Medicare Physician Payments: Concerns about Spending Target	 
System Prompt Interest in Considering Reforms (08-OCT-04,	 
GAO-05-85).							 
                                                                 
Concerns were raised about the current system Medicare uses to	 
determine annual changes to physician fees--the sustainable	 
growth rate (SGR) system--when fees were reduced by 5.4 percent  
in 2002. Subsequent administrative and legislative actions	 
modified or overrode the SGR system, resulting in fee increases  
for 2003, 2004, and 2005. However, projected fee reductions for  
2006-2012 have raised new concerns about the SGR system.	 
Policymakers are considering whether to eliminate spending	 
targets or modify them. The Medicare Prescription Drug, 	 
Improvement, and Modernization Act of 2003 (MMA) required that	 
GAO study SGR and potential alternatives to the system. This	 
report examines (1) how the SGR system is designed to control	 
spending for physician services, (2) what concerns have been	 
raised about the SGR system and its components, (3) what affects 
the stability and predictability of physician fee updates under  
the SGR system, and (4) what alternatives to the current SGR	 
system exist. GAO reviewed relevant laws and regulations and	 
interviewed officials and organizations representing physicians. 
On the basis of this information, GAO identified potential	 
alternatives to the SGR system and requested illustrative	 
simulations of fee updates and spending on physician services	 
from the Centers for Medicare & Medicaid Services (CMS).	 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-85						        
    ACCNO:   A13032						        
  TITLE:     Medicare Physician Payments: Concerns about Spending     
Target System Prompt Interest in Considering Reforms		 
     DATE:   10/08/2004
  SUBJECT:   Medicare Physician Payments 				        
                                                                 

******************************************************************
** 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-05-85

                 United States Government Accountability Office

GAO

                       Report to Congressional Committees

October 2004

                                    MEDICARE
                                   PHYSICIAN
                                    PAYMENTS

  Concerns about Spending Target System Prompt Interest in Considering Reforms

                                       a

GAO-05-85

[IMG]

October 2004

MEDICARE PHYSICIAN PAYMENTS

Concerns about Spending Target System Prompt Interest in Considering Reforms

What GAO Found

To moderate Medicare spending for physician services, the SGR system sets
spending targets and adjusts physician fees based on the extent to which
actual spending aligns with specified targets. If growth in the number of
services provided to each beneficiary-referred to as volume-and in the
average complexity and costliness of services-referred to as intensity-is
high enough to cause spending to exceed the SGR target, fee updates are
set lower than inflation in the cost of operating a medical practice. A
wide enough gap between spending and the target results in fee reductions.

Physician groups are dissatisfied with SGR as a system to update physician
fees. For example, they question the fairness of including rapidly growing
spending for physician-administered drugs in the SGR system's definition
of physician services expenditures. The groups also contend that the
allowance for growth in volume and intensity is too low and lacks the
flexibility to allow for factors outside physicians' control.

Fee updates under the SGR system have varied widely within an allowed
range largely because of annual fluctuations in the growth of the volume
and intensity of services that physicians provide to beneficiaries.
Certain system design features, such as the use of cumulative spending
targets and the need to estimate data, also reduce the stability and
predictability of updates. However, MMA's revision of the allowance for
growth in volume and intensity of services from an annual change to a
10-year moving average will help to make future updates more stable and
predictable.

Possible alternatives to the SGR system cluster around the two broad
approaches under consideration: (1) end the use of spending targets and
separate fee updates from explicit efforts to moderate spending growth or
(2) retain spending targets but modify the current SGR system to address
perceived shortcomings. CMS projects that either of the two approaches
will result in higher aggregate spending, thereby increasing the
difficulty of addressing Medicare's long-run financial challenges. The
first approach emphasizes stable fee updates, while the second approach
automatically adjusts fee updates if spending growth deviates from a
predetermined target. While seeking to pay physicians appropriately, it is
important to consider how modifications or alterations to the SGR system
would affect the longterm sustainability and affordability of the Medicare
program. In this context, the choice between the two approaches may hinge
on whether primary consideration should be given to stable fee increases
or to the need for fiscal discipline within the Medicare program.

CMS agreed with the concluding observations in the draft report. Groups
representing physicians commented that overall, the draft report offered a
good analysis of problems with the SGR system, but did not fully reflect
their concerns. We modified the draft as appropriate.

                 United States Government Accountability Office

Contents

Letter

Results in Brief
Background
SGR System Designed to Adjust Fee Updates to Bring Actual

Spending for Physician Services in Line with Spending Targets Various
Concerns Raised about SGR System and Its Components Variable Growth in
Provision of Physician Services and Certain

SGR System Design Elements Reduce Stability and Predictability of
Physician Fee Updates Alternatives for Updating Physician Fees Would
Eliminate

Spending Targets or Revise Current SGR System Concluding Observations
Agency and Industry Comments and Our Evaluation

                                       1

                                      4 6

                                     12 17

21

27 56 57

Appendix I Calculation of the Performance Adjustment Factor

Appendix II~Corrections to Prior Estimates Caused the SGR System's
Cumulative Targets to Produce Negative Updates

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

Tables

Table 1: CMS's Estimate of the 2005 Sustainable Growth Rate and Its
Determinants, as of March 2004 13 Table 2: Estimated 2005 Fee Update, as
of March 2004 16

Figures

Figure 1: Average Annual Percentage Change in Medicare Spending for
Physician Services per Beneficiary, 1980-2003 8 Figure 2: Growth in Volume
and Intensity of Medicare Physician Services per Beneficiary, 1980-2003 10
Figure 3: Projected MEI and Fee Update under Current Law 24

Figure 4: Annual Percentage Change in Real GDP Per Capita and

10-Year Moving Average Change in Real GDP Per Capita,

1992-2003 26 Figure 5: Projected MEI and Fee Updates under Current Law and

under Option of Eliminating Spending Targets and Tying

Fee Updates to MEI 30 Figure 6: Projected Aggregate Spending under Current
Law and

under Option of Eliminating Spending Targets and Tying

Fee Updates to MEI 31 Figure 7: Projected Real Spending for Physician Fee
Schedule

Services per FFS Beneficiary under Current Law and

under Option of Eliminating Spending Targets and Tying

Fee Updates to MEI 32 Figure 8: Projected MEI and Fee Updates under
Current Law and

under Option of Not Including Any Part B Drugs in the

SGR System Beginning in 2005 35 Figure 9: Projected Aggregate Spending
under Current Law and

under Option of Not Including Any Part B Drugs in the

SGR System Beginning in 2005 36 Figure 10: Projected Real Spending for
Physician Fee Schedule

Services per FFS Beneficiary under Current Law and

under Option of Not Including Any Part B Drugs in the

SGR System Beginning in 2005 37 Figure 11: Projected MEI and Fee Updates
under Current Law and

under Option of Resetting the Cumulative Spending

Target Equal to Cumulative Actual Spending as of 2006 39 Figure 12:
Projected Aggregate Spending under Current Law and

under Option of Resetting the Cumulative Spending

Target Equal to Cumulative Actual Spending as of 2006 40 Figure 13:
Projected Real Spending for Physician Fee Schedule

Services per FFS Beneficiary under Current Law and

under Option of Resetting the Cumulative Spending

Target Equal to Cumulative Actual Spending as of 2006 41 Figure 14:
Projected MEI and Fee Updates under Current Law and

under Option of Eliminating the Cumulative Aspect of

Spending Targets 43 Figure 15: Projected Aggregate Spending under Current
Law and

under Option of Eliminating the Cumulative Aspect of

Spending Targets 44

Figure 16: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Option of Eliminating the
Cumulative Aspect of Spending Targets 45

Figure 17: Projected MEI and Fee Updates under Current Law and under
Option of Increasing Volume and Intensity Growth Allowance to GDP Plus 1
Percentage Point 47

Figure 18: Projected Aggregate Spending under Current Law and under Option
of Increasing Volume and Intensity Growth Allowance to GDP Plus 1
Percentage Point 48

Figure 19: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Option of Increasing Volume
and Intensity Growth Allowance to GDP Plus 1 Percentage Point 49

Figure 20: Projected MEI and Fee Updates under Current Law and under
Combination of Options of Resetting the Cumulative Spending Target Equal
to Cumulative Actual Spending as of 2006, Not Including Any Part B Drugs
in the SGR System Beginning in 2005, and Using GDP Plus 1 Percentage Point
51

Figure 21: Projected Aggregate Spending under Current Law and under
Combination of Options of Resetting the Cumulative Spending Target Equal
to Cumulative Actual Spending as of 2006, Not Including Any Part B Drugs
in the SGR System Beginning in 2005, and Using GDP Plus 1 Percentage Point
52

Figure 22: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Combination of Options of
Resetting the Cumulative Spending Target Equal to Cumulative Actual
Spending as of 2006, Not Including Any Part B Drugs in the SGR System
Beginning in 2005, and Using GDP Plus 1 Percentage Point 53

Figure 23: Projected MEI and Fee Updates under Current Law and under
Combination of Options of Not Including Any Part B Drugs in the SGR System
Beginning in 2005 and Eliminating the Cumulative Aspect of Spending
Targets 54

Figure 24: Projected Aggregate Spending under Current Law and under
Combination of Options of Not Including Any Part B Drugs in the SGR System
Beginning in 2005 and Eliminating the Cumulative Aspect of Spending
Targets 55

Figure 25: Projected Real Spending for Physician Fee Schedule

Services per FFS Beneficiary under Current Law and

under Combination of Options of Not Including Any Part

B Drugs in the SGR System Beginning in 2005 and

Eliminating the Cumulative Aspect of Spending Targets 56 Figure 26:
Formula Used to Determine the Performance

Adjustment Factor in 2005 63 Figure 27: Percentage Change in MEI, SGR Fee
Schedule Update,

and Medicare Physician Services Spending per

Beneficiary, 1998-2005 65

Abbreviations

BBA Balanced Budget Act of 1997
BBRA Medicare, Medicaid, and SCHIP Balanced Budget

Refinement Act of 1999 CBO Congressional Budget Office CMS Centers for
Medicare & Medicaid Services CPI-U consumer price index for urban
consumers ESRD end-stage renal disease FFS fee-for-service GDP gross
domestic product HHS Department of Health and Human Services HI Hospital
Insurance MedPAC Medicare Payment Advisory Commission MEI Medicare
Economic Index MMA Medicare Prescription Drug, Improvement, and

Modernization Act of 2003 MVPS Medicare volume performance standard OACT
Office of the Actuary PAF performance adjustment factor PPRC Physician
Payment Review Commission SGR sustainable growth rate SMI Supplementary
Medical Insurance

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 8, 2004

Congressional Committees

Physicians and others raised concerns about the current system Medicare
uses to determine annual changes to physician fees when those fees were
reduced by 5.4 percent in 2002. This reduction was triggered, in part,
because spending on physician services had exceeded predetermined spending
targets and Medicare's system for updating fees-the sustainable growth
rate (SGR) system-called for a reduction in fees to impose fiscal
discipline.1 Subsequent administrative and legislative actions modified or
overrode the SGR system, resulting in fee increases for 2003, 2004, and
2005. Absent additional action, however, fees are expected to fall by
approximately 5 percent each year beginning in 2006 and continuing through
2012 as the SGR system attempts to offset previous excess spending and
align actual spending with the system's spending targets. According to
physician groups, such a decline in fees would likely discourage many
physicians from treating Medicare beneficiaries. As a result of these
concerns, policymakers are interested in considering the appropriateness
of current spending targets and the SGR system as a method for determining
physician fee updates. Essentially, they are considering whether to
eliminate spending targets or retain them, while making modifications to
the system.

Although the current focus of concern is largely on the potential for
declining physician fees, the historic challenge for Medicare has been to
find ways to moderate the rapid growth in spending for physician services
under the Medicare Supplementary Medical Insurance (SMI)-or Part B-
program. In the 1980s, attempts to moderate spending by limiting physician
fees without addressing aggregate expenditures for physician services were
unsuccessful because increases in the number of services physicians
provided per beneficiary-known as volume-and the average complexity and
costliness of those services-known as intensity- continued to drive up
spending. As a result, in the Omnibus Budget Reconciliation Act of 1989,2
the Congress required the establishment of a

1The SGR system reduced fees by 4.8 percent. Additional adjustments
resulted in a total fee reduction of 5.4 percent.

2See Pub. L. No. 101-239, S:6102, 103 Stat. 2106, 2169-89.

national Medicare physician fee schedule and a system for annually
updating fees that included spending targets. The fee schedule and
spending targets first affected physician fees in 1992. The SGR system,
Medicare's current system for updating physician fees, was established in
the Balanced Budget Act of 1997 (BBA) and was implemented in 1998.3  Both
the SGR and its predecessor system provided for cumulative fee updates
that generally exceeded cumulative increases in physicians' cost of
providing services.4 Since the establishment of the national fee schedule
and spending targets, the growth in spending for Medicare physician
services has slowed substantially. Nonetheless, recent increases in
physician expenditures due to volume and intensity growth are a reminder
that the historic challenge of moderating spending growth has not
disappeared.

