted the
four rates by their respective utilization as reported on freestanding
hospice cost reports in each year, accounting for the annual October 1
update in the payment rates.

category. Because RHC and CHC costs are reported in aggregate on the
hospice cost report, we could only report hospices' average combined RHC
and CHC per diem costs.14 We refer to combined RHC and CHC hospice costs
and payment rates as home care costs and payment rates. Because of data
limitations in the Medicare hospice cost report and claims data, we could
not compare actual payments to costs.15 We did not apply the wage
adjustment, which generally adjusts payments up for urban areas and down
for rural areas to account for geographic differences in wages, to the
payment rates. We estimated per diem costs for all freestanding hospices;
for freestanding hospices of different sizes, with size defined by the
number of days of care a hospice provided during the year; and for rural
and urban freestanding hospices.16

To evaluate the appropriateness of the per diem payment methodology, we
compared the proportion of total costs attributable to different services
provided during RHC days when the payment rate was first determined, in
1983, to the proportion in 2001. We used the 2001 freestanding hospice
cost reports to calculate proportions of cost for 2001. Our estimates of
RHC costs in 2001 include CHC costs because RHC and CHC costs are reported
in aggregate on the hospice cost report.17 Using 2002 proprietary data on
patient visits collected by a large, for-profit hospice with multiple
freestanding facilities, which we refer to as patient-specific visit data,
we analyzed the types of visits provided to patients over the course of
their hospice stays. We used these proprietary data because Medicare
hospice cost reports and claims do not contain information on the visits
hospices provide to Medicare patients each day. We confirmed that the
distribution of visits in 2002 was consistent with the distribution in
1997 and 1999 for this hospice. We also calculated descriptive statistics
for this hospice and

14To calculate an unadjusted payment rate that encompassed RHC and CHC, we
weighted the individual rates of these two categories by their respective
utilization as reported on freestanding hospice cost reports in each year.

15Unlike for other providers, Medicare's hospice cost reports do not
include Medicare payment information. In addition, Medicare hospice claims
contain only the total payment for all services provided during the
billing period, including physician services, not the payment for each
hospice payment category.

16We define a hospice as urban if it was located in a county that was in a
metropolitan statistical area and as rural if it was located in a county
outside a metropolitan statistical area, as determined by the Office of
Management and Budget as of June 30, 1999.

17It is likely that CHC costs were a very small proportion of home care
costs, as CHC days accounted for just over 1 percent of total hospice days
in 2001. We refer to proportions of combined RHC and CHC costs as
proportions of RHC costs.

  Results in Brief

compared them to statistics calculated using the Medicare hospice claims.
We determined these data were suitable for our purposes. We interviewed
officials at CMS, two national hospice associations, 18 hospices and
several national independent and academic hospice researchers, and we
conducted a site visit to a freestanding hospice with an inpatient unit.
Our methodology is detailed in appendix I. We did our work from January
2003 through October 2004 in accordance with generally accepted government
auditing standards.

We determined that for freestanding hospices, the unadjusted per diem
payment rate across the four payment categories was about 8 percent higher
than estimated average per diem costs in 2000, and over 10 percent higher
in 2001. We estimate that the home care (RHC and CHC) per diem payment
rate was almost 10 percent higher than average home care per diem costs in
2000, and over 12 percent higher in 2001. The IRC payment rate was almost
53 percent lower than average IRC per diem costs in 2000, and 61 percent
lower in 2001. In both years, the GIC payment rate was about 7 percent
higher than average GIC per diem costs. We could not compare actual
payments to costs because of limitations in the hospice cost report and
claims data. In 2000, we estimate that average per diem costs for small
freestanding hospices were over 13 percent higher than for medium
freestanding hospices and almost 7 percent higher than for large
freestanding hospices. In 2001, average per diem costs for small
freestanding hospices were over 15 percent higher than for medium
freestanding hospices and almost 8 percent higher than for large
freestanding hospices. With the exception of average GIC per diem costs in
2000, small freestanding hospices had higher per diem costs for each
payment category in both years. Small freestanding hospices are more
likely than other hospices to be located in rural areas, and thus are more
likely to receive lower Medicare payments because the wage index
adjustment generally reduces the payments for providers in rural areas.

Several aspects of the hospice per diem payment methodology may not
reflect how hospices currently deliver services. For freestanding
hospices, the share of total costs accounted for by various services
provided during RHC days was considerably different in 2001 than when the
rates were developed in 1983. This suggests that the services delivered or
the resources necessary for those services have changed. The proportions
of RHC costs attributable to nursing, social services, drugs, and medical
equipment have increased, while the proportions attributable to home
health aide services, supplies, and outpatient services have decreased. In
addition, our analysis of proprietary 2002 patient-specific visit data

showed that visit frequency varied during the hospice stay, yet the
payment rates remain the same throughout the stay for each payment
category. Also, the mean length of stay has decreased. Additionally,
hospice officials told us that some CHC and IRC payment policy
requirements limited their ability to deliver services, although data were
not available to determine whether these requirements influenced the
delivery of care to hospice patients. Finally, the annual aggregate cap
was intended to help limit Medicare hospice spending, but the cap amount
was not based on actual hospice costs, and for each year from 1999 through
2002, few hospices reached it.

