oversight

Military Retiree Health Benefits: Enrollment Low in Federal Employee Health Plans under DOD Demonstration

Published by the Government Accountability Office on 2003-06-06.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

             United States General Accounting Office

GAO          Report to Congressional Committees




June 2003
             MILITARY RETIREE
             HEALTH BENEFITS
             Enrollment Low in
             Federal Employee
             Health Plans under
             DOD Demonstration




GAO-03-547
                                                June 2003


                                                MILITARY RETIREE HEALTH BENEFITS

                                                Enrollment Low in Federal Employee
Highlights of GAO-03-547, a report to           Health Plans under DOD Demonstration
Congressional Committees




Prior to 2001, military retirees who            Enrollment in the DOD-FEHBP demonstration was low, peaking at 5.5
turned age 65 and became eligible               percent of eligible beneficiaries in 2001 (7,521 enrollees) and then falling to
for Medicare lost most of their                 3.2 percent in 2002, after the introduction of comprehensive health coverage
Department of Defense (DOD)                     for all Medicare-eligible military retirees. Enrollment was considerably
health benefits. The DOD-Federal                greater in Puerto Rico, where it reached 30 percent in 2002. Most retirees
Employees Health Benefits
Program (FEHBP) demonstration
                                                who knew about the demonstration and did not enroll said they were
was one of several demonstrations               satisfied with their current coverage, which had better benefits and lower
established to examine alternatives             costs than the coverage they could obtain from FEHBP. Some of these
for addressing retirees’ lack of                retirees cited, for example, not being able to continue getting prescriptions
Medicare supplemental coverage.                 filled at military treatment facilities if they enrolled in the demonstration.
The demonstration was mandated                  For those who enrolled, the factors that encouraged them to do so included
by the Strom Thurmond National                  the view that FEHBP offered retirees better benefits, particularly
Defense Authorization Act for                   prescription drugs, than were available from their current coverage, as well
Fiscal Year 1999 (NDAA 1999),                   as the lack of any existing coverage.
which also required GAO to
evaluate the demonstration. GAO                 Monthly premiums charged to enrollees for individual policies in the
assessed enrollment in the
demonstration and the premiums
                                                demonstration varied widely—from $65 to $208 in 2000—with those plans
set by demonstration plans. To do               that had lower premiums and were better known to eligible beneficiaries,
this, GAO, in collaboration with the            capturing the most enrollees. In setting premiums initially, plans had little
Office of Personnel Management                  information about the health and probable cost of care for eligible
(OPM) and DOD, conducted a                      beneficiaries. Demonstration enrollees proved to have lower average health
survey of enrollees and eligible                care costs than either their counterparts in the civilian FEHBP or those
nonenrollees. GAO also examined                 eligible for the demonstration who did not enroll. Plans enrolled similar
DOD enrollment data, Medicare                   proportions of beneficiaries in poor health, regardless of whether they
and OPM claims data, and OPM                    charged higher, lower, or the same premiums for the demonstration as for
premiums data.                                  the civilian FEHBP.

                                                In commenting on a draft of the report, DOD concurred with the overall
                                                findings but disagreed with the description of the demonstration’s impact on
                                                DOD’s budget as small. As noted in the draft report, DOD’s costs for the
                                                demonstration relative to its total health care budget were less than 0.1
                                                percent of that budget. OPM declined to comment.

                                                DOD-FEHBP Demonstration Enrollment, 2000-2002




www.gao.gov/cgi-bin/getrpt?GAO-03-547.

To view the full product, including the scope
and methodology, click on the link above.
For more information, contact Marjorie E.
                                                Note: Enrollment is expressed as a percentage of eligible beneficiaries.
Kanof (202) 512-7101.
Contents


Letter                                                                                      1
               Results in Brief                                                             3
               Background                                                                   5
               Enrollment Was Low, Largely Due to Beneficiaries’ Satisfaction
                 with Existing Coverage                                                   13
               Premiums Varied Widely, Reflecting Plans’ Different Assessments
                 of Demonstration Risk                                                    20
               Impact of Demonstration on DOD Was Limited Due to Small Size
                 and Low Enrollment, but Impact on Enrollees Was Greater                  26
               Agency Comments                                                            27

Appendix I     GAO-DOD-OPM Survey of Military Retirees and
               Others Eligible for the DOD-FEHBP Demonstration                            30



Appendix II    Data, Methods, and Models Used in Analyzing
               Factors Affecting DOD-FEHBP Demonstration
               Enrollment                                                                 35



Appendix III   Enrollment in the DOD-FEHBP Demonstration                                  43



Appendix IV    DOD’s Approach to Informing Beneficiaries about
               the DOD-FEHBP Demonstration                                                45



Appendix V     Enrollees’ and Nonenrollees’ Reasons for Joining
               or Not Joining a DOD-FEHBP Demonstration Plan                              50



Appendix VI    Comments from the Department of Defense                                    52




               Page i                             GAO-03-547 Military Retiree Health Benefits
Appendix VII           GAO Contacts and Staff Acknowledgments                                      53
                       GAO Contacts                                                                53
                       Acknowledgments                                                             53

Related GAO Products                                                                               54



Tables
                       Table 1: Number of Eligible Beneficiaries by DOD-FEHBP
                                 Demonstration Site, 2000-2002                                     11
                       Table 2: Monthly Premiums Charged to Enrollees for Individual
                                 Policies in the DOD-FEHBP Demonstration, 2000                     21
                       Table 3: Average Spending on Medicare-covered Services for
                                 Retirees Eligible for the DOD-FEHBP Demonstration—by
                                 Enrollment Status, 2000                                           25
                       Table 4: Major Survey Sections and Topics Covered                           31
                       Table 5: Survey Responses and Nonresponses                                  33
                       Table 6: Population, Sample Size, and Response Rate, by DOD-
                                 FEHBP Demonstration Site and Enrollee Status, 2000                34
                       Table 7: Estimated Effects of Selected Factors on Whether
                                 Eligible Retirees Knew about the DOD-FEHBP
                                 Demonstration                                                     37
                       Table 8: Estimated Effects of Selected Factors on Whether
                                 Eligible Retirees Enrolled in an FEHBP Plan                       39
                       Table 9: Health Status Comparisons of DOD-FEHBP
                                 Demonstration Enrollees with Eligible Retirees Who Did
                                 Not Enroll and with Civilian FEHBP Retirees, Based on
                                 PIP-DCG Scores                                                    42
                       Table 10: Enrollment in the DOD-FEHBP Demonstration, 2000                   43
                       Table 11: Enrollment in the DOD-FEHBP Demonstration, 2001                   43
                       Table 12: Enrollment in the DOD-FEHBP Demonstration, 2002                   44
                       Table 13: Beneficiaries Who Recalled Receiving DOD-FEHBP
                                 Demonstration Mailings and Who Found Them Useful                  46
                       Table 14: Beneficiaries’ Sources of Information about the DOD-
                                 FEHBP Demonstration                                               48
                       Table 15: Sources of Information for Eligible Beneficiaries about
                                 Specific FEHBP Plans                                              49
                       Table 16: Survey Responses by Enrollees to the Question “Why Did
                                 You Join a DOD-FEHBP Health Plan?”                                50
                       Table 17: Survey Responses by Nonenrollees to the Question “Why
                                 Didn’t You Join a DOD-FEHBP Health Plan?”                         51


                       Page ii                             GAO-03-547 Military Retiree Health Benefits
Figures
          Figure 1: DOD-FEHBP Demonstration-wide Enrollment, 2000-2002                              14
          Figure 2: DOD-FEHBP Demonstration Enrollment on the Mainland
                    and in Puerto Rico, 2000-2002                                                   15
          Figure 3: DOD-FEHBP Demonstration Enrollment by Type of
                    Previous Health Coverage, 2000                                                  19
          Figure 4: Comparison of Premiums for the DOD-FEHBP
                    Demonstration with Civilian FEHBP Premiums, 2000                                23




          Abbreviations

          CMS               Centers for Medicare & Medicaid Services
          DOD               Department of Defense
          FAQ               frequently asked questions
          FEHBP             Federal Employees Health Benefits Program
          HMO               health maintenance organization
          MTF               military treatment facility
          NMOP              National Mail Order Pharmacy
          OBRA 1990         Omnibus Budget Reconciliation Act of 1990
          OPM               Office of Personnel Management
          PIP-DCG           Principal Inpatient Diagnostic Cost Group
          POS               point-of-service
          PPO               preferred provider organization
          SNF               skilled nursing facility
          TFL               TRICARE For Life
          VA                Department of Veterans Affairs



          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. It may contain copyrighted graphics, images or other materials.
          Permission from the copyright holder may be necessary should you wish to reproduce
          copyrighted materials separately from GAO’s product.




          Page iii                                     GAO-03-547 Military Retiree Health Benefits
United States General Accounting Office
Washington, DC 20548




                                   June 6, 2003

                                   Congressional Committees

                                   Prior to 2001, military retirees who turned age 65 and became eligible for
                                   Medicare lost most of their Department of Defense (DOD) health care
                                   benefits. DOD did not offer its military retirees1 Medicare supplemental
                                   coverage, which some private employers make available to their retirees.
                                   Such coverage pays for Medicare deductibles and copayments as well as
                                   certain items not covered by Medicare, including most outpatient
                                   prescription drugs. Military retirees age 65 and over could obtain free care
                                   from the more than 600 military treatment facilities (MTF), but only if
                                   space was available after beneficiaries under age 65 had been treated.
                                   Older retirees could also get prescription drugs at no charge from MTF
                                   pharmacies if the drugs were stocked by the MTFs, although only about 40
                                   percent of retirees age 65 and over lived close to an MTF.