The Medicare Prescription Drug, Improvement, and Modernization Act of 2003
(MMA) required us to study certain adjustments to physician fees,
including the SGR system and alternatives to the system.5 As discussed
with the committees of jurisdiction, this report examines (1) how the SGR
system is designed to control spending for physician services, (2) what
concerns have been raised about the SGR system and its components, (3)
what affects the stability and predictability of physician fee updates
under the SGR system, and (4) what alternatives to the current SGR system
exist.

In addressing these objectives, we analyzed Medicare expenditure data from
the Medicare Trustees' 1998 and 2004 annual reports.6 We also reviewed
laws and regulations pertaining to the SGR system and its predecessor
spending target system and interviewed officials at the

3See Pub. L. No. 105-33, S:4503, 111 Stat. 251, 433-34. BBA set a specific
fee update for 1998. See BBA, S:4505, 111 Stat. 435-39. Physician fees
were first affected by the SGR system in 1999.

4Specifically, from 1992 through 2001, fee updates resulting from the SGR
and its predecessor system, increased by 39.7 percent, whereas input
prices increased by 25.9 percent. These updates do not reflect other
required adjustments, such as those for legislated changes and for budget
neutrality.

5See Pub. L. No. 108-173, S:953, 117 Stat. 2066, 2427-28.

6Boards of Trustees of the Federal Hospital Insurance and Federal
Supplementary Medical Insurance Trust Funds, 1998 Annual Report of the
Board of Trustees of the Federal Supplementary Medical Insurance Trust
Fund (Washington, D.C.: Apr. 28, 1998), and 2004 Annual Report of the
Board of Trustees of the Federal Hospital Insurance and Federal
Supplementary Medical Insurance Trust Funds (Washington, D.C.: Mar. 23,
2004).

Centers for Medicare & Medicaid Services (CMS), the agency responsible for
administering Medicare; the Congressional Budget Office (CBO); the
Medicare Payment Advisory Commission (MedPAC);7 and organizations
representing physicians, including the American Medical Association, the
Medical Group Management Association, the Alliance for Specialty Medicine,
and the American College of Physicians. On the basis of these document
reviews and interviews, we identified potential alternatives to the SGR
system. We requested illustrative simulations of fee updates and total
spending under these alternatives from the CMS Office of the Actuary
(OACT).8 Total spending includes expenditures from all sources-that is,
government outlays and beneficiary spending, including monthly Part B
premiums, deductibles, and coinsurance payments.9 Because the simulation
estimates produced by CMS OACT include total spending from all sources,
the estimated spending changes will differ from CBO's cost estimates for
the same alternatives. CBO, which is responsible for estimating how
legislated changes would affect federal spending, does not include
beneficiary spending when it estimates the cost of SGR alternatives. CMS
OACT and CBO estimates may also differ as the result of differences in the
underlying assumptions used by the two agencies. Our analyses apply only
to spending affected by the SGR system-that is, physician spending in the
traditional fee-for-service (FFS) program. We assessed the reliability of
the Medicare expenditure data and data used for the simulations under
alternatives to the SGR system by interviewing agency officials
knowledgeable about the data and who are responsible for producing the
projections for the SGR system. We determined that the data were
sufficiently reliable for the purposes of our study. We performed this
work from January 2004 through September 2004 in accordance with generally
accepted government auditing standards.

7MedPAC is an independent federal body that advises the Congress on issues
affecting the Medicare program.

8CMS OACT has the program responsibility to calculate Medicare's spending
targets for physician services and annual physician fee updates. In
producing these simulations, CMS OACT used the agency's assumptions
regarding the various factors that affect the SGR system, such as
projected fee-for-service enrollment.

9The Part B premium amount is adjusted each year so that expected premium
revenues equal 25 percent of expected Part B spending. Beneficiaries must
pay coinsurance-usually 20 percent-for most Part B services.

Results in Brief

To help impose fiscal discipline and moderate Medicare spending for
physician services, the SGR system sets spending targets and adjusts fees
paid to physicians based on the extent to which actual spending aligns
with specified targets. SGR system targets are designed to allow real
spending per beneficiary-that is, spending per beneficiary adjusted for
the estimated underlying cost of providing physician services-to grow at
the same rate that the national economy (as measured by the rate that real
gross domestic product (GDP)) grows over time on a per capita basis-
currently estimated to be about 2.3 percent annually.10 If Medicare
spending for physician services remains on target, the annual increase in
physician fees is set equal to the estimated change in physicians' cost of
providing services.11 However, if growth in the volume and intensity of
services provided is high enough to cause spending to exceed the SGR
system target, future fee updates are set below the estimated increase in
physicians' average cost for providing services-in other words, physicians
receive fee increases that are lower than the Medicare Economic Index
(MEI). If the gap between spending and the target is wide enough, the SGR
system results in fee reductions. Conversely, if volume and intensity
growth is low enough to cause spending to fall below the target, the SGR
system benefits physicians by producing fee increases that exceed the
change in their cost of providing services. Under the SGR system's
cumulative spending targets, excess spending that is not offset in one
year accumulates in succeeding years until it is recouped.

Physician groups are dissatisfied with SGR as a system to update physician
fees and have raised various concerns about its components. In general,
they note that expenditures for physician services constitute Medicare's
only spending that is subject to a target system. Physician groups report
that under this system, fee updates-which are explicitly linked to
spending controls-have caused payment rates in recent years to fall behind
physicians' cost of providing services. Among specific concerns, physician
groups question the fairness of reducing fee updates for physician
services to offset rapidly growing expenditures for certain outpatient
drugs that are covered by Medicare Part B and that are largely physician
administered. The groups also contend that the SGR system's

10This rate incorporates the 10-year moving average of real GDP per
capita.

11The change in the cost of providing physician services is measured by
the Medicare Economic Index (MEI). MEI measures input prices for resources
needed to provide physician services. It is designed to estimate the
increase in the total cost for the average physician to operate a medical
practice.

allowance for spending growth due to volume and intensity increases-the
growth rate of real GDP per capita-is too low and inflexible. Physician
groups contend that as a result, factors outside physicians' control-such
as any future declines in the FFS population's average health status and
introduction of new, effective medical technology-may cause spending to
exceed the SGR system targets and thus lead to reduced fee updates.
Additional concerns include whether CMS's method used to account for
spending increases due to changes in laws and regulations-which can change
payments or expand the extent and number of Medicare-covered services-is
sufficiently complete, accurate, and transparent.

For several reasons, fee updates under the SGR system have varied- within
a specified range-and have been difficult to predict accurately.12 A
principal cause of variation within this range has been annual
fluctuations in the growth of the volume and intensity of services that
physicians provide to beneficiaries. Since the SGR system was implemented
in 1998, volume and intensity growth has ranged from 1.2 percent in 1999
to 6.1 percent in 2002. Two system design characteristics also reduce the
stability and predictability of updates. First, the SGR system is designed
to respond to fluctuating volume and intensity growth by adjusting fee
updates to keep cumulative spending in line with the targets. Attempting
to control cumulative spending tends to amplify the variation in annual
updates. For example, if spending has exceeded the spending target, the
SGR system must reduce future updates both to slow future spending growth
and to recoup previous excess spending. Second, uncertainty in estimates
of data used in the SGR system makes long-term estimates of fee updates
less predictable and causes updates to vary from year to year as new data
become available and estimates of data used in the SGR system are revised.

Alternatives to the SGR system we identified cluster around the two
approaches that policymakers are considering. One approach would end the
use of spending targets-separating fee updates from efforts to moderate
spending growth. MedPAC is a proponent of this approach and since 2001 has
recommended tying fee updates to estimated changes in physicians' cost of
providing services. It has further recommended that Medicare seek to
control spending growth by, among other things,

12The SGR system permits annual physician fee updates to vary by as much
as 7 percent below to 3 percent above the estimated change in physicians'
cost of providing services as measured by MEI.

identifying and addressing the utilization of rapidly growing services,
such as diagnostic imaging. The other approach includes alternatives that
would retain spending targets but modify the current SGR system to address
perceived shortcomings. These modifications could include removing the
Part B prescription drug expenditures that are currently counted in the
SGR system; resetting the targets by not requiring the system to recoup
previous excess spending; using annual, rather than cumulative, targets to
dampen the fluctuation in fee updates; and modifying the allowance for
increased spending due to volume and intensity growth. The advantage of
eliminating spending targets would be greater fee update stability and
predictability, whereas the advantage of retaining spending targets as
part of the system for updating fees is that the system would
automatically work to moderate spending if volume and intensity growth
began to increase above allowable rates. However, either approach compared
to current law, under which fees are projected to be reduced by as much as
5 percent or more for several years, will be very expensive-ranging from 4
percent to 23 percent higher cumulative spending over the 10-year period
from 2005 to 2014. Given the importance of the long-term sustainability
and affordability of the Medicare program, examining the impact of
spending over a longer period may be appropriate when contemplating
modifications or alternatives to the SGR system.

CMS agreed with our concluding observations and expressed its commitment
to pay physicians appropriately to ensure that Medicare beneficiaries have
access to high-quality health care. Groups representing physicians
commented that overall, a draft of our report offered a good analysis of
problems with the SGR system, but indicated it did not fully reflect the
extent of their concerns. Some of the issues the groups raised were
outside the scope of our report. We modified the report as appropriate.

Background 	Medicare spending per beneficiary on physician services has
varied substantially-both among geographic areas and in its growth over
time. The geographic variation in spending-unrelated to beneficiary health
status or outcomes-provides evidence that health needs alone do not
determine spending. Consequently, policymakers have deemed it both
reasonable and desirable to question the appropriateness of current and
projected physician services spending and to explicitly consider the
affordability of such spending when setting physician fees. The
implementation of a national fee schedule and spending targets in 1992,
for example, was designed, in part, to address issues of affordability and
program sustainability by slowing spending growth. Moderating this

growth remains part of the larger effort to ensure future Medicare program
sustainability.

Some Spending on Physician Services May Be Unnecessary, as Suggested by
Unwarranted Regional Variation in Use of Physician Services

In 1989, the Physician Payment Review Commission (PPRC) reported that from
1979 through 1989 (the decade prior to the establishment of spending
targets), Medicare spending on physician services per beneficiary more
than tripled, rising much more rapidly than general inflation.13 At that
time, PPRC recommended an expenditure target for controlling aggregate
spending on physician services. The target was to apply initially to all
physician services nationally and later to evolve to separate targets for
regions, categories of physician services, or both.

Then, as now, utilization of physician services varied widely by
geographic area, while the Medicare patient populations in these areas
differed little from one another in their illnesses. Some studies report
that variation in service use indicates that in some parts of the country
compared with others, there was either overuse or underuse of services.
Recent studies of Medicare expenditures show that regional variation in
the use of medical services remains and that the spending disparities
among areas are explained by physicians' discretionary practices rather
than by differences in patient populations' health status.14

Physician Service  Three periods from 1980 to the present describe
Medicare's recent Expenditures Have Grown experience in spending for
physician services. Figure 1 shows growth in Less Rapidly after Medicare
spending per beneficiary for physician services during the three

periods. In the first period, 1980 through 1991, Medicare's payment
ratesSpending Targets and Fee

for physician services were based on historical charges for these
services,Schedule Were Established and limits were placed on fees and fee
updates but not on aggregate

13PPRC, established by the Congress in the Consolidated Omnibus Budget
Reconciliation Act of 1985, Pub. L. No. 99-272, S:9305, 100 Stat. 82,
190-91 (1986), was charged with advising the Congress on methods to reform
payment to physicians under the Medicare program and with making
recommendations annually. Subsequent legislation expanded PPRC's
responsibilities to include, among other things, setting standards for
expenditure growth and updating fees and monitoring beneficiary access and
financial liability. In 1997, BBA dissolved PPRC and the Prospective
Payment Assessment Commission and formed MedPAC. BBA, S:4022, 111 Stat.
350-355.

14John E. Wennberg, Elliot S. Fisher, and Jonathan S. Skinner, "Geography
And The Debate Over Medicare Reform," Health Affairs Web Exclusive,
February 13, 2002; E.S. Fisher et al., "The Implications of Regional
Variations in Medicare Spending, Part 1: The Content, Quality, and
Accessibility of Care," Annals of Internal Medicine (2003): 273-287; and
E.S. Fisher et al.,"The Implications of Regional Variations in Medicare
Spending, Part 2: Health Outcomes and Satisfaction with Care," Annals of
Internal Medicine (2003): 288-298.

spending. In the 1992 through 1997 period, physician services were paid
under a national fee schedule, and the first spending target system-called
the Medicare volume performance standard (MVPS)-set an allowable growth
rate for aggregate spending that was used to adjust physician fees. From
1998 on, services continue to be paid under a fee schedule and the SGR
system replaced the MVPS system and uses a different method to set an
acceptable growth rate for aggregate spending.