We recommend that the Administrator of CMS collect comprehensive,
patient-specific data on the utilization and cost of hospice visits and
services. Using these data, the Administrator should determine whether the
hospice payment methodology and payment categories need modifications,
including any special adjustments for small providers. The Administrator
should implement modifications that would not require a change in Medicare
law, and submit a legislative proposal to the Congress for those that do.
In commenting on a draft of this report, CMS stated that it agreed with
our recommendations and intends to use our findings to supplement and
reinforce preliminary evaluations the agency has made and future studies
that are planned. In responding to our recommendation that it collect
comprehensive, patient-specific data on hospice visits and services and
the costs of these services and reevaluate the payment methodology, CMS
stated that it recognized the need for these types of analyses but that
its research funding is limited. In its comments, CMS also stated that it
believed additional information is needed before modifying the payment
methodology because of the higher average costs of small hospices. We have
clarified our conclusion to indicate the need for comprehensive,
patient-specific data on the visits and services delivered by hospices and
the costs of these services to inform any changes to the payment
methodology. Industry representatives who reviewed a copy of this draft
generally agreed with our findings and recommendations.

Background 	Medicare hospice benefit services include nursing services,
services provided by a physician to a hospice, drugs and medical supplies
necessary for treating pain and other symptoms of a terminal illness, as
well as dietary, spiritual, and bereavement counseling; medical social
worker services; homemaker services; and short-term inpatient care both to
provide respite for caregivers and to treat a patient's symptoms.
Volunteers are an important resource in delivering hospice care; Medicare
requires each hospice to have volunteers provide services equal to at
least

5 percent of the total paid patient care hours.18 The specific services
that a patient should receive are outlined in a plan of care, vary based
on the type and intensity of the patient's symptoms and psychosocial needs
and the needs of the patient's caregiver, and may vary throughout the
hospice stay as the patient's condition changes.

                              Hospice Eligibility

To be eligible for the Medicare hospice benefit, a patient must be
certified by a physician as having a life expectancy of 6 months or less
if his or her terminal illness runs its normal course. A Medicare patient
who elects hospice care must waive Medicare coverage for all other
services related to the terminal illness, although the patient retains
coverage for services to treat other conditions. A patient may opt out of
the hospice benefit and return to traditional Medicare at any time; a
patient may also reelect hospice coverage at a later date. While there is
no limit on the number of days an individual can receive hospice care, the
prognosis of the patient's terminal illness must be reaffirmed after the
first 90 days, the first 180 days, and then every 60 days thereafter.

    Medicare Hospice Payment

Under the law, Medicare pays hospices a daily rate that covers all
services provided to the patient. HCFA developed four hospice per diem
payment categories, which reflect the intensity of the services and the
location of service delivery. In 1986, annual updates to the rates for the
four payment categories were set in law. A typical day of care provided in
a patient's residence is paid as RHC, and in 2001, the vast majority of
hospice care days, 96 percent, were billed as RHC (see table 1). Unless a
hospice provides CHC, IRC, or GIC, it is paid the RHC rate for each day
the patient is under its care. Hospice care delivered during periods of
crisis can be paid as CHC if the care is provided in the home for at least
8 hours within a 24-hour period beginning at midnight and at least half
the care hours are delivered by a nurse. To provide respite for primary
caregivers, IRC can be provided for up to 5 consecutive days in an
inpatient setting. Inpatient care for symptoms that cannot be treated in
the patient's residence is paid as GIC. Hospices provide the care in their
own inpatient units or arrange with hospitals, skilled nursing facilities,
or other inpatient facilities to provide these services. The payment rate
is adjusted by a wage index, which varies based on the patient's
residence, to account for geographic differences in wage costs.

1842 C.F.R. S: 418.70(e) (2003).

Table 1: Utilization and Payment Rates for Each Hospice Payment Category

                            Percentage of total  
           Payment category  hospice days, 2001            Payment rate, 2004 
                        RHC                 96.0                  $118.08/day 
                        CHC                  1.2 $28.72/hr, up to $689.18/day 
                        IRC                  0.2                  $122.15/day 
                        GIC                  2.6                  $525.28/day 

Source: GAO analysis of CMS data and Medicare payment rates.

The hospice payment categories and their corresponding payment rates were
developed from cost data from the 26 hospices that participated in the
1980 to 1982 Medicare demonstration. To calculate the payment rates, HCFA
used cost data to identify the cost factors that contributed to providing
hospice services and summed the mean cost per day of each cost factor for
each of the four categories of hospice care. The costs of bereavement and
volunteer services were not included in the rates.19

By law, hospices are subject to an annual aggregate Medicare payment cap
that was meant to ensure that payments for hospice care would not exceed
what Medicare would have paid if patients had been treated in a
traditional setting, such as a hospital.20 Total annual payments to a
hospice may not exceed a per-patient amount multiplied by the number of
Medicare patients who received care from that hospice during the year. The
2004 cap amount for the 12 months beginning November 1, 2003, is
$19,635.67 per Medicare patient.21

19By statute, hospices must provide bereavement counseling, although
Medicare is prohibited from paying hospices for such services (42 U.S.C.
S:S: 1395x(dd)(1)(H), 1395x(dd)(2)(A)(i), and 1395f(i)(1)(A) (2000)). The
statute also specifies that hospices must use volunteers in the provision
of care (42 U.S.C. S: 1395x(dd)(2)(E)(i) (2000)). HCFA, however, did not
include volunteer costs (such as volunteer training and program
coordination costs) in setting the payment rates.

2042 U.S.C. S: 1395f(i)(2) (2000).

21Each year, the cap amount is updated by the annual percentage change in
the medical care component of the Consumer Price Index for all urban
consumers.

