Ambulance Services: Medicare Payments Can Be Better Targeted to  
Trips in Less Densely Populated Rural Areas (19-SEP-03, 	 
GAO-03-986).							 
                                                                 
The Centers for Medicare & Medicaid Services (CMS) recently	 
implemented a Medicare ambulance fee schedule in which providers 
are paid a base payment per trip plus a mileage payment. An	 
adjustment is made to the mileage rate for rural trips to account
for higher costs. CMS has stated that this rural adjustment may  
not sufficiently target providers serving sparsely populated	 
rural areas. The Medicare, Medicaid, and SCHIP Benefits 	 
Improvement and Protection Act of 2000 (BIPA) directed GAO to	 
examine rural ambulance costs. GAO identified factors that affect
ambulance costs per trip, examined how these factors varied	 
across geographic areas, and analyzed whether Medicare payments  
account for geographic cost differences. GAO used survey data on 
ambulance providers and Medicare claims data.			 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-03-986 					        
    ACCNO:   A08516						        
  TITLE:     Ambulance Services: Medicare Payments Can Be Better      
Targeted to Trips in Less Densely Populated Rural Areas 	 
     DATE:   09/19/2003 
  SUBJECT:   Cost analysis					 
	     Ground transportation operations			 
	     Health care costs					 
	     Health care programs				 
	     Medical fees					 
	     Medical services rates				 
	     Transportation costs				 
	     Managed health care				 
	     Emergency medical services 			 
	     Rural areas					 

******************************************************************
** 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-03-986

                                       A

Letter

September 19, 2003 Congressional Committees In 2001, Medicare paid over $2
billion for over 10 million ambulance trips for its 40 million elderly and
disabled beneficiaries. Ambulance providers 1 that deliver Medicare-
covered services range from small community, onevehicle

operations staffed entirely by volunteers to large privately owned firms
or government agencies that operate many vehicles and rely on paid staff.
Medicare covers medically necessary ambulance services when other means of
transportation, such as a wheelchair van or a taxi, are

inadvisable, given the beneficiary*s medical condition at the time.
Medically necessary ambulance trips include both emergency care, such as
responses to 911 calls, and nonemergency care, such as transfers from one
hospital to another. In 2002, the Centers for Medicare & Medicaid Services
(CMS) implemented a congressionally mandated ambulance fee schedule that
substantially

changed the way Medicare pays for ambulance services. Under the fee
schedule, providers receive a base payment per trip, which varies by the
kind of service provided, and a mileage payment, which varies by the
length of the trip. A rural adjustment, which is applied to the mileage
payment, increases the payment for trips that begin in rural areas,
generally defined as areas outside of metropolitan areas. CMS has stated
that this approach to a rural adjustment was the only one feasible at the
time the agency was developing the fee schedule. However, as we have
stated before 2 and as

CMS has acknowledged, this adjustment may not sufficiently target the
increased payments to providers serving sparsely populated rural areas.
These providers may incur higher per trip costs than other providers

because of their low volume of ambulance trips. We have recommended that
CMS develop a more refined rural adjustment, and CMS is exploring
alternative approaches to adjusting payments for rural ambulance trips.

1 The Centers for Medicare & Medicaid Services uses the term *provider* to
refer to institutional providers of ambulance services, including
hospitals and skilled nursing facilities, and uses the term *supplier* to
refer to freestanding ambulance providers* that is, those not associated
with a hospital, skilled nursing facility, or other facility. In this
report, unless otherwise indicated, we use the term *provider* to refer to
all organizations that

provide Medicare ground ambulance services* both institutional and
freestanding. 2 U. S. General Accounting Office, Rural Ambulances:
Medicare Fee Schedule Payments Could Be Better Targeted, HEHS- 00- 115
(Washington, D. C.: July 17, 2000).

Developing a Medicare payment method for ambulance services that maintains
beneficiary access to these services has been complicated by the wide
variation in ambulance providers, their volume of trips, and the areas
they serve. For example, ambulance providers in rural areas where there
are few people are likely to be idle more often than providers in more

densely populated areas, and may need to earn more per trip to maintain
ambulances and crews. As a result, payments that are appropriate for
providers serving densely populated urban areas may not be appropriate for
those serving less densely populated rural areas. Recognizing cost
differences across providers in the payment method is important because
many providers rely on Medicare revenue and their continuing availability
is critical to ensuring beneficiaries* access to services.

The Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act
of 2000 (BIPA) directed us to examine the cost of ambulance services in
rural areas. 3 This report (1) identifies the factors that account for
differences in ambulance providers* costs per trip, and (2) analyzes the
geographic differences, particularly among rural areas, in the factors
affecting ambulance providers* costs and whether Medicare*s payments for
ambulance services under the fee schedule account for geographic
differences in costs.

Our analysis focused on ground ambulance services and did not include air
ambulance services, which account for less than 1 percent of annual
Medicare- covered ambulance trips. To identify the factors that influence
ambulance providers* costs, we used data from the 1999 National Survey of
Ambulance Providers, conducted by the Project HOPE Center for Health
Affairs. 4 This survey, the sole national data source on the costs of
providing ambulance services, obtained responses from a nationally
representative sample of 421 ground ambulance providers that participated
in Medicare. To determine how the factors affecting ambulance providers*
costs vary

geographically, we used data from the 2001 Area Resource File (ARF), which
is maintained by the Health Resources and Services Administration (HRSA),
to examine the characteristics of urban and rural counties, such as their
population and land area. In addition, we used Medicare claims data

for 2001 to determine the number and length of Medicare- covered 3 Pub. L.
No. 106- 554, Appendix F, S:221( b), 114 Stat. 2763A- 463, 486- 87 (2000).
4 The Project HOPE Center for Health Affairs is a nonprofit health policy
research organization.

ambulance trips delivered in urban and rural counties, and the number of
providers that served those areas. 5 We also used these data to estimate
Medicare*s average payments for ambulance trips under the fee schedule.
Finally, we interviewed experts from eight industry and professional
organizations as well as several ambulance providers. These data were
adequate for addressing the issues in this report. Where appropriate, we
examined the data for implausible values and tested the data for internal
consistency. For more details on our data and methods, see appendix I. We
performed our work from November 2001 through September 2003 in accordance
with generally accepted government auditing standards.

Results in Brief Ambulance trip volume is the key factor affecting
differences in ambulance providers* average cost per trip. The majority of
ambulance providers* total

costs are related to their need to have ambulances and crew available when
an ambulance is required. As long as a provider has excess capacity, these
readiness- related costs are fixed and do not increase with the number of
trips. Consequently, providers that make fewer trips tend to have a higher
cost per trip than those that make many trips. For example, providers
surveyed by Project HOPE that averaged 3 or fewer trips per day had a cost
per trip that was nearly twice as high as the average cost per trip among
providers that averaged 9 to 12 trips per day. The modest variation in
Medicare payments to ambulance providers that

serve rural counties probably does not fully reflect their differences in
cost per trip because the key factor affecting provider costs* the number
of trips* varies widely across rural counties. In 2001, trip volume was
much

lower in the least densely populated quarter of rural counties than in the
most densely populated quarter. Medicare per- trip payments are somewhat
higher on average for trips provided in the least densely populated rural

counties than for trips in other rural counties. However, the modestly
higher payments are unlikely to fully account for the higher cost per trip
of low- volume providers, which are most likely to serve the least densely
populated rural counties. The Medicare payment differences are due to the

greater length and the resulting higher mileage payments for trips in the
least densely populated rural counties. The cost differences, however, are
5 Our analysis focused on ambulance services paid under Medicare*s
ambulance fee schedule and therefore excluded services paid for by
Medicare managed care organizations. Further, our analysis did not include
any ambulance services for Medicare beneficiaries that were

not billed to Medicare.

due to a higher fixed cost per trip, for which Medicare*s base rates are
intended to compensate.

We recommend that the Administrator of CMS better target the Medicare
rural payment adjustment to trips provided in rural counties with
particularly low population density by adjusting the base rates, rather
than the mileage rate, for ground ambulance services provided in those
counties.

In written comments on a draft of this report, CMS stated that the report
will be useful as the agency develops a proposed rule to address
appropriate payment for ambulance services furnished in rural, lowvolume
areas. The eight ambulance associations that commented on the draft report
generally agreed with our findings and recommendation. However, four of
the associations raised concerns about using counties to identify rural
areas when targeting rural payments.

Background In recent years, the ambulance industry has experienced several
changes. In 2002, CMS implemented a new Medicare fee schedule for
ambulance services, replacing the previous system that paid providers on a
reasonable cost or reasonable charge basis. In addition, according to
industry experts, many volunteer providers have reported greater
difficulty maintaining adequate staff. Rural providers in particular have
begun to rely more heavily on paid staff. Experts also told us that while
many rural volunteer providers have not billed Medicare* or have billed
nominal amounts* more of these providers have begun billing for services.