                                   To gather information on alternative ways of addressing military retirees’
                                   lack of Medicare supplemental coverage, Congress established several
                                   demonstrations that allowed Medicare-eligible military retirees to enroll in
                                   DOD-sponsored health care programs.2 One of those demonstrations was
                                   the DOD Federal Employees Health Benefits Program (FEHBP)
                                   demonstration (“the demonstration”),3 which lasted from 2000 through
                                   2002. Under the demonstration, military retirees and several smaller
                                   groups of beneficiaries4—such as certain former spouses of active duty


                                   1
                                    Our use of the term “military retirees” includes their dependents and survivors age 65 and
                                   over.
                                   2
                                    The Medicare subvention demonstration allowed retirees to enroll in new DOD-run
                                   Medicare managed care plans at six sites. See U.S. General Accounting Office, Medicare
                                   Subvention Demonstration: Pilot Satisfies Enrollees, Raises Cost and Management
                                   Issues for DOD Health Care, GAO-02-284 (Washington, D.C.: Feb. 11, 2002). Another
                                   demonstration, called TRICARE Senior Supplement, used TRICARE—the DOD health care
                                   program covering military personnel, younger retirees, and their dependents—to
                                   supplement retirees’ Medicare coverage.
                                   3
                                    The demonstration was created by the Strom Thurmond National Defense Authorization
                                   Act for Fiscal Year 1999, (NDAA 1999) Pub. L. No. 105-261, § 721, 112 Stat. 1920, 2061
                                   (1998) (codified at 10 U.S.C. § 1108) (2000).
                                   4
                                    In this report, the term “beneficiaries” refers to all those eligible for the demonstration:
                                   retirees, their spouses and other dependents, and other beneficiaries designated by law. It
                                   includes some persons under age 65.



                                   Page 1                                         GAO-03-547 Military Retiree Health Benefits
military personnel and retirees—could purchase coverage from one of the
private health plans that participate in FEHBP, the federal government’s
health insurance program for civilian employees and retirees. DOD
subsidized this retiree health coverage, paying up to three-quarters of the
premium. Enrollees could no longer use MTFs or military pharmacies. The
demonstration was open to about 120,000 of the more than 1.5 million
military retirees and dependents age 65 and over.5 It initially included
retirees and other eligible beneficiaries in eight geographic areas and
expanded in 2001 to include two additional areas.

The law establishing the demonstration (the Strom Thurmond National
Defense Authorization Act for Fiscal Year 1999 (NDAA 1999)) directed us
to examine a number of topics relating to enrollment and the
demonstration’s effects on beneficiaries and DOD.6 Specifically, this report
addresses (1) enrollment in the demonstration and the factors that
influenced whether military retirees enrolled, (2) the premiums set by
FEHBP plans for the demonstration and their strategies for setting
premiums, and (3) any effects that the demonstration project had on DOD
and beneficiaries—enrollees and nonenrollees.

To address these topics, we, in cooperation with DOD and the Office of
Personnel Management (OPM), which administers FEHBP, surveyed
between May and August 2000 a representative sample of about 5,600
persons eligible for the demonstration, of whom 85 percent responded. To
analyze factors affecting enrollment, we obtained survey information from
both enrollees and nonenrollees on health status, insurance coverage, and
other factors potentially affecting their enrollment decisions. We also
obtained information from DOD on persons eligible for the demonstration
and their use of military health care. We obtained information from
Quotesmith Inc. on premiums for private Medigap insurance plans that
supplement Medicare and are sold directly to individuals. To assess the
premiums offered by FEHBP plans, we obtained information from OPM on
premiums in the demonstration and in the civilian FEHBP. To obtain
information on whether demonstration enrollees were sicker than others,
we used Medicare claims on the diagnoses and costs of enrollees, eligible
nonenrollees, and civilian FEHBP enrollees age 65 and over who lived
near the demonstration sites. To examine the costs of demonstration


5
 In addition, the demonstration was open to approximately 17,000 eligible beneficiaries
under age 65.
6
10 U.S.C. § 1108(k) (2000).




Page 2                                       GAO-03-547 Military Retiree Health Benefits
                   enrollees, we obtained information from OPM and from Medicare claims.
                   We restricted some analyses to retirees age 65 and over for two reasons.
                   First, these retirees constituted 85 percent of all enrollees. Second, cost
                   and diagnostic information was available for these retirees but not for
                   beneficiaries under age 65. We also interviewed representatives of military
                   retiree associations as well as DOD and OPM officials. (See app. I for a
                   discussion of our survey methods and app. II for a discussion of our
                   methods of analyzing health status and factors affecting enrollment,
                   including tests of statistical significance.) We found that the size and
                   design of the demonstration were adequate for us to evaluate its effects
                   and answer the questions that Congress asked. We performed our work in
                   phases from November 1999 through May 2003. In 1999 and 2000, we
                   observed the initial planning and implementation of the demonstration,
                   and in 2000 we conducted the GAO-OPM-DOD survey. At the end of 2002
                   and in 2003, after the demonstration had ended, we conducted additional
                   analyses. We completed our work in accordance with generally accepted
                   government auditing standards.


                   Enrollment in the DOD-FEHBP demonstration peaked at 5.5 percent of
Results in Brief   potential beneficiaries in 2001 (7,521 enrollees) and then fell to 3.2 percent
                   in 2002, after the introduction of comprehensive health coverage—
                   TRICARE For Life (TFL) and the senior pharmacy benefit—for Medicare-
                   eligible military retirees.7 Enrollment was considerably greater in Puerto
                   Rico than on the mainland,8 reaching 30 percent in 2002. Most retirees who
                   knew about the demonstration and did not enroll said they were satisfied
                   with their current coverage—it had better benefits and lower costs than
                   the coverage they could obtain through the demonstration. Many
                   nonenrollees also cited not being able to continue getting prescriptions
                   filled at no charge at MTFs if they enrolled. Among the relatively small
                   proportion of people who did enroll, factors that encouraged their
                   enrollment included their view that the demonstration offered better
                   benefits, such as prescription drugs, than were available to them from
                   other plans, and their lack of existing coverage, such as employer-
                   sponsored insurance or a Medicare managed care plan. These factors also




                   7
                    The Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001 allowed
                   Medicare-eligible retirees to begin participating in TRICARE in 2001. Pub. L. No. 106-398, §
                   712, 114 Stat.1645A, 1554A-176 (2000).
                   8
                   The mainland refers to the 48 contiguous states.




                   Page 3                                        GAO-03-547 Military Retiree Health Benefits
help explain the high enrollment in Puerto Rico, where the share of
retirees without existing coverage was much greater than on the mainland.

Premiums charged enrollees in the demonstration varied widely—from
$65 to $208 monthly for an individual policy in 2000—with those plans that
had lower premiums and greater name recognition capturing the largest
number of enrollees. In setting premiums, plans had little information
about the health and probable cost of military beneficiaries. Plans adopted
two different strategies to reduce their financial burden if they attracted
sick, costly enrollees. One strategy kept premiums relatively low—at or
near premiums in the civilian FEHBP, with the intent of attracting a
representative mix of enrollees. The second strategy was to charge higher
premiums than in the civilian program, which tended to discourage
enrollment and provided a financial cushion in case those beneficiaries
who enrolled proved costly. However, plans following the two different
strategies attracted about the same proportion of enrollees who were in
poor health. In addition, demonstration enrollees were on average less
sick and younger than either their counterparts in the civilian program or
demonstration nonenrollees. During the first year of the demonstration,
enrolled retirees’ health care was considerably less expensive per person
than the health care for their counterparts in the civilian FEHBP—$3,529
(excluding prescription drugs) compared to $5,313. Premiums for
individual policies rose on average in 2001, but they fell in 2002, the first
time that a full year’s information on enrollees’ costs was available when
OPM and the plans negotiated premiums.

The demonstration’s impact on DOD’s budget, MTFs, and military
beneficiary access to military health care was small, although its impact on
beneficiaries who enrolled was considerable. The limited impact on DOD’s
budget and MTFs was due in part to the demonstration’s small number of
potential beneficiaries, relative to the more than 1.5 million military
retirees age 65 and over, and in part to the small proportion that actually
enrolled. For enrollees, the demonstration substantially expanded their
choice of health care options.

In commenting on a draft of this report, DOD said that it concurred with
our overall findings but disagreed with our description of the
demonstration’s impact on DOD’s budget as small. DOD’s costs for the
demonstration relative to its total health care budget were less than 0.1
percent of that budget. DOD provided technical comments that we
incorporated as appropriate. OPM declined to comment.




Page 4                                GAO-03-547 Military Retiree Health Benefits
             Medicare is generally the primary source of health insurance for people
Background   age 65 and over. However, traditional Medicare leaves beneficiaries liable
             for considerable out-of-pocket costs, and most beneficiaries have
             supplemental coverage. Military retirees can also obtain some care from
             MTFs and, since October 1, 2001, DOD has provided comprehensive
             supplemental coverage to its retirees age 65 and over. Civilian federal
             retirees and dependents age 65 and over can obtain supplemental
             coverage from FEHBP. The demonstration tested extending this coverage
             to military retirees age 65 and over, and their dependents.


Medicare     Medicare, a federally financed health insurance program for persons age
             65 and older, some people with disabilities, and people with end-stage
             kidney disease, is typically the primary source of health insurance for
             persons age 65 and over. Eligible Medicare beneficiaries are automatically
             covered by part A, which includes inpatient hospital and hospice care,
             most skilled nursing facility (SNF) care, and some home health care.9 They
             can also pay a monthly premium ($54 in 2002) to join part B, which covers
             physician and outpatient services as well as those home health services
             not covered under part A. Outpatient prescription drugs are generally not
             covered.10 Under traditional fee-for-service Medicare, beneficiaries choose
             their own providers and Medicare reimburses those providers on a fee-for-
             service basis. Beneficiaries who receive care through traditional Medicare
             are responsible for paying a share of the costs for most services.

             The alternative to traditional Medicare, Medicare+Choice, offers
             beneficiaries the option of enrolling in private managed care plans and
             other private health plans. In 1999, before the demonstration started,
             about 16 percent of all Medicare beneficiaries were enrolled in a
             Medicare+Choice plan; by 2002, the final year of the demonstration,
             enrollment had fallen to 12 percent. Medicare+Choice plans cover all basic


             9
              U.S. citizens and permanent residents are generally eligible for Medicare part A without
             having to pay a premium if they or their spouse worked for at least 10 years in Medicare-
             covered employment. Certain other persons with disabilities or end-stage kidney disease
             are also covered. Work by members of the armed services has been considered Medicare-
             covered employment since 1966, when Medicare was established.
             10
              Medicare generally covers outpatient prescription drugs only if they cannot be self-
             administered and are related to a physician’s services, such as cancer chemotherapy, or are
             provided in conjunction with covered durable medical equipment, such as inhalation drugs
             used with a nebulizer. In addition, Medicare covers selected immunizations and certain
             drugs that can be self-administered, such as blood clotting factors and some oral drugs
             used in association with cancer treatment and immunosuppressive therapy.