Figure 1: Average Annual Percentage Change in Medicare Spending for
Physician Services per Beneficiary, 1980-2003

Percentage 14

12 11.6

10

8

6

4

2

0 Pretargets MVPS SGR 1980-1991 1992-1997 1998-2003

Source: GAO analysis of data from CMS and the Boards of Trustees of the
Federal Hospital Insurance (HI) and Supplementary Medical Insurance (SMI)
Trust Funds.

Notes: Spending changes for 1980 through 1991 are for the years ending
June 30 and represent average Medicare spending for beneficiaries in the
traditional FFS program, net of beneficiary cost sharing. Spending for
end-stage renal disease (ESRD) patients is not included. Spending changes
for 1992 through 1997 and 1998 through 2003 are for calendar years and
represent changes in total allowed charges-Medicare spending, including
beneficiary cost sharing-for beneficiaries in the traditional FFS program.

In the 1980s, Medicare paid physicians on the basis of "reasonable
charge," defined as the lowest of the physician's actual charge, the
customary charge (the amount the physician usually charged for the
service), or the prevailing charge (based on comparable physicians'
customary charges). Under this system, payment inconsistencies existed
among physicians by services, specialties, and locations. The system also
had an inflationary bias, as a rise in customary charges could increase
prevailing charges over

time.15 During this decade, expenditures for physician services grew
rapidly: from 1980 through 1991, Medicare spending per beneficiary for
physician services grew at an average annual rate of 11.6 percent.
Although the Congress froze fees or limited fee increases in the 1980s,
spending continued to rise because there were no limits on growth in the
volume and intensity of services physicians provided to beneficiaries.

Recognizing that the expenditure growth of the 1980s was not sustainable,
the Congress reformed the way Medicare paid for physician services in the
traditional FFS program by requiring the establishment of a national fee
schedule for physician services and a system for controlling aggregate
physician service spending, MVPS. The establishment of a fee schedule in
1992 was an attempt to break the link between physicians' charges and
Medicare payments. The fee schedule was designed to pay for services based
on the relative resources used by physicians to provide different types of
care and to address the inflationary bias of the charge-based system. The
adoption of a spending target system was an attempt to control spending
growth attributable to increases in the volume and intensity of physician
services.

Under MVPS, a performance standard for a given year was set, indicating a
growth rate for expenditures that should not be exceeded. The extent to
which actual expenditure growth fell above or below the performance
standard helped to determine the update to physician fees 2 years later.
For example, in 1993, CMS compared actual spending in 1992 with the
performance standard for 1992; the difference largely determined the
update to physician fees in 1994.16 The performance standard was based on
changes in four factors: the number of FFS Medicare beneficiaries,
practice cost inflation, the historical growth in volume and intensity,
and

17

laws and regulations that could affect spending for physician services.

From 1992 through 1997-the period that MVPS was used to set fee
updates-annual spending growth for physician services was far lower than
in the preceding decade. The decline in spending growth during this

15Beginning in 1975, increases in prevailing charges were limited to the
change in MEI.

16Under MVPS, the fee updates depended on both the change in MEI and the
difference between actual spending and the performance standard.

17Inflation was measured as a weighted average of input price increases,
estimated by MEI for physician services and the consumer price index for
urban consumers (CPI-U) for laboratory services.

period was the result, in large part, of slower volume and intensity
growth. For example, from 1985 through 1991, spending per beneficiary grew
at an average annual rate of 10.8 percent; during that period, volume and
intensity of service use per beneficiary rose an average 7 percent
annually. From 1992 through 1997, the growth in spending per beneficiary
fell to 4.4 percent; during that period, average annual growth in volume
and intensity of service use per beneficiary fell to 1 percent. (See fig.
2.)

Figure 2: Growth in Volume and Intensity of Medicare Physician Services
per Beneficiary, 1980-2003

Percentage 10 9 8 7 6 5 4 3 2 1 0 -1

9.7

1980 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
1999 2000 2001 2002 2003

                              Charge-based system

                             Fee schedule and MVPS

                              Fee schedule and SGR

Source: GAO analysis of data from CMS and the Boards of Trustees of the
Federal HI and SMI Trust Funds.

Notes: Data are for beneficiaries in the traditional FFS program only.
Data for ESRD patients are not included. From 1980 through 1992, volume
and intensity of services changes are based on Medicare outlays for all
physician services. From 1993 through 2003, volume and intensity of
services changes are based on Medicare outlays for physician services
covered by the fee schedule.

Concerns about the MVPS spending targets arose in 1995 when physician fees
were expected to fall over time unless there were continual declines in
the volume and intensity of services provided.18, 19 In response to the

18Physician Payment Review Commission, 1995 Annual Report to Congress
(Washington, D.C.: 1995).

system's perceived shortcomings, the Congress took action in BBA in 1997
to replace it with the SGR system.20 In 1998 and 1999, the first 2 years
of the SGR system, volume and intensity growth remained similar to the
rate under MVPS. However, from 2000 through 2003, volume and intensity
growth rose at an average annual rate of about 5 percent.21 Over the 1998-
2003 SGR system period,22 the average growth in volume and intensity of
services per Medicare beneficiary was higher than the average for the
1992-1997 MVPS period-but substantially below that experienced before
spending targets were introduced. Since the introduction of the SGR
system, total spending on physician services is projected to grow by an
average of 8 percent a year from 2000 through 2005.

Controlling Spending for Physician Services Part of Larger Challenge to
Maintain Fiscal Discipline in Medicare

In 2003, Medicare spending for physician services totaled nearly $48
billion,23 which accounted for about one-sixth of program spending
overall. We and others have argued for the need for additional fiscal
discipline in Medicare.24 Within the next 10 years, the federal budget
will experience significant increases in spending pressure, due primarily
to known demographic trends and rising health care costs. Expected
technological advances-involving new drugs and diagnostic procedures,
among other things-may improve health outcomes but will likely increase
the price tag of a Medicare program that is already unsustainable

19The MVPS spending target was based, in part, on a 5-year historical
trend in volume and intensity reduced by a specified number of percentage
points. Because of this design and the fact that volume and intensity
growth dropped dramatically after the adoption of the MVPS system, the
target for future volume and intensity increases fell too.

20The SGR system was revised by the Medicare, Medicaid, and SCHIP Balanced
Budget Refinement Act of 1999 (BBRA), (Pub. L. No. 106-113, App. F,
S:211(b), 113 Stat. 1501A-321, 348-49) and by MMA (see S:601(b), 117 Stat.
2301).

21This recent growth in volume and intensity for physician services is
higher than the 3 percent a year that CMS OACT is projecting for 2005
through 2014.

22In 2002, a year in which physicians' fees fell by 5.4 percent, volume
and intensity grew by 6.1 percent, the largest growth in a single year
since the fee schedule and spending targets were introduced.

23This figure does not include spending associated with Medicare's private
plan option.

24GAO, Medicare: Financial Challenges and Considerations for Reform,
GAO-03-577T (Washington, D.C.: Apr. 10, 2003); Congressional Budget
Office, Medicare's Long-Term Financial Condition, testimony before the
Joint Economic Committee (Apr. 10, 2003); Office of Management and Budget,
Analytical Perspectives, Budget of the United States Government, Fiscal
Year 2005 (Washington, D.C.: Feb. 2, 2004); and Boards of Trustees of the
Federal Hospital Insurance and the Federal Supplementary Medical Insurance
Trust Funds, 2004 Report of the Boards of Trustees of the Federal Hospital
Insurance and the Federal Supplementary Medical Insurance Trust Funds.

in its present form. In light of physician service expenditures'
significant contribution to aggregate spending, containing their growth
plays an important role in helping to address the program's long-range and
fundamental financing problem.

SGR System Designed to Adjust Fee Updates to Bring Actual Spending for
Physician Services in Line with Spending Targets

The SGR system is designed to impose fiscal discipline and to moderate
spending for physician services by adjusting annual fee updates to bring
spending in line with targets. The SGR system, similar to the predecessor
MVPS system, relies on spending targets because earlier attempts to
achieve fiscal discipline through limits on fee increases did not control
the spending that resulted from volume and intensity growth. The SGR
system uses a formula specified in statute to establish each year an
allowed spending growth rate, a spending target, and a fee update. Like
MVPS, the SGR system includes an allowance for volume and intensity
increases but, unlike MVPS, ties the allowance to a measure of the growth
of the national economy.

Spending Targets for Physician Services Used to Encourage Fiscal
Discipline

As noted, spending targets were established-first under MVPS and later
under the SGR system-because policymakers contended that the fee schedule
alone would not have adequately constrained expenditure growth for
physician services. The fee schedule limits payment for individual
services but does not moderate spending growth resulting from volume and
intensity increases. Although the SGR system's spending target does not
cap expenditures for physician services, it serves as a budgetary control
by automatically lowering fee updates in response to excess spending due
to volume and intensity growth. In addition, reduced fee updates serve as
a signal to physicians collectively and to the Congress that spending due
to volume and intensity has increased more than allowed.

An additional reason for spending targets was advanced by PPRC in its 1995
report to the Congress.25 PPRC explained that spending targets were
intended, in part, to create a collective incentive for physicians.
Specifically, the report stated that spending targets "provid[e] the
medical profession with a collective incentive to reduce inappropriate
care by, for instance, developing and disseminating practice guidelines
that promote cost-effective practice styles."

25Physician Payment Review Commission, 1995 Annual Report to Congress.

SGR System Sets Allowable Spending Growth and Targets for Physician
Services

Every year, CMS must estimate the allowed rate of increase in spending for
physician services and use that rate to construct the annual spending
target for the following calendar year.26 The sustainable growth rate is
the product of the estimated percentage change in (1) input prices for
physicians' services; 27, 28 (2) the average number of Medicare
beneficiaries in traditional FFS; (3) national economic output, as
measured by real (inflation-adjusted) GDP per capita; and (4) expected
expenditures for physician services resulting from changes in laws or
regulations. CMS's current estimate of the sustainable growth rate for
2005 is 4.6 percent, based on the agency's estimates of the four factors.
(See table 1.)

Table 1: CMS's Estimate of the 2005 Sustainable Growth Rate and Its
Determinants, as of March 2004

Estimated percentage change

                      Sustainable growth rate determinants

a

Input prices for physician services

Traditional FFS Medicare enrollment

Real GDP per capita

Expenditures for physician services resulting from changes in laws and
regulations

                  Estimated 2005 sustainable growth rate 4.6b

Source: CMS OACT.

aFor purposes of the sustainable growth rate, physician services include
services paid for by the fee schedule as well as laboratory services and
certain Medicare-covered Part B outpatient drugs.

bThe sustainable growth rate is computed as the product of the percentage
change in the four factors. The percentage changes are expressed in
decimal form relative to 1.0. For example, a percentage change of 2.6
percent is expressed as 1.026. Therefore, the sustainable growth rate is
computed as (1.026) x (0.998) x (1.022) x (1.0) = 1.046 , or 4.6 percent.

26This allowed rate is the sustainable growth rate from which the SGR
system derives its name. For the purposes of this report, we use the
abbreviation SGR when referring to the system and the full term of
"sustainable growth rate" when referring to the allowed rate of increase.

27CMS calculates changes in physician input prices based on the growth in
the costs of providing physician services as measured by MEI, growth in
the costs of providing laboratory tests as measured by CPI-U, and growth
in the cost of Medicare Part B prescription drugs included in SGR
spending.

28Under the SGR and MVPS systems, the Secretary of Health and Human
Services defined "physician services" to include "services and supplies
incident to physicians' services," such as laboratory tests and most Part
B prescription drugs.

To set each year's spending target, the SGR system increases the previous
year's spending target by the sustainable growth rate for the given
year.29  For example, target spending for 2004 was $77.3 billion. Target
spending for 2005 is 4.6 percent higher, or $80.9 billion. Under the SGR
system, every annual target depends on the targets set in all previous
years since the base year 1996.30 BBRA required CMS, in calculating each
year's SGR system spending target, to first revise the targets set for the
2 previous years using the most recent available data for all elements of
the target- that is, revisions to figures for input prices for physician
services, FFS beneficiary enrollment, real GDP per capita, and
expenditures due to relevant new laws and regulations.31, 32

SGR System Adjusts Fee Updates to Align Spending with Target

Every fall, CMS determines whether the fee update for the following
calendar year must be adjusted to help align spending with targets.33 To
do so, the agency compares actual spending, measured cumulatively since
1996, to the cumulative value of the annual targets, measured over the
same period. If the two are equal, the fee update is set to equal the
estimated increase in physicians' average cost of providing services-as
measured by MEI. Otherwise, a performance adjustment factor (PAF) is used
to increase or decrease the update relative to MEI in order to help bring
spending back in line with the targets. (See app. I for the formula used
to calculate the PAF.) The PAF is subject to limits and may not cause the
update to be set at more than 3 percent above MEI or 7 percent below MEI.
In part because of these limits, adjustments to realign actual cumulative
spending with cumulative targets are spread out over more than 1 year.34

29The SGR system changed from a fiscal year basis to a calendar year basis
in 2000.

