Veterans' Health Care: Most Care Provided Through Non-VA Programs
(Briefing Report, 04/25/94, GAO/HEHS-94-104BR).

When the Department of Veterans Affairs' (VA) health care system was
established in 1930, neither public nor private health insurance
programs were available to American veterans.  With the subsequent
growth of public and private health insurance programs, most veterans
today have alternatives to VA health care.  National health care reform
could further reduce the number of veterans lacking health insurance.
This briefing report determines (1) how many veterans are receiving
services under other federal health programs and the cost of providing
those services and (2) how many veterans using VA services are eligible
to receive care under other federal programs.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  HEHS-94-104BR
     TITLE:  Veterans' Health Care: Most Care Provided Through Non-VA 
             Programs
      DATE:  04/25/94
   SUBJECT:  Veterans benefits
             Veterans hospitals
             Health care planning
             Health care cost control
             Health insurance
             Health care programs
             Medicare programs
             Health services administration
             Employee medical benefits
IDENTIFIER:  Health Security Act
             Clinton Health Care Plan
             National Health Care Reform Initiative
             CHAMPUS
             Civilian Health and Medical Program of the Uniformed 
             Services
             Federal Employees Health Benefits Program
             
**************************************************************************
* This file contains an ASCII representation of the text of a GAO        *
* report.  Delineations within the text indicating chapter titles,       *
* headings, and bullets are preserved.  Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are           *
* identified by double and single lines.  The numbers on the right end   *
* of these lines indicate the position of each of the subsections in the *
* document outline.  These numbers do NOT correspond with the page       *
* numbers of the printed 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.                                          *
*                                                                        *
* A printed copy of this report may be obtained from the GAO Document    *
* Distribution Facility by calling (202) 512-6000, by faxing your        *
* request to (301) 258-4066, or by writing to P.O. Box 6015,             *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time.                                    *
**************************************************************************


Cover
================================================================ COVER


Briefing Report to the Ranking Minority Member, Committee on
Veterans' Affairs,
U.S.  Senate

April 1994

VETERANS' HEALTH CARE - MOST CARE
PROVIDED THROUGH NON-VA PROGRAMS

GAO/HEHS-94-104BR

VA Health Care


Abbreviations
=============================================================== ABBREV

  BIRLS - Beneficiary Identification and Records Locator Subsystem
  CHAMPUS - Civilian Health and Medical Program of the Uniformed
     Services
  CHAMPVA - Civilian Health and Medical Program of the Department of
     Veterans Affairs
  CPDF - current personnel data file
  DEERS - Defense Enrollment Eligibility Reporting System
  DOD - Department of Defense
  FEHBP - Federal Employees Health Benefits Program
  HCFA - Health Care Financing Administration
  HISKEW - Health Insurance Skeleton Eligibility Writeoff File
  MADRS - Medicare Automated Data Retrieval System
  OPC - outpatient clinic system
  OPM - Office of Personnel Management
  PTF - patient treatment file
  SC - service-connected
  SIPP - Survey of Income and Program Participation
  SSN - social security number
  VA - Department of Veterans Affairs

Letter
=============================================================== LETTER


B-254211

April 25, 1994

The Honorable Frank H.  Murkowski
Ranking Minority Member
Committee on Veterans' Affairs
United States Senate

Dear Senator Murkowski: 

When the Department of Veterans Affairs' (VA) health care system was
established in 1930, neither public nor private health insurance
programs were available to meet the health care needs of America's
veterans.  But with the subsequent growth of public and private
health insurance programs, most veterans now have one or more
alternatives to VA health care.  Reforms of the nation's health
financing system such as those currently being considered could
further reduce the number of veterans without health insurance. 

When veterans have multiple health care options, changes in one
program can have unforeseen repercussions on other programs.  For
example, we reported in June 1992 that implementation of a universal
care program could reduce demand for VA hospital care by almost 50
percent.\1

Similarly, significant changes in an existing program, such as adding
benefits or increasing cost sharing, could affect future demand for
services under other programs. 

To help insure that the potential effect of fundamental changes in
the availability of health coverage under other public or private
programs on demand for VA services is considered in planning the
future of the VA health care system, you asked that we

  assess how VA health benefits currently compare to those available
     under other health programs,

  determine how many veterans are receiving services under other
     federal health programs and the cost of providing those
     services, and

  determine how many veterans using VA services are eligible to
     receive care under other federal programs. 

Your first question was addressed in our July 29, 1993, report Health
Care:  Comparison of VA Benefits With Other Public and Private
Programs (GAO/HRD-93-94).  We reported that VA's complex eligibility
and entitlement provisions place more restrictions on the
availability of services than do other health care programs.  Once in
the VA system, however, veterans are generally offered a more
extensive array of services, fewer limitations in terms of the
duration and number of visits or services covered, and less cost
sharing than are available under most public and private health
benefits programs. 

This report addresses the two remaining questions and, as agreed with
your staff, discusses the potential implications of the data
presented on VA's role under a reformed health care system.  It
summarizes and expands on information provided in our April 15, 1994,
briefing of your staff. 

In summary, we found the following: 

  Nine out of 10 veterans have non-VA health care coverage.  Overall,
     about 81 percent have private health insurance and almost 26
     percent are Medicare-eligible.  Over 20 percent of veterans have
     both Medicare and private health insurance in addition to their
     VA coverage.  (See section 1.)

  Veterans with Medicare coverage are unlikely to use VA services. 
     Of Medicare-eligible veterans, almost 62 percent used Medicare
     but not VA services in 1990.  By contrast, fewer than 7 percent
     of Medicare-eligible veterans used only VA services.  Finally,
     fewer than 8 percent used both VA and Medicare services.\2 (See
     section 2.)

  Seven out of 10 federal dollars spent on veterans' health care come
     from non-VA programs.  Because Medicare-eligible veterans tend
     to use their Medicare coverage rather than VA, Medicare
     accounted for about $20.6 billion of the $36 billion in federal
     expenditures we examined for veterans health care in 1990
     compared with about $10.9 billion under VA.  Other federal
     programs accounted for about $4.5 billion.  (See section 3.)

  Expenditures on veterans' health care through private health
     insurance likely exceed those under VA.  If veterans' use of
     private insurance is similar to that of the general population,
     then payments for veterans likely amounted to over $22 billion. 
     (See section 3.)

  Veterans using VA services tend to have lower incomes and less
     private health insurance coverage than nonusers.  Of the over
     2.2 million veterans (about 8 percent of the nation's 28.2
     million veterans) using VA services, over half reported having
     incomes under $10,000.  More than 4 users in 10 had neither
     public nor private health insurance coverage.  Still,
     Medicare-eligible veterans accounted for almost half of the VA
     use in 1990.  (See section 4.)

  Health reform could reduce VA's role as a safety net for acute-care
     services.  Many VA users who have no public or private health
     insurance may leave the VA system if given a choice.  In
     addition, proposed changes in Medicare benefits, such as adding
     prescription drug coverage and expanded long-term care services,
     might cause more veterans to rely on Medicare for most, if not
     all, of their health care.  Without changes in the VA system, a
     significant portion of VA's acute-care work load may be lost. 
     (See section 5.)