    Changes in Hospice Care

Hospice patients, services, and providers have changed since the
demonstration. For example, the mean patient length of stay at hospices
participating in the Medicare demonstration from 1980 to 1982, was 70
days; in 2001, the mean length of stay was about 50 days. While
demonstration costs were based only on Medicare patients with cancer
diagnoses, patients with noncancer diagnoses, who may require a different
mix of services, represented approximately half of all hospice patients in
2000. In addition, hospice providers have stated that advances in
end-oflife care, notably new, more costly pain-management drugs and
palliative chemotherapy and radiation, have increased the costs of
providing care to certain types of patients.

The mix of hospice providers today differs from the provider types in the
demonstration. In the demonstration, the predominant type of hospice
provider was hospital-based, whereas in 2001, the predominant type was
freestanding (see fig. 1). More recently, the proportion of for-profit
hospices increased from almost 13 percent of all hospices in 1992 to
almost 28 percent in 2001, and the percentage of hospices serving patients
primarily living in rural areas rose from 32 in 1992 to 38 in 2001. In
addition, the number of hospices participating in Medicare grew from 1,208
in 1992 to 2,275 in 2002, the most recent data available.

Figure 1: Hospice Provider Types Participating in the 1980-1982
Demonstration and in 2001 (Percentages)

Percentages 45

42 42

40

35

30

25

20

15

10

5

0

Freestanding Home health Hospital Skilled nursing agency facility

Type of hospice

1980-1982

2001

Sources: 66 Fed. Reg. 49,454 (2001) and Vincent Mor, David S. Greer, and
Robert Kastenbaum, The Hospice Experiment (Baltimore: The Johns Hopkins
University Press, 1988), 33-34.

  Overall Medicare Hospice Payment Rate Higher Than Freestanding Hospices'
  Estimated Average Costs, but Relationship Varied by Payment Category and
  Hospice Characteristics

We determined that for freestanding hospices, the unadjusted per diem
payment rate across the four payment categories was about 8 percent higher
than estimated average per diem costs in 2000, and over 10 percent higher
in 2001. For the payment categories, we estimate that the home care (RHC
and CHC) per diem payment rate was almost 10 percent higher than average
home care per diem costs in 2000, and over 12 percent higher in 2001. We
estimate that the IRC payment rate was almost 53 percent lower than
average IRC per diem costs in 2000, and 61 percent lower in 2001. In both
years, the GIC payment rate was about 7 percent higher than average GIC
per diem costs. In 2000, we estimate that average per diem costs for small
hospices were over 13 percent higher than for medium hospices and almost 7
percent higher than for large hospices. In 2001, average per diem costs
for small hospices were over 15 percent higher than for medium hospices
and almost 8 percent higher than for large hospices. With the exception of
average GIC per diem costs in 2000, small hospices also had higher average
per diem costs than medium or large hospices for each payment category.

    Unadjusted Payment Rate Higher Than Freestanding Hospices' Estimated Average
    Costs

Medicare's hospice payment rate, across the four payment categories and
unadjusted for geographic differences in wages, was higher than
freestanding hospices' estimated average per diem cost. The unadjusted
payment rate was about 8 percent higher than average per diem costs in
2000, and over 10 percent higher in 2001 (see fig. 2). The 25 percent of
hospices with the lowest average per diem costs had costs that were at
least 27 percent below the unadjusted payment rate in 2000, and at least
31 percent below the unadjusted payment rate in 2001. However, in 2000,
average per diem costs for almost 34 percent of freestanding hospices and,
in 2001, almost 32 percent of freestanding hospices, were higher than the
unadjusted per diem rate. The costs of individual hospices differ
depending on the mix of services provided. In addition, the payments to
individual hospices differ because of the wage adjustment and the mix of
payment categories billed.

Figure 2: Unadjusted Per Diem Payment Rate and Estimated Average Per Diem
Costs for Freestanding Hospices, 2000 and 2001

Dollars 120 116

100

80

60

40

20

0 2000 2001 Year

Unadjusted per diem payment rate

Estimated average per diem cost

Sources: GAO analysis of CMS data and Medicare hospice payment rates.

Note: The unadjusted payment rate equals the rates of the four payment
categories weighted by their respective utilization as reported in
freestanding hospice cost reports in each year, accounting for the annual
October 1 update in the payment rates.

We could not determine the relationship between payments and actual costs
for individual hospices because of data limitations in the hospice cost
reports and claims data. Unlike those for other providers, Medicare's
hospice cost reports do not include Medicare payment information. In
addition, Medicare hospice claims data contain only the total payment for
all services provided during the billing period, including physician
services, not the payment for each hospice payment category.

    Relationship between Estimated Costs and Unadjusted Payment Rates Varied
    among Payment Categories

The specific relationship between payment rates and costs for freestanding
hospices varied among payment categories. For home care (RHC and CHC)
days, we estimate that in 2000, the unadjusted per diem payment rate for
freestanding hospices was almost 10 percent higher than the average per
diem cost of over $92. In 2001, the per diem payment rate was over 12
percent higher than the average home care per diem cost of over $96.
Nonetheless, about 35 percent of freestanding hospices in 2000, and over
32 percent in 2001, had average home care per diem costs that were higher
than the home care per diem payment rate.