30The base year is the 12-month period ending March 31, 1997.

31See BBRA, S:211(b), 113 Stat. 1501A348-49.

32Revisions to targets first affected fee updates in 2001. In setting the
target for that year, CMS revised only the 2000 SGR target. According to
CMS, the agency was not authorized to revise the 1998 or 1999 SGR targets.

33Estimates of the fee update for the following year are made in the
spring. The final fee update is announced in November.

34The formula used in the SGR system spreads the recoupment of excess
spending over several years. Statutory limits on the PAF can increase the
time necessary to recoup excess spending.

For example, in projecting the update for 2005, CMS estimated that
cumulative spending from 1996 through 2004 ($543.8 billion) exceeded the
cumulative value of the annual targets during the same period ($531.9
billion) by approximately $11.9 billion. The estimated $11.9 billion
difference is due to two components of accumulated excess spending and
needs to be offset: first, excessive growth in volume and intensity in
2003 and 2004, and second, the additional spending attributable to MMA.
MMA replaced a fee reduction for 2004 with a minimum 1.5 percent increase,
but it did not adjust SGR system targets to account for the additional
spending.35 Because of the large discrepancy between spending and the
target, the SGR system calls for the maximum PAF reduction. In conjunction
with an estimated MEI of 2.8 percent, the application of the PAF would
produce a fee update of negative 4.4 percent.36 In addition to MEI and the
PAF, fees are sometimes subject to other adjustments, including those set
by law. For 2005, there is an additional adjustment of 0.8 percent, which
results in an estimated SGR system fee update for 2005 of negative 3.6
percent. (See table 2.) However, this negative update will be overridden
by an MMA-specified minimum update of 1.5 percent for 2005. The resulting
fee update is applied to the fee schedule's "conversion factor," a dollar
amount that translates each service's relative value into an

37, 38

actual disbursement amount.

35See MMA, S:601(a)(1), 117 Stat. at 2300.

36For 2005, the product of the change in input prices multiplied by the
PAF is equal to

[(1+.028) x (1 - 0.07)] -1 = -0.044, or -4.4 percent.

37The fee for each service is determined using a resource-based relative
value scale in which the resources required for a service are valued in
relation to the resources required to provide all other physician services
adjusted for the differences in the costs of providing services across
geographic areas. To arrive at a fee, the service's relative value is
multiplied by the dollar conversion factor.

38The update to the dollar conversion factor represents the aggregate of
increases and decreases across all services. Because the relative value of
individual services can change yearly, fee changes for specific services
may be different than the overall fee update.

     Table 2: Estimated 2005 Fee Update, as of March 2004 Percentage change

          Estimated 2005 fee update under SGR systema Update based on:

o  Change in input prices for physician services (MEI)b

o  PAF

o  BBRA required adjustmentc

2005 fee update as specified in MMAd

Source: CMS OACT.

aUpdate is computed as the product of the change in input prices, the PAF,
and the BBRA-required adjustment; all are expressed in decimal form. That
is, 2005 fee update = [(1+0.028) x (1 - 0.07) x (1+ 0.008)] -1 = -0.036 or
-3.6 percent.

bFor purposes of the fee update, physician services include only services
paid for under the fee schedule.

cThis adjustment, required by the Medicare, Medicaid, and SCHIP Balanced
Budget Refinement Act of 1999 (BBRA), maintains the budget neutrality of a
technical change to the conversion factor.

dThe actual fee update for 2005 is the greater of the calculated update of
-3.6 percent or 1.5 percent as legislated by MMA.

Under SGR's system of cumulative spending targets, excess spending that is
not offset in one year accumulates in succeeding years until it is
recouped. For 2005, MMA increased actual spending but did not adjust the
target for this additional spending. Now the gap between actual spending
and the target will result in an additional deficit that under the SGR
system will have to be recouped through negative updates in future years.

SGR System Ties Allowed Increases in Volume and Intensity to Growth in
National Economy

The parameters of the SGR system allow spending due to the volume and
intensity of physician services to increase, but limit that growth to the
same rate that the national economy (GDP) grows in real terms (that is,
adjusted for inflation) over time on a per capita basis. Under the SGR
system, if the volume and intensity of physician service use grows faster
than the national economy, the annual increase in physician fees will be
less than the estimated increase in the cost of providing services.
Conversely, if volume and intensity grows more slowly, the SGR system
permits physicians to benefit from fee increases that exceed the increased
cost of providing services. To reduce the effect of yearly business cycles
on physician fees, MMA required that economic growth be measured as the
10-year moving average change in real GDP per capita for each year

beginning in 2003.39 This measure is projected to range from 2.1 percent
to 2.5 percent during the 2005 through 2014 period.

When the SGR system was established, GDP growth was seen as a benchmark
that would allow for affordable increases in volume and intensity and also
one that represented a significant improvement over the benchmark included
in the previous MVPS system. In its 1995 annual report to the Congress,
PPRC stated that limiting real expenditure growth to 1 or 2 percentage
points above GDP would be a "realistic and affordable goal."40 Ultimately,
BBA specified the growth rate of GDP alone. This limit was an indicator of
what the nation could afford to spend on volume and intensity increases.
Whether this rate is a sufficient and appropriate allowance for volume and
intensity increases is uncertain. Currently, volume and intensity is
projected to grow by more than 4 percent per year, whereas the allowance
for this growth under the SGR system is about 2.3 percent annually. Such
excess volume and intensity growth is a key contributing factor to
negative fee updates.

Various Concerns Raised about SGR System and Its Components

Physician groups are dissatisfied with SGR as a system to update physician
fees and have raised various concerns about its key components. Noting
that physicians are uniquely subject to a system of fee updates that are
explicitly linked to spending controls, the groups contend that the SGR
system has caused payment rates in recent years to fall behind physicians'
cost of providing services. The groups' concerns with specific SGR system
components center on the following issues: the fairness of including
Medicare-covered outpatient drugs in the calculation of physician service
expenditures; the appropriateness of tying allowable volume and intensity
increases to the average growth in real GDP per capita; and the
completeness, accuracy, and transparency of the method used to account for
spending increases due to changes in laws and regulations.

39See MMA, S:601(b), 117 Stat. at 2301. 40Physician Payment Review
Commission, 1995 Annual Report to Congress.

Physicians Dissatisfied That Medicare Spending for Physician Services Is
Subject to Spending Targets

Physician groups are concerned that physicians are the only Medicare
provider type whose annual payment updates are subject to a spending
target system. Payment rate updates for hospitals and other institutional
providers, they note, are typically based on changes in the cost of
providing services. However, as CBO, MedPAC, and others have noted,
physicians are different from other providers in certain ways, which helps
to provide a rationale for the application of targets solely to physician
expenditures. Specifically, they note that physicians determine the
services they deliver to their patients and influence the care delivered
by other providers. In addition, under Medicare payment policies,
physicians receive a separate payment for each service they provide. Thus,
they can boost income by increasing the volume or intensity of services
they provide. For example, a physician may follow up a patient's visit by
scheduling another visit, even when such a follow-up visit is
discretionary and could be substituted with a telephone call. In contrast,
Medicare typically pays institutional providers a fixed amount for a
bundle of services; under this arrangement, no inherent incentive exists
to provide extra services, as doing so would not increase payments.

Physicians Question Fairness of Including

Part B Outpatient Drugs in Calculation of Physician Service Expenditures

One of physician groups' chief concerns is that through fee schedule
updates, the SGR system holds physicians accountable for the escalating
growth in Medicare expenditures for the majority of Part B-covered
drugs.41 (Drugs included in the SGR system are largely physician
administered and do not include all Part B-covered drugs.) The groups
contend that the SGR system should not include these drugs in the
calculation of aggregate physician service expenditures or the spending
targets. Although the targets account for increases in the drugs' prices,
the targets do not explicitly account for increases in their utilization
or the substitution of more expensive drugs for less expensive ones.
Physician groups note that the use of the outpatient drugs currently
covered by Medicare is largely nondiscretionary and that physicians should
not be penalized for prescribing these drugs. To the extent that
expenditures for these Medicare-covered outpatient drugs grow faster than
real GDP per capita-which is the SGR system's allowance for volume and
intensity increases-other physician spending must grow more slowly or
aggregate

41Most of the Part B drugs that Medicare covers fall into three
categories: those typically provided in a physician office setting (such
as chemotherapy drugs), those administered through a durable medical
equipment item (such as a respiratory drug given in conjunction with a
nebulizer), and those that are patient-administered and covered explicitly
by statute (such as certain immunosuppressives).

spending will exceed the targets and fee updates for physician services
will be reduced.

In 2002, Medicare covered approximately 450 outpatient prescription drugs.
The drugs that account for most of Medicare's Part B drug expenditures are
physician administered, such as those for cancer chemotherapy, accounting
for 80 percent of total Medicare spending for Part B drugs in 2001. In
2001, oncologists submitted about 42 percent of prescription drug claims,
while urologists accounted for 17 percent.

Part B prescription drugs are not covered by the physician fee schedule,42
but the expenditures for most Part B drugs are included in the SGR system
expenditures because, at the time spending targets were first introduced,
the Secretary of Health and Human Services (HHS) included these drugs as
services and supplies "incident to" physicians' services. Since that time,
Medicare spending for all Part B drugs has grown substantially, from about
$700 million in 1992 to an estimated $8.5 billion in 2002. Much of the
spending growth has resulted from increases in utilization and the
substitution of newer, more expensive medications. Because SGR-covered
Part B drug expenditures have grown more rapidly than other physician
service expenditures, drug expenditures as a proportion of allowable
spending under the targets have grown from 8.7 percent in 2002 to an
estimated 12.3 percent in 2004. Such rapid growth in drug expenditures
increases the likelihood that actual spending will exceed SGR system
targets. Moreover, because only payments for services included in the
physician fee schedule are offset when physician service spending deviates
from the spending targets, the increase in the share of total expenditures
attributed to prescription drugs magnifies the adjustment that must be
made to the update to bring spending in line with the targets.

42In general, payment for covered outpatient prescription drugs is made
under Medicare Part B and is equal to either 85 percent or 95 percent of
the average wholesale price, depending on the drug. MMA provided for the
implementation of a new payment methodology beginning in 2005. See MMA,
S:303, 117 Stat. 2233-2255. The legislation also establishes a new
voluntary prescription drug benefit program under a new Part D of Title
XVIII of the Social Security Act that will be effective January 1, 2006.

Physicians Concerned That Key Spending Drivers Are Not Included in SGR
System's Allowance for Volume and Intensity Growth

Physician groups have expressed concern that the SGR system's allowance
for volume and intensity growth-the 10-year moving average growth in real
GDP per capita-is both too low and inflexible. They contend that tying the
allowance to GDP results in targets that do not adequately account for
appropriate increases in the demand for physician services and changes in
medical practice, such as the following:

o  	A sicker beneficiary population. Physician groups reason that although
health status drives demand for services, the GDP growth allowance would
not account for any increases in physician spending that could be due to
greater care demands per beneficiary.

o  	Technological advances. The groups note that new, expensive medical
technologies can provide meaningful health gains for Medicare
beneficiaries but that these technology costs are likely to grow faster
than GDP.

o  	Site-of-service shifts. The groups note that patients with complex
conditions formerly treated in hospitals are increasingly treated in
physician offices and that treating such patients, who may require
frequent office visits and costly procedures, is likely to contribute to
volume and intensity growth.

The MVPS system provided an explicit opportunity to address some of these
concerns procedurally. In addition to the allowance for volume and
intensity growth specified in statute, the MVPS system also provided
specific authority for the HHS Secretary to recommend revising the allowed
increase based on factors such as changes in technology and concerns about
access to physician services. Under the MVPS system, the Secretary never
exercised the authority to make recommendations other than implementing
the MVPS default formula, but it still remained an option.

Transparency Lacking in Process for Estimating Changes in Medicare
Spending for Physician Services due to Laws and Regulations

The SGR system is designed to account for changes in law and regulation
that could affect aggregate spending for physician services. For example,
for 2005, CMS estimates that increased spending resulting from MMA's
coverage of a preventive physical examination for new beneficiaries,
cardiovascular screening blood tests, and diabetes screening tests, among
other new increases, will be almost fully offset by new MMA-required
payment adjustments for Part B drugs, which will lower physician service
spending. Physician groups we spoke with contend that the process for
developing such estimates may not be accurate or complete.

Assessing the accuracy and completeness of these estimates is difficult,
as CMS's process for identifying the applicable statutory and regulatory
changes and the methods used to arrive at dollar estimates are not fully
transparent. Either data are lacking to quantify the effects of changes or
consensus is lacking on the assumptions and interpretations made about the
changes and their effects. Currently, CMS does not use a formal mechanism
for soliciting input from physician groups or other experts before
obtaining public comment when future fees are announced in the Federal
Register.43 Physician groups contend that at least including physician
representatives in the process of assessing changes in laws and
regulations would improve CMS's analysis of effects and would be more
efficient than waiting for the public comment period.