  President Clinton's proposed Health Security Act is the only major
     health reform proposal that would change the role of the VA
     health care system.  The proposed act would (1) transform the VA
     system into a series of managed care plans to compete with
     private sector plans and (2) expand entitlement to free
     comprehensive acute-care services.  Veterans enrolling in other
     health plans would be required to pay up to 20 percent of the
     cost of their insurance premiums and any copayments and
     deductibles.

     Currently, about 445,000 veterans with service-connected
     disabilities rated at 50 percent or higher are entitled to free
     comprehensive care at VA facilities.\3 Millions of other
     veterans are eligible for free care, but are entitled to only
     selected services, such as inpatient hospital care or outpatient
     treatment for their service-connected disabilities.  Under the
     proposed Health Security Act, about 9 million veterans would be
     entitled to free comprehensive care if they enrolled in a VA
     health plan.  Many factors could affect the number of veterans
     choosing VA plans, including the extent of covered services.  To
     the extent the VA plan offers free coverage for services not
     generally covered under competing plans, such as long-term
     nursing home care, the number of veterans choosing VA plans, and
     the government's cost to operate those plans, will increase. 
     (See section 5.)

  Several other options exist for restructuring the VA health care
     system.  These include (1) maintaining a smaller direct-delivery
     system strictly for veterans; (2) opening the VA system to other
     federal beneficiaries (such as dependents of military personnel)
     to maintain work loads; (3) converting some existing facilities
     to other uses such as long-term care; (4) merging the VA system
     with one or more of the other federal health care systems, such
     as that of the Department of Defense; (5) eliminating the
     separate VA health care system and meeting the nation's
     commitment to veterans by supplementing the coverage available
     under a universal care program; and (6) contracting to provide
     hospital services to private sector managed care plans.  VA
     officials said that all of these options were considered in
     developing the proposed Health Security Act.  They said that the
     act would give VA the flexibility to adopt any of the options
     other than (1) merging the VA system with another federal health
     care delivery system or (2) eliminating the separate VA system
     and supplementing coverage available under a universal care
     program.  (See section 5.)

In summary, VA plays a small but nevertheless important role in
meeting the health care needs of America's veterans.  VA's role in
meeting the needs of service-connected veterans would not be affected
by any health reform proposal but its role as a safety-net provider
for uninsured veterans would be reduced if any of the major health
reform proposals is adopted.  As a result, VA is likely to face a
significant decline in the use of its acute care services unless
there are changes either in veterans' health benefits or in the VA
health care system.  The challenge facing the Congress and VA is to
identify the most appropriate role for the VA under a reformed health
care system. 

Data on veterans' health care coverages under public and private
programs were obtained through analysis of the Survey of Income and
Program Participation (SIPP).\4 Information on veterans' utilization
of health care services under federal health care programs and
expenditures on those services was developed through computer matches
of eligibility and payment records from those programs.  For most
programs, this report reflects actual federal expenditures on
veterans' health care in 1990.  We used 1990 as a base year for our
analyses because it was the most recent year for which data were
available for all of the federal health programs studied.  Because VA
does not have a complete database of veterans, and Medicare records
do not contain a veteran identifier, we obtained actual expenditures
for about 60 percent of the Medicare-eligible veterans and projected
total Medicare expenditures.  We supplemented the demographic data
available through these computer matches with data from VA's Survey
of Medical System Users (see p.  53).  Section 6 of this briefing
report provides detailed information on the scope and methodology of
our analyses. 

We did not obtain formal agency comments on this briefing report;
however, we did discuss a draft of the briefing report with
responsible VA officials and have included their comments where
appropriate. 

As agreed with your office, we are providing copies of this briefing
report to the Chairman, Senate Committee on Veterans' Affairs; the
Chairmen and Ranking Minority Members of the House Committee on
Veterans' Affairs and the Senate and House Committees on
Appropriations; the Secretary of Veterans Affairs; and other
interested parties.  Copies will be available to others upon request. 
Please call me at (202) 512-7101 if you or your staff have any
questions.  Major contributors to this briefing report are listed in
appendix I. 

Sincerely yours,

David P.  Baine
Director, Federal Health Care
 Delivery Issues


--------------------
\1 VA Health Care:  Alternative Insurance Reduces Demand for VA Care
(GAO/HRD-92-79, June 30, 1992). 

\2 The remaining 24 percent did not use services under either
program. 

\3 Excluding nursing home care, which is optional for all veterans,
and care for a routine pregnancy. 

\4 SIPP is a nationwide longitudinal survey based on a statistical
sample of residents of about 22,000 noninstitutional living quarters. 
It covers such areas as income, assets, employment, health insurance
coverage, veteran status, and eligibility for participation in
various government programs.  The SIPP estimate results in a
95-percent confidence interval of 26.9 to 29.5 million veterans in
1990.

VA uses a lower estimate of 26.9 million veterans in 1990 based on
its projections of 1980 census data.  We use the SIPP figure
throughout this report because some of the analyses of alternate
coverage are derived from SIPP. 


MOST VETERANS HAVE MULTIPLE HEALTH
CARE COVERAGE
============================================================ Chapter 1

With the creation and expansion of public and private health benefits
programs, about 25.6 million of the nation's estimated 28.2 million
veterans (almost 91 percent) had public and/or private health care
coverage in 1990 in addition to their VA coverage (see figure 1.1). 
Over 81 percent of veterans (22.9 million) had private health
insurance; 26 percent (7.4 million) had Medicare coverage; 5.1
percent (1.4 million) had coverage under the Civilian Health and
Medical Program of the Uniformed Services (CHAMPUS); and 1.6 percent
(0.4 million) had Medicaid coverage. 

Over 22 percent of veterans had coverage under more than one other
health benefits program.  Such veterans were primarily
Medicare-eligible veterans who also had private health insurance,
most likely Medicare supplemental policies (commonly referred to as
Medigap policies).  Nearly 79 percent of the Medicare-eligible
veterans also had private health insurance coverage. 

   Figure 1.1:  Veterans' Sources
   of Health Care Coverage (1990)

   (See figure in printed
   edition.)

Note:  Veterans covered by CHAMPUS are also eligible for care in DOD
health care facilities on a space available basis.  Veterans losing
CHAMPUS coverage upon becoming Medicare-eligible can still use DOD
facilities on a space-available basis. 


   COMPENSATION AND PENSION STATUS
   OF VETERANS
---------------------------------------------------------- Chapter 1:1

Veterans eligible for health benefits under Medicare, the Department
of Defense (DOD), CHAMPUS, or the Federal Employees Health Benefits
Program (FEHBP) are more likely to have service-connected
disabilities than other veterans.  Although fewer than 8 percent of
the overall veteran population have compensable service-connected
disabilities, over 26 percent of military retirees have such
disabilities.  Similarly, over 13 percent of Medicare-eligible
veterans and 16 percent of those covered under FEHBP have compensable
service-connected disabilities.\1

Although 2 percent of veterans receive VA pensions, few veterans
eligible for DOD/CHAMPUS and FEHBP receive VA pensions.  On the other
hand, almost 6 percent of Medicare-eligible veterans receive VA
pensions.  (See fig.  1.2.)

   Figure 1.2:  Compensation and
   Pension Status of Veterans, by
   Federal Health Program
   Eligibility (1990)

   (See figure in printed
   edition.)

Note:  SC = service-connected. 