We estimate that in 2000, the unadjusted IRC per diem payment rate for
freestanding hospices was almost 53 percent lower than the average IRC per
diem cost of about $218. In 2001, the IRC per diem payment rate was over
61 percent lower than the average IRC per diem cost of over $279. However,
the GIC per diem payment rate was higher than average GIC per diem costs
for freestanding hospices; it was over 7 percent higher than costs in both
years. In addition, average per diem costs for IRC and GIC varied widely
among freestanding hospices. Our estimates of average IRC and GIC per diem
costs may understate actual costs because of data limitations.

IRC costs may be much higher than the IRC payment rate because the hospice
continues to provide services and visits to the patient in addition to
paying the inpatient facility.22 Our analysis of the proprietary 2002
patient-specific visit data found that the number and type of visits
provided per day to patients during IRC days were comparable to the number
and type of visits per day to patients during RHC days. In 2001, IRC
accounted for 0.2 percent of hospice days of care.

22During GIC, hospices also provide services and visits to the patient in
addition to paying the inpatient facility.

    Small Hospices Had Higher Per Diem Costs and Likely Lower Payments

We estimate that for 2000 and 2001, small freestanding hospices had higher
average per diem costs than medium and large freestanding hospices.23 In
2000, average per diem costs for small hospices were more than 13 percent
higher than for medium hospices and almost 7 percent higher than for large
hospices. In 2001, average per diem costs for small hospices were more
than 15 percent higher than for medium hospices and almost 8 percent
higher than for large hospices (see table 2). With the exception of
average GIC per diem costs in 2000, small hospices' average per diem costs
were higher than medium and large hospices' costs for each individual
payment category for both years. Cost disparities across providers of
different sizes were greatest for IRC and GIC.

Table 2: Estimated Average Per Diem Costs by Payment Category for
Freestanding Hospices, by Size, 2001

                                            Home care IRC average GIC average 
           Hospice size Average costs a average costs       costs       costs 
                  Small          $116            $112        $692        $564 
                 Medium           101              94         254 
                  Large           108              96         264 

Source: GAO analysis of CMS data.

Notes: The size of the hospice is based on the number of days of care it
provided during the year. Small hospices were those that reported total
days of care less than the 25th percentile of all hospices' total days of
care. Medium hospices were those that reported total days of care equal to
or greater than the 25th percentile and less than or equal to the 75th
percentile of all hospices' total days of care. Large hospices were those
that reported total days of care greater than the 75th percentile of all
hospices' total days of care.

aHome care costs include RHC and CHC costs.

As small freestanding hospices are more likely than other hospices to be
located in rural areas, they are more likely to receive lower Medicare
payments because the wage index adjustment generally reduces the payment
rates for providers in rural areas. In 2001, 60 percent of small
freestanding hospices were located in rural areas, while 35 percent of

23The size of the hospice in each year is based on the number of days of
care it provided that year. Small hospices were those that reported total
days of care less than the 25th percentile of all hospices' total days of
care. Medium hospices were those that reported total days of care equal to
or greater than the 25th percentile and less than or equal to the 75th
percentile of all hospices' total days of care. Large hospices were those
that reported total days of care greater than the 75th percentile of all
hospices' total days of care.

medium freestanding hospices and 10 percent of large freestanding hospices
were located in rural areas.24

  Structure of Hospice Payment Methodology May Not Be Consistent with Current
  Service Delivery

The structure of the hospice payment system may not reflect how hospices
currently deliver services. For example, our analysis of the relative
costs for freestanding hospices for different services provided during RHC
days, the most common payment category, showed they have changed
considerably since the payment rate was initially calculated, suggesting
that the services delivered or the resources necessary for those services
have changed over the years. In addition, our analysis of proprietary 2002
patient-specific visit data showed that visit frequency varied during the
hospice stay, although the rate for each payment category does not. Also,
the mean length of stay has decreased. Hospice officials raised concerns
about some of the payment policy requirements for CHC and IRC, although
our analysis of the limited available data could not confirm that the
requirements restrict hospices' ability to provide care. Finally, the
annual aggregate cap was intended to help limit Medicare spending for all
hospices, but it was not based on actual hospice costs, and for each year
from 1999 through 2002, few hospices reached it.

    Relative Costs of Services Provided during RHC Differ from When Payment Rate
    Was Developed

The relative costs of services in 2001 have changed considerably since the
payment rate was developed in 1983, suggesting that the services delivered
or the resources necessary for those services have changed over time.
Specifically, the proportions of RHC costs attributable to nursing, drugs,
social services, and durable medical equipment (DME)25 have increased,
while the proportions attributable to home health aide services, supplies,
and outpatient services have decreased (see fig. 3).26 In our analysis,
this pattern is present across freestanding hospices of all sizes and
locations. The largest cost increase occurred for drugs, which rose from 3
to 15 percent of RHC costs over this period. Hospice officials we spoke
with

24While urban freestanding hospices also had higher estimated average per
diem costs compared to rural freestanding hospices in our analysis, urban
hospices generally receive higher payments than rural hospices to account
for the higher wages they must pay.

25DME is equipment prescribed by a physician for the medical purpose of
addressing a patient's illness or injury that can withstand repeated use,
such as respirators, crutches, and oxygen equipment.

26RHC costs for 2001 are slightly overestimated, as they include CHC
costs. CHC days accounted for just over 1 percent of total hospice days in
2001.

stated that this increase was due in part to the introduction of new, more
costly medications. Some stated that drugs have become one of their
greatest cost pressures.