Variable Growth in Provision of Physician Services and Certain SGR System
Design Elements Reduce Stability and Predictability of Physician Fee
Updates

Fee updates under the SGR system have varied widely within an allowed
range, principally because of annual fluctuations in the growth of the
volume and intensity of services that physicians provide to beneficiaries.
Two of the SGR system's design characteristics-the cumulative nature of
spending targets and the use of estimated data elements in the spending
target-also serve to reduce the stability and predictability of updates.
The MMA provision that revised the allowance for growth in service volume
and intensity from real GDP per capita growth rates each year to a 10-year
moving average will reduce some of the swings in future SGR system
updates.

Fluctuating Volume and Annual fluctuations in the growth of the volume and
intensity of services Intensity Growth Is a that physicians provide to
beneficiaries have been a principal cause of the Principal Cause of
instability of physician fee updates. Since the SGR system was

implemented in 1998, volume and intensity growth has ranged
fromInstability of Fee Updates

1.2 percent in 1999 to 6.1 percent in 2002. (See fig. 2.) It is uncertain
how

much physicians' discretion in the provision of their services contributes

to the fluctuation in volume and intensity growth.

43CMS is required to publish the final conversion factor update for the
upcoming calendar year by November 1. In the period prior to publishing
the final update-a period that usually runs from August to October-CMS
collects public comments in response to its proposed rule. It is at this
time that physician groups are able to submit formal comments on CMS's
estimate of this factor.

Several studies have found that physicians respond to reduced fee updates
by increasing the volume and intensity of services they provide to help
maintain their total Medicare income.44 In estimating future spending and
fee updates, both CMS and CBO assume that physicians will compensate,
through volume and intensity increases, for a portion of any fee
reductions. Consequently, both CMS and CBO project that for example, a 1
percent fee reduction would cause aggregate spending to fall by less than
1 percent. In addition, CBO assumes that physicians will respond to fee
increases by reducing volume and intensity.

Physician groups contend that volume and intensity growth is a necessary
response to increased demand caused by factors outside of physicians'
control as noted earlier, such as the declining health status of Medicare
beneficiaries, Medicare coverage of new benefits, and changing medical
technology and practices that encourage beneficiaries to schedule more
appointments with physicians. As long as the contributing factors are not
fully understood and predictable, unexpected volume and intensity
fluctuations will result in uncertain fee updates year to year.

SGR System's Cumulative Targets Increase Potential Fluctuation of
Physician Fee Updates

The cumulative nature of the SGR system's spending targets increases the
potential fluctuation of physician fee updates, as the system requires
that excess spending in any year be recouped in future years.
Conceptually, this means that if actual spending has exceeded the SGR
system targets, fee updates in future years must be lowered sufficiently
to both offset the accumulated excess spending and slow expected spending
for the coming year. Conversely, the system also requires that if spending
were to fall short of the targets, fees would need to be increased so that
future spending would be raised to align with target spending.

Estimation of the 2005 fee update illustrates how excess spending that is
not addressed affects future fee updates. In 2004 actual expenditures
under the SGR system are estimated to be $83.4 billion, whereas target
expenditures for the same year will be $77.3 billion. As a result, 2005
fee updates need to offset a $6.1 billion deficit from excess spending in
2004 (plus accumulated excess spending of $5.8 billion in past years) and
to

44CMS OACT has analyzed the results of these studies. See Office of the
Actuary, Centers for Medicare & Medicaid Services, "Physician Volume and
Intensity Response Memorandum," August 13, 1998.

realign the year's expected spending with target spending.45 Because the
SGR system is designed to offset accumulated excess spending over a period
of years, the deficit for 2004 and preceding years will reduce fee updates
for multiple years.

According to projections made by CMS OACT, maximum fee reductions will be
in effect from 2006 through 2012. Fee updates will be positive in 2014.
(See fig. 3.)

45The 2005 fee update will be higher than allowed by the SGR system owing
to an MMA minimum update of 1.5 percent.

Figure 3: Projected MEI and Fee Update under Current Law

Percentage 6

4

2

0

-2

-4

-6 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected MEI Projected fee update under current law Source: CMS OACT.

                     Note: Projections are as of July 2004.

Uncertainty in Estimates of Underlying SGR System Data Elements Decreases
Stability and Predictability of Physician Fee Updates

The stability of fee updates under the SGR system depends, in part, on
CMS's ability to accurately estimate current spending and annual changes
in the four factors that determine the sustainable growth rate: input
prices, FFS enrollment, the 10-year moving average of real GDP per capita,
and expenditures due to changes in laws and regulations. If reality proves
different from these estimates, then the estimates are revised to
incorporate more complete data, thereby contributing to the year to year
fluctuation in fee updates. For example, in the fall of 2004, when CMS
determines the update for 2005, the agency must estimate cumulative
expenditures through the end of 2004 based on incomplete data. If actual
spending is underestimated, the 2005 update will be set higher than it
would have been set without the estimation error. This underestimate will
be corrected, because in setting a fee update, the SGR system requires CMS
to revise the spending estimates and the sustainable growth rates for the
2 preceding years. Therefore, when more complete spending data become
available, the agency will revise its previous cumulative spending
estimates through 2004 and reduce future fee updates relative to what they
would have been if spending had not been underestimated.

Uncertainty in long-term projections of FFS enrollment, in conjunction
with the cumulative nature of the SGR system's targets, makes long-term
estimates of fee updates less predictable. Because the SGR system offsets
accumulated excess spending by reducing the update for the fee paid for
each service, a decline in the number of services results in less spending
being offset. For example, currently, CMS estimates that over the next 10
years, enrollment in FFS will decline as more beneficiaries join private
plans. CMS projects that the percentage of Medicare beneficiaries in the
FFS program will decline from about 85 percent in 2005 to 67 percent in
2014. With fewer beneficiaries in FFS, fewer services would be provided.
Therefore, the SGR system would call for more severe update reductions to
offset accumulated excess spending relative to what would have occurred if
FFS enrollment had remained stable. In contrast, CBO projected that FFS
enrollment will increase over the 10-year period at about the same rate as
the increase in overall Medicare enrollment.46 With more beneficiaries in
FFS, and thus more services provided, update reductions would not need to
be as severe to offset accumulated excess spending. Therefore, under CBO's
FFS projection, positive fee updates would be expected to return sooner
than under CMS's FFS projection.

46In its March 2004 baseline CBO projected the percentage of beneficiaries
in FFS would remain relatively flat at about 86 percent to 87 percent over
the 2005-2014 period.

Switching to the 10-Year  MMA changed the SGR system formula to use a
10-year moving average of Moving Average of Real  real GDP per capita,
which is currently 2.3 percent. As noted in our 2002 GDP Per Capita Will 
testimony, this change will eliminate much of the cyclical variation in
this

factor that occurred in previous years under the formula when the
SGRIncrease Stability and system target was tied to the yearly change in
real GDP per capita.47 (See Predictability fig. 4.) Including a more
stable measure of economic growth in the SGR system formula will help
increase the stability of fee updates.

Figure 4: Annual Percentage Change in Real GDP Per Capita and 10-Year
Moving Average Change in Real GDP Per Capita, 1992-2003

Percentage

          1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Annual change in the 10-year moving average of real GDP per capita

                      Annual change in real GDP per capita

Source: GAO's analysis of data from the Bureau of Economic Analysis and
the U.S. Census Bureau.

47GAO, Medicare Physician Payments: Spending Targets Encourage Fiscal
Discipline, Modifications Could Stabilize Fees, GAO-02-441T (Washington,
D.C.: Feb. 14, 2002).

Alternatives for Updating Physician Fees Would Eliminate Spending Targets
or Revise Current SGR System

The projected sustained period of declining physician fees and the
potential for beneficiaries' access to physician services to be disrupted
have heightened interest in alternatives for the current SGR system. In
general, potential alternatives we identified cluster around two
approaches. One approach would end the use of spending targets as a method
for updating physician fees and encouraging fiscal discipline. The other
approach would retain spending targets but modify the current SGR system
to address perceived shortcomings. These modifications could include one
or more of the following options: removing the Part B prescription drug
expenditures that are currently counted in the SGR system; resetting the
targets and not requiring the system to recoup previous excess spending;
using annual, rather than cumulative, targets; raising the allowance for
increased spending due to volume and intensity growth; and permitting some
flexibility in setting the volume and intensity allowance.

The alternatives discussed in this section-intended to be illustrative-
would all increase fees and thus aggregate spending-both government
outlays and beneficiary cost sharing-for physician services relative to
projected spending under current law.48 Most changes would require new
legislation; one exception is the removal of Part B prescription drugs
from the spending targets, which could be done administratively. We used
CMS OACT's projections to provide a sense of the magnitude of the effect
that potential alternatives might have on physician fee updates, aggregate
spending for physician services, and real spending for physician fee
schedule services per beneficiary (indicating a level of services
beneficiaries receive excluding prescription drugs and other
non-fee-schedule services, such as laboratory tests).49, 50 To simplify
comparisons among the discussed alternatives and current law, all of the
projections

48The projection under current law, which is used as a comparison to
projections under various options, assumes that the fee updates determined
by the SGR system will not be altered by any legislative action. However,
many parties, such as the Medicare Trustees, believe it is unlikely that
the projected negative fee updates will be allowed to take effect.

49The projections are included to aid comparisons among the various
options and are not intended to serve as predictions for what would occur
if the SGR system was replaced or modified. In addition, there is a degree
of uncertainty surrounding any projection and that uncertainty tends to
increase with the number of years for which the projection is made.

50For some of these options we present, CBO has developed budget scores,
which are specific cost estimates that include only federal expenditures
and exclude spending from other sources, such as beneficiary cost sharing.
When available, we present CBO's cost estimates for the options.

use the same assumptions regarding volume and intensity growth for
physician services and future FFS enrollment.51

Eliminate Spending Targets, Base Fee Updates on Physician Cost Increases

In its March 2001 report to the Congress, MedPAC recommended eliminating
the SGR system of spending targets and replacing it with an approach that
would base annual fee updates on changes in the cost of efficiently
providing care.52 Under this approach, efforts to control aggregate
spending would be separate from the mechanism used to update fees. The
advantage of eliminating spending targets would be greater fee update
stability. However, CMS OACT estimates that this approach, compared with
the current law projection, would result in cumulative expenditures that
are 22 percent greater over a 10-year period.

MedPAC reported that its recommendation could be implemented, in part, by
basing the update on forecast changes in MEI. It suggested that other
adjustments to the update might be necessary, for example, to ensure
overall payment adequacy or correct for previous MEI forecast errors. In
subsequent annual reports to the Congress, MedPAC has continued to
recommend a physician fee update based on MEI.53 In its March 2004 report,
for example, MedPAC stated that current Medicare payments for physician
services were adequate and recommended an update of approximately 2.6
percent for 2005 to "help maintain physician willingness and ability to
furnish services to Medicare beneficiaries."54 MedPAC's recommendation
contrasts with the 1.5 percent minimum update provided for by MMA and the
negative 3.6 percent update specified by the SGR system. In 2004
testimony, MedPAC stated that fee updates for physician services should
not be automatic, but should be informed by changes in beneficiaries'
access to services, the quality of services provided, the appropriateness
of cost increases, and other factors.

51Volume and intensity growth for physician services alone is projected to
be 3 percent per year. Overall volume and intensity growth-that is,
including outpatient prescription drugs and other services included under
the SGR system-is projected at about 4 percent per year.

52Medicare Payment Advisory Commission, Report to the Congress: Medicare
Payment Policy (Washington, D.C.: March 2001).

53Medicare Payment Advisory Commission, Report to the Congress: Medicare
Payment Policy (Washington, D.C.: March 2002, 2003, and 2004).

54Medicare Payment Advisory Commission, Report to the Congress: Medicare
Payment Policy (Washington, D.C.: March 2004).

Basing the update on MEI would result in positive and relatively stable
fee updates. (See fig. 5.) According to CMS OACT simulations, such an
approach would likely produce fee updates that ranged from 2.1 percent to
2.4 percent over the period from 2006 through 2014. Because physician fees
would increase each year during the entire period, rather than decreasing
each year until positive updates returned in 2014 as they would under the
current SGR system, Medicare spending for physician services would rise.
For the 10-year period from 2005 through 2014, CMS OACT estimates that
this approach would result in cumulative expenditures that are 22 percent
greater than projected under current law.55 (See fig. 6.) CMS OACT
projects that under current law the net present value of total Medicare
spending (both federal and beneficiary) over the next 75 years on all Part
B services will be $16.9 trillion. If physician fee updates are based on
the change in MEI, CMS OACT estimates that the net present value of total
Medicare spending (both federal and beneficiary) over the next 75 years on
Part B services would equal $19.1 trillion. Real spending per beneficiary
would increase from $2,157 in 2005 to $2,802 in 2014, compared with real
spending per beneficiary under current law, which would decrease to $1,774
in 2014. (See fig. 7.)