Characteristics of Medicare Recently, both the number of ambulance
providers that bill Medicare and Ambulance Providers

the number of ambulance trips paid for by Medicare have increased. From
1998 to 2001, the number of ambulance providers that billed Medicare
increased from just under 9,300 to over 9,700, and the total number of
trips paid for by Medicare rose from roughly 8 million to over 10 million.

Medicare ambulance providers include a wide variety of provider types. In
1998, about 8,200 freestanding providers and 1,100 hospitals and other
institution- based providers billed Medicare for ground trips.
Freestanding providers are a diverse group, including private for- profit,
not- for- profit, and public entities. They range from small community
one- vehicle operations to large fire and rescue departments serving major
metropolitan

areas. They include operations staffed almost entirely by community
volunteers, public ventures that include a mix of volunteer and paid
professional staff, and private firms that use only paid staff. In 1998,

volunteer staff accounted for 80 percent or more of full- time equivalent
personnel for over one- third of Medicare ambulance providers. 6 About
one- third of freestanding Medicare ambulance providers are managed by
local fire departments.

Medicare ambulance providers also vary in the types of services they
provide. Some deliver only basic life support (BLS) while others deliver
advanced life support (ALS) services. 7 In addition to responding to
emergencies, ambulance providers may provide nonemergency

transportation, such as transfers from one hospital to another. For some
ambulance providers, nonemergency trips account for a significant share of
their trips; for others, such trips account for few or none of their
trips. Some ambulance providers are the sole providers serving their

communities, while others operate in areas with multiple ambulance
providers. Medicare ambulance providers also differ in the percentage of
their trips covered by Medicare and in their reliance on Medicare revenue.
In 1998, Medicare beneficiaries on average accounted for about half of the
total trips by providers that billed Medicare. However, Medicare
beneficiaries accounted for less than one- quarter of trips for 13 percent
of Medicare providers, and accounted for over 80 percent of annual trips
for 9 percent of providers. On average, Medicare revenue accounted for 41
percent of providers* cash receipts. 8 Other sources of ambulance
providers* revenue include local tax subsidies and payments from private
insurers, Medicaid, and individuals.

Requirements affecting ambulance providers vary by location. States and
localities may require certain training for ambulance staff, establish
maximum payment rates that licensed providers are allowed to charge, or
specify response times through contracts with providers. Some

6 See Penny E. Mohr and others, Findings from the 1999 National Survey of
Ambulance Providers (Bethesda, Md.: 2000), p. 13. 7 ALS services are
provided by personnel with advanced training and involve assessments by
them or the provision of advanced interventions or procedures. 8 See Mohr
and others, pp. 19- 20.

jurisdictions* such as those that provide financial support to ambulance
providers* prohibit providers from billing for services. In addition, some
communities require all ambulance providers to maintain ALS capacity on
all vehicles. Medicare Payment for

CMS recently implemented a Medicare fee schedule that changed the way
Ambulance Services

Medicare pays for ambulance services. 9 The fee schedule, mandated by the
Balanced Budget Act of 1997 (BBA), recognizes seven levels of ground
ambulance services, ranging from BLS services to specialty care
transports.

(See table 1.) Under the previous payment system, Medicare paid
institutional providers on a reasonable cost basis and freestanding
providers on a reasonable charge basis. This approach led to wide

differences in payments across providers for the same services. The new
fee schedule standardized payment rates across provider types by applying
the same payment rates to both institutional and freestanding providers.
10 The fee schedule*s payment rates are updated annually. Medicare*s

payment is based on the lesser of the actual charge or the applicable fee
schedule amount. 9 The fee schedule took effect April 1, 2002. The BBA had
directed that the fee schedule be effective for services furnished on or
after January 1, 2000. CMS stated that several factors* other statutory
obligations, the scope of systems changes required to implement the fee
schedule, and the need to ensure that its computer systems were compliant
with

Year 2000 (Y2K) requirements* delayed implementation of the fee schedule.
67 Fed. Reg. 9100. A federal district court order, issued on January 16,
2003, required the Department of Health and Human Services (HHS) to adopt
a fee schedule for freestanding providers for the period of January 1,
2000, through March 31, 2002. Lifestar Amb. Serv., Inc. v. U. S., 211 F.
R. D. 688 (M. D. Ga. 2003). On April 16, 2003, CMS published a notice in
the Federal Register stating that Medicare*s fee schedule would apply to
ambulance trips provided by freestanding providers during that period. 68
Fed. Reg. 18654. Payments would be adjusted

retroactively. HHS has appealed the court*s decision. 10 Several other
changes to the way ambulance services are paid for were introduced with
the fee schedule. For example, in accordance with the BBA, all providers
must accept the Medicare payment amount as full payment for covered
services and may only collect allowed cost- sharing amounts from
beneficiaries. In addition, Medicare will pay a BLS rate for services
furnished at the BLS level even when an ALS vehicle is used. This latter
provision will be phased in over several years. As specified by BIPA,
ambulance providers that are critical access hospitals or entities owned
and operated by them are exempt from the fee schedule and paid on a
reasonable cost basis if there is no other ambulance provider within 35
miles. Critical access hospitals are small, isolated hospitals that have
an annual average length of stay of 4 days or less.

Table 1: Ground Ambulance Services Covered by Medicare*s Ambulance Fee
Schedule Level of ambulance service Definition Staffing requirements Base
rate (2003) a

Basic life support Transportation by ambulance and the

The ambulance must be staffed by an $172.42 (BLS) provision of medically
necessary supplies and

individual who is qualified in services, including the provision of BLS
accordance with state and local laws ambulance services as defined by the
state. as an emergency medical technician basic (EMT- Basic).

BLS- emergency BLS level of service provided in immediate Same as BLS
$275.87 response to a 911 call or the equivalent. Advanced life support,

Transportation by ambulance and the ALS personnel are individuals trained

$206.90 level 1 (ALS1) provision of medically necessary supplies and

to the level of the emergency medical services, including an ALS
assessment by technician- intermediate (EMT ALS personnel or the provision
of at least one Intermediate) or EMT- Paramedic. c ALS intervention. b
ALS1- emergency ALS level of service provided in immediate Same as ALS1
$327.60

response to a 911 call or the equivalent. ALS2 Transportation by ambulance
and the Same as ALS1 $474.16 provision of medically necessary supplies and
services, including (1) at least three separate administrations of one or
more medications by intravenous push/ bolus or by continuous

infusion, d or (2) the provision of at least one of seven ALS2 procedures.
e Paramedic ALS

EMT- Paramedic services furnished by a EMT- Paramedic $301.74 intercept
provider that does not furnish the ground

ambulance transport, provided the services meet certain requirements. f

Specialty care Hospital- to- hospital transportation of a

SCT is necessary when a patient*s $560.37 transport (SCT) critically
injured or ill patient, including the

condition requires ongoing care that provision of medically necessary
supplies and must be furnished by one or more services, at a level of
service beyond the health professionals in an appropriate scope of the
EMT- Paramedic.

specialty area, such as nursing or respiratory care, or a paramedic with
additional training.

Source: CMS. Note: GAO summary based on information in CMS*s final rule as
published in the Federal Register, a subsequent program memorandum
regarding definitions of ambulance services, and the ambulance fee
schedule public use file for calendar year 2003. See 67 Fed. Reg. 9100,
CMS*s Program Memorandum AB- 02- 130 (Sept. 27, 2002). a This is the base
rate for each level of service prior to the geographic adjustment for
differences in

wages across areas. b An ALS assessment is an assessment performed by an
ALS crew as part of an emergency response

that was necessary because the patient*s reported condition at the time of
dispatch was such that only an ALS crew was qualified to perform the
assessment. An ALS assessment does not necessarily result in a
determination that the patient requires an ALS level of service. An ALS
intervention is a procedure that is, in accordance with state and local
laws, beyond the scope of practice of an EMTBasic.

c An EMT- Intermediate is an individual who is qualified, in accordance
with state and local laws, as an EMT- Basic and who is also certified, in
accordance with those laws, to perform essential advanced techniques and
to administer a limited number of medications. An EMT- Paramedic is an
individual who has the qualifications of an EMT- Intermediate and, in
accordance with state and local laws, has enhanced skills that include
being able to perform additional interventions and administer additional
medications. d This excludes certain solutions.

e These include chest decompression, cardiac pacing, surgical airway, and
other procedures. f Paramedic ALS intercept services are most often
furnished for an emergency ambulance trip in which a local volunteer
ambulance that can furnish only BLS services is dispatched to transport a
beneficiary. If the beneficiary needs ALS services, another provider
dispatches a paramedic to meet the BLS ambulance at the scene or enroute
to the hospital. The ALS paramedics then provide ALS services for

the beneficiary. In general, Medicare payment may be made only to the
provider furnishing the trip. However, the BBA provided that payments also
could be made for the ALS provider under limited circumstances. CMS has
stated that New York is the only state in which providers meet the
statutory requirements for Medicare payment.