             Page 5                                       GAO-03-547 Military Retiree Health Benefits
                        Medicare benefits, and many also offer additional benefits such as
                        prescription drugs, although most plans place a limit on the amount of
                        drug costs they cover. These plans typically do not pay if their members
                        use providers who are not in their plans, and plan members may have to
                        obtain approval from their primary care doctors before they see
                        specialists. Members of Medicare+Choice plans generally pay less out of
                        pocket than they would under traditional Medicare.11


Medicare Supplemental   Medicare’s traditional fee-for-service benefit package and cost-sharing
Coverage                requirements leave beneficiaries liable for significant out-of-pocket costs,
                        and most beneficiaries in traditional fee-for-service Medicare have
                        supplemental coverage. This coverage typically pays part of Medicare’s
                        deductibles, coinsurance, and copayments, and may also provide benefits
                        that Medicare does not cover—notably, outpatient prescription drugs.
                        Major sources of supplemental coverage include employer-sponsored
                        insurance, the standard Medigap policies sold by private insurers to
                        individuals, and Medicaid.

                        Employer-sponsored insurance. About one-third of Medicare’s
                        beneficiaries have employer-sponsored supplemental coverage. These
                        plans, which typically have cost-sharing requirements, pay for some costs
                        not covered by Medicare, including part of the cost of prescription drugs.12

                        Medigap. About one-quarter of Medicare’s beneficiaries have Medigap, the
                        only supplemental coverage option available to all beneficiaries when they
                        initially enroll in Medicare. Prior to 1992, insurers were free to establish
                        the benefits for Medigap policies. The Omnibus Budget Reconciliation Act
                        of 1990 (OBRA 1990) required that beginning in 1992, Medigap policies be
                        standardized, and OBRA authorized 10 different benefit packages, known
                        as plans A through J, that insurers could offer.13 The most popular Medigap
                        policy is plan F, which covers Medicare coinsurance and deductibles, but
                        not prescription drugs. It had an average annual premium per person of



                        11
                         See U.S. General Accounting Office, Medicare+Choice: Selected Program Requirements
                        and Other Entities’ Standards for HMOs, GAO-03-180 (Washington, D.C.: Oct. 31, 2002).
                        12
                         Employer-sponsored health benefits have declined over the last decade and continue to
                        erode. See U.S. General Accounting Office, Retiree Health Insurance: Gaps in Coverage
                        and Availability, GAO-02-178T (Washington, D.C.: Nov. 1, 2001).
                        13
                         The Balanced Budget Act of 1997 permitted insurers to offer high deductible versions of
                        existing F and J plans. Pub. L. No. 105-33, § 4032. 111 Stat.251, 359 (1997).




                        Page 6                                      GAO-03-547 Military Retiree Health Benefits
                               about $1,200 in 1999, although in some cases plan F cost twice that
                               amount. Among the least popular Medigap policies are those offering
                               prescription drug coverage. These policies are the most expensive of the
                               10 standard policies—they averaged about $1,600 in 1999, and some cost
                               over $5,000. Beneficiaries with these policies pay most of the cost of drugs
                               because the Medigap drug benefit has a deductible and high cost sharing
                               and does not reimburse policyholders for drug expenses above a set
                               limit.14


Health Care for Military       DOD provides health care to active-duty military personnel and retirees,
Retirees                       and to eligible dependents and survivors through its TRICARE program.15
                               Prior to 2001, retirees lost most of their military health coverage when
                               they turned age 65, although they could still use MTFs when space was
                               available, and they could obtain prescription drugs without charge from
                               MTF pharmacies.16 In the Floyd D. Spence National Defense Authorization
                               Act for Fiscal Year 2001 (NDAA 2001), Congress established two new
                               benefits to supplement military retirees’ Medicare coverage:

                           •   Pharmacy benefit. Effective April 1, 2001, military retirees age 65 and
                               over were given access to prescription drugs through TRICARE’s National
                               Mail Order Pharmacy (NMOP) and civilian pharmacies. Retirees make
                               lower copayments for prescription drugs purchased through NMOP than
                               at civilian pharmacies. Retirees continue to have access to free
                               prescription drugs at MTF pharmacies.
                           •   TFL. Effective October 1, 2001, military retirees age 65 and over who were
                               enrolled in Medicare part B became eligible for TFL. As a result, DOD is
                               now a secondary payer for these retirees’ Medicare-covered services,
                               paying all of their required cost sharing. TFL also offers certain benefits
                               not covered by Medicare, including catastrophic coverage. Retirees can
                               continue to use MTFs without charge on a “space available” basis.



                               14
                                See U.S. General Accounting Office, Medigap: Current Policies Contain Coverage Gaps,
                               Undermine Cost Control Incentives, GAO-02-533T (Washington, D.C.: Mar. 14, 2002) and
                               Medigap Insurance: Plans Are Widely Available but Have Limited Benefits and May
                               Have High Costs, GAO-01-941 (Washington, D.C.: July 31, 2001).
                               15
                                DOD also provides health care to retired reserve service members and their families as
                               well as Medal of Honor recipients and their families.
                               16
                                 Retirees could obtain prescription drugs from an MTF only if the drugs were stocked by
                               the MTF. In addition, over 400,000 beneficiaries age 65 and over were eligible for the mail
                               order and retail pharmacy benefit as a result of the Base Realignment and Closure (BRAC)
                               actions.




                               Page 7                                       GAO-03-547 Military Retiree Health Benefits
        In fiscal year 1999, before TFL was established, DOD’s annual
        appropriations for health care were about $16 billion, of which over
        $1 billion funded the care of military retirees age 65 and over. In fiscal year
        2002, DOD’s annual health care appropriations totaled about $24 billion, of
        which over $5 billion funded the care of retirees age 65 and over who used
        TFL, the pharmacy benefit, and MTF care.

        In addition to their DOD coverage, military retirees—but generally not
        their dependents—can use Department of Veterans Affairs (VA) facilities.
        There are 163 VA medical centers throughout the country that provide
        inpatient and outpatient care as well as over 850 outpatient clinics. VA
        care is free to veterans with certain service-connected disabilities or low
        incomes;17 other veterans are eligible for care but have lower priority than
        those with service-connected disabilities or low incomes and are required
        to make copayments.


FEHBP   FEHBP, the health insurance program administered by OPM for federal
        civilian employees and retirees, covered about 8.3 million people in 2002.
        Civilian employees become eligible for FEHBP when hired by the federal
        government. Employees and retirees can purchase health insurance from a
        variety of private plans, including both managed care and fee-for-service
        plans, that offer a broad range of benefits, including prescription drugs.
        Insurers offer both self-only plans and family plans, which also cover the
        policyholders’ dependents. Some plans also offer two levels of benefits: a
        standard option and a high option, which has more benefits, less cost
        sharing, or both.18 For retirees age 65 and over, FEHBP supplements
        Medicare, paying beneficiaries’ Medicare deductibles and coinsurance in
        addition to paying some costs not covered by Medicare, such as part of the
        cost of prescription drugs.19




        17
          Veterans with a service-connected disability rating of 50 percent or more qualify for free
        health care in VA facilities. Their treatment may be for conditions unrelated to military
        service. The disability rating is based on an evaluation that represents the average loss in
        earning capacity associated with the severity of physical and mental conditions.
        Individuals’ ratings range from 0 percent to 100 percent.
        18
         Some plans refer to the two options as the basic option and the standard option.
        19
         See U.S. General Accounting Office, Federal Employees’ Health Plans: Premium Growth
        and OPM’s Role in Negotiating Benefits, GAO-03-236 (Washington, D.C.: Dec. 31, 2002).




        Page 8                                        GAO-03-547 Military Retiree Health Benefits
Over two-thirds of FEHBP policyholders are in national plans; the
remainder are in local plans. National plans include plans that are
available to all civilian employees and retirees as well as plans that are
available only to particular groups, for example, foreign service
employees. In the FEHBP, the largest national plan is Blue Cross Blue
Shield, accounting for about 45 percent of those insured by an FEHBP
plan.20 Other national plans account for about 24 percent of insured
individuals. The national plans are all preferred provider organizations
(PPO) in which enrollees use doctors, hospitals, and other providers that
belong to the plan’s network, but are allowed to use providers outside of
the network for an additional cost. Local plans, which operate in selected
geographic areas and are mostly managed care, cover the remaining 32
percent of people insured by the FEHBP.

Civilian employees who enroll in FEHBP can change plans during an
annual enrollment period. During this period, which runs from mid-
November to mid-December, beneficiaries eligible for FEHBP can select
new plans for the forthcoming calendar year. To assist these beneficiaries
in selecting plans, OPM provides general information on FEHBP through
brochures and its Web site. Also, as part of this information campaign,
plans’ representatives may visit government agencies to participate in
health fairs, where they provide detailed information about their specific
health plans to government employees.

The premiums charged by these plans, which are negotiated annually
between OPM and the plans, depend on the benefits offered by the plan,
the type of plan—fee-for-service or managed care—and the plan’s out-of-
pocket costs for the enrollee. Plans may propose changes to benefits as
well as changes in out-of-pocket payments by enrollees. OPM and the
plans negotiate these changes and take them into account when
negotiating premiums. Fee-for-service plans must base their rates on the
claims experience of their FEHBP enrollees, while adjusting for changes in
benefits and out-of-pocket payments, and must provide OPM with data to
justify their proposed rates. Managed care plans must give FEHBP the best
rate that they offer to groups of similar size in the private sector under
similar conditions, with adjustments to account for differences in the
demographic characteristics of FEHBP enrollees and the benefits



20
  Blue Cross Blue Shield is a consortium of local Blue Cross Blue Shield plans across the
country. It charges the same premium in all locations and distributes that premium to its
local plans, without any adjustment for local variations in health care costs.




Page 9                                       GAO-03-547 Military Retiree Health Benefits
                provided.21 The government pays a maximum of 72 percent of the weighted
                average premium of all plans and no more than 75 percent of any plan’s
                premium. Unlike most other plans, including employer-sponsored
                insurance and Medigap, FEHBP plans charge the same premium to all
                enrollees, regardless of age. As a result, persons over age 65, for whom the
                FEHBP plan supplements Medicare, pay the same rate as those under age
                65, for whom the FEHBP plan is the primary insurer.