--------------------
\1 Additional veterans have noncompensable "0" percent
service-connected disabilities.  VA does not, however, maintain a
database of such veterans.  VA estimates that about 1.6 million
veterans have "0" percent service-connected disabilities.  Throughout
this report, we show only those veterans with compensable
service-connected disabilities. 


   VETERANS WITH SERVICE-CONNECTED
   DISABILITIES OR LOW INCOMES
   MORE LIKELY TO HAVE OTHER
   FEDERAL COVERAGE
---------------------------------------------------------- Chapter 1:2

Veterans with service-connected disabilities rated at 50 percent or
higher and low-income veterans receiving VA pensions are
significantly more likely to have health care coverage under other
federal programs than are other veterans with incomes above the
pension level.  Of veterans with disabilities rated at 50 percent or
higher, 80 percent have coverage under other federal health care
programs, such as Medicare or CHAMPUS; over 40 percent of those have
multiple coverages under federal health programs.  Similarly, over
three-fourths of VA pension recipients are eligible for Medicare;
few, however, have multiple program eligibilities.  By contrast, only
3 out of 10 nonservice-connected veterans with incomes above the
pension level are covered under other federal health programs,
primarily Medicare. 

Figures 1.3 through 1.6 provide additional information on the federal
health care coverages of veterans by category. 

   Figure 1.3:  Additional
   Federally Sponsored Health
   Coverages of Veterans with
   Service-Connected Disabilities
   Rated at 50 percent or Higher
   (1990)

   (See figure in printed
   edition.)

Notes:  Of the 443,338 service-connected veterans with disabilities
rated at 50 percent or higher, 271,433 were Medicare-eligible.  Of
these, 62,491 also had DOD but not FEHBP coverage, 12,731 also had
FEHBP but not DOD coverage, and 1,591 also had both DOD and FEHBP
coverage.

Of the 171,905 service-connected veterans who were not
Medicare-eligible, 11,771 had FEHBP coverage, 65,096 had DOD/CHAMPUS
coverage, and 7,903 had both FEHBP and DOD/CHAMPUS coverage.

Figure does not include information on coverage under Medicaid or the
Indian Health Service. 

   Figure 1.4:  Additional
   Federally Sponsored Health
   Coverages for Veterans with
   Service-Connected Disabilities
   Rated at Lower Than 50 Percent
   (1990)

   (See figure in printed
   edition.)

Notes:  Of the 1,665,359 service-connected veterans with disabilities
rated at lower than 50 percent, 715,748 were Medicare-eligible.  Of
these, 65,861 also had DOD (but not FEHBP) coverage, 59,973 had FEHBP
(but not DOD) coverage, and 6,261 had both DOD and FEHBP coverage.

Of the 949,611 veterans not Medicare-eligible, 98,148 had FEHBP (but
not DOD/CHAMPUS) coverage, 204,816 had DOD/CHAMPUS (but not FEHBP)
coverage, and 39,892 had both DOD/CHAMPUS and FEHBP coverage.

Figure does not include information on coverage under Medicaid or the
Indian Health Service. 

   Figure 1.5:  Additional
   Federally Sponsored Health
   Coverages of Veterans Receiving
   VA Pensions (1990)

   (See figure in printed
   edition.)

Notes:  Of the 555,839 VA pension recipients, 426,823 were
Medicare-eligible, including 761 who also had DOD coverage, 924 who
also had FEHBP coverage, and 15 who had both DOD and FEHBP coverage.

Of the remaining 129,016 pension recipients, 742 had FEHBP (but not
DOD/CHAMPUS) coverage and 432 had DOD coverage; 10 veterans had both
DOD/CHAMPUS and FEHBP coverage.

Figure does not include information on coverage under Medicaid or the
Indian Health Service. 

   Figure 1.6:  Additional
   Federally Sponsored Health
   Coverages of Veterans Not
   Receiving VA Compensation or
   Pension Payments (1990)

   (See figure in printed
   edition.)

Notes:  Of the 25,527,883 veterans not receiving VA compensation or
pension payments, 5,985,827 were Medicare-eligible.  Of these,
367,191 also had DOD (but not FEHBP) coverage; 310,499 had FEHBP (but
not DOD) coverage; and 36,567 had both DOD and FEHBP coverage.

Of the remaining 19,542,056 veterans, 837,765 had FEHBP (but not
DOD/CHAMPUS) coverage, 758,491 had DOD/CHAMPUS (but not FEHBP)
coverage, and 73,799 had both FEHBP and DOD/CHAMPUS coverage.

Figure does not include information on coverage under Medicaid or the
Indian Health Service. 


MOST MEDICARE-ELIGIBLE VETERANS
RELY ON MEDICARE
============================================================ Chapter 2

Most Medicare-eligible veterans rely on Medicare rather than VA to
meet their health care needs.  However, veterans are more likely to
seek certain types of care from VA, such as inpatient psychiatric
care where there is only limited coverage under Medicare.  Like those
individuals without Medicare coverage, veterans with
service-connected disabilities rated at 50 percent or higher or
receiving VA pensions are more likely to seek care from VA than are
other groups of Medicare-eligible veterans. 

Almost 62 percent of the 7.4 million Medicare-eligible veterans used
Medicare but no VA services during 1990.  By contrast, fewer than 7
percent used VA but no Medicare services.  Finally, fewer than 8
percent used both Medicare and VA services during 1990 (fig.  2.1).\1

   Figure 2.1:  Use of VA and
   Medicare Services by
   Medicare-Eligible Veterans
   (1990)

   (See figure in printed
   edition.)

Note:  A total of about 7.4 million veterans were Medicare-eligible
in 1990. 


--------------------
\1 The remaining 24 percent used neither Medicare nor VA services
during 1990. 


   MEDICARE-ELIGIBLE VETERANS WITH
   SERVICE- CONNECTED DISABILITIES
   MOST LIKELY TO USE VA
---------------------------------------------------------- Chapter 2:1

Medicare-eligible veterans with service-connected disabilities rated
at 50 percent or higher were the most likely to use VA health care
services.  Veterans with service-connected disabilities rated at 50
percent or higher used VA at a greater rate than they used
Medicare--about 34 percent used VA but no Medicare services, 30
percent used both VA and Medicare services; and 24 percent used
Medicare but no VA services during 1990.\2

Medicare-eligible veterans receiving VA pensions were also more
likely to use VA services than those nonservice-connected veterans
with incomes above the pension level.  One possible explanation of
this higher usage by VA pension recipients may be the lack of Medigap
insurance or other resources to help pay the copayments and
deductibles under Medicare.  While data were not generally available
on veterans' Medigap coverage, we found that VA usage by
Medicare-eligible veterans with private health insurance coverage
under FEHBP was much lower than that of comparable veterans without
FEHBP coverage. 

Figure 2.2 provides additional information on the use of VA inpatient
and outpatient services by Medicare-eligible veterans.  Figures 2.3
through 2.6 break down the use of VA and Medicare by category of
veteran (that is, service-connected disabilities rated 50 percent or
higher, service-connected disabilities rated lower than 50 percent,
VA pension recipients, or veterans not receiving compensation or
pension payments.). 

   Figure 2.2:  Types of VA
   Services Used by
   Medicare-Eligible Veterans, by
   Veteran Category (1990)

   (See figure in printed
   edition.)

Notes:  Percentages may not add due to rounding. 