For most ambulance services, the fee schedule payment is the sum of a base
payment and a payment for mileage. 11  The base payment for a trip, which
is intended to pay for fixed costs

such as staff and equipment, reflects both a base rate and a geographic
modifier. The base rate varies by the level of ambulance service provided.
The geographic modifier, which is applied to 70 percent of the base rate,
is intended to account for wage differences across areas. 12  The mileage
payment reflects both the length of a trip and the per- mile

payment rate. For trips in which the beneficiary is picked up in an urban
area, the per- mile rate is $5.53. Because of the fee schedule*s rural
adjustment, the per- mile rate for rural trips is 150 percent of the urban
mileage rate for each of the first 17 miles ($ 8. 30) and 125 percent of
the urban mileage rate for miles 18 through 50 ($ 6.91). 13 The urban
mileage rate applies to every mile over 50 miles. The mileage payment
applies

11 For paramedic ALS intercept services, there is no separate payment for
mileage. 12 The modifier is the same as that applied to the practice
expense component of Medicare*s physician fee schedule. 13 The mileage
rate increase for the first 17 miles was not mandated by law, but was
specified in CMS*s final rule. The increase for miles 18 through 50 is a
temporary increase mandated by law. BIPA required that, for miles 18
through 50 of a rural trip, the mileage rate should be increased by at
least half as much as the mileage rate increase established for the first
17 miles of a rural trip. BIPA stated that this increase would apply to
ground ambulance

services provided on or after July 1, 2001, and before January 1, 2004.
Pub. L. No. 106- 554, S:221, 114 Stat. 2763A- 463, 486.

only to *loaded miles** the miles the beneficiary is transported by
ambulance. Under the fee schedule, rural areas are defined as areas
outside of metropolitan statistical areas (MSA) and New England County
Metropolitan Areas, as well as parts of MSAs that are identified as rural
by the Goldsmith modification. 14 MSAs are groups of counties containing a
core of at least 50,000 people, together with adjacent areas that have a
high degree of economic and social integration with that core. 15 The
Goldsmith

modification identifies small towns and rural areas within large
metropolitan counties that are isolated from central areas by distance or
other features, such as mountains. About one- quarter of the roughly 3,100
counties in the United States are in MSAs, and about 75 of those counties
have areas that are identified as rural under the Goldsmith modification.

The ambulance fee schedule will be phased in over several years. During
this period, payments will be based in part on the fee schedule*s
servicespecific payment rates and in part on the amounts that Medicare
would have paid under the prior payment system. The proportion of the
payment based on the fee schedule will increase each year until 2006, when
provider payments will be based entirely on the fee schedule. 16 In 2003,
payments are based on 40 percent of the fee schedule payment and 60
percent of the rates under the prior system. 17 14 CMS has stated that it
could not easily adopt and implement other methods for recognizing

geographic differences in population density within the constraints
necessary to implement the fee schedule in a timely manner. 67 Fed. Reg.
9100.

15 CMS currently uses the MSA definitions established by the Office of
Management and Budget (OMB) and in effect prior to June 6, 2003, when OMB
announced revised definitions based on the 2000 census.

16 Medicare payment for ambulance services is based on the lesser of the
actual charge or the applicable fee schedule amount. During the transition
period, the applicable fee schedule amount is a blended payment, not the
fee schedule payment. 17 In response to the federal district court order
that required HHS to adopt a fee schedule for

freestanding providers for the period of January 1, 2000, to March 31,
2002, CMS issued a notice in the Federal Register that specified blended
payments applicable to 2000, 2001, and early 2002. 68 Fed. Reg. 18654- 55.

Ambulance Providers* Trip volume is the major determinant of differences
across providers in the

Trip Volume Is the average cost per trip. Ambulance providers* total costs
primarily reflect readiness* having an ambulance and crew available when
emergency calls

Main Factor Affecting are received. These readiness- related costs are
fixed costs, meaning that

Their Cost Per Trip they do not increase with the number of trips
provided, as long as the

provider has the excess capacity to make additional trips. Consequently,
providers that can spread these fixed costs across more trips have a lower
average cost per trip than providers that make fewer trips. Providers*
Total Costs Are

The majority of ambulance providers* total costs are related to readiness*
Predominantly Readiness

the need to have an ambulance and crew available when emergency calls
Related, and Do Not Vary

are received. Readiness- related costs include costs of labor, vehicles,
With Trip Volume building space, and administration, as well as the cost
of any back- up

vehicles and crew, which constitute a reserve that permits responses to
multiple simultaneous calls as well as scheduled maintenance on other
vehicles. (See table 2.) Readiness- related costs are fixed, meaning that
they do not vary with the number of trips a provider makes, as long as the

provider has excess capacity. For example, total vehicle costs do not
increase significantly when a provider makes more trips. Likewise,
building and administrative costs are largely unaffected by trip volume.
However, if a provider were to add another ambulance and crew to respond

to higher volume, its fixed costs would rise substantially. Tabl e 2:
Ambulance Providers* Cost Components, 1998 Cost component Percentage of
total costs

Labor costs 65 Administrative costs 14 Vehicle and equipment costs 11
Building costs 6 Supply costs 3

Total costs 100

Source: Project HOPE. Note: GAO analysis of data from Project HOPE*s
National Survey of Ambulance Providers. Fuel costs are included in the
vehicle and equipment costs category. These data are for full cost
providers. Those that did not report costs for all five components were
excluded. Full cost providers are defined as those that have 80 percent or
more of their staff comprised of paid employees rather than volunteers,
and that pay for 80 percent or more of their garage and office space.
These data are for the fiscal year preceding the survey, which for most
providers included 6 months or more of calendar year 1998. Cost component
categories do not add to 100 percent due to rounding.

In contrast, an ambulance provider*s costs for fuel and supplies (such as
drugs and oxygen) are variable because they increase with the number of
trips. These costs, however, account for a small fraction of ambulance
providers* total costs. 18

Fewer Trips Linked to a Providers that make fewer trips tend to have a
higher cost per trip than

Higher Cost Per Trip those that make more trips. Figure 1 illustrates the
average relationship between ambulance providers* cost per trip and their
total trip volume, for

providers that made 5, 000 or fewer trips. 19 As trip volume increases,
the cost per trip decreases. Our statistical analysis considered other
factors that affect providers* costs, notably trip length, but trip volume
was most strongly related to the cost per trip.

18 In the available data on ambulance costs, fuel costs are not always
reported separately, so they are included in the vehicle and equipment
costs category. 19 Total trip volume includes all of a provider*s trips,
not just those covered by Medicare.

Figure 1: The Relationship between Cost Per Trip and Total Ambulance Trip
Volume for Full Cost Providers With 5, 000 or Fewer Annual Trips, 1998

Above average cost per

trip Below average cost per

trip 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 5,000 Total
trips

Predicted cost per trip Average cost per trip for all providers Providers'
actual cost per trip Source: Project HOPE.

Note: GAO analysis of data from Project HOPE*s National Survey of
Ambulance Providers. The curve represents the predicted average cost per
trip, based on our statistical analysis of providers* total costs,
controlling for variation in type of service and trip volume, both of
which were statistically significant. Total trip volume includes all of a
provider*s trips, not just those covered by Medicare. These data are for
full cost providers. Full cost providers are defined as those that have 80
percent or more of their staff comprised of paid employees rather than
volunteers, and that pay for 80 percent or more of their garage and office
space. Providers with over 5,000 trips were excluded. We found similar
results when we analyzed all full cost providers. These data are for the
fiscal year preceding the survey, which for most providers included 6
months or more of calendar year 1998.