The FEHBP       The FEHBP demonstration allowed eligible beneficiaries in the
Demonstration   demonstration sites to enroll in an FEHBP plan. The demonstration ran for
                3 years, from January 1, 2000, through December 31, 2002. The law that
                established the demonstration capped enrollment at 66,000 beneficiaries
                and specified that DOD and OPM should jointly select from 6 to 10 sites.
                Initially, the agencies selected 8 sites that had about 69,000 eligible
                beneficiaries according to DOD’s calculation for 2000.22 (See table 1.) Four
                sites had MTFs, and 1 site—Dover—also participated in the subvention
                demonstration.23 Two other sites, which had about 57,000 eligible
                beneficiaries, were added in 2001. Demonstration enrollees received the
                same benefits as civilian FEHBP enrollees, but could no longer use MTFs
                or MTF pharmacies.




                21
                 These private sector groups are referred to as similarly sized subscriber groups.
                22
                  More recent DOD data indicate that the number of eligible beneficiaries was
                approximately 80,000 in the 8 original sites. (See app. III.) This substantial increase in
                eligible beneficiaries, compared to the initial figure, resulted from corrections that DOD
                made to its eligibility and enrollment database. We used the lower figure in implementing
                the sampling strategy for our survey because it was the only information available at the
                time of the survey. To maintain consistency, all analyses for 2000 use the original (lower)
                DOD figure.
                23
                 The law establishing the FEHBP demonstration required that at least one site contain an
                MTF, one site not contain an MTF, one site be a participant in the DOD Medicare
                subvention demonstration, and no TRICARE region have more than one FEHBP
                demonstration site. 10 U.S.C. § 1108(C) (2000).




                Page 10                                       GAO-03-547 Military Retiree Health Benefits
    Table 1: Number of Eligible Beneficiaries by DOD-FEHBP Demonstration Site,
    2000-2002

        Site                                                     2000             2001             2002
        With MTF:
           Camp Pendleton, Calif.                             24,907            27,328           27,287
                       a
           Dover, Del.                                         4,384             4,868            4,867
           Fort Knox, Ky.                                      7,757             9,121            9,113
           Puerto Rico                                         6,907             9,401            9,453
        No MTF:
           Dallas, Tex.                                       13,607            16,159           16,133
           Greensboro, N.C.                                    3,278             4,033            4,024
           Humboldt County, Calif.                             2,919             3,461            3,454
           New Orleans, La.                                    5,083             6,095            6,085
           Adair County, Iowa                                                   29,584           29,530
           Coffee County, Ga.                                                   27,329           27,284
                              b
        Total—initial 8 sites                                 68,842
        Total—10 sites                                                        137,379          137,230

    Source: DOD.

    Note: The 2000 data are as of January 1, 2000, 2001 data are as of March 14, 2001, and 2002 data
    are as of February 21, 2002.
    a
    Dover also participated in the DOD Medicare subvention demonstration.
    b
     DOD initially calculated that there were 68,842 beneficiaries in the original 8 sites. Based on this
    figure, the demonstration including the two new sites had approximately 126,000 eligible
    beneficiaries. The higher numbers in 2001 and 2002 resulted from corrections that DOD made to its
    eligibility and enrollment database.


    Military retirees age 65 and over and their dependents age 65 and over
    were permitted to enroll in either self-only or family FEHBP plans.
    Dependents who were under age 65 could be covered only if the eligible
    retiree chose a family plan. Several other groups were permitted to enroll
    including:

•   unremarried former spouses of a member or former member of the armed
    forces entitled to military retiree health care,
•   dependents of a deceased member or former member of the armed forces
    entitled to military retiree health care, and
•   dependents of a member of the armed services who died while on active
    duty for more than 30 days.




    Page 11                                            GAO-03-547 Military Retiree Health Benefits
About 13 percent of those eligible for the demonstration were under age
65.24

DOD, with assistance from OPM, was responsible for providing eligible
beneficiaries information on the demonstration. A description of this
information campaign is in appendix IV.

The demonstration guaranteed enrollees who dropped their Medigap
policies the right to resume their coverage under 4 of the 10 standard
Medigap policies—plans A, B, C, and F—at the end of the demonstration.
However, demonstration enrollees who held any other standard Medigap
policies, or Medigap policies obtained before the standard plans were
established, were not given the right to regain the policies. Enrollees who
dropped their employer-sponsored retiree health coverage had no
guarantee that they could regain it.

Each plan was required by OPM to offer the same package of benefits to
demonstration enrollees that it offered in the civilian FEHBP, and plans
operating in the demonstration sites were generally required to participate
in the demonstration. Fee-for-service plans that limit enrollment to
specific groups, such as foreign service employees, did not participate. In
addition, health maintenance organizations (HMO) and point-of-service
(POS) plans were not required to participate if their civilian FEHBP
enrollment was less than 300 or their service area overlapped only a small
part of the demonstration site.25 Thirty-one local plans participated in the
demonstration in 2000; for another 14 local plans participation was
optional, and none of these participated.

The law established a separate risk pool for the demonstration, so any
losses from the demonstration were not covered at the expense of persons
insured under the civilian FEHBP.26 As a result, plans had to establish
separate reserves for the demonstration and were allowed to charge



24
 Persons eligible for the civilian FEHBP were not eligible for the demonstration.
25
 HMOs are comprehensive medical plans that coordinate health care through a network of
physicians and hospitals. A POS option provides enrollees with a choice of using the plan’s
health care providers or paying higher fees to see providers outside of the plan’s network.
26
  A risk pool is the group of people with respect to whom the premium is set. In the
FEHBP, premiums depend upon the expected claims or costs of those enrolled. The
FEHBP demonstration required that expected costs for the demonstration enrollees and
for civilian FEHBP enrollees be calculated separately. 10 U.S.C. § 1108(h) (2000).




Page 12                                      GAO-03-547 Military Retiree Health Benefits
                             different premiums in the demonstration than they charged in the civilian
                             program.


                             Enrollment in the demonstration was low, although enrollment in Puerto
Enrollment Was Low,          Rico was substantially higher than on the U.S. mainland. Among eligible
Largely Due to               beneficiaries who knew about the demonstration yet chose not to enroll,
                             most were satisfied with their existing health care coverage and preferred
Beneficiaries’               it to the demonstration’s benefits. Lack of knowledge about the
Satisfaction with            demonstration accounted for only a small part of the low enrollment.
                             Although most eligible retirees did not enroll in a demonstration plan,
Existing Coverage            several factors encouraged enrollment. Some retirees took the view that
                             the demonstration plans’ benefits, notably prescription drug coverage,
                             were better than available alternatives. Other retirees mentioned lack of
                             satisfactory alternative coverage. In particular, retirees who were not
                             covered by an existing Medicare+Choice or employer-sponsored health
                             plan were much more likely to enroll. The higher enrollment in Puerto
                             Rico reflected a higher proportion of retirees there who considered the
                             demonstration’s benefits—ranging from drug coverage to choice of
                             doctors—better than what they had. The higher enrollment in Puerto Rico
                             also reflected in part Puerto Rico’s greater share of retirees without
                             existing coverage, such as an employer-sponsored plan.


Enrollment Rate Low on       While some military retiree organizations as well as a large FEHBP plan
U.S. Mainland, Far Greater   predicted at the start of the demonstration that enrollment would reach 25
in Puerto Rico               percent or more of eligible beneficiaries, demonstration-wide enrollment
                             was 3.6 percent in 2000 and 5.5 percent in 2001.27 In 2002, following the
                             introduction of the senior pharmacy benefit and TFL the previous year,
                             demonstration-wide enrollment fell to 3.2 percent. (See fig. 1.) The
                             demonstration’s enrollment peaked at 7,521 beneficiaries, and by 2002 had
                             declined to 4,367 of the 137,230 eligible beneficiaries.28




                             27
                               Enrollment as a percentage of eligible beneficiaries in 2000 is based on DOD’s initial
                             figure of 68,842 eligible beneficiaries.
                             28
                              Enrollment for 2000 was as of January 1, 2000, enrollment for 2001 was as of March 14,
                             2001, and enrollment for 2002 was as of February 21, 2002.




                             Page 13                                       GAO-03-547 Military Retiree Health Benefits
Figure 1: DOD-FEHBP Demonstration-wide Enrollment, 2000-2002

 12 Percentage



 10



     8



     6               5.5


     4   3.6
                                    3.2


     2



     0
         2000        2001          2002
Source: DOD.

Note: GAO analysis of DOD data. Enrollment is expressed as a percentage of eligible beneficiaries.


These low demonstration-wide enrollment rates masked a sizeable
difference in enrollment between the mainland sites and Puerto Rico. In
2000, enrollment in Puerto Rico was 13.2 percent of eligible
beneficiaries—about five times the rate on the mainland. By 2001, Puerto
Rico’s enrollment had climbed to 28.6 percent. Unlike 2002 enrollment on
the mainland, which declined, enrollment in Puerto Rico that year rose
slightly, to 30 percent. (See fig. 2.) Among the mainland sites, there were
also sizeable differences in enrollment, ranging from 1.3 percent in Dover,
Delaware, in 2001, to 8.8 percent in Humboldt County, California, that
year. Enrollment at all mainland sites declined in 2002.29




29
 See app. III for enrollment by site.




Page 14                                          GAO-03-547 Military Retiree Health Benefits
                             Figure 2: DOD-FEHBP Demonstration Enrollment on the Mainland and in Puerto
                             Rico, 2000-2002

                              35 Percentage


                                                                                        30.1
                              30                                   28.6


                              25


                              20


                              15
                                      13.2

                              10


                                  5                                              3.8
                                                    2.6
                                                                                                      1.2
                                  0
                                             2000                         2001                 2002

                                             Puerto Rico

                                             U.S. mainland sites

                             Source: DOD.

                             Note: GAO analysis of DOD data. Enrollment is expressed as a percentage of eligible beneficiaries.




Nonenrollees Emphasized      Retirees who knew about the demonstration and did not enroll cited many
Better Benefits and Lower    reasons for their decision, notably that their existing coverage’s benefits—
Costs of Existing Coverage   in particular its prescription drug benefit—and costs were more attractive
                             than those of the demonstration.30 In addition, nonenrollees expressed
                             several concerns, including uncertainty about whether they could regain
                             their Medicare supplemental coverage after the demonstration ended.