  Figure 3: Costs of Services Provided during RHC as a Percentage of Total RHC
                              Costs, 1983 and 2001

Percentage

                                       43

Nursing

    Drugs Social services Home health aide DME Supplies Outpatient services

Service

1983

2001 Sources: GAO analysis of CMS data and 1983 hospice final rule (48
Fed. Reg. 56,008 (1983)).

Notes: Percentages from 1983 are based on cost data collected in the 1980
to 1982 hospice demonstration. RHC costs for 2001 are slightly
overestimated, as they include CHC costs. CHC days accounted for just over
1 percent of total hospice days in 2001.

Fewer Days of Lower Visit Hospice visits are particularly concentrated at
the beginning and end of a Frequency with Shorter hospice stay, yet the
payment rate of each category does not vary Mean Length of Stay throughout
a hospice stay. Our analysis of the 2002 patient-specific visit

data showed that patients have a higher mean number of visits per day

during the first, and especially the last, week of a stay.27 As a result,
the costs of care are higher both at the beginning and end of a hospice
stay. Officials from almost all hospices with whom we spoke also reported
this pattern. They told us that at the beginning of a hospice stay they
provide more visits because the patient's symptoms, including pain, must
be stabilized and the family must be educated about the patient's care.
Near the end of life, hospice officials indicated that the patient's
symptoms and needs change, usually requiring more hospice management, and
the family often needs additional psychosocial support. Our analysis of
the 2002 patient-specific visit data showed that patients with a length of
stay of 2 weeks or less had a higher mean number of visits per day than
patients with a length of stay greater than 2 weeks.28 Hospice officials
we spoke with stated that patients who are in hospice care a short time
are relatively more costly on a per diem basis because there are fewer
days of lower visit frequency to balance the higher costs of the days with
more visits at the beginning and end of the stay.

In 1983, the Medicare hospice per diem payment amounts accounted for the
variation in daily hospice costs because they were based on the mean daily
costs incurred by the hospices in the demonstration over a mean hospice
stay of 70 days. However, hospice stays are considerably shorter now; the
mean length of stay was 50 days in 2001. Mean daily costs may now be very
different because of the change in the length of stay. No data are
available, however, to compare costs at different points during a stay or
for stays of different lengths.

    Providers Raised Concerns about Payment Category Requirements

Hospice officials we spoke with raised concerns about some of the policy
requirements for particular payment categories, although our analysis of
the available data could not confirm their concerns. For example, to bill
for CHC, Medicare requires that a nurse provide at least half of billed
CHC

27We also analyzed patient visit frequency by certain patient
characteristics. Using 2002 patient-specific visit data, we found no
substantial differences in the mean number of visits per day among
patients of different ages, diagnoses, numbers of secondary conditions, or
residential locations. (The patient-specific visit data categorized
patients as residing in their home, a nursing home, an assisted living
facility, or a hospital. In this analysis, patients residing in a hospital
had a slightly lower mean number of visits per day than patients residing
in other locations.)

28We also found that patients with a length of stay of 2 weeks or less had
a higher proportion of CHC days, representing the delivery of more
intensive, and therefore more costly, services, compared to patients with
longer stays.

hours. Hospice officials stated that this could restrict the hospice's
ability to provide the most appropriate care when a social worker was a
more appropriate caregiver than a nurse.

The officials were also concerned that the 8-hour minimum required for
billing CHC payment, counted from midnight of one day until midnight of
the next, could restrict their ability to bill for CHC. For example, if a
patient dies in less than 8 hours or the hospice provides 8 hours of
services over 2 calendar days, the hospice must bill for RHC. Our analysis
of 2001 Medicare hospice claims indicated that the mean number of hours
provided on a CHC day was 18 hours, considerably above the 8-hour minimum.
Similarly, our analysis of the 2002 patient-specific visit data from one
large, freestanding hospice showed a mean of 20 hours provided on each CHC
day. Therefore, instances of continuous care hours that fall just short of
8 hours, for which a hospice cannot bill CHC hours, do not occur often
based on the patient-specific visit and claims data.

Hospice officials we spoke with also stated that the statutory requirement
that respite care be provided in an inpatient setting might hinder its
use.29 Specifically, they stated that while primary caregiver respite is
important, enabling patients to remain at home rather than moving them to
an inpatient facility is also important; primary caregivers may not take
respite in order to avoid moving the patient to an inpatient facility. Few
hospices we spoke with currently provide home respite care for extended
periods. They said this is largely because the costs are higher than the
RHC payment rate, which is the payment category the hospices must bill for
these services. Data related to home respite care are not available,
although it is likely that the costs of providing 24 hours of home respite
care would be higher than RHC costs.

Few Hospices Affected by According to our analysis of data from the
regional home health

Annual Aggregate Cap 	intermediaries, the contractors responsible for
processing and paying Medicare hospice claims, less than 2 percent of all
hospice providers reached the annual aggregate payment cap each year from
1999 through 2002.30 In 1982, the Congress required HCFA to calculate a
cap that limited

2942 U.S.C. 1395x(dd)(1)(G) (2000).

30One regional home health intermediary had jurisdiction over at least 71
percent of the hospices that reached the cap each year from 1999 through
2002. This intermediary had no additional information on why these
hospices reached the cap.

a hospice's total payments to a specific per-patient amount based on the
Medicare costs incurred for patients with cancer during the last 6 months
of life.31 However, a subsequent law enacted before the hospice benefit
was implemented set a per-patient cap amount that was not based on the
cost data;32 for the 12 months beginning November 1, 2003, the cap was
$19,635.67 per Medicare patient. The cap is intended to ensure that
payments for hospice care do not exceed what Medicare would have spent if
patients had been treated in a traditional setting, such as a hospital.
However, it affects few hospices, and therefore may not represent a
meaningful limit. Hospice officials we spoke with who discussed the cap
said it did not affect them.