In addition, we found that providers surveyed by Project HOPE that
averaged 3 or fewer trips per day had an average cost per trip that was
nearly twice as high as the cost per trip among those that averaged 9 to
12 trips per day. 20 (See table 3.) Providers that averaged 4 to 8 trips
per day

had a cost per trip that was 1. 3 times as high as the average cost among
providers with 9 to 12 trips per day. Tabl e 3: Relative Cost Per Trip for
Full Cost Ambulance Providers, 1998

Providers* average number of total trips Cost per trip relative to the
average for per day (range) providers with 9 to 12 trips per day

3 or fewer 1.94 4 to 8 1.30 9 to 12 1.00 Source: Project HOPE. Note: GAO
analysis of data from Project HOPE*s National Survey of Ambulance
Providers. The relative cost per trip is the ratio of the average cost per
trip for each group of providers to the average cost per trip of providers
that averaged 9 to 12 trips per day. Total trip volume includes all of a
provider*s trips, not just those covered by Medicare. These data are for
full cost providers. Full cost providers are defined as those that have 80
percent or more of their staff comprised of paid employees rather than
volunteers, and that pay for 80 percent or more of their garage and office
space. Providers that had daily trip volumes outside the ranges shown
above were excluded. The cost differences are statistically significant at
the .05 level between the 3 trips or fewer group of providers and the
other two groups (4 to 8 trips and 9 to 12 trips). The cost difference
between the second and third groups was statistically significant at the
.10 level. Significance was assessed using a one- tailed test. These data
are for the fiscal year preceding the survey, which for most providers
included 6 months or more of calendar year 1998.

20 These data are for full cost providers, defined as those that have 80
percent or more of their staff comprised of paid employees rather than
volunteers, and that pay for 80 percent or more of their garage and office
space. These data are for the fiscal year preceding the survey, which for
most providers included 6 months or more of calendar year 1998.

Medicare Ambulance Although Medicare*s payments generally are higher for
trips originating in Payments for Trips in

the least densely populated rural counties than in other counties, the
payment differential is probably not large enough to account for the
higher Rural Counties Are

costs incurred by low- volume providers likely to serve these areas. Far
Unlikely to Fully

fewer Medicare- covered ambulance trips are typically provided in rural
Reflect Differences in

counties than in urban counties. Trip volume also varies widely across
rural counties, with the least densely populated generally having

Providers* Cost Per substantially fewer trips than the most densely
populated. This suggests

Trip that the cost per trip is likely higher for providers serving the
least densely populated rural counties. Ambulance providers on average are
paid more

for trips originating in the least densely populated rural counties than
for those in the most densely populated rural counties, but the payment
differences are modest and unlikely to reflect the higher cost per trip of
low- volume providers.

Rural and Urban Counties Rural counties, as defined by Medicare*s
ambulance fee schedule, tend to

Differ in Ambulance Trip have a much lower volume of ambulance trips than
counties defined as

urban. 21 In 2001, rural counties averaged about 1,200 Medicare- covered
Volume and Population

trips (both emergency and nonemergency), while urban counties averaged
Density about 9,100 trips. The lower number of trips in rural counties
suggests that providers that serve these areas likely have a higher cost
per trip than other providers.

The difference in the volume of Medicare ambulance trips provided in rural
and urban counties largely reflects differences in their population
density. Not surprisingly, the number of Medicare ambulance trips in a
county is

strongly related to its population, with counties with fewer residents
having fewer trips. Trip volume is also related to a county*s land area,
although to a lesser extent. 22 Population density* the ratio of
population to land area* reflects both of these measures. (See table 4.)

21 We classified counties as urban if they were in an MSA and as rural if
they were not in an MSA, using 2001 Area Resource File data. The roughly
75 urban counties that contain areas identified as rural by the Goldsmith
modification are included in the urban county group. We used the
beneficiary*s address as a proxy for where each trip originated, since the
2001 national claims files did not contain that information. (See app. I
for details.) 22 Larger counties have somewhat fewer trips than smaller
counties, after accounting for county population.

Table 4: Characteristics of Urban and Rural Counties, 2001 Average Average
Average number of length of

population Medicare Medicare

density Average ambulance

ambulance (persons/ Average land area trips

trips (miles) sq. mile) population (sq. miles)

Urban counties 9,144 14 747 276,791 844

Rural counties 1,153 23 47 23,942 1,132

Sources: HRSA and CMS. Note: GAO analysis of HRSA and CMS data. We
classified counties as urban if they were in an MSA and as rural if they
were not in an MSA. The roughly 75 urban counties that contain rural areas
as identified by the Goldsmith modification are included in the urban
county group. We used the beneficiary*s address as a proxy for where each
trip originated.

Dominant Providers in Less The number of Medicare ambulance trips provided
in rural counties varies

Densely Populated Rural markedly with population density, with the least
densely populated rural Counties Provide Fewer

counties tending to have fewer trips than other rural counties. For Trips
example, the quarter of rural counties that are the most densely
populated, with 52 or more persons per square mile, averaged over 2,200
Medicare trips in 2001. 23 (See table 5.) In contrast, only about 300
Medicare trips, on

average, were made in the quarter of rural counties that are the least
densely populated, with 11 or fewer persons per square mile. Even fewer
Medicare trips* only about 200* were made in frontier counties, which are
counties with 6 or fewer persons per square mile. 24 This suggests that
the cost per trip is likely higher for providers serving the least densely
populated rural counties.

23 We also grouped counties according to their Medicare population
density. See app. II. 24 *Frontier* is a term used to describe counties
with very low population density, and in most cases frontier counties are
defined as those with six or fewer persons per square mile.

Table 5: Average Number of Medicare Ambulance Trips, Population and Land
Area, by Counties Grouped by Population Density, 2001

Average number of Medicare Average land Number of

ambulance Average area County categories counties

trips population (sq. miles)

Urban counties 854 9,144 276, 791 844 Rural counties 2,273 1,153 23, 942
1,132 52+ persons/ sq. mile 569 2,254 45, 612 502 30- 51 persons/ sq. mile
568 1,290 25, 351 654 12- 29 persons/ sq. mile 568 771 16, 744 898 0- 11
persons/ sq. mile 568 296 8,021 2,477

7- 11 persons/ sq. mile 182 470 12, 288 1,491 0- 6 persons/ sq. mile 386
214 6,009 2,942 Sources: HRSA and CMS. Note: GAO analysis of HRSA and CMS
data. We classified counties as urban if they were in an MSA and as rural
if they were not in an MSA. The roughly 75 urban counties that contain
rural areas as identified by the Goldsmith modification are included in
the urban county group. Rural counties are grouped by quartiles of total
county population density. The first quartile (0- 11 persons per square
mile) is further divided into frontier counties (0- 6 persons per square
mile) and nonfrontier (7- 11 persons per square mile). We used the
beneficiary*s address as a proxy for where each trip originated.

The dominant providers in the least densely populated rural counties tend
to have far fewer trips than the dominant providers serving other rural
counties. Overall, rural counties vary little in the number of providers
serving them. However, in most rural counties, one or two providers
dominate, delivering the bulk of Medicare trips, with others having a much
smaller share. We found that in 2001, about 70 percent of the trips in a
rural county were typically supplied by two providers. The number of trips
made by these dominant providers varied with counties* population density.
In the quarter of rural counties with the lowest population density, the
median number of Medicare trips made by each of the top two

providers* in all of the counties they served* was 275. 25 (See table 6.)
In contrast, the median number of Medicare trips made by the top two
providers was much higher* over 2,100 trips* in the quarter of rural
counties that were the most densely populated. 26

25 Providers* trips are not necessarily limited to one county, since
providers may serve multiple counties. 26 The total number of trips made
by these providers would be expected to be double the

number of Medicare trips, since Medicare beneficiaries account on average
for roughly half of providers* total trip volume.

Tabl e 6: Characteristics of Rural Counties and Their Ambulance Providers,
by Counties Grouped by Population Density, 2001 Percentage of Number of

a county*s Medicare Medicare

ambulance Number of

ambulance trips in all Medicare trips covered

counties for providers

by the top 2 each of the serving a

providers in a top 2 Number of

county a county

providers County categories counties

(median) (median)

(median)

Rural 2,273 5 70 1,100 52+ persons/ sq. mile 569 8 68 2,168 30- 51
persons/ sq. mile 568 6 70 1,422 12- 29 persons/ sq. mile 568 6 69 832 0-
11 persons/ sq. mile 568 4 74 275

7- 11 persons/ sq. mile 182 5 71 433 0- 6 persons/ sq. mile 386 4 75 215
Sources: HRSA and CMS. Note: GAO analysis of HRSA and CMS data. We
classified counties as rural if they were not in an MSA. Rural counties
are grouped by quartiles of total county population density. The first
quartile (0- 11

persons per square mile) is further divided into frontier counties (0- 6
persons per square mile) and nonfrontier counties (7- 11 persons per
square mile). We used the beneficiary*s address as a proxy for where each
trip originated. a Providers that delivered less than 1 percent of their
total Medicare trips in a county were excluded

from the count of providers serving that county.