                         •   Benefits of existing coverage. Almost two-thirds of nonenrollees who
                             knew about the demonstration reported that they were satisfied with their
                             existing employer-sponsored or other health coverage.31 For the majority


                             30
                              Only nonenrollees who knew about the demonstration (44 percent of eligible
                             beneficiaries) were asked to give their reasons for not enrolling. Because respondents to
                             our survey gave multiple reasons for not enrolling, percentages reported concerning
                             benefits, prescription drugs, and other reasons add to more than 100 percent.
                             31
                              Satisfaction with existing coverage was a much less important reason for not enrolling in
                             Puerto Rico than on the mainland. In Puerto Rico, 28 percent of nonenrollees were
                             satisfied with their existing coverage, compared to 66 percent of nonenrollees on the
                             mainland.



                             Page 15                                                   GAO-03-547 Military Retiree Health Benefits
                                 of nonenrollees with private employer-sponsored coverage, the
                                 demonstration’s benefits were no better than those offered by their
                                 current plan.
                             •   Costs of existing coverage. Nearly 30 percent of nonenrollees who
                                 knew about the demonstration stated that its plans were too costly.32 This
                                 was likely a significant concern for retirees interested in a managed care
                                 plan, such as a Medicare+Choice plan, whose premiums were generally
                                 lower than demonstration plans.
                             •   Prescription drugs and availability of doctors. In explaining their
                                 decision not to enroll, many eligible beneficiaries who knew about the
                                 demonstration focused on limitations of specific features of the benefits
                                 package that they said were less attractive than similar features of their
                                 existing coverage. More than one-quarter of nonenrollees cited not being
                                 able to continue getting prescriptions filled without charge at MTF
                                 pharmacies if they enrolled. More than one-quarter also said their decision
                                 at least partly reflected not being able to keep their current doctors if they
                                 enrolled. These nonenrollees may have been considering joining one of the
                                 demonstration’s managed care plans, which generally limit the number of
                                 doctors included in their provider networks. Otherwise, they would have
                                 been able to keep their doctors, because PPOs, while encouraging the use
                                 of network doctors, permit individuals to select their own doctors at an
                                 additional cost.
                             •   Uncertainty. About one-fourth of nonenrollees said they were uncertain
                                 about the viability of the demonstration and wanted to wait to see how it
                                 worked out. In addition, more than 20 percent of nonenrollees were
                                 concerned that the demonstration was temporary and would end in 3
                                 years. Furthermore, some nonenrollees who looked beyond the
                                 demonstration period expressed uncertainty about what their coverage
                                 would be after the demonstration ended: Roughly one-quarter expressed
                                 concern that joining a demonstration plan meant risking the future loss of
                                 other coverage—either Medigap or employer-sponsored insurance.
                                 Finally, about one-quarter of nonenrollees were uncertain about how the
                                 demonstration would mesh with Medicare.


Lack of Knowledge about          Lack of knowledge—although common among eligible retirees—was only
Demonstration Accounted          a small factor in explaining low enrollment. If everyone eligible for the
for Only Small Part of Low       demonstration had known about it, enrollment might have doubled, but
                                 would still have been low. DOD undertook an extensive information
Enrollment                       campaign, intended to inform all eligible beneficiaries about the


                                 32
                                  See app. V for a complete list of reasons given.




                                 Page 16                                       GAO-03-547 Military Retiree Health Benefits
                              demonstration, but nearly 54 percent of those eligible for the
                              demonstration did not know about it at the time of our survey (May
                              through August 2000). Of those who knew about the demonstration, only
                              7.4 percent enrolled. Those who did not know about the demonstration
                              were different in several respects from those who did: They were more
                              likely to be single, female, African American, older than age 75, to have
                              annual income of $40,000 or less, to live an hour or more from an MTF, not
                              covered by employer-sponsored health insurance, not officers, not to
                              belong to military retiree organizations and to live in the demonstration
                              areas of Camp Pendleton, California, Dallas, Texas, and Fort Knox,
                              Kentucky.

                              Accounting for the different characteristics of those retirees who knew
                              about the demonstration and those who did not, we found that roughly 7
                              percent of those who did not know about the demonstration would have
                              enrolled in 2000 if they had known about it. As a result, we estimate that
                              demonstration-wide enrollment would have been about 7 percent if all
                              eligible retirees knew about the demonstration. (See app. II.)

                              Comparison of enrollment in Puerto Rico and the mainland sites also
                              suggests that, among the factors that led to low enrollment, knowledge
                              about the demonstration was not decisive. In 2000, fewer people in Puerto
                              Rico reported knowing about the demonstration than on the mainland (35
                              percent versus 47 percent). Nonetheless, enrollment in Puerto Rico was
                              much higher.


Factors Spurring              In making the decision to enroll, retirees were attracted to an FEHBP plan
Enrollment Included           if it had better benefits—particularly prescription drug coverage—or
Favorable Assessment of       lower costs than their current coverage or other available coverage.
                              Among those who knew about the demonstration, retirees who enrolled
FEHBP and Lack of             were typically positive about one or both of the following:
Existing Coverage
                          •   Better FEHBP benefits. Two-thirds of enrollees cited their
                              demonstration plan’s benefits package as a reason to enroll, with just over
                              half saying the benefits package was better than other coverage available
                              to them. Nearly two-thirds of enrollees mentioned the better coverage of
                              prescription drugs offered by their demonstration plan. Furthermore, the
                              inclusiveness of FEHBP plans’ networks of providers mattered to a
                              majority of enrollees: More than three-fifths mentioned as a reason for
                              enrolling that they could keep their current doctors under the
                              demonstration.




                              Page 17                               GAO-03-547 Military Retiree Health Benefits
•   Lower demonstration plan costs. Among enrollees, about 62 percent
    said that their demonstration FEHBP plan was less costly than other
    coverage they could buy.

    Beneficiaries’ favorable assessments of FEHBP—and their enrollment in
    the demonstration—were related to whether they lacked alternative
    coverage to traditional Medicare and, if they had such coverage, to the
    type of coverage. In 2000, among those who lacked employer-sponsored
    coverage or a Medicare+Choice plan, or lived more than an hour’s travel
    time from an MTF, about 15 percent enrolled. By contrast, among those
    who had such coverage, or had MTF access, 4 percent enrolled.

    In particular, enrollment in an FEHBP plan was more likely for retirees
    who lacked either Medicare+Choice or employer-sponsored coverage.

•   Lack of Medicare+Choice. Controlling for other factors affecting
    enrollment, those who did not use Medicare+Choice were much more
    likely to enroll in a demonstration plan than those who did. (See fig. 3.)
    Several reasons may account for this. First, in contrast to fee-for-service
    Medicare, Medicare+Choice plans are often less costly out-of-pocket,
    typically requiring no deductibles and lower cost sharing for physician
    visits and other outpatient services. Second, unlike fee-for-service
    Medicare, many Medicare+Choice plans offered a prescription drug
    benefit. Third, while Medicare+Choice plan benefits were similar to those
    offered by demonstration FEHBP plans, Medicare+Choice premiums were
    typically less than those charged by the more popular demonstration
    plans, including Blue Cross Blue Shield, the most popular demonstration
    plan on the mainland.
•   Lack of employer-sponsored coverage. Retirees who did not have
    employer-sponsored health coverage were also more likely to join a
    demonstration plan. Of those who did not have employer-sponsored
    coverage, 8.6 percent enrolled in the demonstration, compared to 4.7
    percent of those who had such coverage. Since benefits in employer-
    sponsored health plans often resemble FEHBP benefits, retirees with
    employer-sponsored coverage would have been less likely to find FEHBP
    plans attractive.33




    33
      Like retirees’ employer-sponsored coverage, those with Medicare+Choice coverage were
    significantly less likely to enroll, while retirees covered by Medicare part B were
    significantly more likely to enroll. (See app. IV.) Part B coverage of enrollees and
    nonenrollees differed slightly: 94.7 percent for enrollees and 92.1 percent for nonenrollees.




    Page 18                                       GAO-03-547 Military Retiree Health Benefits
Retirees with another type of alternative coverage, Medigap, responded
differently to the demonstration. Unlike the pattern with other types of
insurance coverage, more of those with a Medigap plan enrolled (9.3
percent) than did those without Medigap (5.6 percent). Medigap plans
generally offered fewer benefits than a demonstration FEHBP plan, but at
the same or higher cost to the retiree. Seven of the 10 types of Medigap
plans available to those eligible for the demonstration do not cover
prescription drugs. As a result of these differences, retirees who were
covered by Medigap policies would have had an incentive to enroll instead
in a demonstration FEHBP plan, which offered drug coverage and other
benefits at a lower premium cost than the most popular Medigap plan.

Figure 3: DOD-FEHBP Demonstration Enrollment by Type of Previous Health
Coverage, 2000

 12 Percentage




                                                                                   9.3
  9       8.6                              8.6




  6                                                                    5.6
                                                       4.7



  3

                      1.8



  0
        Not       Enrolled            Did not have      Had       Did not have     Had
      enrolled       in                employer-     employer-     Medigap       Medigap
         in       Medicare+            sponsored     sponsored       policy       policy
      Medicare+    Choice              insurance     insurance
       Choice
Source: CMS and GAO-DOD-OPM survey.

Note: GAO analysis of CMS and GAO-DOD-OPM survey data. Enrollment is expressed, for
employer-sponsored coverage, as a percentage of eligible beneficiaries who knew about the
demonstration and, for Medicare+Choice enrollment and Medigap coverage, as a percentage of
eligible retirees who knew about the demonstration. An eligible beneficiary or retiree may have more
than one type of coverage.


Like the lack of Medicare+Choice or employer-sponsored coverage, lack
of nearby MTF care stimulated enrollment. While living more than an hour
from an MTF was associated with higher demonstration enrollment, MTF
care may have served some retirees as a satisfactory supplement to



Page 19                                              GAO-03-547 Military Retiree Health Benefits
                           Medicare-covered care, making demonstration FEHBP plans less
                           attractive to them. Of eligible retirees who knew of the demonstration and
                           lived within 1 hour of an MTF, 3.7 percent enrolled, compared to 11.1
                           percent of those who lived more than 1 hour away.