   Figure 2.3:  Use of VA and
   Medicare Services by
   Service-Connected Veterans with
   Disabilities Rated at 50
   Percent or Higher (1990)

   (See figure in printed
   edition.)

   Figure 2.4:  Use of VA and
   Medicare Services by Veterans
   With Service-Connected
   Disabilities Rated at Lower
   Than 50 Percent (1990)

   (See figure in printed
   edition.)

   Figure 2.5:  Use of VA and
   Medicare Services by VA Pension
   Recipients (1990)

   (See figure in printed
   edition.)

   Figure 2.6:  Use of VA and
   Medicare Services by Veterans
   Not Receiving VA Compensation
   or Pension Payments (1990)

   (See figure in printed
   edition.)


--------------------
\2 About 11 percent did not use either VA or Medicare services during
1990. 


   USE OF SELECTED INPATIENT
   SERVICES BY MEDICARE-ELIGIBLE
   VETERANS, BY SOURCE OF CARE
---------------------------------------------------------- Chapter 2:2

Medicare-eligible veterans were generally more likely to use VA for
inpatient psychiatric care than for inpatient medical/surgical care
and nursing home care.  Service-connected veterans with disabilities
rated at 50 percent or higher were the only category of veterans more
likely to use VA than Medicare for inpatient medical/surgical care
and nursing home care.  By contrast, all categories of veterans,
except those not receiving compensation or pension benefits, were
more likely to use VA than Medicare for inpatient psychiatric care. 

The greater reliance of severely disabled service-connected veterans
on VA even when Medicare coverage is fairly extensive is clearly
demonstrated through examination of inpatient medical/surgical care. 
Of veterans hospitalized for medical/surgical care during 1990, about
90 percent of those who were Medicare-eligible nonservice-connected
veterans with incomes above the pension level relied on Medicare and
not VA to pay their bills.  By contrast, about 47 percent of
service-connected veterans with disabilities rated at over 50 percent
relied on VA to provide their inpatient medical/surgical care. 

Figures 2.7 through 2.9 provide additional information on the use of
VA and Medicare inpatient services by Medicare-eligible veterans. 

   Figure 2.7:  Inpatient
   Medical/Surgical Care of
   Medicare-Eligible Veterans, by
   Source of Care (1990)

   (See figure in printed
   edition.)

Note:  SC = Service-connected disability. 

   Figure 2.8:  Psychiatric
   Hospital Use by
   Medicare-Eligible Veterans, by
   Source of Care (1990)

   (See figure in printed
   edition.)

Note:  SC = Service-connected disability. 

   Figure 2.9:  Nursing Home Use
   by Medicare-Eligible Veterans,
   by Source of Care (1990)

   (See figure in printed
   edition.)

Note:  SC = Service-connected disability. 


   VA ELIGIBILITY DOES NOT APPEAR
   TO REDUCE MEDICARE
   USE/EXPENDITURES
---------------------------------------------------------- Chapter 2:3

Eligibility for VA health care does not appear to reduce Medicare
utilization or expenditures.  We found no significant differences in
(1) the percentage of beneficiaries using Medicare services during
1990 or (2) the Medicare expenditures per beneficiary for
Medicare-eligible male veterans and all Medicare-eligible males age
65 and over.\3 Additional details of this comparison are contained on
page 52 in Section 6. 


--------------------
\3 We compared veterans to the general male Medicare-eligible
population age 65 and over because about 95 percent of veterans over
the age of 65 are males. 


MOST VETERANS' HEALTH CARE
FINANCED UNDER FEDERAL PROGRAMS
OTHER THAN VA
============================================================ Chapter 3

About 70 percent of the more than $36 billion in federal expenditures
on veterans' health care in 1990 was paid through programs other than
VA.\1 Private health insurance and state and local government
programs also play important roles in meeting the health care needs
of veterans, but expenditures on veterans' health care under these
programs were not readily available. 

Medicare accounted for about 57 percent ($20.6 billion) of the
federal expenditures for veterans' health care in 1990 compared with
VA's approximately 30 percent (about $10.9 billion).  Other federal
programs accounted for the remaining 13 percent of federal
expenditures (see fig.  3.1). 

   Figure 3.1:  Federal
   Expenditures on Veterans'
   Health Care, by Source of
   Payment (1990)

   (See figure in printed
   edition.)

Notes:  Medicaid and Indian Health Service expenditures could not be
readily determined.

FEHBP expenditures are based on the government's share of premiums.

Identifiable federal expenditures for veterans' health care totaled
over $36 billion in 1990. 


--------------------
\1 This is a conservative estimate--we could not readily determine
federal expenditures on veterans' health care provided through
Medicaid and the Indian Health Service. 


   MEDICARE REIMBURSEMENTS FOR
   VETERANS' HEALTH CARE
---------------------------------------------------------- Chapter 3:1

Over half of the $20.6 billion spent on veterans' health care through
the Medicare program was spent on inpatient medical and surgical
care.  Another $1.7 billion was spent on outpatient hospital care and
$6.0 billion for physicians and laboratories.  (See fig.  3.2.)

   Figure 3.2:  Medicare
   Reimbursements for Veterans'
   Health Care, by Type of Care
   (1990)

   (See figure in printed
   edition.)

Almost four times as many Medicare-eligible veterans used physician
and laboratory services, and twice as many used outpatient hospital
services as inpatient medical and surgical care.  (See figure 3.3.)

   Figure 3.3:  Number of Veterans
   Using Medicare Services, by
   Type of Service (1990)

   (See figure in printed
   edition.)

Figures 3.4 and 3.5 provide information on the average Medicare
expenditures per veteran user and per Medicare-eligible veteran,
respectively. 

   Figure 3.4:  Average Medicare
   Reimbursements per Veteran
   User, by Type of Service (1990)

   (See figure in printed
   edition.)

   Figure 3.5:  Average Medicare
   Reimbursements per
   Medicare-Eligible Veteran, by
   Type of Service (1990)

   (See figure in printed
   edition.)


   EXPENDITURES UNDER PRIVATE
   INSURANCE LIKELY TO EXCEED VA
   EXPENDITURES
---------------------------------------------------------- Chapter 3:2

Although expenditures on veterans' health care through private health
insurance are not readily available, they likely exceed those under
VA.  Data published by the Health Insurance Association of America,
an industry trade association, indicate that private health insurance
paid claims of about $980 per insured person in 1989.\2 If veterans'
use of private insurance is similar to that of the general
population, then private health insurance payments for veterans
likely amounted to over $22 billion. 


--------------------
\2 The Association's 1991 Source Book of Health Insurance Data
reports that 189 million people were covered by private health
insurance in 1989 and claim payments totaled $185.3 billion ($980 per
insured). 


   STATE AND LOCAL PROGRAMS ALSO
   PAY FOR SERVICES FOR VETERANS
---------------------------------------------------------- Chapter 3:3

Finally, state and local programs likely play a significant role in
meeting the health care needs of veterans.  Fewer than half of the
veterans with no health insurance coverage have ever used a VA
hospital.  VA officials believe many of those veterans may be unable
to use VA services because of geographic inaccessibility on VA's
complex eligibility and entitlement provisions.  It is not clear to
what extent such veterans are obtaining care from other sources but
public hospitals and clinics are likely providing services to many
uninsured veterans. 