Medicare Ambulance Ambulance providers on average are paid 16 percent more
for trips

Payments Are Somewhat originating in the least densely populated quarter
of rural counties than for

trips in the most densely populated quarter. 27 (See table 7.) Payments
for Higher for Trips in Less

those trips are higher because the trips are generally longer, resulting
in a Densely Populated Rural

higher mileage payment. In 2001, while trips that began in the most
densely Counties

populated quarter of rural counties averaged 18 miles, trips in the least
densely populated quarter averaged 30 miles. The rural adjustment, which
provides a higher per- mile rate for the first 50 miles of rural trips,
also contributed to the higher mileage payments.

27 Payment estimates were calculated by applying 100 percent of the 2003
Medicare ambulance fee schedule rates to Medicare ground ambulance trips
delivered in 2001.

The modest difference in Medicare payment across rural counties is dwarfed
by the difference in trip volume: The difference in trip volume between
the least and most densely populated quarters of rural counties is nearly
eightfold. 28 Because trip volume is an indicator of costs, the Medicare
payment differences likely do not fully reflect differences across rural
counties in providers* cost per trip. 29 Tabl e 7: Average Number of
Medicare Ambulance Trips, Trip Length, and Estimates

of Average Medicare Payment a per Ambulance Trip, by Rural Counties
Grouped by Population Density Average Average Average number of length of
Medicare Medicare

Medicare payment per Number of ambulance ambulance ambulance County
categories counties

trips trips (miles)

trip

Rural counties 2,273 1,153 23 $463 52+ persons/ sq. mile 569 2,254 18 $434
30- 51 persons/ sq. mile 568 1,290 21 $446 12- 29 persons/ sq. mile 568
771 25 $465 0- 11 persons/ sq. mile 568 296 30 $505

7- 11 persons/ sq. mile 182 470 27 $490 0- 6 persons/ sq. mile 386 214 31
$512 Sources: HRSA and CMS. Note: GAO analysis of HRSA and CMS data. We
classified counties as urban if they were in an MSA and as rural if they
were not in an MSA. The roughly 75 urban counties that contain rural areas
as identified by the Goldsmith modification are included in the urban
county group. Rural counties are grouped by quartiles of total county
population density. The first quartile (0- 11 persons per square mile) is
further divided into frontier counties (0- 6 persons per square mile) and
nonfrontier counties (7- 11 persons per square mile). We used the
beneficiary*s address as a proxy for where each trip originated. a Payment
estimates were calculated by applying 100 percent of the 2003 Medicare
ambulance fee

schedule rates to Medicare ground ambulance trips delivered in 2001. These
estimates reflect the mix of ambulance services provided in the different
county categories as well as the geographic adjustment to account for wage
differences across areas.

28 There is a similar difference in trip volume for the dominant providers
serving these counties. 29 Differences in trip volume likely result in
smaller differences in cost per trip.

Conclusions Refining Medicare*s ambulance fee schedule to adequately
account for cost differences in providing ambulance services across
various geographic

areas is important to ensuring beneficiaries* access to services. Access
is a particular concern in rural areas, since providers* cost per trip is
likely to be higher because they provide fewer trips. Moreover, our
analysis shows that the cost per trip is likely to be highest in the least
densely populated rural counties. While the fee schedule incorporates a
rural adjustment to raise payments for trips provided in rural areas, its
definition of *rural* is broad. As a result, the fee schedule*s rural
payment adjustment does not sufficiently target trips provided in the
least densely populated rural counties. In implementing the fee schedule,
CMS adjusted the mileage rate for rural

trips to account for the higher cost per trip of providers serving rural
areas. However, trip volume is a better indicator of providers* cost per
trip than is trip length. Thus, adjusting the base rates for rural trips*
the portion of Medicare*s payment that is designed to pay for providers*
fixed costs* is a more appropriate way of accounting for rural low- volume
providers* higher cost per trip than adjusting the mileage rate.

Recommendation for To help ensure that Medicare beneficiaries* access to
ambulance services is

Executive Action adequate, we recommend that the Administrator of CMS
better target the

rural payment adjustment to trips provided in rural counties with
particularly low population density by adjusting the base rates, rather
than the mileage rate, for ground ambulance services provided in those
counties.

Agency and External We received written comments on a draft of this report
from CMS. We also

received comments from eight ambulance associations: American Reviewer
Comments Ambulance Association, 30 American Hospital Association,
Association of and Our Evaluation

Air Medical Services, National Association of State Emergency Medical
Services Directors, National Volunteer Fire Council, Rural EMS Advocate,
American College of Emergency Physicians, and the National Association of
EMS Physicians.

30 The American Ambulance Association represents many types of ambulance
providers. In this section we refer to it as representing ground ambulance
providers because they account for the majority of the association*s
members.

CMS CMS stated that the report will be useful as the agency develops a
proposed rule to address appropriate payment for ambulance services
furnished in

rural, low- volume areas. CMS also noted that the report reflects the
complexity of the issues and the need for careful analysis to ensure that
payment adjustments are made only for those ambulance providers that
require additional payment because of their low volume, rather than, for
example, inefficiency or competition from another provider. CMS*s comments
appear in appendix III. CMS also provided technical comments, which we
incorporated as appropriate. Ambulance Associations The associations that
reviewed the draft report generally agreed with our

findings and recommendation. All of the associations agreed with the need
to address the higher cost of providing ambulance services in rural areas.
Six agreed that an area*s ambulance trip volume reflects its population
density, while the remaining two associations did not address this issue.
The majority of the associations agreed that CMS should adjust the base
rates to recognize the higher cost per trip of providing ambulance
services in areas with low population density. However, three associations
went further, proposing to use both mileage and base rates to address the
higher costs in rural areas. While supporting the principle of paying
higher base rates to providers in rural areas where costs are high, the
state EMS directors* and EMS physicians* associations were concerned that
higher payments for rural providers could be at the expense of other
providers.

Four associations raised concerns about using counties as the geographic
areas for applying the adjustment. These associations said that a system
that used counties would not accurately target rural ambulance payments.
Three of these associations noted that, because counties may include both

densely and sparsely populated areas, a system that used counties could
overpay some providers and underpay others. They proposed using zip codes
as the geographic areas for assessing population density and

applying the adjustment. The rural ambulance association, in particular,
also advocated the use of multiple rural categories based on population
density to adjust payments for rural trips. The fourth association
emphasized the need for a system that ensures that all areas with
sufficiently low population density are eligible for an appropriate
payment

adjustment.

Our Response GAO and the ambulance associations agree with the need to
adjust payments for rural trips and that the adjustment should be applied
to the base rate. With respect to adjusting payments for rural trips in
low population density areas, we believe the adjustment should be applied
to the base rate. We believe that the mileage rate for any trip, rural or
urban, is best suited to compensating ambulance providers for costs that
vary with

trip length. As stated in the report, a base rate adjustment is a more
appropriate way of accounting for rural low- volume providers* higher
costs per trip because base rates reflect fixed costs, and because trip
volume is a better indicator of providers* cost per trip than is trip
length. With respect to possible payment reductions for other providers,
implementing our recommendation could have this effect. If a revised rural
adjustment is implemented in a way to keep total Medicare expenditures the
same, some providers could face lower payments.

With respect to the geographic unit used to identify trips for the rural
adjustment, we agree that, since counties are relatively large geographic
units, it is possible for trips in some areas to be overpaid and others
underpaid. Moreover, in principle, a rural classification system that uses
a smaller geographic unit, such as zip codes, might better target payments
to trips in areas with low population density. Yet our analysis indicates
that zip codes do not explain variation in trip volume as well as
counties. Further, county boundaries tend to be more stable over time than
zip code boundaries. In addition, a variety of technical difficulties
hinder the use of zip codes for ambulance payments, including the absence
of zip codes for some rural areas. With respect to multiple adjustment
categories, we did not address whether there should be a single adjustment
or whether there should be multiple adjustment amounts to reflect
differing levels of population density. A decision on single or multiple
categories would require balancing increased precision with increased
complexity.

We are sending copies of this report to the Administrator of CMS,
appropriate congressional committees, and other interested parties. We
will also make copies available to others upon request. This report is
also

available at no charge on GAO*s Web site at http:// www. gao. gov.

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

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

List of Committees

The Honorable Charles E. Grassley Chairman The Honorable Max Baucus
Ranking Minority Member Committee on Finance United States Senate The
Honorable W. J. *Billy* Tauzin Chairman The Honorable John D. Dingell
Ranking Minority Member Committee on Energy and Commerce House of
Representatives The Honorable William M. Thomas Chairman The Honorable
Charles B. Rangel Ranking Minority Member Committee on Ways and Means
House of Representatives

Appendi Appendi xes x I

Data and Methods 1999 National Survey of Ambulance Providers. To identify
the factors that influenced ambulance provider costs, we used the 1999
National Survey of Ambulance Providers. This survey, conducted by the
Project HOPE Center for Health Affairs under the sponsorship of the
American Ambulance Association, is the only nationally representative
source for ambulance providers* costs. Project HOPE selected a stratified
random sample of providers that had billed Medicare in 1997, obtained 421
completed questionnaires, and reported a response rate of 56 percent. 1
The survey included questions on costs, total number of trips by type of
service, geographic location, and total mileage.