Higher Enrollment in       Higher enrollment in Puerto Rico than on the mainland reflected in part
Puerto Rico Associated     the more widespread lack of satisfactory alternative health coverage in
with Greater Lack of       Puerto Rico compared to the mainland. In Puerto Rico, of those who knew
                           of the demonstration, the share of eligible retirees with employer-
Satisfactory Alternative   sponsored health coverage (14 percent) was about half that on the
Coverage                   mainland (27 percent). In addition, before September 2001, no
                           Medicare+Choice plan was available in Puerto Rico. By contrast, in
                           mainland sites where Medicare+Choice plans were available, their
                           attractive cost sharing and other benefits discouraged retirees from
                           enrolling in demonstration plans. Other factors associated with Puerto
                           Rico’s high enrollment and cited by enrollees there included the
                           demonstration plan’s better benefits package—especially prescription
                           drug coverage—compared to many retirees’ alternatives, the
                           demonstration plan’s broader choice of doctors, and the plan’s reputation
                           for quality of care.34


                           The premiums charged by the demonstration plans varied widely,
Premiums Varied            reflecting differences in how they dealt with the concern that the
Widely, Reflecting         demonstration would attract a disproportionate number of sick, high-cost
                           enrollees. To address these concerns, plans generally followed one of two
Plans’ Different           strategies. Most plans charged higher premiums than those they charged
Assessments of             to their civilian FEHBP enrollees—a strategy that could have provided a
                           financial cushion and possibly discouraged enrollment. A small number of
Demonstration Risk         plans set premiums at or near their premiums for the civilian FEHBP with
                           the aim of attracting a mix of enrollees who would not be
                           disproportionately sick. Plans’ underlying concern that they would attract
                           a sicker population was not borne out. In the first year of the
                           demonstration, for example, on average health care for demonstration
                           retirees was 50 percent less expensive per enrollee than the care for their
                           civilian FEHBP counterparts.




                           34
                            There was only one local plan in the demonstration in Puerto Rico: Triple-S.




                           Page 20                                      GAO-03-547 Military Retiree Health Benefits
Plans’ Premiums Varied     Demonstration plans charged widely varying premiums to enrollees, with
Widely, and Plans with     the most popular plans offering some of the lowest premiums. In 2000,
Lower Premiums Attracted   national plans’ monthly premiums for individual coverage ranged from $65
                           for Blue Cross Blue Shield to $208 for the Alliance Health Plans. Among
the Most Enrollees         local plans—most of which were managed care—monthly premiums for
                           individual coverage ranged from $43 for NYLCare Health Plans of the
                           Southwest to $280 for Aetna U.S. Healthcare. Not surprisingly, few
                           enrollees selected the more expensive plans.35 The two most popular plans
                           were Blue Cross Blue Shield and Triple-S; the latter offered a POS in
                           Puerto Rico. Both plans had relatively low monthly premiums—the
                           Triple-S premium charged to individuals was $54 in the demonstration’s
                           first year. Average premiums for national plans were about $20 higher than
                           for local plans, which were largely managed care plans. (See table 2.)

                           Table 2: Monthly Premiums Charged to Enrollees for Individual Policies in the DOD-
                           FEHBP Demonstration, 2000

                            Type of plan              Plan or group of plans                Enrollee share of premium
                            National plans
                                                      Blue Cross Blue Shield                                              $65
                                                      GEHA Benefit Plan                                                    99
                                                      Other national plans—average                                        142
                            National plan average                                                                         125
                            Local plans
                                                      Triple-S                                                            54
                                                      Other fee-for-service plans—
                                                      average                                                           78
                                                      Managed care plans—average                                       107
                            Local plan average                                                                         103
                            Average of all plans                                                                      $107

                           Source: OPM.

                           Note: GAO analysis of OPM premium data. Premiums are for a standard option individual policy
                           unless only one option was available.


                           Some plans in the demonstration were well known in their market areas,
                           while others—especially those open only to government employees—
                           likely had much lower name recognition. Before the demonstration
                           started, OPM officials told us that they expected beneficiaries to be
                           unfamiliar with many of the plans included in the demonstration. These


                           35
                            For example, the 10 percent of plans with the highest premiums attracted 0.1 percent of
                           enrollees.




                           Page 21                                         GAO-03-547 Military Retiree Health Benefits
                            officials said that beneficiaries were likely to have only experience with or
                            knowledge of Blue Cross Blue Shield and, possibly, some local HMOs. The
                            success of Blue Cross Blue Shield relative to other national plans in
                            attracting enrollees appears to support their view, as does Triple-S’s
                            success in Puerto Rico, where it is one of the island’s largest insurers. In
                            2000, Blue Cross Blue Shield was the most popular plan in the
                            demonstration, with 42 percent of demonstration-wide enrollment and 68
                            percent of enrollment on the mainland. Among national plans, the GEHA
                            Benefit Plan (known as GEHA) was a distant second with 4 percent of
                            enrollment. The other five national plans together captured less than 1
                            percent of all demonstration enrollment. Among local plans, Triple-S was
                            most successful, capturing 96 percent of enrollment in Puerto Rico and 38
                            percent of enrollment demonstration-wide. The other local plans, taken
                            together, accounted for about 14 percent of demonstration-wide
                            enrollment.


Plans’ Premium Strategies   Several factors contributed to plans’ concern that they would attract
Diverged despite Common     sicker—and therefore more costly—enrollees in the demonstration. Plans
Concerns about Attracting   did not have the information that they usually use to set premiums—
                            claims history for fee-for-service plans and premiums charged to
Sicker Enrollees            comparable private sector groups for managed care plans. Moreover,
                            according to officials, some plans were reluctant to assume that
                            demonstration enrollees would be similar to their counterparts in the
                            civilian FEHBP. A representative from one of the large plans noted that
                            the small size of the demonstration was also a concern. The number of
                            people eligible for the demonstration (approaching 140,000, when the
                            demonstration was expanded in 2001) was quite small compared to the
                            number of people in the civilian program (8.5 million in 2001). If only a
                            small number of people enrolled in a plan, one costly case could result in
                            losses, because claims could exceed premiums.

                            In response to the concern that the demonstration might attract a
                            disproportionate number of sick enrollees, plans developed two different
                            strategies for setting premiums. Plans in one group, including Blue Cross
                            Blue Shield and GEHA, kept their demonstration premiums at or near
                            those they charged in the civilian FEHBP. Representatives of one plan
                            explained that it could have priced high, but they believed that would have
                            resulted in low enrollment and might have attracted a disproportionate
                            number of sick—and therefore costly—enrollees. Instead, by keeping their
                            premium at the same level as in the civilian program, these plan officials
                            hoped to make their plan attractive to those who were in good health as



                            Page 22                               GAO-03-547 Military Retiree Health Benefits
well as to those who were not. Such a balanced mix of enrollees would
increase the likelihood that a plan’s revenues would exceed its costs.

By contrast, some plans charged higher premiums in the demonstration—
in some cases, 100 percent higher—than in the civilian FEHBP. Setting
higher premiums might provide plans with a financial cushion to deal with
potential high-cost enrollees. While higher premiums might have
discouraged enrollment and reduced plans’ exposure to high-cost patients,
this strategy carried the risk that those beneficiaries willing to pay very
high premiums might be sick, high-cost patients.

More than four-fifths of plans chose the second strategy, charging higher
premiums in the demonstration than in the civilian FEHBP. In 2000, only
two plans—both local plans—charged enrollees less in the demonstration
than in the civilian program for individual, standard option policies; these
represented about 6 percent of all plans. By contrast, three plans—about 9
percent of all plans—set premiums at least twice as high as premiums in
the civilian FEHBP. (See fig. 4.)

Figure 4: Comparison of Premiums for the DOD-FEHBP Demonstration with Civilian
FEHBP Premiums, 2000


                         6.1                         Lower premium for demonstration
                 9.1                                 Demonstration premium at least
                                                     twice as high


       21.2
                                  33.3               Demonstration premium
                                                     0 to 19 percent higher



               18.2
                          12.1                       Demonstration premium
                                                     20 to 49 percent higher

                                                     Demonstration premium
                                                     50 to 99 percent higher

                                                     Demonstration premium
                                                     100 to 199 percent higher
Source: OPM.

Note: GAO analysis of OPM premium data. Premiums are for a standard option individual policy
unless only one option was available.




Page 23                                         GAO-03-547 Military Retiree Health Benefits
Military Retirees Who         The demonstration did not attract sicker, more costly enrollees—instead,
Enrolled in Demonstration     military retirees who enrolled were less sick on average than eligible
Not as Sick as Other          nonenrollees.36 We found that, as scored by a standard method to assess
                              patients’ health, older retirees who enrolled in the demonstration were an
Retirees                      estimated 13 percent less sick than eligible nonenrollees. At each site
                              enrollees were, on average, less sick than nonenrollees. In the GAO-DOD-
                              OPM survey, fewer enrollees on the U.S. mainland (33 percent) reported
                              that they or their spouses were in fair or poor health compared to
                              nonenrollees (40 percent). Retirees who enrolled in demonstration plans
                              had scores that indicated they were, on average, 19 percent less sick than
                              civilian FEHBP enrollees in these plans.

                              Plans’ divergent strategies for setting premiums resulted in similar mixes
                              of enrollees. Blue Cross Blue Shield and GEHA, both of which did not
                              increase premiums, attracted about the same proportion of individuals in
                              poor health as plans on the mainland that raised premiums.


Demonstration Enrollees       During 2000, the first year of the demonstration, enrolled retirees’ health
Less Expensive than           care was 28 percent less expensive—as measured by Medicare claims—
Eligible Nonenrollees and     than that of eligible nonenrolled retirees and one-third less expensive than
                              that of their FEHBP counterparts.37 (See table 3.) The demonstration
Much Less Expensive than      enrollees’ average age (71.8 years) was lower than eligible nonenrollees’
Their Civilian FEHBP          average age (73.1 years), which in turn was lower than the average age of
Counterparts, Leading to      civilian FEHBP retirees (75.2 years) in the demonstration areas. OPM has
Reduced Premiums for          obtained from the three largest plans claims information that includes the
Most Plans in Final Year of   cost of drugs and other services not covered by Medicare. These claims
Demonstration                 show a similar pattern: Demonstration enrollees were considerably less
                              expensive than enrollees in the civilian FEHBP.