VA PROVIDES A SAFETY NET FOR SOME
VETERANS BUT MOST VA USERS HAVE
OTHER COVERAGE
============================================================ Chapter 4

VA continues to play a critical role in meeting the health care needs
of some veterans by serving as a safety net for veterans with limited
health care options.  Veterans who use VA facilities tend to have
lower incomes and less private health insurance coverage than the
overall veteran population.\1 For example, only about 33 percent of
veterans using VA hospitals have private health insurance compared
with 81 percent of the overall veteran population identified through
our Survey of Income and Program Participation analysis.  Similarly,
over half of VA users had incomes of lower than $10,000.  More than 4
users in 10 were found to be medically indigent in VA's Survey of
Medical System Users, having neither private nor public insurance
that would enable them to pay their health care bills. 

Our analyses confirm that veterans with low incomes (that is, those
receiving VA pensions) are more likely to use VA services than are
higher-income veterans.  For example, among non-Medicare-eligible
veterans, about 23 percent of those receiving VA pensions used VA
hospital services during 1990 compared with about 8 percent of
nonservice-connected veterans not receiving VA pensions. 

Still, over 56 percent of the veterans who used VA health care
services in 1990 had other federal health care coverage.  In fact,
veterans using VA health services are more likely to have other
federal health care coverage than those veterans who do not use VA. 
For example, 47 percent of the veterans who used the VA system in
1990 were Medicare-eligible, even though only 26 percent of all
veterans were Medicare-eligible.  One possible explanation for this
is that VA users tend to be older, have lower incomes, and have less
private insurance coverage than the overall veteran population. 

Figures 4.1 through 4.6 contain additional information on VA health
care users, including their compensation and pension status, Medicare
eligibility, and use of services by veterans not Medicare-eligible. 

   Figure 4.1:  Health Care
   Options of Veterans Using VA
   Health Care During 1990

   (See figure in printed
   edition.)

Notes:  Percents do not add to 100 because some veterans have
multiple coverages.

Data on private health insurance coverage and the percentage of
veterans with no health care options are from the VA Survey of
Medical System Users.

Data on Medicaid and the Indian Health Service were not readily
available. 

   Figure 4.2:  VA Users, by
   Category of Veteran (1990)

   (See figure in printed
   edition.)

   Figure 4.3:  Percentage of
   Veterans Using VA Health Care
   Services, by Medicare and VA
   Status (1990)

   (See figure in printed
   edition.)

Note:  SC = Service-connected disability. 

   Figure 4.4:  Types of VA
   Services Used by Veterans Not
   Eligible for Medicare, by
   Category of Veteran (1990)

   (See figure in printed
   edition.)

Note:  SC = Service-connected disability. 

   Figure 4.5:  Use of VA Services
   by Veterans With FEHBP but Not
   Medicare Coverage, by Category
   of Veteran (1990)

   (See figure in printed
   edition.)

Note:  SC = Service-connected disability. 

   Figure 4.6:  Use of VA Services
   by Veterans Eligible for
   CHAMPUS but Not Medicare, by
   Category of Veteran (1990)

   (See figure in printed
   edition.)

Note:  SC = Service-connected disability. 


--------------------
\1 Based on VA's Survey of Medical System Users, conducted between
August 1988 and May 1989.  VA surveyed 2,865 veterans who had been
inpatients in a VA medical center during fiscal year 1987.  The
survey developed a sociodemographic profile of VA medical system
users including age, income, and insurance coverage. 


IMPLICATIONS OF HEALTH REFORM FOR
THE VA HEALTH CARE SYSTEM
============================================================ Chapter 5


   HEALTH REFORM COULD FURTHER
   REDUCE VA ROLE AS A SAFETY NET
---------------------------------------------------------- Chapter 5:1

Reform of the nations' health financing system to reduce the number
of Americans who lack coverage of basic acute health care services
could further reduce VA's role as a safety net for low-income and
uninsured veterans.  Similarly, expansion of Medicare benefits could
result in decreased demand for those types of VA services.  Without
fundamental changes in the structure of the veterans' health care
system, VA could lose much of its current acute care work load as a
result of health reform.  Health reform is, however, unlikely to
affect demand for those types of services, such as long-term
psychiatric care, that are not extensively covered under other public
and private insurance and would not change VA's obligation to meet
the health care needs of service-connected veterans. 

Although 7 out of 10 federal dollars spent on veterans' health care
in 1990 came from Medicare and other non-VA programs, VA continues to
serve as a safety net for veterans without alternative coverage or
without the resources to pay for copayments and deductibles.  About
40 percent of veterans using VA have neither public nor private
insurance to help pay for their care in private sector facilities. 
We previously reported that many of these veterans, given health
insurance, would likely choose to obtain much of their future care
from private sector facilities.\1

But given that over 56 percent of VA users do have other federal
health care coverage, it is also important to explore the factors
that contribute to their decisions to use VA care.  Cost-sharing
appears to be an important determinant of VA use as 21 percent of
Medicare-eligible VA pension recipients (a proxy for low-income) used
VA but no Medicare services (and another 22 percent used both VA and
Medicare services) whereas only 4 percent of nonservice-connected
veterans not receiving VA pensions used VA but no Medicare services
(and another 5 percent used a combination of VA and Medicare
services). 

A second major determinant appears to be covered services.  Virtually
all public and private health insurance provides coverage of acute
inpatient medical/surgical care.  Not surprisingly then, most
Medicare-eligible veterans relied on Medicare to meet such acute care
needs.  Those most likely to use VA were severely disabled veterans
and VA pension recipients who may be unable to afford Medicare
copayments and deductibles. 

By contrast, all categories of Medicare-eligible veterans, except
those not receiving compensation or pension payments, tended to use
VA for psychiatric care, presumably because of the limited coverage
of inpatient psychiatric care under Medicare.  Another area where
higher-income Medicare-eligible veterans made significant use of VA
was for nursing home care.  While such veterans relied primarily on
Medicare, which has a very limited nursing home program with high
cost sharing, about 12 percent relied on VA. 

Thus, the extent to which veterans gaining health insurance coverage
through health reforms reduce their use of the VA system will, in
large measure, depend on the services covered and the out-of-pocket
costs that would be incurred.  In other words, if low-income veterans
are given health insurance through health reform, but would be
required to pay significant copayments or deductibles to obtain care
from private sector facilities, many might choose to stay with the VA
system.  Others will likely stay with VA for specific services not
covered by their insurance. 

Just as many veterans gaining health insurance through health reform
might reduce their use of VA care, those VA users with health
insurance might reduce their use of VA services if their insurance
coverage is improved or cost sharing reduced.  For example, the 8
percent of Medicare-eligible veterans who used both VA and Medicare
services might shift more toward Medicare usage if the Medicare
program adds prescription drug coverage or expanded long-term care
services. 

Similar to the fact that veterans with private insurance tend to use
private sector providers rather than VA, our current work shows that
Medicare-eligible veterans generally choose to use those benefits to
pay for health care services from private sector facilities rather
than obtaining care from VA.  Fewer than 7 percent of
Medicare-eligible veterans used VA but no Medicare services during
1990. 

Clearly, without fundamental changes in the VA health care system, VA
could face significant decreases in its acute care work load. 