We took several steps to ensure that the Project HOPE data were suitable
for our analysis. We examined the accuracy and completeness of the data by
testing for implausible values and internal consistency. In addition, we

questioned an anomalous result in Project HOPE*s initial analysis of its
data, which raised concerns about the credibility of the data: emergency
advanced life support (ALS) trips cost less than nonemergency basic life
support (BLS) trips. 2 In response, Project HOPE provided us with
information about its subsequent analysis, which showed the expected
result* ALS trips cost more than BLS trips, after controlling for
providers* volume. This result resolved our major concern about the data.

We limited our analysis of the factors affecting differences in providers*
costs to full cost providers* those providers that paid for 80 percent or
more of their staff and paid for 80 percent or more of their office and
garage space. The costs reported by these providers are more likely to
reflect the full cost of providing ambulance services. We also excluded
ambulance providers that were part of fire departments, because about half

could not separate ambulance costs from other costs. Finally, we excluded
one provider that reported implausible values. After these exclusions, we
had 114 cases for analysis. Certain analyses that did not pertain to all
full cost providers used a smaller number of cases. (See tables 8 and 9.)
1 See Penny E. Mohr and others, Findings from the 1999 National Survey of
Ambulance Providers (Bethesda, Md.: 2000), p. 11. 2 See Mohr and others,
p. 25.

Tabl e 8: Full Cost Ambulance Providers by Average Number of Trips Per
Day, 1998 Providers* average number of total trips per day Percentage of
full cost providers

3 or fewer 22 4 to 8 27 9 to 12 9 13 to 20 7

21 or more 35 Tot al 100

Source: Project HOPE. Note: GAO analysis of data from Project HOPE*s
National Survey of Ambulance Providers. Total trip volume includes all of
a provider*s trips, not just those covered by Medicare. Full cost
providers are defined as those that have 80 percent or more of their staff
comprised of paid employees rather than volunteers, and that pay for 80
percent or more of their garage and office space. These data are for the
fiscal year preceding the survey, which for most providers included 6
months or more of calendar year 1998. The number of full cost providers is
114.

Tabl e 9: Full Cost Ambulance Providers by Average Number of Trips Per
Year, 1998 Providers* average number of total trips per year Percentage of
full cost providers

5,000 or fewer 58 5,001 to 10,000 20 More than 10, 000 22

Tot al 100

Source: Project HOPE. Note: GAO analysis of data from Project HOPE*s
National Survey of Ambulance Providers. Total trip volume includes all of
a provider*s trips, not just those covered by Medicare. Full cost
providers are defined as those that have 80 percent or more of their staff
comprised of paid employees rather than volunteers, and that pay for 80
percent or more of their garage and office space. These data are for the
fiscal year preceding the survey, which for most providers included 6
months or more of calendar year 1998. The number of full cost providers is
114.

Area Resource File. The Area Resource File (ARF), which is maintained by
the Health Resources and Services Administration (HRSA), is a countybased
health resources information database that contains data from many
sources, including the U. S. Census. From the 2001 ARF, we obtained county
data on land area in 1990 and total population in 2000, which we

used to calculate population density. We also obtained data on the number
of persons age 65 and over in each county in 1999, which we used as a
proxy for Medicare beneficiaries. The ARF is a standard data source that
is

well documented and widely used, so we did not independently verify its
accuracy or completeness.

Medicare claims files. We used Medicare claims data to determine the
volume and length of all ground- based Medicare- covered trips, as well as
Medicare*s payments for those trips. We used the 2001 national claims
history 100 percent nearline file for physicians and suppliers to identify
claims for ambulance services by freestanding providers, and the 2001
outpatient 100 percent standard analytic file to identify claims for
ambulance services by institutional providers. We used the zip code of the
beneficiary*s primary address as a proxy for the point where the ambulance
picked up the beneficiary because the point of pickup is not recorded in
the 2001 data. 3 Although we did not independently verify the reliability
of the national claims files, we screened the files and excluded claims
that were denied, claims that were superseded by an adjustment claim, and
claims for

services in other years. We retained all final claims for 2001. Provider
interviews. To gain an understanding of the ambulance industry, we
interviewed experts from eight industry and professional organizations. We
also interviewed several individual ambulance providers. Factors affecting
ambulance providers* costs. To examine the effect of

selected factors on ambulance providers* costs, we analyzed the Project
HOPE survey data using a simplified version of a model reported by Project
HOPE. 4 In our model, the natural logarithm of total costs is a function
of the number of trips, the number of trips squared, and the proportion of
the total trips that are ALS. 5 We tested a number of additional terms,
including length of trips, but they were all either statistically
insignificant or significant but with very small effects. We restricted
our model to

3 Beginning January 1, 2001, CMS required ambulance providers to include
the point of pickup zip code on all claims, and on April 1, 2002, began
using it for payment, to determine whether the rural adjustment should be
applied. Although the CMS contractors* fiscal intermediaries and carriers*
that pay the claims have had the point of pick- up zip code in their data
bases, it was not incorporated into the national claims history files
until April 1, 2003.

4 See Mohr and others, pp. A1- A3. 5 The Project HOPE model resembles a
type of model frequently used by health services researchers for analyzing
costs of health care. For example, see T. Grannemann and others,
*Estimating Hospital Costs,* Journal of Health Economics, vol. 5, no. 2
(1986), and J. Nyman, *The Marginal Cost of Nursing Home Care,* Journal of
Health Economics, vol. 7, no. 4 (1988).

providers with 5, 000 or fewer total trips per year because we were
primarily interested in rural providers, which generally have fewer trips.
6 However, our sensitivity analyses showed that the results were broadly

similar when the model was applied to all full cost providers. 7 Our model
has an adjusted R 2 of 0.48, indicating that the model explains 48 percent
of the variance in costs. In general, when trip volume declines, the
estimated cost per trip increases, although less than proportionately.
That is, a 10 percent decrease in trip volume is associated with an
increase in cost per trip of less than 10 percent.

Analysis of variation in factors affecting costs across geographic areas.
To examine differences between urban and rural areas in factors affecting
ambulance costs, we grouped counties with similar characteristics. We
followed CMS in classifying counties in metropolitan statistical areas
(MSA) as urban counties and counties outside MSAs as rural. 8 However, for
our analysis we did not apply the Goldsmith modification that CMS uses to
identify as rural certain areas within MSAs. 9 These rural areas are
typically small, so we did not treat them as rural counties because that
would distort our urban and rural comparisons. Our sensitivity analyses
determined that our findings would have been generally the same if we had
considered these areas as rural counties, although in some cases the
differences between urban and rural counties would have been heightened.

To examine differences among rural counties, we grouped them based on
their population density. Population density* the ratio of population to
land area* is a commonly used measure of rurality. We used population
density to group counties into quartiles, and then divided the least
densely populated quartile of rural counties into frontier counties* those
with six or fewer persons per square mile* and nonfrontier counties,
because of our interest in the most sparsely populated rural areas. Using
this

6 Although some rural providers have more than 5,000 total trips per year,
most have less than 5,000 total trips per year. 7 The coefficients had the
same signs, although they differed in magnitude.

8 MSAs are groups of counties containing a core of at least 50,000 people,
together with adjacent areas that have a high degree of economic and
social integration with that core. New England County Metropolitan Areas
are considered urban.

9 The Goldsmith modification identifies small towns and rural areas within
large metropolitan counties that are isolated from central areas by
distance or other features, such as mountains. CMS uses a Goldsmith
modification based on 1980 census data.

grouping, we found that ambulance trip volume decreased steadily from the
most densely populated rural counties to the least densely populated. We
also examined several other classification systems: urban influence codes
(UIC), which classify counties based on each county*s largest city and its
proximity to other areas with large, urban populations; rural- urban
continuum codes (RUCC), which classify metropolitan counties by the size
of the urban area and nonurban counties by the size of the urban
population and proximity to a metropolitan area; and rural- urban
commuting areas (RUCA), which classify census tracts using patterns of
urbanization, population density, and daily commuting patterns, and then
map the census tracts into zip codes. 10 These systems are more complex

than the system we used, and we found that they did not help explain
variation in trip volume as well as counties grouped by population
density.