                                  Conclusions

CMS has not evaluated the hospice per diem payment rates and methodology
since they were developed to determine the relationship between payments
and costs and whether the per diem methodology is consistent with current
patterns of care. There are several indications that hospice payments may
not be appropriately distributed across days of care or types of
providers. The type of care provided during a hospice stay appears to be
different than when the hospice per diem payment rates and methodology
were developed. Comprehensive data are not available, however, to evaluate
the number of visits or costs of services provided during a Medicare
hospice stay. While our analysis of the limited data available indicates
that the overall Medicare payment rate across all payment categories was
above estimated costs, IRC costs were considerably above the payment rate.
Further, small freestanding hospices had substantially higher average per
diem costs than other hospices. As a result, a comprehensive analysis of
patient-specific data may show that modifications to the hospice payment
methodology are warranted. Because the payment rates for the four hospice
payment categories, the per diem methodology, and the cap are set by law,
CMS's ability to make modifications to the payment approach is limited.

31TEFRA, S: 122(i)(1), 96 Stat. 358-359. 3242 U.S.C. S: 1395f(i)(2)
(2000).

  Recommendations for Executive Action

Agency and External Reviewer Comments and Our Evaluation

We recommend the following three actions. First, we recommend that the
Administrator of CMS collect comprehensive, patient-specific data on the
visits and services being delivered by hospices and the costs of these
services. Second, using these data, the Administrator should determine
whether the hospice payment methodology and payment categories need to be
modified, including any special adjustments for small providers. Third,
the Administrator should implement those modifications that would not
require a change in Medicare law and submit a legislative proposal to the
Congress for those that do.

We received written comments on a draft of this report from CMS (see app.
II). We also received oral comments from two groups representing industry
organizations, the Hospice Association of American (HAA) and the National
Hospice and Palliative Care Organization (NHPCO), as well as from the
large, for-profit hospice that provided the patient-specific visit data.

    CMS Comments and Our Evaluation

In commenting on a draft of this report, CMS stated that it agreed with
our recommendations and intends to use our findings to supplement and
reinforce preliminary evaluations the agency has made and future studies
that are planned. In responding to our recommendation that it collect
comprehensive, patient-specific data on hospice visits and services and
the costs of these services, CMS stated that it recognized the need for
this type of analysis. It stated that collection of these data would
require additional research funding, and it is uncertain when such funding
would be available. CMS noted that it has initiated efforts to collect
data on costs with the recent establishment of the hospice cost reports.
CMS stated that it hoped the recommendations in our report could help the
agency in developing a comprehensive research strategy for the hospice
benefit.

In responding to our recommendation that CMS determine whether the hospice
payment methodology and payment categories need to be modified, including
any adjustments for small providers, CMS agreed that the methodology
implemented in 1983 was based on a delivery model that may have changed
since that time. It concurred that the methodology should be reevaluated
to determine its current appropriateness. It again stated that research
funding is limited. CMS agreed that the costs of drugs and other
therapies, the number of hospice beneficiaries with noncancer diagnoses,
and the mean length of stay have all changed since 1983. CMS stated that
we did not demonstrate in the draft report that the provision of these and
other therapies have increased the cost of providing care beyond the
present payment. In the draft report, we stated that there may

be problems with the distribution of hospice payments, but that
comprehensive data are not available to evaluate the number of visits or
costs of services provided during a Medicare hospice stay. As noted in the
draft report, the overall payment rate across all types of care is higher
than our estimate of hospices' overall costs. In its comments, CMS also
raised concerns that we implied that payment methodology changes be made
for small hospices before CMS collects comprehensive data. We have
clarified our conclusion to indicate the need for comprehensive,
patient-specific data on the visits and services delivered by hospices and
the costs of these services to inform any changes to the payment
methodology.

In response to our recommendation that CMS should submit a legislative
proposal to the Congress to implement those modifications that would
require a change in Medicare law, CMS stated that should it determine
changes are necessary, it would evaluate those changes as part of its
overall legislative strategy.

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

    Industry Comments and Our Evaluation

The external reviewers generally agreed with our findings and
recommendations. Comments on specific portions of the draft report
centered on two areas: our scope and methodology and the hospice payment
methodology.

Regarding our scope and methodology, HAA and NHPCO were concerned that we
based our findings on Medicare freestanding hospice cost reports that had
not been audited. The large, for-profit hospice noted that the cost report
is complex and that hospices' accounting systems are not generally
compatible with its structure. Similarly, HAA and NHPCO stated that
hospices may not have had sufficient experience with completing the cost
reports at the time of our review. NHPCO stated that our exclusion of
hospice cost reports with fewer than 11 total patients or an average of
less than 1 patient per day might have excluded a substantial number of
cost reports. In addition, HAA and NHPCO recommended that we include
bereavement counseling costs in our per diem cost calculation. They stated
that although Medicare is precluded from paying hospices for bereavement
counseling, it is a required service, and excluding it from the per diem
cost calculation may misrepresent the amount by which payment rates
exceeded hospice costs.