                              36
                               We assessed enrollees’ health prior to the demonstration, using the Principal Inpatient
                              Diagnostic Cost Group, (PIP-DCG), which relies on diagnoses from inpatient hospital stays
                              and other patient characteristics. See app. II for discussion of the method and our results.
                              37
                               “Their FEHBP counterparts” refers to civilian retirees who were Medicare-eligible and
                              enrolled in FEHBP plans.




                              Page 24                                       GAO-03-547 Military Retiree Health Benefits
Table 3: Average Spending on Medicare-covered Services for Retirees Eligible for
the DOD-FEHBP Demonstration—by Enrollment Status, 2000

                               Demonstration                 Eligible         Civilian FEHBP
 Spending                          enrollees             nonenrollees                 retirees
 Medicare                            $3,174                   $4,412                    $4,785
 Coinsurance                             213                      315                      344
 Deductible                              142                      169                      184
 Total                               $3,529                   $4,896                    $5,313

Sources: CMS, DOD, and OPM.

Note: GAO analysis of CMS Medicare claims data, DOD enrollment data, and OPM enrollment data.
As of January 1, 2000, the average age of demonstration enrollees was 71.8 years; of eligible
nonenrollees, 73.1 years; and of civilian FEHBP retirees, 75.2 years.


Although demonstration enrollees’ costs were lower than those of their
FEHBP counterparts in the first year, demonstration premiums generally
remained higher than premiums for the civilian FEHBP. In 2001, the
second year of the demonstration, only a limited portion of the first year’s
claims was available when OPM and the plans negotiated the premiums, so
the lower demonstration costs had no effect on setting 2001 premiums.
Demonstration premiums in 2001 increased more rapidly than the civilian
premium charged by the same plans: a 30 percent average increase in the
demonstration for individual policies compared to a 9 percent increase for
civilians in the same plans. In 2002, the third year, when both the plans and
OPM were able to examine a complete set of claims for the first year
before setting premiums, the pattern was reversed: On average, the
demonstration premiums for individual policies fell more than 2 percent
while the civilian premiums rose by 13 percent. However, on average, 2002
premiums remained higher in the demonstration than in the civilian
FEHBP. Blue Cross Blue Shield was an exception, charging a higher
monthly premium for an individual policy to civilian enrollees ($89) in
2000 than to demonstration enrollees ($74).




Page 25                                        GAO-03-547 Military Retiree Health Benefits
                             Because the demonstration was open to only a small number of military
Impact of                    retirees—and only a small fraction of those enrolled—the demonstration
Demonstration on             had little impact on DOD, nonenrollees, and MTFs. However, the impact
                             on enrolled retirees was greater. If the FEHBP option were made
DOD Was Limited Due          permanent, the impact on DOD, nonenrollees, and MTFs would depend on
to Small Size and Low        the number of enrollees.
Enrollment, but
Impact on Enrollees
Was Greater

DOD Little Affected by       Because of its small size, the demonstration had little impact on DOD’s
Demonstration, Due           budget. About 140,000 of the more than 8 million people served by the
Primarily to Its Size, but   DOD health system were eligible for the demonstration in its last 2 years.
                             Enrollment at its highest was 7,521—about 5.5 percent of eligible
Enrollees More Affected      beneficiaries. DOD’s expenditures on enrollees’ premiums that year
                             totaled about $17 million—roughly 0.1 percent of its total health care
                             budget.38 Under the demonstration, DOD was responsible for about 71
                             percent of each individual’s premium, whereas under TFL it is responsible
                             for the entire cost of roughly similar Medicare supplemental coverage.39

                             Probably because of its small size, the demonstration had no observable
                             impact on either the ability of MTFs to assist in the training and readiness40
                             of military health care personnel or on nonenrollees’ access to MTF care.
                             Officials at the four MTFs in demonstration sites told us that they had seen
                             no impact from the demonstration on either MTFs or nonenrollees’ access
                             to care.




                             38
                               We were not able to adjust DOD expenditures to account for any reductions in the cost of
                             prescription drugs and MTF care due to the demonstration. While some military retirees
                             who enrolled were diverted from military to civilian care, the numbers were small and any
                             reductions in MTF costs could not be separated from other factors affecting DOD
                             expenditures. In addition, according to DOD, its costs for the demonstration were $28
                             million for FEHBP premiums and $11 million for administration, when measured over 3
                             years. These costs averaged less than 0.1 percent of the DOD health care budget over the
                             life of the demonstration.
                             39
                              TFL pays for Medicare-covered services not paid for by Medicare, as well as certain other
                             services.
                             40
                              Readiness refers to the capability of the military health system to provide medical support
                             of military deployments, from small humanitarian engagements to major military actions.




                             Page 26                                      GAO-03-547 Military Retiree Health Benefits
                       Since enrollees were typically attracted to the demonstration by both its
                       benefits and its relatively low costs, the impact on those who enrolled was
                       necessarily substantial. In the first 2 years, the demonstration provided
                       enrollees with better supplemental coverage, which was less costly or had
                       better benefits, or both. In the third year of the demonstration, after TFL
                       and the retirees’ pharmacy benefit were introduced and enrollment
                       declined, the number of beneficiaries affected by the demonstration
                       decreased. TFL entitled military retirees to low-cost, comprehensive
                       coverage, making the more expensive FEHBP unattractive. The average
                       enrollee premium for an individual policy in the demonstration’s third year
                       was $109 per month. In comparison, to obtain similar coverage under the
                       the combined TFL-pharmacy benefit, the only requirement was to pay the
                       monthly Medicare part B premium of $54. Further, pharmacy out-of-
                       pocket costs under TFL are less than those in the most popular FEHBP
                       plan.


Impact of Permanent    The impact on DOD of a permanent FEHBP option for military retirees
FEHBP Option Would     nationwide would depend on the number of retirees who enrolled. For
Depend on Enrollment   example, if the same percentage of eligible retirees who enrolled in 2002—
                       after TFL and the retirees’ pharmacy benefit were introduced—enrolled in
                       FEHBP, enrollment would be roughly 20,000 of the more than 1.5 million
                       military retirees. As retirees’ experience with TFL grows, their interest in
                       an FEHBP alternative may decline further. As long as enrollment in a
                       permanent FEHBP option remains small, the impact on DOD’s ability to
                       provide care at MTFs and on MTF readiness would also likely be small.


                       We provided DOD and OPM with the opportunity to comment on a draft of
Agency Comments        this report. In its written comments DOD stated that, overall, it concurred
                       with our findings. However, DOD differed with our description of the
                       demonstration’s impact on DOD’s budget as small. In contrast, DOD
                       described these costs of the 3-year demonstration–$28 million for FEHBP
                       premiums and $11 million for administration—as substantial. While we do
                       not disagree with these dollar-cost figures and have included them in this
                       report, we consider them to be small when compared to DOD’s health care
                       budget, which ranged from about $18 billion in fiscal year 2000 to about
                       $24 billion in fiscal year 2002. For example, as we report, DOD’s premium
                       costs for the demonstration during 2001, when enrollment peaked, were
                       about $17 million—less than 0.1 percent of DOD’s health care budget.
                       Although DOD’s cost per enrollee in the demonstration was substantial,
                       the number of enrollees was small, resulting in the demonstration’s total
                       cost to DOD being small. DOD’s comments appear in appendix VI. DOD


                       Page 27                              GAO-03-547 Military Retiree Health Benefits
also provided technical comments, which we incorporated as appropriate.
OPM declined to comment.


We are sending copies of this report to the Secretary of Defense and the
Director of the Office of Personnel Management. We will make copies
available to others upon request. In addition, this report will be available
at no charge on GAO’s Web site at http://www.gao.gov.

If you or your staffs have questions about this report, please contact me at
(202) 512-7101. Other GAO contacts and staff acknowledgments are listed
in appendix VII.




Marjorie E. Kanof
Director, Health Care—Clinical
 and Military Health Care Issues




Page 28                                GAO-03-547 Military Retiree Health Benefits
List of Committees

The Honorable John Warner
Chairman
The Honorable Carl Levin
Ranking Minority Member
Committee on Armed Services
United States Senate

The Honorable Peter G. Fitzgerald
Chairman
The Honorable Daniel K. Akaka
Ranking Minority Member
Subcommittee on Financial Management,
 the Budget, and International Security
Committee on Governmental Affairs
United States Senate

The Honorable Duncan L. Hunter
Chairman
The Honorable Ike Skelton
Ranking Minority Member
Committee on Armed Services
House of Representatives

The Honorable Jo Ann S. Davis
Chair
The Honorable Danny K. Davis
Ranking Minority Member
Subcommittee on Civil Service and
 Agency Organization
Committee on Government Reform
House of Representatives




Page 29                             GAO-03-547 Military Retiree Health Benefits
                       Appendix I: GAO-DOD-OPM Survey of
Appendix I: GAO-DOD-OPM Survey of
                       Military Retirees and Others Eligible for the
                       DOD-FEHBP Demonstration


Military Retirees and Others Eligible for the
DOD-FEHBP Demonstration
                       To determine why those eligible for the Federal Employees Health
                       Benefits Program (FEHBP) demonstration enrolled or did not enroll in an
                       FEHBP plan, we co-sponsored with the Department of Defense (DOD) and
                       the Office of Personnel Management (OPM) a mail survey of eligible
                       beneficiaries—military retirees and others eligible to participate in the
                       demonstration. The survey was fielded during the first year of the
                       demonstration, from May to August 2000, and was sent to a sample of
                       eligible beneficiaries, both those who enrolled and those who did not
                       enroll, at each of the eight demonstration sites operating at that time. The
                       survey was designed to be statistically representative of eligible
                       beneficiaries, enrollees, nonenrollees, and sites, and to facilitate valid
                       comparisons between enrollees and nonenrollees.


Questionnaire Design   In constructing the questionnaire, we developed questions pertaining to
                       individuals’ previous use of health care services, access to and satisfaction
                       with care, health status, knowledge of the demonstration, reasons for
                       enrolling or not enrolling in the demonstration, and other topics. Because
                       eligible beneficiaries could choose FEHBP plans that also covered their
                       family members, we included questions about spouses and dependent
                       children. DOD and OPM officials and staff members from Westat, the DOD
                       subcontractor with responsibility for administering the survey, provided
                       input on the questionnaire’s content and format. After pretesting the
                       questionnaire with a group of military retirees and their family members,
                       the final questionnaire included the topic areas shown in table 4. We also
                       produced a Spanish version of the questionnaire that was mailed to
                       beneficiaries living in Puerto Rico.