To confirm the effect of population density on trip volume, we did several
additional analyses. We regressed counties* annual volume of Medicare
trips (expressed as natural logarithms) on population and land area
(expressed as natural logarithms). Population had a positive effect on the
number of trips, while land area had a negative effect. An increase of 1
percent in population increased the number of trips by about 1 percent in
a

county, while an increase of 1 percent in land area decreased the number
of trips by about 0.1 percent. 11 Population density combines the two
effects: a 1 percent increase in population density increases the number
of trips by 0.7 percent. 12

10 For more information on UICs, see http:// www. ers. usda. gov/
Briefing/ Rurality/ urbaninf/; for more information on RUCCs, see http://
www. ers. usda. gov/ Briefing/ Rurality/ ruralurbcon/; and for more
information on RUCAs, see

http:// www. fammed. washington. edu/ wwamirhrc/ rucas/ rucas. html. 11
The adjusted R 2 for the model is 0.74. The adjusted R 2 is a measure of
the proportion of the variation in the dependent variable (the natural
logarithm of trips) accounted for by the independent variables (the
natural logarithms of land area and population).

12 The adjusted R 2 for the model is 0. 60.

Characteristics of Rural Counties Grouped by

Appendi x II

Medicare Population Density Total population density is strongly related
to Medicare population density. (See table 10.) For example, 525 rural
counties with the lowest total population density were also lowest in
terms of Medicare population density. In total, 83 percent of all rural
counties were in the same density quartile, regardless of whether total
population or Medicare population was used to group rural counties. Our
results with respect to county characteristics and ambulance services
would have been similar had we used Medicare population density to group
counties rather than total population density. (See tables 11, 12, and
13.)

Tabl e 10: Rural Counties Grouped by Total Population Density and by
Medicare Population Density, 2001

Medicare population density 7.6 +

4.3- 7. 5 1.8- 4.2 0- 1.7 Medicare

Medicare Medicare

Medicare Total population beneficiaries/

beneficiaries/ beneficiaries/

beneficiaries/ density

sq. mile sq. mile

sq. mile sq. mile

52+ people/ sq. mile 505 61 3 0 30- 51 people/ sq. mile 64 423 79 2 12- 29
people/ sq. mile 0 83 444 41 0- 11 people/ sq. mile 0 0 43 525

Source: HRSA. Note: GAO analysis of the 2001 Area Resource File. Bolded
numbers refer to the number of counties that fall in the same quarter of
rural counties* whether rural counties are grouped by total population
density or Medicare population density. We used the number of persons age
65 and over in each county in 1999 as a proxy for the number of Medicare
beneficiaries.

Table 11: Average Number of Medicare Ambulance Trips, Population and Land
Area, by Counties Grouped by Medicare Population Density, 2001

Average number of Medicare

Average Number of ambulance

Average land area

County categories counties trips population

(sq. miles)

Urban counties 854 9,144 276,791 844 Rural counties 2,273 1,153 23, 942
1,132 7.6+ Medicare beneficiaries/ sq. mile 569 2,279 45, 362 517

4.3- 7.5 Medicare beneficiaries/ sq. mile 567 1,265 24, 836 646

1.8- 4.2 Medicare beneficiaries/ sq. mile 569 752 16, 584 841

0- 1.7 Medicare beneficiaries/ sq. mile 568 315 8,961 2,527

Sources: HRSA and CMS. Note: GAO analysis of HRSA and CMS data. We
classified counties as urban if they were in an MSA and as rural if they
were not in an MSA. The roughly 75 urban counties that contain rural areas
as identified by the Goldsmith modification are included in the urban
county group. We used the beneficiary*s address as a proxy for where each
trip originated. We used the number of persons age 65 and over in each
county in 1999 as a proxy for the number of Medicare beneficiaries.

Tabl e 12: Characteristics of Rural Counties and Their Ambulance
Providers, by Counties Grouped by Medicare Population Density, 2001
Percentage of a county*s Medicare

Number of Number of

ambulance Medicare

Medicare trips covered

ambulance trips providers

by the top 2 in all counties serving a

providers in for each of the

Number of county a

a county top 2 providers

County categories counties (median)

(median) (median)

Rural counties 2,273 5 70 1,100 7.6+ Medicare beneficiaries/ sq. mile 569
9 67 2,080

4.3- 7.5 Medicare beneficiaries/ sq. mile 567 6 69 1,366

1.8- 4.2 Medicare beneficiaries/ sq. mile 569 6 71 834

0- 1.7 Medicare beneficiaries/ sq. mile 568 4 74 308

Sources: HRSA and CMS. Note: GAO analysis of HRSA and CMS data. We
classified counties as rural if they were not in an MSA. We used the
beneficiary*s address as a proxy for where each trip originated. We used
the

number of persons age 65 and over in each county in 1999 as a proxy for
the number of Medicare beneficiaries. a Providers that delivered less than
1 percent of their total Medicare trips in a county were excluded

from the count of providers serving that county.

Tabl e 13: Average Number of Medicare Ambulance Trips, Trip Length, and
Estimates of Average Medicare Payment a per Ambulance Trip, by Rural
Counties Grouped by Medicare Population Density

Average Average Average number of length of Medicare Medicare

Medicare payment per Number of

ambulance ambulance ambulance County categories counties

trips trips (miles)

trip

Rural counties 2,273 1,153 23 $463 7.6+ Medicare beneficiaries/ sq. mile
569 2,279 19 $434

4.3- 7.5 Medicare beneficiaries/ sq. mile 567 1,265 21 $448

1.8- 4.2 Medicare beneficiaries/ sq. mile 569 752 25 $467

0- 1.7 Medicare beneficiaries/ sq. mile 568 315 29 $501

Sources: HRSA and CMS. Note: GAO analysis of HRSA and CMS data. We
classified counties as urban if they were in an MSA and as rural if they
were not in an MSA. We used the beneficiary*s address as a proxy for where
each trip originated. We used the number of persons age 65 and over in
each county in 1999 as a proxy for the number of Medicare beneficiaries. a
Payment estimates were calculated by applying 100 percent of the 2003
Medicare ambulance fee

schedule rates to Medicare ground ambulance trips delivered in 2001. These
estimates reflect the mix of ambulance services provided in the different
county categories as well as the geographic adjustment to account for wage
differences across areas.

Comments from the Centers for Medicare &

Appendi x III Medicaid Services

Appendi x IV

GAO Contacts and Staff Acknowledgments GAO Contacts Jonathan Ratner, (202)
512- 7107 Phyllis Thorburn, (202) 512- 7012 Acknowledgments Major
contributors to this report were Martha Kelly, Robin Burke, Eric

Wedum, Michael Kendix, and Jessica Farb.

Related GAO Products

Ambulance Services: Changes Needed to Improve Medicare Payment Policies
and Coverage Decisions. GAO- 03- 244T. Washington, D. C.: November 15,
2001.

Rural Ambulances: Medicare Fee Schedule Payments Could Be Better Targeted.
GAO/ HEHS- 00- 115. Washington, D. C.: July 17, 2000.

(290155)

GAO*s Mission The General Accounting 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.

Obtaining Copies of 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 GAO Reports and

files of current reports and testimony and an expanding archive of older
Testimony

products. The Web site features a search engine to help you locate
documents using key words and phrases. You can print these documents in
their entirety, including charts and other graphics.

Each day, GAO issues a list of newly released reports, testimony, and
correspondence. GAO posts this list, known as *Today*s Reports,* on its
Web site daily. The list contains links to the full- text document files.
To have GAO e- mail this

list to you every afternoon, go to www. gao. gov and select *Subscribe to
e- mail alerts* under the *Order GAO Products* heading. Order by Mail or
Phone The first copy of each printed report is free. Additional copies are
$2 each. A check

or money order should be made out to the Superintendent of Documents. GAO
also accepts VISA and Mastercard. Orders for 100 or more copies mailed to
a single address are discounted 25 percent. Orders should be sent to:

U. S. 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

To Report Fraud, Contact:

Waste, and Abuse in Web site: www. gao. gov/ fraudnet/ fraudnet. htm E-
mail: fraudnet@ gao. gov Federal Programs

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

Report to Congressional Committees

September 2003 AMBULANCE SERVICES Medicare Payments Can Be Better Targeted
to Trips in Less Densely Populated Rural Areas

GAO- 03- 986

Contents Letter 1

Results in Brief 3 Background 4 Ambulance Providers* Trip Volume Is the
Main Factor Affecting

Their Cost Per Trip 10 Medicare Ambulance Payments for Trips in Rural
Counties Are

Unlikely to Fully Reflect Differences in Providers* Cost Per Trip 14
Conclusions 20 Recommendation for Executive Action 20 Agency and External
Reviewer Comments and Our Evaluation 20

Appendixes

Appendix I: Data and Methods 25

Appendix II: Characteristics of Rural Counties Grouped by Medicare
Population Density 30

Appendix III: Comments from the Centers for Medicare & Medicaid Services
34

Appendix IV: GAO Contacts and Staff Acknowledgments 35 GAO Contacts 35
Acknowledgments 35