--------------------
\1 See VA Health Care:  Alternative Insurance Reduces Demand for VA
Care (GAO/HRD-92-79, June 30, 1992). 


   HEALTH SECURITY ACT ONLY MAJOR
   HEALTH REFORM PROPOSAL THAT
   WOULD CHANGE THE ROLE OF THE VA
   SYSTEM
---------------------------------------------------------- Chapter 5:2

Although many veterans would likely continue to seek treatment at VA
facilities under a universal health care system, the magnitude of the
likely decline in demand for VA-sponsored care suggests that plans
for restructuring the VA health care system be developed as part of a
national health care reform initiative.  Of the major health reform
bills that have been introduced, however, only one--the President's
Health Security Act--contains plans for restructuring the VA health
care system. 

The proposed Health Security Act would (1) transform VA facilities
into a series of managed care plans to compete with private sector
plans and (2) expand entitlement to free comprehensive acute care
services for veterans choosing the VA plan.  Additional VA services
not covered under the comprehensive benefits plan would continue to
be offered to all veterans under existing eligibility and entitlement
provisions, subject in most cases to the availability of resources. 
VA would also be given the authority to provide services to the
veterans' dependents. 

Currently, about 445,000 veterans with service-connected disabilities
rated at 50 percent or higher are entitled to free comprehensive
health care services from VA.\2 Millions of other veterans are
eligible for free care, but entitled only to certain services, such
as inpatient hospital care or outpatient treatment for their
service-connected disabilities.  Under the proposed Health Security
Act, about 9 million veterans would be entitled to free comprehensive
care if they enrolled in a VA health plan.  Veterans enrolling in
other health plans would be required to pay up to 20 percent of the
cost of their insurance premiums and any copayments and deductibles. 

In transforming the VA system into a series of managed care plans, VA
expects to build or lease hundreds of additional outpatient
facilities and to contract for services for the dependents of
veterans. 

Other provisions of the proposed Health Security Act might limit the
effectiveness of the VA provisions in attracting veterans.  First,
low-income individuals would obtain subsidies regardless of which
health plan they choose.  This could largely negate the incentive for
low-income veterans to enroll in the VA plan.  Second, Medicare would
be expanded to offer a prescription drug benefit and expanded
long-term care benefits.  This could reduce the use of such VA
services by Medicare-eligible veterans. 

Under the proposed system, all veterans, including those currently
without public or private insurance, would have comprehensive acute
care services with or without a VA managed care plan.  Those with low
incomes would receive subsidies to help pay for their insurance
coverage.  But veterans have other health care needs, such as
substance abuse treatment, vision care, dental care, long-term
psychiatric care, and long-term nursing home care, that would not be
extensively covered under the managed care plans.  Implementation of
a universal coverage program would give VA the opportunity to shift
resources and programs from basic acute care services to meeting
veterans' remaining health care needs. 


--------------------
\2 The provision of nursing home care is an optional benefit for all
veterans. 


   OTHER OPTIONS EXIST FOR
   RESTRUCTURING THE VA HEALTH
   CARE SYSTEM
---------------------------------------------------------- Chapter 5:3

Converting VA facilities into managed care plans is but one option
for preserving veterans' health care benefits.  As we pointed out in
our December 1992 report, Veterans Affairs Issues (GAO/OGC-93-21TR),
other options that could be considered include

  maintaining a smaller direct delivery system strictly for veterans,
     but focusing on those services, such as treatment of spinal cord
     injuries and service-connected disabilities, that may not be
     adequately covered under a reformed national health care system;

  maintaining the current direct delivery system but opening the
     system to other federal beneficiaries to maintain work loads;

  converting some existing facilities to other uses, such as
     long-term psychiatric care, nursing home care, housing for
     homeless veterans, or AIDS treatment facilities;

  merging the VA system with one or more of the other federal health
     care systems, such as the DOD health care system; or

  eliminating the separate VA health care system and meeting the
     nation's commitment to veterans by supplementing the coverage
     available under a national health care reform initiative. 

Another option under a managed care approach such as the one proposed
by the Health Security Act would be to preserve VA facilities by
contracting to provide services to private sector health plans.  Such
an option might enable VA to preserve or expand its inpatient work
load without undertaking a costly expansion of VA facilities.  Such
contracts could be restricted to treatment of veterans or be expanded
to other federal beneficiaries. 

VA officials said that all of the above options were considered in
developing the proposed Health Security Act.  They said that the act
would give VA the flexibility to adopt any of the options other than
(1) merging the VA system with another federal health care delivery
system or (2) eliminating the separate VA system and supplementing
coverage available under a universal care program. 


SCOPE AND METHODOLOGY
============================================================ Chapter 6


   DETERMINING THE NUMBER OF
   VETERANS WITH INSURANCE
   COVERAGE
---------------------------------------------------------- Chapter 6:1

To estimate the number of veterans covered by federal and private
health benefits programs, we analyzed data contained in the Census
Bureau's 1990 Survey of Income and Program Participation (SIPP).\1
SIPP is a nationwide longitudinal survey based on a statistical
sample of about 22,000 noninstitutional households and covers such
areas as health insurance coverage, income, assets, veteran status,
and eligibility for participation in various government programs. 

Because a statistical sample rather than the entire population was
surveyed, each estimate from the survey has a standard error
associated with it.  By using each estimate and its standard error,
we calculated a 95-percent confidence interval around each estimate. 
This means there is a 95-percent chance that the actual population
total of interest falls within that interval. 


--------------------
\1 We used the Panel Wave 1 Rectangular Core File.  The Wave 1 SIPP
survey consists of four rotational groups, each interviewed in a
different month from February through May 1990.  For each group, the
reference period for reporting is the four calendar months preceding
the interview month and, thus, one of the reference months in each
rotational group overlaps.  The common month for Wave 1--the month
occurring in all the reference periods--is January.  We chose January
as the reference month for our analysis because this ensured only one
record for each sampled person and provided the largest population. 


      ESTIMATING THE VETERAN
      POPULATION
-------------------------------------------------------- Chapter 6:1.1

We extracted the records for each rotational group with the reference
month equal to January 1990.  We estimated the veteran population on
the basis of the number of people who responded "yes" to the question
"Did this person ever serve on active duty in the U.S.  Armed
Forces?" and "no" to the question "Is this person currently in the
Armed Forces?" There were 6,080 such respondents.  On that basis, we
estimated that the total veteran population was 28,192,419.  The
95-percent confidence interval around this estimate is 26,865,627 to
29,519,211. 


      ESTIMATING MEDICARE COVERAGE
-------------------------------------------------------- Chapter 6:1.2

We estimated the number of veterans covered by Medicare on the basis
of the number of veterans in the sample who responded "yes" to the
question "Is this person covered by Medicare?" to estimate the number
of veterans covered by Medicare.  There were 1,592 such veterans.  On
that basis, we estimated that the total Medicare-eligible veteran
population was 7,399,831.  The 95-percent confidence interval around
this estimate is 6,762,565 to 8,037,097. 


      ESTIMATING MEDICAID COVERAGE
-------------------------------------------------------- Chapter 6:1.3

We estimated the number of veterans covered by Medicaid on the basis
of the number of veterans in the sample who responded "yes" to the
question "Was this person covered by Medicaid for the month?" There
were 95 such veterans.  On that basis, we estimated that the total
Medicaid-eligible veteran population was 439,446.  The 95-percent
confidence interval around this estimate is 280,863 to 598,029. 