55In May 2004 testimony before the Subcommittee on Health, House Committee
on Energy and Commerce, CBO estimated that if this option went into effect
in 2005, it would raise net federal mandatory outlays by about $95 billion
over the 2005-2014 period. CBO's estimates differ from those of CMS OACT
in that CBO's estimates exclude beneficiary cost sharing and are based on
different underlying assumptions about the various factors that affect the
SGR system.

Figure 5: Projected MEI and Fee Updates under Current Law and under Option
of Eliminating Spending Targets and Tying Fee Updates to MEI

Percentage 6

4

2

0

-2

-4

-6 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected MEI Projected fee update under option Projected fee update under
current law Source: CMS OACT.

                     Note: Projections are as of July 2004.

Figure 6: Projected Aggregate Spending under Current Law and under Option
of Eliminating Spending Targets and Tying Fee Updates to MEI

Dollars in billions 250

200

150

22% greater cumulative spending

100

50

0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected spending under option Projected spending under current law
Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Aggregate spending includes all
expenditures for physician services-both government outlays and
beneficiary cost sharing.

Figure 7: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Option of Eliminating Spending
Targets and Tying Fee Updates to MEI

Dollars

                                     3,000

                                     2,500

                                     2,000

                                     1,500

                                     1,000

                                      500

0

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected real spending for physician services per FFS beneficiary under
current law

Projected real spending for physician services per FFS beneficiary under
option

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Real spending per beneficiary for
physician fee schedule services includes both government outlays and
beneficiary cost sharing and is adjusted by MEI. Spending for
non-fee-schedule services-laboratory services and certain Medicare-covered
Part B outpatient drugs-is excluded.

Although MedPAC's recommended update approach would limit annual increases
in the price Medicare pays for each service, the approach does not contain
an explicit mechanism for constraining aggregate spending resulting from
increases in the volume and intensity of services physician provide. In
2001, when MedPAC first recommended eliminating the SGR system, it stated
that volume and intensity increases had not been a major concern since
1992. It added, however, that if volume and intensity growth reemerged as
a concern, Medicare might address the problem by trying to achieve
appropriate use of services through outcomes and effectiveness research,
disseminating practice guidelines and other tools for applying this
research, and developing evidence-based measures to assess the application
of the research findings.

Since MedPAC's 2001 report, volume and intensity growth has increased
considerably. (See fig. 2.) Subsequent MedPAC reports and testimony have
discussed trends in the use of physician services and have identified
particular services-such as diagnostic imaging-that are growing rapidly,

but the reports have not made recommendations for addressing volume and
intensity growth. However, in 2004 testimony, MedPAC stated that it
planned to study the efficacy of private insurers' strategies for
controlling spending for high-growth services and whether Medicare might
be able to

56

emulate them.

Retain Spending Targets, Modify Current SGR System

Another approach for addressing the perceived shortcoming of the current
SGR system would retain spending targets but modify one or more elements
of the system. The key distinction of this approach, in contrast to basing
updates on MEI, is that fiscal controls designed to moderate spending
would continue to be integral to the system used to update fees. The
advantage of retaining spending targets as part of the system for updating
fees is that the system would automatically work to moderate spending if
volume and intensity growth began to increase above allowable rates.
Although many options are possible under this approach, six are discussed
below. All six would produce fee updates that are higher during the
10-year period from 2005 through 2014 than those projected under current
law but would also result in higher aggregate spending ranging from 4
percent to 23 percent more, depending on the modification.

56Medicare Payment Advisory Commission, Payment for Physician Services in
the Medicare Program, testimony before the Subcommittee on Health, House
Committee on Energy and Commerce (May 5, 2004).

Remove Part B Drugs from the SGR System

The Secretary of HHS could, under current authority, consider excluding
Part B drugs from the definition of services furnished incident to
physician services for purposes of the SGR system. As discussed earlier,
expenditures for these drugs have been growing rapidly, which, in turn,
has put downward pressure on the fees paid to Medicare physicians.
However, according to CMS OACT simulations, removing Part B drugs from the
SGR system beginning in 2005 would not prevent several years of fee
declines and would not decrease the volatility in the updates. Fees would
decline by about 5 percent per year from 2006 through 2010. (See fig. 8.)
There would be a positive update in 2011-3 years earlier than is projected
under current law. From 2012 through 2014, fees would increase by
approximately 5 percent per year. CMS OACT estimates that removing Part B
drugs from the SGR system would result in cumulative spending over the
10-year period from 2005 through 2014 that is 5 percent higher than is
projected under current law.57 (See fig. 9.) Real spending per beneficiary
would increase from $2,157 in 2005 to $2,240 in 2014, compared with real
spending per beneficiary under current law, which would decrease to $1,774
in 2014. (See fig. 10.)

57In May 2004 testimony, CBO estimated that this option would raise net
federal mandatory outlays by about $15 billion through 2014. CBO's
estimates differ from those of CMS OACT in that CBO's estimates exclude
beneficiary cost sharing and are based on different underlying assumptions
about the various factors that affect the SGR system.

Figure 8: Projected MEI and Fee Updates under Current Law and under Option
of Not Including Any Part B Drugs in the SGR System Beginning in 2005

Percentage

6

4

2

0

-2

-4

-6 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected MEI

Projected fee update under option

Projected fee update under current law Source: CMS OACT.

Notes: Projections are as of July 2004. Projected updates under the option
will be equal to projected updates under current law through 2010.

Figure 9: Projected Aggregate Spending under Current Law and under Option
of Not Including Any Part B Drugs in the SGR System Beginning in 2005

Dollars in billions 250

200

150

5% higher cumulative spending

100

50

0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

                        Projected spending under option

                      Projected spending under current law

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Aggregate spending includes all
expenditures for physician services-both government outlays and
beneficiary cost sharing. The line depicting the projection under current
law includes all spending included in the SGR system. Under the option of
not including any Part B drugs in the SGR system, spending for Part B
drugs will occur even though it is not included in the SGR formula. To
ensure comparability between the two projections, we included aggregate
spending for both remaining SGR-covered services and Part B drugs in the
line depicting the projection under the option.

Figure 10: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Option of Not Including Any
Part B Drugs in the SGR System Beginning in 2005

Dollars

3,000

2,500

2,000

1,500

1,000

500

0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Base Future SGR System Targets on Actual Spending from a Recent Year

Projected real spending for physician services per FFS beneficiary under
current law

Projected real spending for physician services per FFS beneficiary under
option

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Real spending per beneficiary for
physician fee schedule services includes both government outlays and
beneficiary cost sharing and is adjusted by MEI. Spending for
non-fee-schedule services-laboratory services and certain Medicare-covered
Part B outpatient drugs-is excluded.

In 2002, we testified that physician spending targets and fees may need to
be adjusted periodically as health needs change, technology improves, or
health care markets evolve.58 Such adjustments could involve specifying a
new base year from which to set future targets. Currently, the SGR system
uses spending from 1996, trended forward by the sustainable growth rate
computed for each year, to determine allowable spending.

MMA avoided a fee decline in 2004, and a projected fee decline for 2005,
by stipulating a minimum update of 1.5 percent in each of those 2 years,
but the law did not similarly adjust the spending targets to account for
the additional spending that would result from the minimum update.
Consequently, under the SGR system the additional MMA spending and

58GAO-02-441T.

other accumulated excess spending will have to be recouped through fee
reductions beginning in 2006. If policymakers believe that the resulting
negative fee updates are inappropriately low, one solution is to use
actual spending from a recent year as a basis for setting future SGR
system targets. Using such an approach, policymakers could essentially
forgive the accumulated excess spending attributable to MMA and other
factors. The effect would be to increase future updates and, as with other
alternatives presented here, overall spending.

According to CMS OACT simulations, forgiving the accumulated excess
spending as of 2005-that is, resetting the cumulative spending target so
that it equals cumulative actual spending-would raise fees in 2006. (See
fig. 11.) However, because volume and intensity growth is projected to
exceed the SGR system's allowance for such growth, negative updates would
return beginning in 2008 and continue through 2013. Resulting cumulative
spending over the 10-year period from 2005 through 2014 would be 13
percent higher than is projected under current law. (See fig. 12.) Real
spending per beneficiary for physician services would grow from $2,157 in
2005 to $2,334 in 2014, compared with real spending per beneficiary under
current law, which would decrease to $1,774 in 2014. (See fig. 13.)

Figure 11: Projected MEI and Fee Updates under Current Law and under
Option of Resetting the Cumulative Spending Target Equal to Cumulative
Actual Spending as of 2006

Percentage 6

4

2

0

-2

-4

-6 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected MEI

Projected fee update under option

Projected fee update under current law Source: CMS OACT.

Notes: Projections are as of July 2004. Projection under option of
eliminating accumulated excess spending assumes that the physician fee
update would equal MEI in 2006.

Figure 12: Projected Aggregate Spending under Current Law and under Option
of Resetting the Cumulative Spending Target Equal to Cumulative Actual
Spending as of 2006

Dollars in billions 250

200

150

cumulative spending

100

50

0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected spending under option Projected spending under current law
Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Aggregate spending includes all
expenditures for physician services-both government outlays and
beneficiary cost sharing. Projection under option of eliminating
accumulated excess spending assumes that the physician fee update would
equal MEI in 2006.

Figure 13: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Option of Resetting the
Cumulative Spending Target Equal to Cumulative Actual Spending as of 2006

Dollars

                                     3,000

                                     2,500

                                     2,000

                                     1,500

                                     1,000

                                      500

0

               2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Eliminate the Cumulative Aspect of Spending Targets

Projected real spending for physician services per FFS beneficiary under
current law

Projected real spending for physician services per FFS beneficiary under
option

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Projection under option of
eliminating accumulated excess spending assumes that the physician fee
update would equal MEI in 2006. Real spending per beneficiary for
physician fee schedule services includes both government outlays and
beneficiary cost sharing and is adjusted by MEI. Spending for
non-fee-schedule services-laboratory services and certain Medicare-covered
Part B outpatient drugs-is excluded.

One option for reducing the fluctuation in fee updates would be to
eliminate the cumulative aspect of the SGR system's spending targets and
return to a system of annual targets, as was used under MVPS. As
previously discussed, the cumulative aspect of the SGR system's spending
targets-although rigorous as a budgetary tool-can produce updates that
swing from the maximum fee reduction to the maximum fee increase. In
contrast, MVPS's annual spending target approach traded off some fiscal
control for increased fee stability. The MVPS update for a year depended,
in part, on whether actual spending 2 years earlier had exceeded or fallen
short of the annual spending target for that year. For example, the MVPS

update for 1996, which was determined in 1995, was affected by the
relationship between actual and target spending in 1994. In principle,
under MVPS excess spending from a single year, up to a limit specified by
its update formula, was required to be recouped. Excess spending that
could not be made up within those limits would, in essence, be forgiven.59

According to CMS OACT simulations, eliminating the cumulative aspect of
the SGR system would result in fee updates that vary less than projected
updates under current law. For example, under an MVPS-like system of
annual targets, from 2006 through 2014, the largest negative update would
be negative 0.6 percent instead of negative 5.0 percent under current law,
and the largest positive update would be 0.9 percent instead of 3.9
percent. (See fig. 14.) Fees would be essentially flat over the period,
instead of swinging from large fee declines to fee increases as they are
expected to do under the SGR system. Relative to spending projected under
current law, under an MVPS-like system total spending would be greater
each year from 2006 through 2014. CMS OACT estimates that cumulative
expenditures over the 10-year period from 2005 through 2014 would be 15
percent higher than under current law. (See fig. 15.) Real spending per
beneficiary would increase from $2,157 in 2005 to $2,442 in 2014, compared
with real spending per beneficiary under current law, which would decrease
to $1,774 in 2014. (See fig. 16.)

59Both the SGR and MVPS systems provided for updates that could exceed MEI
if spending fell below their respective targets.

Figure 14: Projected MEI and Fee Updates under Current Law and under
Option of Eliminating the Cumulative Aspect of Spending Targets

Percentage
6

4

2

0

-2

-4

-6
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected MEI
Projected fee update under option

Projected fee update under current law

Source: CMS OACT. Note: Projections are as of July 2004.

Figure 15: Projected Aggregate Spending under Current Law and under Option
of Eliminating the Cumulative Aspect of Spending Targets

Dollars in billions 250

200

150

15% higher cumulative spending

100

50

0
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected spending under option Projected spending under current law

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Aggregate spending includes all
expenditures for physician services-both government outlays and
beneficiary cost sharing.

Figure 16: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Option of Eliminating the
Cumulative Aspect of Spending Targets

Dollars

                                     3,000

                                     2,500

                                     2,000

                                     1,500

                                     1,000

                                      500

0

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected real spending for physician services per FFS beneficiary under
current law

Projected real spending for physician services per FFS beneficiary under
option

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Real spending per beneficiary for
physician fee schedule services includes both government outlays and
beneficiary cost sharing and is adjusted by MEI. Spending for
non-fee-schedule services-laboratory services and certain Medicare-covered
Part B outpatient drugs-is excluded.