Related GAO Products 36 Tables Table 1: Ground Ambulance Services Covered
by Medicare*s

Ambulance Fee Schedule 7 Table 2: Ambulance Providers* Cost Components,
1998 10 Table 3: Relative Cost Per Trip for Full Cost Ambulance Providers,

1998 13 Table 4: Characteristics of Urban and Rural Counties, 2001 15
Table 5: Average Number of Medicare Ambulance Trips, Population

and Land Area, by Counties Grouped by Population Density, 2001 16 Table 6:
Characteristics of Rural Counties and Their Ambulance

Providers, by Counties Grouped by Population Density, 2001 18

Table 7: Average Number of Medicare Ambulance Trips, Trip Length, and
Estimates of Average Medicare Payment a per Ambulance Trip, by Rural
Counties Grouped by Population Density 19 Table 8: Full Cost Ambulance
Providers by Average Number of

Trips Per Day, 1998 26 Table 9: Full Cost Ambulance Providers by Average
Number of

Trips Per Year, 1998 26 Table 10: Rural Counties Grouped by Total
Population Density and

by Medicare Population Density, 2001 30 Table 11: Average Number of
Medicare Ambulance Trips, Population

and Land Area, by Counties Grouped by Medicare Population Density, 2001 31
Table 12: Characteristics of Rural Counties and Their Ambulance

Providers, by Counties Grouped by Medicare Population Density, 2001 32
Table 13: Average Number of Medicare Ambulance Trips, Trip

Length, and Estimates of Average Medicare Payment a per Ambulance Trip, by
Rural Counties Grouped by Medicare Population Density 33

Figure Figure 1: The Relationship between Cost Per Trip and Total
Ambulance Trip Volume for Full Cost Providers With

5,000 or Fewer Annual Trips, 1998 12

Abbreviations

ALS advanced life support ARF Area Resource File BBA Balanced Budget Act
of 1997 BIPA Medicare, Medicaid, and SCHIP Benefits Improvement and

Protection Act of 2000 BLS basic life support CMS Centers for Medicare &
Medicaid Services EMT emergency medical technician

HCFA Health Care Financing Administration HHS Department of Health and
Human Services HRSA Health Resources and Services Administration MSA
metropolitan statistical area OMB Office of Management and Budget

RUCA rural- urban commuting area RUCC rural- urban continuum code SCT
specialty care transport UIC urban influence code Y2K Year 2000 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.

a

GAO United States General Accounting Office

Trip volume is the key factor affecting differences in ambulance
providers* cost per trip. Ambulance providers* total costs primarily
reflect readiness* the need to have an ambulance and crew available when
emergency calls are received. Readiness- related costs are fixed, meaning
that they do not increase with the number of trips provided, as long as a
provider has excess capacity. As a result, providers that make fewer trips
tend to have a higher cost per trip than those that make more trips. We
also found that the length

of providers* trips had little effect on their cost per trip. The modest
variation in Medicare payments to ambulance providers that serve rural
counties probably does not fully reflect their differences in costs
because the key factor affecting provider costs* the number of trips*
varies widely across rural counties. In 2001, the least densely populated
quarter of

rural counties averaged far fewer trips than the most densely populated
quarter. This suggests that the cost per trip is likely higher for
providers serving the least populated rural counties. On average ambulance
providers are paid somewhat more for trips in the least densely populated
rural counties than for those in other rural counties. However, those
payment differences are dwarfed by the difference in trip volume. Because
trip volume is a strong indicator of costs, the Medicare payment
differences across rural counties likely do not fully reflect differences
in providers* cost per trip.

In implementing the fee schedule, CMS adjusted the mileage rate for rural
trips to account for the higher cost per trip of providers serving rural
areas. However, trip volume is a better indicator of providers* cost per
trip than is trip length. Thus, adjusting the base rates for rural trips*
the portion of Medicare*s payment that is designed to pay for providers*
fixed costs* is a more appropriate way of accounting for rural low- volume
providers* higher cost per trip than adjusting the mileage rate. The
Centers for Medicare & Medicaid Services (CMS) recently

implemented a Medicare ambulance fee schedule in which providers are paid
a base payment per trip plus a mileage payment. An adjustment is made to
the mileage rate for rural trips to

account for higher costs. CMS has stated that this rural adjustment may
not sufficiently target

providers serving sparsely populated rural areas. The Medicare, Medicaid,
and SCHIP Benefits Improvement and

Protection Act of 2000 (BIPA) directed GAO to examine rural ambulance
costs. GAO identified factors that affect ambulance costs per trip,
examined how these

factors varied across geographic areas, and analyzed whether Medicare
payments account for geographic cost differences. GAO used survey data on
ambulance providers and Medicare claims

data. GAO recommends that CMS better target the rural adjustment to trips
in less densely populated rural

counties by adjusting the base rates for ground ambulance services
provided in those counties. CMS stated that the report will be useful as
the agency develops a proposed rule to address appropriate payment for
ambulance services furnished in rural, low- volume areas.

www. gao. gov/ cgi- bin/ getrpt? GAO- 03- 986. To view the full product,
including the scope and methodology, click on the link above. For more
information, contact Laura A. Dummit, (202) 512- 7114. Highlights of GAO-
03- 986, a report to

congressional committees

September 2003

AMBULANCE SERVICES

Medicare Payments Can Be Better Targeted to Trips in Less Densely
Populated Rural Areas

Page i GAO- 03- 986 Rural Ambulance Services

Contents

Page ii GAO- 03- 986 Rural Ambulance Services

Contents

Page iii GAO- 03- 986 Rural Ambulance Services

Page 1 GAO- 03- 986 Rural Ambulance Services United States General
Accounting Office Washington, D. C. 20548

Page 1 GAO- 03- 986 Rural Ambulance Services

A

Page 2 GAO- 03- 986 Rural Ambulance Services

Page 3 GAO- 03- 986 Rural Ambulance Services

Page 4 GAO- 03- 986 Rural Ambulance Services

Page 5 GAO- 03- 986 Rural Ambulance Services

Page 6 GAO- 03- 986 Rural Ambulance Services

Page 7 GAO- 03- 986 Rural Ambulance Services

Page 8 GAO- 03- 986 Rural Ambulance Services

Page 9 GAO- 03- 986 Rural Ambulance Services

Page 10 GAO- 03- 986 Rural Ambulance Services

Page 11 GAO- 03- 986 Rural Ambulance Services

Page 12 GAO- 03- 986 Rural Ambulance Services

Page 13 GAO- 03- 986 Rural Ambulance Services

Page 14 GAO- 03- 986 Rural Ambulance Services

Page 15 GAO- 03- 986 Rural Ambulance Services

Page 16 GAO- 03- 986 Rural Ambulance Services

Page 17 GAO- 03- 986 Rural Ambulance Services

Page 18 GAO- 03- 986 Rural Ambulance Services

Page 19 GAO- 03- 986 Rural Ambulance Services

Page 20 GAO- 03- 986 Rural Ambulance Services

Page 21 GAO- 03- 986 Rural Ambulance Services

Page 22 GAO- 03- 986 Rural Ambulance Services

Page 23 GAO- 03- 986 Rural Ambulance Services

Page 24 GAO- 03- 986 Rural Ambulance Services

Page 25 GAO- 03- 986 Rural Ambulance Services

Appendix I

Appendix I Data and Methods

Page 26 GAO- 03- 986 Rural Ambulance Services

Appendix I Data and Methods

Page 27 GAO- 03- 986 Rural Ambulance Services

Appendix I Data and Methods

Page 28 GAO- 03- 986 Rural Ambulance Services

Appendix I Data and Methods

Page 29 GAO- 03- 986 Rural Ambulance Services

Page 30 GAO- 03- 986 Rural Ambulance Services

Appendix II

Appendix II Characteristics of Rural Counties Grouped by Medicare
Population Density

Page 31 GAO- 03- 986 Rural Ambulance Services

Appendix II Characteristics of Rural Counties Grouped by Medicare
Population Density

Page 32 GAO- 03- 986 Rural Ambulance Services

Appendix II Characteristics of Rural Counties Grouped by Medicare
Population Density

Page 33 GAO- 03- 986 Rural Ambulance Services

Page 34 GAO- 03- 986 Rural Ambulance Services

Appendix III

Page 35 GAO- 03- 986 Rural Ambulance Services

Appendix IV

Page 36 GAO- 03- 986 Rural Ambulance Services

United States General Accounting Office Washington, D. C. 20548- 0001
Official Business Penalty for Private Use $300 Address Service Requested

Presorted Standard Postage & Fees Paid

GAO Permit No. GI00
*** End of document. ***