      ESTIMATING CHAMPUS COVERAGE
-------------------------------------------------------- Chapter 6:1.4

We estimated the number of veterans covered by CHAMPUS on the basis
of the number of veterans in the sample who responded "yes" to the
question "CHAMPUS/CHAMPVA coverage?" Because CHAMPUS coverage ends at
age 65, we eliminated those veterans over 65 from our determination
of CHAMPUS coverage.  In addition, CHAMPVA coverage was naturally
eliminated because it covers dependents of veterans not veterans. 
There were 283 veterans who met the above criteria.  On that basis,
we estimated that the total CHAMPUS-eligible veteran population was
1,432,796.  The 95-percent confidence interval around this estimate
is 1,147,286 to 1,718,306. 


      ESTIMATING PRIVATE HEALTH
      INSURANCE COVERAGE
-------------------------------------------------------- Chapter 6:1.5

We estimated the number of veterans covered by private insurance on
the basis of the number of veterans in the sample who responded "yes"
to the question "Did this person have health insurance coverage this
month?" There were 4,972 such veterans.  On that basis, we estimated
that the total privately insured veteran population was 22,924,828. 
The 95-percent confidence interval around this estimate is 21,712,928
to 24,136,728. 


      ESTIMATING THE NUMBER OF
      VETERANS WITHOUT PUBLIC OR
      PRIVATE HEALTH INSURANCE
-------------------------------------------------------- Chapter 6:1.6

We estimated the number of veterans without private or public health
insurance on the basis of the number of veterans in the sample who
did not answer "yes" to any of the above questions.  There were 549
such veterans.  On that basis, we estimated that the total uninsured
veteran population was 2,600,145.  The 95-percent confidence interval
around this estimate is 2,172,502 to 3,027,788. 

In addition to the above estimates, we developed estimates of
multiple coverages following the same techniques. 


   ESTIMATING EXPENDITURES ON
   VETERANS' HEALTH CARE UNDER
   FEDERAL HEALTH PROGRAMS AND
   ANALYZING VA AND MEDICARE
   UTILIZATION
---------------------------------------------------------- Chapter 6:2

To determine federal expenditures for veterans' health care under VA
and other federal health programs and analyze veterans' utilization
of health services under VA and Medicare, we obtained and analyzed
fiscal year 1990 (1) Medicare eligibility and payment data from the
Health Care Financing Administration, which administers Medicare; (2)
cost and eligibility data from the Department of Defense and its
Office of CHAMPUS, which administer the DOD health care system and
CHAMPUS, respectively; (3) employee and annuitant records from the
Office of Personnel Management, which administers the Federal
Employees Health Benefits Program;\2 and (4) cost and utilization
data from VA. 

While the statistical and financial reporting systems of DOD,
CHAMPUS, and VA contained sufficient program eligibility and cost
information to enable us to estimate how much was spent on veterans'
health care under those programs, HCFA and OPM information systems
did not.  For example, Medicare's eligibility records did not contain
a veteran indicator. 

As a result, we developed a database of veterans (hereafter referred
to as "LIVEVETS") and performed a series of computerized tape matches
to arrive at Medicare and FEHBP expenditures for veterans' health
care in 1990.  The database was also used to analyze the number of
veterans who used VA health care services in 1990 who also had
coverage under other public health benefits programs and use of
Medicare and VA health care services by Medicare-eligible veterans
that same year. 


--------------------
\2 FEHBP data on active postal workers was obtained from the United
States Postal Service. 


      CREATION OF LIVEVETS
-------------------------------------------------------- Chapter 6:2.1

We created LIVEVETS because no complete database of veterans exists. 
VA officials told us that VA's Beneficiary Identification and Records
Locator Subsystem (BIRLS) was the most complete database of veterans. 
Even so, over 11 million of the veteran records had no social
security number (SSN)--about 40 percent of over 29 million records in
the file.  In addition, the database did not reflect all veteran
deaths. 

We did some preliminary comparisons of the numbers of usable records
(that is, those with SSNs) in BIRLS with VA and SIPP projections of
the number of veterans by age group.  BIRLS appears to be a fairly
complete database for younger veterans (VA now receives copies of
military discharge papers and creates a BIRLS record) but an
incomplete database for older veterans.  There were SSNs for only
about 49 percent of the estimated number of veterans aged 65 or
older. 

In order to develop the LIVEVETS file, we obtained information from
the following databases (some are extracts and some are complete
databases): 

  BIRLS;

  Social Security Death Master File to identify veterans who were
     deceased before 1990;

  patient treatment file (PTF) to identify inpatients at VA medical
     facilities;

  outpatient clinic system (OPC) to identify patients treated on an
     outpatient basis at VA facilities;

  Active Employee Reference File from the United States Postal
     Service to identify active postal service employees who are
     veterans and their participation in FEHBP;

  Current Personnel Data File (CPDF) from OPM to identify all other
     active federal employees and their FEHBP coverage;

  Annuity Roll from OPM to identify all retired federal and postal
     employees who are veterans and their FEHBP coverage;

  Defense Enrollment Eligibility Reporting System (DEERS) from DOD to
     identify retired military personnel (veterans) eligible for
     treatment in the DOD direct care system or through CHAMPUS; and

  Health Insurance Skeleton Eligibility Writeoff File (HISKEW) from
     HCFA to identify veterans eligible for Medicare part A and/or
     part B. 

First, we eliminated those BIRLS records with no SSNs.  This reduced
the size of the file from 29,401,503 records to 18,141,393 records. 

Second, we matched the remaining records against the Social Security
Death Master File and eliminated 354,227 deceased veterans.  This
further reduced the LIVEVETS file to 17,787,166 records. 

Third, to enhance the LIVEVETS file and determine how many veterans
were eligible for health care under the various federal health care
programs, we conducted matches of the LIVEVETS file to the other
files mentioned above, adding veterans that were not originally on
BIRLS and noting which veterans were eligible for health care
services from each program during 1990.  This process resulted in the
addition of 1,332,129 additional SSNs to the LIVEVETS file as
follows: 

  PTF (574,282 records already in LIVEVETS; 9,239 veterans added),

  OPC (1,310,444 records already in LIVEVETS; 343,477 veterans
     added),\3

  USPS (200,227 records already in LIVEVETS; 90,656 records added),

  CPDF (433,001 records already in LIVEVETS; 203,870 records added),

  Annuity (465,523 records already in LIVEVETS; 437,686 records
     added, and

  DEERS (1,443,976 records already in LIVEVETS; 247,201 records
     added). 

Fourth, the resulting 19,119,295 veteran SSNs in the LIVEVETS file
were matched against Medicare's HISKEW file, yielding 4,411,558
Medicare-eligible veterans. 

Fifth, we matched the Medicare-eligible veterans against the Medicare
Automated Data Retrieval System (MADRS) to obtain the number of
veterans receiving Medicare covered services and the cost of those
services.  The match yielded 3,059,855 Medicare users. 

Finally, to enable us to analyze differences in VA and Medicare
utilization based on the existence and extent of service-connected
disabilities and receipt of a VA pension, we matched the LIVEVETS
file against the Compensation and Pension Minimaster file to mark all
pension recipients and veterans with a service-connected disability
including the percentage rating of that disability. 