Modify Allowance for Volume If policymakers agree with physician groups
that the current SGR system's

and Intensity Growth	allowance for volume and intensity growth does not
adequately account for appropriate spending increases that result from
technological innovation or changes in medical practice, the allowance
could be increased by some factor above the percentage change in real GDP
per capita. As stated earlier, the current SGR system's allowance for
volume and intensity growth is approximately 2.3 percent per year-the
10-year moving average in real GDP per capita-while projected volume and
intensity growth is higher-about 3 percent per year for physician services
alone, and about 4 percent per year including Part B drugs. To offset the
increased spending associated with the higher volume and intensity growth,
the SGR system will reduce updates below the increase in MEI. In its 1997
report to the Congress, PPRC recommended adopting an

allowance equal to real GDP per capita plus 1 or 2 percentage points "to
allow for advancements in medical capabilities."60

According to CMS OACT simulations, increasing the allowance for volume and
intensity growth to GDP plus 1 percentage point would likely produce
positive fee updates beginning in 2012-2 years earlier than is projected
under current law. (See fig. 17.) Because fee updates would be on average
greater than under current law during the 10-year period from 2005 through
2014, Medicare spending for physician services would rise. CMS OACT
estimates that cumulative expenditures over the 10-year period would
increase by 4 percent more than under current law.61  (See fig. 18.) Real
spending per beneficiary would change little from $2,157 in 2005 to $2,158
in 2014, compared with real spending per beneficiary under current law,
which would decrease to $1,774 in 2014. (See fig. 19.)

60Physician Payment Review Commission, 1997 Annual Report to Congress
(Washington, D.C.: 1997), 248.

61In May 2004 testimony, CBO estimated that this option would raise net
federal mandatory outlays by about $35 billion over the 2008-2014 period.
CBO's estimates differ from those of CMS OACT in that CBO's estimates
exclude beneficiary cost sharing and are based on different underlying
assumptions about the various factors that affect the SGR system.

Figure 17: Projected MEI and Fee Updates under Current Law and under
Option of Increasing Volume and Intensity Growth Allowance to GDP Plus 1
Percentage Point

Percentage 6

4

2

0

-2

-4

-6 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

                                 Projected MEI

                       Projected fee update under option

Projected fee update under current law Source: CMS OACT.

Notes: Projections are as of July 2004. Projected updates under the option
will be equal to projected updates under current law through 2010.

Figure 18: Projected Aggregate Spending under Current Law and under Option
of Increasing Volume and Intensity Growth Allowance to GDP Plus 1
Percentage Point

Dollars in billions 250

200

150

cumulative spending

100

50

0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

                        Projected spending under option

                      Projected spending under current law

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Aggregate spending includes all
expenditures for physician services-both government outlays and
beneficiary cost sharing.

Figure 19: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Option of Increasing Volume
and Intensity Growth Allowance to GDP Plus 1 Percentage Point

Dollars

                                     3,000

                                     2,500

                                     2,000

                                     1,500

                                     1,000

                                      500

0

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected real spending for physician services per FFS beneficiary under
current law

Projected real spending for physician services per FFS beneficiary under
option

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Real spending per beneficiary for
physician fee schedule services includes both government outlays and
beneficiary cost sharing and is adjusted by MEI. Spending for
non-fee-schedule services-laboratory services and certain Medicare-covered
Part B outpatient drugs-is excluded.

Congress could also modify the SGR system's allowance for volume and
intensity growth by providing flexibility similar to that afforded by the
MVPS system. Although that earlier system of spending targets specified a
default volume and intensity increase, it also allowed the HHS Secretary
to recommend a different increase if changes in medical technology,
beneficiary access to physician services, or other factors warranted an
allowance that was higher or lower than the default increase.

Combine Options	Two alternatives illustrate the effects of combining
individual options. For example, together the Congress and CMS could
modify the SGR system by removing Part B drugs, resetting the base, and
increasing allowed volume and intensity growth to GDP plus 1 percentage
point.62 According to CMS

62We use GDP plus 1 percentage point as the allowance for volume and
intensity growth for illustrative purposes only.

OACT simulations, this combination of options would result in positive
updates ranging from 2.2 percent to 2.8 percent for the 2006-2014 period.
(See fig. 20.) CMS OACT projects that the combined options would increase
aggregate spending by 23 percent over the 10-year period (see fig. 21.)
and that real spending per beneficiary for physician services would
increase from $2,157 to $2,866, compared with real spending per
beneficiary under current law, which would decrease to $1,774 in 2014.
(See fig. 22.)

Figure 20: Projected MEI and Fee Updates under Current Law and under
Combination of Options of Resetting the Cumulative Spending Target Equal
to Cumulative Actual Spending as of 2006, Not Including Any Part B Drugs
in the SGR System Beginning in 2005, and Using GDP Plus 1 Percentage Point

Percentage 6

4

2

0

-2

-4

-6 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected MEI

Projected fee update under option

Projected fee update under current law

Source: CMS OACT.

Notes: Projections are as of July 2004. Projection under combination of
options of resetting the cumulative spending target equal to cumulative
actual spending as of 2005, removing Part B drugs from the SGR system
beginning in 2005, and using GDP plus 1 percentage point assumes that the
physician fee update would equal MEI in 2006.

Figure 21: Projected Aggregate Spending under Current Law and under
Combination of Options of Resetting the Cumulative Spending Target Equal
to Cumulative Actual Spending as of 2006, Not Including Any Part B Drugs
in the SGR System Beginning in 2005, and Using GDP Plus 1 Percentage Point

Dollars in billions

250

200

150

23 percent

higher cumulative

spending

100

50

0

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

                        Projected spending under option

                      Projected spending under current law

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Aggregate spending includes all
expenditures for physician services-both government outlays and
beneficiary cost sharing. Projection under combination of options of
resetting the cumulative spending target equal to cumulative actual
spending as of 2005, not including any Part B drugs in the SGR system
beginning in 2005, and using GDP plus 1 percentage point assumes that the
physician fee update would equal MEI in 2006. The line depicting the
projection under current law includes all spending included in the SGR
system. Under the option of not including any Part B drugs in the SGR
system, spending for Part B drugs will occur even though it is not
included in the SGR formula. To ensure comparability between the two
projections, we included aggregate spending for both remaining SGR-covered
services and Part B drugs in the line depicting the projection under the
option.

Figure 22: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Combination of Options of
Resetting the Cumulative Spending Target Equal to Cumulative Actual
Spending as of 2006, Not Including Any Part B Drugs in the SGR System
Beginning in 2005, and Using GDP Plus 1 Percentage Point

Dollars

                                     3,000

                                     2,500

                                     2,000

                                     1,500

                                     1,000

                                      500

0

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Projected real spending for physician services per FFS beneficiary under
current law

Projected real spending for physician services per FFS beneficiary under
option

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Projection under combination of
options of resetting the cumulative spending target equal to cumulative
actual spending as of 2005, removing Part B drugs from the SGR system
beginning in 2005, and using GDP plus 1 percentage point assumes that the
physician fee update would equal MEI in 2006. Real spending per
beneficiary for physician fee schedule services includes both government
outlays and beneficiary cost sharing and is adjusted by MEI. Spending for
non-fee-schedule services-laboratory services and certain Medicare-covered
Part B outpatient drugs-is excluded.

Another example of combined options could involve removing Part B drugs
and implementing an MVPS-like system of annual targets, but not increasing
the volume and intensity allowance. CMS OACT simulations project that this
combination would result in fee updates that range from 0.8 percent to 1.3
percent over the period from 2006 through 2014. (See fig. 23.) Over the
10-year period from 2005 through 2014, cumulative spending for physician
services would exceed those projected under current law by 18 percent.
(See fig. 24.) Real spending per beneficiary for physician services would
increase from $2,157 to $2,615, compared with real spending per
beneficiary under current law, which would decrease to $1,774 in 2014.
(See fig. 25.)

Figure 23: Projected MEI and Fee Updates under Current Law and under
Combination of Options of Not Including Any Part B Drugs in the SGR System
Beginning in 2005 and Eliminating the Cumulative Aspect of Spending
Targets

Percentage 6

4

2

0

-2

-4

-6 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

                                 Projected MEI

                       Projected fee update under option

                     Projected fee update under current law

Source: CMS OACT.

Note: Projections are as of July 2004.

Figure 24: Projected Aggregate Spending under Current Law and under
Combination of Options of Not Including Any Part B Drugs in the SGR System
Beginning in 2005 and Eliminating the Cumulative Aspect of Spending
Targets

Dollars in billions

250

200

150 18% higher cumulative spending

100

50

0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

                        Projected spending under option

                      Projected spending under current law

Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Aggregate spending includes all
expenditures for physician services-both government outlays and
beneficiary cost sharing. The line depicting the projection under current
law includes all spending included in the SGR system. Under the option of
not including any Part B drugs in the SGR system, spending for Part B
drugs will occur even though it is not included in the SGR formula. To
ensure comparability between the two projections, we included aggregate
spending for both remaining SGR-covered services and Part B drugs in the
line depicting the projection under the option.

Figure 25: Projected Real Spending for Physician Fee Schedule Services per
FFS Beneficiary under Current Law and under Combination of Options of Not
Including Any Part B Drugs in the SGR System Beginning in 2005 and
Eliminating the Cumulative Aspect of Spending Targets

Dollars

3,000

2,500

2,000

1,500

1,000 500 0

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Projected real spending
for physician services per FFS beneficiary under current law Projected
real spending for physician services per FFS beneficiary under option

                  Source: GAO analysis of data from CMS OACT.

Notes: Projections are as of July 2004. Real spending per beneficiary for
physician fee schedule services includes both government outlays and
beneficiary cost sharing and is adjusted by MEI. Spending for
non-fee-schedule services-laboratory services and certain Medicare-covered
Part B outpatient drugs-is excluded.

Concluding Observations

Medicare faces the challenge of moderating the growth in spending for
physician services while ensuring that physicians are paid fairly so that
beneficiaries have appropriate access to their services. Under the current
SGR system, fees are projected to fall by about 5 percent per year for the
next several years. Total payments to physicians will continue to rise
because of expected increases in volume and intensity. However, on a per
capita basis, real spending per beneficiary will decline, raising concerns
that a sustained period of falling fees could discourage some physicians
from participating in the Medicare program and serving beneficiaries.
These concerns have prompted policymakers to consider alternative
approaches for updating physician fees.

One approach under consideration for solving the problem of declining fees
is for Medicare to abandon the use of spending targets and separate the
program's attempts to control spending from its method for adjusting
physician fees each year. This is the approach that has been recommended
by MedPAC. Although projected future fee increases would be positive and
relatively stable, eliminating spending targets would increase spending.
The extent to which spending growth would be moderated would depend upon
the efficacy of separate efforts to address growth in volume and
intensity.

Similarly, the other approach of retaining spending targets but modifying
the SGR system to overcome its current perceived shortcomings, would also
increase spending. These alternative approaches could also be augmented by
separate efforts to moderate spending. Alternatives under this approach
seek to preserve the fiscal discipline of spending targets while providing
for reasonable fee updates. These alternative approaches could also be
augmented by other efforts to moderate spending. To the extent that the
growth in spending is moderated, physicians would benefit from an increase
in fees that would be triggered under a spending target system.

Almost any change to the SGR system is likely to increase Medicare
spending above the amount that is currently projected. Either of the two
broad types of approaches discussed above-replacing the SGR system and
revising the SGR system-could be implemented in a way that would likely
generate positive fee updates. Therefore, the choice between the two
approaches under consideration may hinge on whether primary importance
should be given to stable fee increases or to the need for fiscal
discipline within the Medicare program.

Agency and Industry Comments and Our Evaluation

Agency Comments 	In written comments on a draft of this report, CMS agreed
with our concluding observations that appropriately updating the physician
payment rates requires a balance between adjusting physician fees in a
stable and predictable manner and encouraging fiscal discipline with
scarce Medicare resources. CMS expressed its commitment to ensuring that
Medicare beneficiaries have access to high-quality health care and noted
that achieving this goal requires paying physicians appropriately.

CMS mentioned several administrative actions it has taken to improve
Medicare's payments to physicians, including specific adjustments to MEI
that have both made the index a more accurate representation of inflation
in physician practice costs and resulted in higher payments to physicians.
In addition, the agency committed to considering further administrative
actions and discussed ongoing efforts to implement various provisions of
MMA that may reduce adverse incentives in the current payment system,
allow the program to pay for higher quality care, and uncover innovative
methods to control spending growth in the future. We have reprinted CMS's
letter in appendix III.