Regarding reviewers' concerns about our use of unaudited cost reports,
BBRA directed us to examine hospice cost factors. Information on these
factors is available only from cost reports, which CMS has not audited. As
stated in the draft report, we assessed the reliability of the cost
reports by comparing descriptive statistics calculated using the cost
reports with those calculated using the Medicare hospice claims, and found
the data suitable for our purposes. Regarding reviewers' concerns about
data we excluded from our analysis, we excluded 51 of 992, or 5 percent,
of freestanding hospice cost reports in 2000, and 48 of 975, or 5 percent,
of freestanding hospice cost reports in 2001, because they had fewer than
11 total patients or an average of less than 1 patient per day. We
excluded these cost reports because we believe that these hospices either
had too few patients to be representative of all hospices, or may have
been reporting data incorrectly. We do not believe that these represent
substantial numbers of cost reports and consider our exclusion criteria
appropriate. Concerning the comments that we should include bereavement
costs in our per diem cost calculation, as stated in the draft report, we
included only Medicare-reimbursable costs in our calculations. If Medicare
cannot, by law, pay hospices for bereavement services, it is inappropriate
to include them in a per diem cost that is compared to a payment rate that
is not designed to cover these costs. In 2001, in comparison to total
Medicare-reimburseable costs, bereavement costs were small; they were
equal to less than 2 percent of total Medicarereimburseable costs.

Reviewers also commented on the hospice payment methodology. NHPCO stated
that costs on the cost report may not reflect the provision of all
services that could potentially be provided because hospices may manage
their costs to more closely approximate the per diem rate. Although the
provision of additional services may be warranted, hospices cannot pay for
them and therefore do not provide them. HAA and NHPCO stated that
instances of CHC provision that fall close to 8 hours may not seem to
occur often because hospices avoid providing CHC if they know they will
not be able to provide at least 8 hours. However, HAA and NHPCO also
stated that data to determine whether this is the case are not available.

Regarding industry comments on hospice costs and the hospice payment
methodology, we acknowledge that hospices may manage their costs to
closely approximate the per diem rate, and that hospices may not provide
CHC if they know they will not be paid for that level of care. Data are
not available to evaluate whether either of these situations occur.

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

We are sending copies of this report to the Administrator of CMS and
appropriate congressional committees. The report is available at no charge
on GAO's Web site at http://www.gao.gov. We will also make copies
available to others on request.

If you or your staffs have any questions, please call me at (202) 512-7119
or
Nancy A. Edwards at (202) 512-3340. Other major contributors to this
report include Beth Cameron Feldpush, Joanna L. Hiatt, and Gordon W.
Richmond.

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

                       Appendix I: Scope and Methodology

To examine hospice costs and Medicare payments, we used 2000 and 2001
Medicare hospice cost reports, the financial documents that hospices
submit annually to the Centers for Medicare & Medicaid Services (CMS), and
2000 and 2001 Medicare hospice claims data, bills submitted by hospices to
receive Medicare payment. We also used proprietary 2002 patient-specific
visit data from a large for-profit hospice, which has been collecting
these data for its internal use since 1994. We interviewed officials from
CMS and one regional home health intermediary, a contractor responsible
for processing and paying Medicare hospice claims, in addition to
officials from AARP, the Hospice Association of America, the National
Hospice and Palliative Care Organization, and the Visiting Nurse
Associations of America. We also spoke with representatives from 18
hospices, several national independent and academic hospice researchers,
and two physicians who provide hospice care. Finally, we conducted a site
visit to a freestanding hospice with an inpatient unit.

To assess the reliability of the cost report data, we compared descriptive
statistics calculated using the cost reports with those calculated using
the Medicare hospice claims data. Because hospices began submitting cost
reports in 1999, we also compared our calculations from the 2000 cost
reports to those from the 2001 cost reports to ensure that hospices had
provided consistent data. To assess the reliability of the claims data, we
compared descriptive statistics calculated using the claims with
statistics published by CMS. To assess the consistency of the 2002
patient-specific visit data, we verified that the distribution of visits
in the 2002 data was similar to the distribution of visits in 1997 and
1999. In addition, before releasing these data to us, the hospice
performed quality assurance edits, which consisted of confirming that the
data provided to us were identical to the data in its database for more
than 20 randomly selected patients. Finally, we calculated descriptive
statistics and compared them with statistics for all hospices calculated
using the Medicare hospice claims. We determined that the cost report,
claims, and patient-specific data were all suitable for our purposes.

The 2000 and 2001 hospice cost reports were the most recent data available
at the time of our analysis. The Medicare payment methodology is the same
for freestanding and facility-based hospices; however, we confined our
analysis to cost reports of freestanding hospices. We excluded hospital-
based and home health agency-based hospices because we found that their
per diem costs were generally much lower than those of freestanding
hospices, which may result from decisions made by these providers in
allocating overhead costs between the hospital or home health agency and
the hospice. For freestanding hospices, the only costs

Appendix I: Scope and Methodology

incurred are for delivering hospice care to patients. We excluded
freestanding cost reports that reported no or low Medicare utilization,
those that had cost reporting periods of fewer than 10 or greater than 14
months, and those outside the 50 states or District of Columbia. We also
excluded cost reports that had fewer than 11 total patients or an average
of less than 1 patient per day, those with no costs, and those reporting
costs outside three standard deviations of the mean. Our final sample
included 82 percent of all freestanding hospice cost reports in 2000 and
80 percent in 2001.