We matched LIVEVETS against the Compensation and Pension Minimaster
file because of concern about the accuracy of the compensation and
pension indicators in BIRLS, OPC, and PTF.  Because the compensation
and pension file is a payment file and likely to be more accurate
than the information contained in the other databases, we used the
compensation and pension indicators from the Compensation and Pension
Minimaster file. 

The match identified 2,664,536 compensation and pension
recipients--443,338 with service-connected disabilities rated at 50
percent or higher, 1,665,359 with service-connected disabilities
rated at lower than 50 percent, and 555,839 pension recipients. 
However, the match also identified 50,495 veterans not on our
LIVEVETS file.  Because these veterans were identified after LIVEVETS
had been matched against Medicare records and account for fewer than
2 percent of veterans contained in the compensation and pension file,
we excluded these veterans from our analysis. 

The final LIVEVETS file contains 19,119,295 veteran records or about
68 percent of the estimated 28.2 million veterans living in 1990. 
The database also contains about 60 percent of all veterans eligible
for Medicare that same year (based on projections of the
Medicare-eligible veteran population from SIPP).  The database
contains all veterans who (1) used VA outpatient clinics or obtained
inpatient care under VA auspices (with the exception of the
approximately 30,000 VA outpatient users discussed in footnote 3),
(2) were retired from the uniformed services, and (3) were either
current or retired federal or postal employees. 


--------------------
\3 Approximately 600,000 veteran SSNs were initially omitted from the
match because of an unnoticed change in coding of the veteran
indicator field in the OPC.  We subsequently matched these records
against LIVEVETS and identified about 30,000 veteran SSNs that should
have been added to LIVEVETS during the initial match or add
processing.  Because this error was not identified until after
LIVEVETS had been matched against Medicare records, and accounted for
fewer than 1.4 percent of VA users (30,000 out of 2.2 million) we
excluded the 30,000 SSNs from our analysis. 


      ESTIMATING MEDICARE
      EXPENDITURES FOR VETERANS'
      HEALTH CARE
-------------------------------------------------------- Chapter 6:2.2

We identified 4,411,558 (60 percent) of the estimated 7,399,831
veterans eligible for Medicare in 1990 through our match of LIVEVETS
to HCFA's HISKEW files.\4 This is consistent with our earlier finding
that most of the BIRLS records with missing SSNs were for older
veterans. 

We compared our Medicare-eligible veteran population to the overall
population of Medicare-eligible males over the age of 65 using three
measures

  rate of use of Medicare services;

  expenditures per beneficiary; and

  expenditures per Medicare user.\5

We found no significant differences between our veteran population
and the overall Medicare population on any of the measures.  The
utilization rate of Medicare eligible beneficiaries (69.4 percent for
our veteran population compared with 71.6 percent for the
Medicare-eligible aged males); the amounts Medicare paid per user
($4,015 for our veteran population compared with $4,018 per
Medicare-eligible aged male user) and the amounts paid per eligible
beneficiary ($2,785 for Medicare eligible veterans compared with
$2,876 per Medicare-eligible aged male) were all similar.  This
allows us to suggest that our 2,988,273 "unknown" Medicare-eligible
veterans used Medicare at the same rate as our 4,411,558 known
Medicare-eligible veterans. 

Accordingly, we projected the amount of expenditures obtained through
our tape matches to the remaining universe of veterans to arrive at
the amount of Medicare expenditures on behalf of all
Medicare-eligible veterans in 1990. 


--------------------
\4 Based on 1990 SIPP data. 

\5 We chose Medicare-eligible males over the age of 65 as our point
of comparison because over 95 percent of all veterans are male and
over 90 percent of Medicare-eligible veterans are over the age of 65. 


      ESTIMATING FEDERAL
      EXPENDITURES FOR FEHBP
-------------------------------------------------------- Chapter 6:2.3

Federal expenditures for FEHBP are the government's share of premiums
paid various plans for covering participating employees.  FEHBP
participants have a wide choice of health plans and the government's
share of premium payments varies among those plans.  We identified
1,498,591 current and retired federal employees who were veterans
participating in FEHBP by matching the LIVEVETS file to OPM and U.S. 
Postal Service tapes containing FEHBP plan enrollment data.  Using a
computerized table containing the government's share of premium
payments for the various FEHBP plans, we then calculated the federal
expenditures on behalf of veterans participating in FEHBP during
calendar year 1990. 


   DEVELOPING DEMOGRAPHIC DATA ON
   VETERANS WHO USED VA FACILITIES
---------------------------------------------------------- Chapter 6:3

Because LIVEVETS does not contain information on private health
insurance coverage other than FEHBP and contains no data on veteran
incomes, we supplemented the data in LIVEVETS with data from VA's
Survey of Medical System Users.  This survey covered 2,865 veterans
who had been inpatients in a VA medical center during fiscal year
1987.  Because it is based on veterans using inpatient services, it
may or may not be representative of veterans who used only VA
outpatient services. 

We did our work between August 1991 and December 1993. 


MAJOR CONTRIBUTORS TO THIS
BRIEFING REPORT
=========================================================== Appendix I

James R.  Linz, Assistant Director, (202) 512-7116
Donald F.  Hass, Evaluator-in-Charge
Linda K.  Sanders
Dorothy M.  Tejada
Donna L.  Berryman
Barbara A.  Johnson


RELATED GAO PRODUCTS
=========================================================== Appendix I

VA Health Care:  Comparison of VA Benefits With Other Public and
Prvate Programs (GAO/HRD-93-94, July 29, 1993). 

VA Health Care:  Potential for Offsetting Long-Term Care Costs
Through Estate Recovery (GAO/HRD/93-68, July 27, 1993). 

Veterans Affairs:  Accessibility of Outpatient Care at VA Medical
Centers (GAO/T-HRD-93-29, July 21, 1993). 

VA Health Care:  Variabilities in Outpatient Care Eligibility and
Rationing Decisions (GAO/HRD-93-106, July 14, 1993). 

VA Health Care:  Veterans' Efforts to Obtain Outpatient Care From
Alternative Sources (GAO/HRD-93-123, June 30, 1993). 

Veterans' Health Care:  Potential Effects of Health Care Reforms on
VA's Major Construction Program (GAO/-T-HRD-93-19, May 6, 1993). 

Veterans' Health Care:  Potential Effects of Health Financing Reforms
on Demand for VA Services (GAO/T-HRD-93-12, Mar.  31, 1993). 

Veterans' Health Care:  Potential Effects of Health Reforms on VA
Construction (GAO/T-HRD-93-7, Mar.  3, 1993). 

VA Health Care:  Actions Needed to Control Major Construction Costs
(GAO/HRD-93-75, Feb.  26, 1993). 

Veterans' Affairs Issues (GAO/OCG-93-21TR, Dec.  1992). 

VA Health Care:  Offsetting Long-Term Care Costs by Adopting State
Copayment Practices (GAO/HRD-92-96, Aug.  12, 1992). 

VA Health Care:  Demonstration Project Concerning Future Structure of
Veterans' Health Reform (GAO/T-HRD-92-53, Aug.  11, 1992). 

VA Health Care:  Alternative Health Insurance Reduces Demand for VA
Health Care (GAO/HRD-92-79, June 30, 1992). 