Industry Association Comments

We obtained oral comments from officials representing the American Medical
Association (AMA), the Medical Group Management Association (MGMA), the
American College of Physicians (ACP), and the Alliance for Specialty
Medicine (ASM). In discussing the draft report with these groups, their
overall reaction was that the report was a good analysis of the problems
with the SGR system; however, they raised a number of concerns about the
draft report. The bulk of their comments focused on OACT's estimates of
aggregate spending on physician services, the SGR system's use of MEI as a
measure of input price inflation for physician services, and the draft's
discussion of physicians' concerns about the SGR system. The rest of their
comments pertained to either issues related to physician behavior or to
topics outside the scope of our review. A summary of the physician groups'
comments and our evaluation is provided below.

Representatives from all four groups commented on CMS OACT's estimates
illustrating each option's additional aggregate spending over a 10-year
period relative to current law spending over the same period. The groups
were confused by the difference between CMS OACT's estimates and CBO's
budget impact estimates, which were available for some of the options.
CBO's budget scores-that is, cost estimates that show the impact of
legislative changes on the federal budget-include only federal
expenditures and exclude spending from other sources, such as beneficiary
cost sharing. In contrast, CMS OACT's aggregate spending estimates include
both federal outlays and beneficiary cost sharing. Because any changes to
the SGR system that result in increased spending would not only affect
taxpayers but also Medicare beneficiaries (through increased cost sharing
and part B premiums), we believe it is appropriate to include the
estimated increase in aggregate spending. Nevertheless, because our focus
is on the relative costliness of each option, we revised the draft to
highlight the proportional difference between current law

spending and the spending estimated for each option. In addition, we now
include CBO's budget scores for each option, where available.

All four physician groups also expressed concern that the draft report did
not discuss the use of MEI as a measure of input price inflation for
physician services. The groups contended that MEI does not contain
sufficiently current data on physician practice costs, stating that it
does not account for or keep pace with the cost of items such as
information technology. Examining MEI and other indices included in the
SGR system was outside the scope of our report. Moreover, in responding to
public comments on a federal regulation, CMS stated that the various
expense categories constituting MEI capture all practice expenses and are
based on the most recent available data. 63

The physician groups commented that the projected payment reductions of 5
percent a year from 2006 through 2014 are unrealistically severe and that
the draft report did not sufficiently emphasize the access problems that
beneficiaries would experience in the event of these cuts. They further
noted that the Congress has regularly made adjustments to the SGR system
and would probably act again. We noted in the draft report that
policymakers, physicians, and others are concerned about the impact that
the projected fee reductions would have on beneficiary access to physician
services, noting that the Medicare Trustees and other parties believe it
is unlikely that the projected fee reductions will take place.

Representatives from both ACP and ASM asserted that we should include a
discussion about the effect of the spending targets on physician behavior
and volume and intensity. They noted that evidence is lacking that
directly correlates the introduction of both spending targets and the
physician fee schedule in 1992 with the corresponding drop in volume and
intensity in that year. They believe this reduction was likely caused by
something other than the spending target, such as initiatives aimed at
correctly coding claims for physician services. ACP stated that for the
Congress to evaluate any alternatives, there must be a discussion of how
the SGR system affects the volume and intensity of physician services. As
noted in the draft report, we do not claim that spending targets and the
fee schedule influenced individual behavior and reduced the volume and
intensity of physician services in the early 1990s, we noted that PPRC
claimed that a spending target system would provide a collective incentive

63See 68 Fed. Reg. 63196, 63239-45.

for physicians to develop practice guidelines and control unnecessary
utilization. Further, in the draft report we described spending targets as
a method for automatically imposing fiscal discipline, not as a tool to
modify the behavior of individual physicians.

Representatives from ACP further noted that while the draft report
included a discussion of geographic variation in physician service use, it
did not mention that the SGR system is a blunt instrument in that it
applies nationally to all physicians. In a year in which fees are reduced,
physicians in regions that could be characterized as low spending would
receive the same fee reduction as physicians in higher-spending regions.
We agree that the SGR system does not distinguish between physicians whose
discretionary practice patterns result in higher Medicare spending and
those physicians whose practice patterns do not. As we stated in the draft
report, at the time PPRC recommended expenditure targets, it initially
envisioned a national target that would apply to all physician services
and later the evolution of separate targets that would apply to regions,
categories of physician services, or both.

The physician groups raised additional topics that were beyond the scope
of our study. For example, AMA contended that Medicare's new preventive
benefits and government-sponsored health campaigns create a
government-induced demand among beneficiaries for services that, in turn,
could increase volume and intensity of service use. To date, studies have
not been conducted on whether new benefits and federal health campaigns
have directly affected Medicare beneficiaries' use of physician services.
Our report notes, however, that the SGR system's allowance for volume and
intensity growth, unlike that of the MVPS system, is inflexible and would
not take such factors into account. ACP noted that increased spending on
physician services may be appropriate, as it may result in other program
savings, such as reduced spending for hospital care. Whether such savings
have been or can be achieved would require research outside this study's
scope.

We are sending copies of this report to the Secretary of Health and Human
Services and interested congressional committees. We will also provide
copies to others on request. In addition, this report is available at no
charge on the GAO Web site at http://www.gao.gov.

If you or your staff have questions about this report, please contact me
at (202) 512-7101 or James Cosgrove at (202) 512-7029. Other contributors
to this report include Jessica Farb, Hannah Fein, and Jennifer Podulka.

A. Bruce Steinwald Director, Health Care-Economic and 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 Joe L. Barton
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

The Honorable Michael Bilirakis
Chairman
The Honorable Sherrod Brown
Ranking Minority Member
Subcommittee on Health
Committee on Energy and Commerce
House of Representatives

The Honorable Nancy L. Johnson
Chairman
The Honorable Pete Stark
Ranking Minority Member
Subcommittee on Health
Committee on Ways and Means
House of Representatives

Appendix I: Calculation of the Performance Adjustment Factor

Each year, CMS follows a statutory formula to compute a performance
adjustment factor (PAF) and determines whether the physician fee update
should be adjusted relative to the percentage change in the Medicare
Economic Index (MEI) and, if so, by how much. (See fig. 26.) The PAF takes
into account the difference between actual and target expenditures. If
spending has equaled the targets, the PAF is equal to 1 and the update
will equal the percentage change in MEI. If spending has been below the
targets, the PAF is greater than 1, thus increasing the update. If
spending has been above the targets, the PAF is less than 1, thus reducing
the update. The PAF is a blend of the relative difference between target
and actual spending in the current year, accounting for 75 percent, and
the relative cumulative difference in expenditures from April 1996 through
the current year, accounting for 33 percent. The weights were developed by
the Centers for Medicare & Medicaid Services' (CMS) Office of the Actuary
(OACT) and included in statute to minimize the volatility of both fee
updates and the time required to align actual spending with the targets.
Applying these weights causes the difference between cumulative actual
expenditures and cumulative target expenditures to be adjusted over
several years rather than during a single year. As a result, the fee
update is less volatile than would be the case if the full adjustment were
made in 1 year. The PAF is subject to statutory limits and may not cause
the fee update to be set at more than 3 percent above MEI or 7 percent
below MEI. These limits may further increase the time necessary to align
spending with targets.

Figure 26: Formula Used to Determine the Performance Adjustment Factor in
2005

PAF2005 =

                               Source: CMS OACT.

Appendix II: Corrections to Prior Estimates Caused the SGR System's Cumulative
Targets to Produce Negative Updates

Since the introduction of the fee schedule in 1992 through 2001,
physicians generally experienced real increases in their fee updates-that
is, fee updates increased more than the increase in the cost of providing
physician services, as measured by MEI. Specifically, during that period,
fee updates increased by 39.7 percent, whereas MEI increased by 25.9
percent. In 2002, however, the sustainable growth rate (SGR) system
reduced fees by 4.8 percent,1 despite an estimated 2.6 percent increase in
the costs of providing physician services. (See fig. 27.)

1Some annual fee updates are adjusted for additional factors. For example,
a budget neutrality adjustment is used to account for changes in the
calculations used to determine the amount of resources associated with
physician services. In 2002, CMS reduced the update by an additional 0.64
percent resulting in a total fee decline of 5.4 percent.

 Appendix II: Corrections to Prior Estimates Caused the SGR System's Cumulative
                      Targets to Produce Negative Updates

Figure 27: Percentage Change in MEI, SGR Fee Schedule Update, and Medicare
Physician Services Spending per Beneficiary, 1998-2005

Percentage

12

                                      10.1

10

8

6

4

2

0

-2

-4

                                      -4.8

-6 1998 1999 2000 2001 2002 2003a 2004b 2005b

MEI

Physician fee update

Spending per beneficiary

Source: GAO analysis of data from the Boards of Trustees of the Federal HI
and SMI Trust Funds.

Notes: Spending per beneficiary represents Medicare spending for
beneficiaries in the traditional FFS program, net of beneficiary cost
sharing. Spending for end-stage renal disease patients is not included.
The physician fee schedule update figures shown do not reflect additional
required adjustments, such as those for legislated changes and for budget
neutrality.

aThe 1.7 percent fee update went into effect in March 2003.

bThe physician fee updates of 1.5 percent for 2004 and 2005 were specified
by the Medicare Prescription Drug, Improvement, and Modernization Act of
2003.

In 2002, corrections to prior estimation errors caused the SGR system's
cumulative targets to begin producing negative updates. The SGR system
reduced fees in 2002 because estimated spending for physician services-
cumulative since 1996-exceeded the target by about $8.9 billion, or 13
percent of projected 2002 spending. In part, the fee reduction occurred
because CMS revised upward its estimates of previous years' actual

Appendix II: Corrections to Prior Estimates Caused the SGR System's
Cumulative Targets to Produce Negative Updates

spending. Specifically, CMS found that its previous estimates had omitted
a portion of actual spending for 1998, 1999, and 2000. In addition, in
2002 CMS lowered the 2 previous years' spending targets based on revised
gross domestic product (GDP) data from the Department of Commerce. Based
on the new higher spending estimates and lower targets, CMS determined
that fees had been too high in 2000 and 2001. In setting the 2002
physician fees, the SGR system reduced fees to recoup previous excess
spending. The update would have been about negative 9 percent if the SGR
system had not limited its decrease to 7 percent below MEI. Because the
previous overpayments were not fully recouped in 2002, and because of
volume and intensity increases, by 2003, physicians were facing several
more years of fee reductions to bring cumulative Medicare spending on
physician services in line with cumulative targets.

Despite its recognition of errors, CMS had determined that its authority
to revise previous spending targets was limited. In 2002, CMS noted that
the 1998 and 1999 spending targets had been based on estimated growth
rates for beneficiary FFS enrollment and real GDP per capita; actual
experience had shown these growth rates to be too low. If the estimates
could have been revised, the targets for those and subsequent years would
have been increased. However, at the time that CMS acknowledged these
errors, the

2

agency concluded that it was not allowed to revise these estimates.
Without such revisions, the cumulative spending targets remained lower
than if errors had not been made.

In late 2002, the estimate of the sustainable growth rate called for a
negative 4.4 percent fee update in 2003. With the passage of the
Consolidated Appropriations Resolution of 2003,3 CMS determined that it
was authorized to correct the 1998 and 1999 spending targets. Because SGR
system targets are cumulative measures, these corrections resulted in an
average 1.4 percent increase in physician fees for services for 2003.4

2BBRA required CMS to use the most recent data to revise the estimates
used to set the spending targets, beginning with the estimated spending
target in 2000. BBRA, S:211(b)(5), 113 Stat. 348-49.

3See Pub. L. No. 108-7, Div. N. Title IV, S:402, 117 Stat. 11, 548.

4The law allowed for a recalculation of prior years' spending targets,
which resulted in a 1.7 increase in fees applied to spending on physician
services provided on or after March 1, 2003. Over 12 months, the increase
averaged 1.4 percent. The Congressional Budget Office estimated that this
provision would increase the baseline for Medicare spending by $800
million in 2003 and $53.4 billion over the 2003-2013 period.

Appendix II: Corrections to Prior Estimates Caused the SGR System's
Cumulative Targets to Produce Negative Updates

The Medicare Prescription Drug, Improvement, and Modernization Act of 2003
(MMA) averted additional fee reductions projected for 2004 and 2005 by
specifying an update to physician fees of no less than 1.5 percent for
those 2 years.5 The MMA increases replaced SGR system fee reductions of
4.5 percent in 2004 and an estimated 3.6 percent in 2005. The fee
increases will result in additional aggregate spending. Because MMA did
not make corresponding revisions to the SGR system's spending targets, its
fee increases will require the SGR system to offset the additional
spending by reducing fees beginning in 2006. In addition, recent growth in
spending due to volume and intensity, which has been larger than SGR
system targets allow, will further compound the excess spending that needs
to be recouped.

                      5See MMA, S:601(a), 117 Stat. 2300.

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

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

GAO's Mission

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.

Obtaining Copies  of 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 GAO
Reports and newly released reports, testimony, and correspondence on its
Web site. To Testimony 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] Federal Programs Automated answering system:
(800) 424-5454 or (202) 512-7470

Congressional 	Gloria Jarmon, Managing Director, [email protected](202)
512-4400 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

                                 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. ***