We calculated freestanding hospices' total Medicare-reimburseable costs by
subtracting nonreimburseable costs, such as bereavement and fundraising,
from total costs. To obtain average per diem costs, we summed total
Medicare-reimburseable costs across all providers and divided by total
hospice days across all providers. In addition, because of the cost report
design, certain inpatient respite care (IRC), general inpatient care
(GIC), and physician costs may be included in our estimate of combined
routine home care (RHC) and continuous home care (CHC), or home care,
costs. As a result, home care costs may be overestimated, which would
result in our understating the amount by which the unadjusted home care
payment rate exceeds average home care per diem costs. Because of the way
cost centers are defined on the cost reports, the costs of IRC and GIC may
be underestimated.

We based the size of a hospice in each year on the number of days of care
it provided that year. Small hospices were those that reported total days
of care less than the 25th percentile of all hospices' total days of care.
Medium hospices were those that reported total days of care equal to or
greater than the 25th percentile and less than or equal to the 75th
percentile of all hospices' total days of care. Large hospices were those
that reported total days of care greater than the 75th percentile of all
hospices' total days of care. We defined a hospice as urban if it was
located in a county that was in a metropolitan statistical area and as
rural if it was located in a county that was not in a metropolitan
statistical area, as determined by the Office of Management and Budget as
of June 30, 1999.

We could not compare the 2000 and 2001 per diem costs we calculated to
actual payments because hospice cost reports do not report Medicare
payment information. In addition, Medicare hospice claims contain only the
total payment for all services provided during the billing period,
including physician services, not the payment for each payment category.
Therefore, we calculated a 2000 and 2001 unadjusted payment rate that

Appendix I: Scope and Methodology

encompassed all payment categories. We did so by weighting the individual
rates of the four payment categories by their respective utilization in
the freestanding hospice cost reports in our final sample in each year.
The costs for home care, combined RHC and CHC, are reported in aggregate
on the hospice cost report. Therefore, we calculated a 2000 and 2001
unadjusted payment rate that encompassed RHC and CHC. We did so by
weighting the individual rates of these two categories by their respective
utilization in the freestanding hospice cost reports in our final sample
in each year. In addition, we weighted the overall unadjusted payment rate
and the unadjusted payment rate for each payment category to account for
the different payment rates in effect during the year. The majority of
freestanding hospices report costs using a calendar year reporting period,
while payment rates are updated on a fiscal year basis, that is, on
October 1 of each year. Therefore, during a calendar year, one payment
rate is in effect from January 1 through September 30 and another from
October 1 through December 31. Our unadjusted payment rates do not account
for the wage adjustment Medicare applies to payments.

To determine the proportion of total cost in 2001 accounted for by each
service, such as nursing or home health aide services, that was included
in the 1983 RHC rate, we grouped the services on the cost report into
categories similar to the 1983 services, and divided by the total cost.
Our estimates of the proportions of 2001 RHC costs include CHC costs
because the costs of RHC and CHC are reported in aggregate on the hospice
cost report. It is likely that CHC costs were a very small proportion of
combined RHC and CHC costs, as CHC days accounted for just over 1 percent
of total hospice days in 2001.

To determine the percentage of total hospice days accounted for by each
payment category and the mean CHC hours per CHC day for all hospices, we
used 2000 and 2001 Medicare hospice claims data, the years that matched
most closely with the cost reports used for our analysis. We excluded from
our analysis patients who were younger than 20 or older than 110 years of
age, who lived outside of the 50 states or the District of Columbia, and
who had total hospice payments that fell below 1 day of care at the lowest
wage-adjusted RHC payment rate and above 1 year of care at the highest
wage-adjusted RHC payment rate. Our final sample included over 98 percent
of all claims in both 2000 and 2001.

To analyze the frequency and types of visits to hospice patients, we used
proprietary 2002 data on Medicare hospice patients collected by a large,
for-profit hospice with multiple freestanding facilities. We determined
the

Appendix I: Scope and Methodology

number of visits per day and the number of nurse, home health aide,
counselor, and other caregiver visits per day for all days and for days
within each of the four payment categories. We also analyzed whether there
were differences in the number of visits per day provided by patient
length of stay, patient residence, diagnosis, number of secondary
conditions, and age and determined the number of visits in the first and
last week of a stay and for the remaining days of a stay.

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

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

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

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

  GAO's Mission

Obtaining Copies of GAO Reports and Testimony

The Government Accountability Office, the audit, evaluation and
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 GAO's Web site (www.gao.gov). Each weekday, GAO posts newly
released reports, testimony, and correspondence on its Web site. To have
GAO e-mail you a list of newly posted products every afternoon, go to
www.gao.gov and select "Subscribe to Updates."

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, Contact:

Waste, and Abuse in Web site: www.gao.gov/fraudnet/fraudnet.htm

E-mail: [email protected] Programs Automated answering system: (800)
424-5454 or (202) 512-7470

Gloria Jarmon, Managing Director, [email protected](202)
512-4400Congressional U.S. Government Accountability Office, 441 G Street
NW, Room 7125 Relations Washington, D.C. 20548

Public Affairs 	Jeff Nelligan, Managing Director, [email protected] (202)
512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

                           PRINTED ON RECYCLED PAPER

                               Presorted Standard
                              Postage & Fees Paid
                                      GAO
                                Permit No. GI00

United States
Government Accountability Office
Washington, D.C. 20548-0001

Official Business
Penalty for Private Use $300

Address Service Requested
*** End of document. ***