                       Page 30                                         GAO-03-547 Military Retiree Health Benefits
                Appendix I: GAO-DOD-OPM Survey of
                Military Retirees and Others Eligible for the
                DOD-FEHBP Demonstration




                Table 4: Major Survey Sections and Topics Covered

                 Section                                    Topics covered
                 Use of Health Care Services in 1999        Health care use, source and use of
                                                            prescription drugs, use of military treatment
                                                            facility (MTF) care, ease of access to MTF
                                                            care, and satisfaction with MTF care.
                 Health Status                              Current health status, health status compared
                                                            to 1 year ago, and need help with personal
                                                            care needs.
                 Family                                     Marital status, spouse’s health care use,
                                                            spouse’s use and source of prescription drugs,
                                                            spouse’s health status, dependent children,
                                                            and dependent children’s health status.
                 Knowledge of the Demonstration and         Knowledge of the demonstration, source of
                 Impact of the Demonstration                knowledge of the demonstration, whether
                 Information Campaign                       demonstration information materials were
                                                            received, usefulness of the information
                                                            materials, use of the toll-free telephone
                                                            service, source of information received about
                                                            individual demonstration plans, usefulness of
                                                            plans’ information, problems with making the
                                                            enrollment decision, reasons for joining the
                                                            demonstration, and reasons for not joining the
                                                            demonstration.
                 Other Insurance Coverage                   Medicare supplemental insurance of self and
                                                            spouse, other insurance coverage, cost of
                                                            insurance coverage, out-of-pocket costs for
                                                            medical services, and prescription drugs.
                 Personal Information                       Zip code, date of birth, sex, membership in a
                                                            military retiree organization, travel time to
                                                            nearest military hospital, rank at retirement,
                                                            race and ethnicity, educational attainment,
                                                            income, and home ownership.

                Source: GAO-DOD-OPM survey.




Sample Design   Working with DOD, OPM, and Westat, we defined the survey population as
                all persons living in the initial eight demonstration sites who were eligible
                to enroll in the demonstration. The population included military retirees,
                their spouses and dependents, and other eligible beneficiaries, such as
                unremarried former spouses, designated by law. We drew the survey
                sample from a database provided by DOD that listed all persons eligible
                for the demonstration as of April 1999.

                We stratified the sample by the eight demonstration sites and by
                enrollment status—enrollees and nonenrollees. Specifically, we used a
                stratified two-stage design in which households were selected within each



                Page 31                                         GAO-03-547 Military Retiree Health Benefits
                 Appendix I: GAO-DOD-OPM Survey of
                 Military Retirees and Others Eligible for the
                 DOD-FEHBP Demonstration




                 of the 16 strata and one eligible person was selected from each household.
                 For the enrollee sample, we selected all enrollees who were the sole
                 enrollee in their households. In households with multiple enrollees, we
                 randomly selected one enrollee to participate. For the nonenrollee sample,
                 first we randomly selected a sample of households from all nonenrollee
                 households and then randomly selected a single person from each those
                 households. We used a modified equal allocation approach, increasing the
                 size of the nonenrollee sample in steps, bringing it successively closer to
                 the sample size that would be obtained through proportional allocation.
                 This modified approach produced the best balance in statistical terms
                 between the gain from the equal allocation approach and the gain from the
                 proportional allocation approach.1 If both an enrollee and a nonenrollee
                 were selected from the same household, the nonenrollee was dropped
                 from the sample and a different nonenrollee was selected. We adjusted the
                 nonenrollee sample size to take account of expected nonresponse. Our
                 final sample included 1,676 out of 2,507 enrollees and 3,971 out of 66,335
                 nonenrollees.


Response Rates   Starting with an overall sample of 5,647 beneficiaries, we obtained usable
                 questionnaires from 4,787 people—an overall response rate of 85 percent.2
                 (See table 5.) Response rates varied across sites, from 76 percent to 85
                 percent among nonenrollees, and from 92 percent to 98 percent among




                 1
                  We considered (1) a proportional (to the population size) allocation across the sites that
                 would provide the greatest precision for population estimates, (2) an equal allocation
                 across the sites that would provide the greatest power to detect differences among the
                 eight sites, and (3) a matched allocation, in which the same number of enrollees and
                 nonenrollees would be selected, and which would provide the greatest power to detect
                 differences between enrollees and nonenrollees. We also examined two blended strategies:
                 one that blended proportional allocation with equal allocation, and another that blended
                 proportional allocation with matched allocation. We conducted a simulation to compare
                 the gain in precision and power—increasing the size of the nonenrollee sample under each
                 blended strategy. Assessing the gains from the two strategies, we determined that the
                 modified equal allocation approach was preferable. We specified the size of the
                 nonenrollee sample that would maximize the probability, at the 5 percent significance
                 level, of detecting a 5 percentage point difference in proportions between enrollees and
                 nonenrollees and a 10 percentage point difference between enrollees and nonenrollees at a
                 given site.
                 2
                  Westat, which fielded the survey, sent initial survey packages to all beneficiaries starting
                 in May 2000. Nonrespondents were sent follow-up reminder postcards as well as additional
                 survey packages as needed. Participants with questions could call toll-free numbers and
                 speak with English- or Spanish-speaking survey staff.




                 Page 32                                         GAO-03-547 Military Retiree Health Benefits
Appendix I: GAO-DOD-OPM Survey of
Military Retirees and Others Eligible for the
DOD-FEHBP Demonstration




enrollees. (See table 6.) At each site, enrollees responded at higher rates
than nonenrollees.

Each of the 16 strata was weighted separately to reflect its population. The
enrollee strata were given smaller sampling weights, reflecting enrollees’
higher response rates and the fact that they were sampled at a higher rate
than nonenrollees. The weights were also adjusted to reflect the variation
in response rates across sites. Finally, the sampling weights were further
adjusted to reflect differences in response rates between male and female
participants in 8 strata.

Table 5: Survey Responses and Nonresponses

 Sample size                                                                         5,647
   Response                                                                          4,787
   Nonresponse                                                                         860
 Overall response rate                                                                85%
 Reason for nonresponse
   Deceased                                                                             27
   Refusal                                                                              36
   Ineligible                                                                           11
   Other nonresponse                                                                   786
 Total not completed                                                                   860

Source: GAO-DOD-OPM survey.

Note: Westat analysis of GAO-DOD-OPM survey.




Page 33                                         GAO-03-547 Military Retiree Health Benefits
                                             Appendix I: GAO-DOD-OPM Survey of
                                             Military Retirees and Others Eligible for the
                                             DOD-FEHBP Demonstration




Table 6: Population, Sample Size, and Response Rate, by DOD-FEHBP Demonstration Site and Enrollee Status, 2000

                                                                                                    Number of            Response rate
 Site and enrollee status                       Population              Sample size               respondents             (percentage)
 Camp Pendleton, Calif.
    Enrollee                                              303                     197                       187                      95
    Nonenrollee                                        24,604                     752                       609                      81
 Dallas, Tex.
    Enrollee                                              520                     350                       323                      92
    Nonenrollee                                        13,087                     731                       618                      85
 Dover, Del.
    Enrollee                                               35                      26                        24                      92
    Nonenrollee                                         4,349                     388                       310                      80
 Fort Knox, Ky.
    Enrollee                                              134                      98                        90                      92
    Nonenrollee                                         7,623                     676                       535                      79
 Greensboro, N.C.
    Enrollee                                              285                     187                       183                      98
    Nonenrollee                                         2,993                     268                       228                      85
 Humboldt County, Calif.
    Enrollee                                              221                     150                       143                      95
    Nonenrollee                                         2,698                     232                       193                      83
 New Orleans, La.
    Enrollee                                               96                      71                        65                      92
    Nonenrollee                                         4,987                     419                       318                      76
 Puerto Rico
    Enrollee                                              913                     597                       561                      94
    Nonenrollee                                         5,994                     505                       400                      79
 All sites
    Enrollee                                            2,507                   1,676                     1,576                      94
    Nonenrollee                                        66,335                   3,971                     3,211                      81
 Total                                                 68,842                   5,647                     4,787                      85

Sources: DOD, OPM, and GAO-DOD-OPM survey.

                                             Note: GAO analysis of DOD and OPM data, and Westat analysis of GAO-DOD-OPM survey.




                                             Page 34                                         GAO-03-547 Military Retiree Health Benefits
                            Appendix II: Data, Methods, and Models Used
Appendix II: Data, Methods, and Models Used
                            in Analyzing Factors Affecting DOD-FEHBP
                            Demonstration Enrollment


in Analyzing Factors Affecting DOD-FEHBP
Demonstration Enrollment
                            In this appendix, we describe the data, methods, and models used to
                            (1) analyze the factors explaining how beneficiaries knew about the
                            demonstration and why they enrolled in it, (2) assess the health of
                            beneficiaries and civilian FEHBP enrollees, and (3) obtain the premiums
                            of Medigap insurance in the demonstration areas.


Analysis of Factors         Our approach to analyzing eligible beneficiaries’ behavior involved two
Affecting Knowledge about   steps: first, analyzing the factors related to whether eligible beneficiaries
the Demonstration and       knew about the demonstration, and second, analyzing the factors related
                            to whether those who knew about the demonstration decided to enroll.
Enrollment
                            Knowledge about the demonstration. To account for differences in
                            beneficiaries’ knowledge about the demonstration, we used individual-
                            level variables as well as variables corresponding to individual sites.1
                            These individual-level categories were demographic and economic
                            variables, such as age and income; health status; other sources of health
                            coverage, such as having employer-sponsored health insurance; and
                            military-related factors. The inclusion of site variables allowed the model
                            to take account of differences across the different sites in beneficiaries’
                            knowledge about the demonstration.

                            We analyzed the extent to which these variables influenced beneficiaries’
                            knowledge about the demonstration using a logistic regression—a
                            standard statistical method of analyzing an either/or (binary) variable. This
                            method yields an estimate of each factor’s effect, controlling for the
                            effects of all other factors in the regression. In our analysis, either a retiree
                            knew about the demonstration or did not. The logistic regression predicts
                            the probability that a beneficiary knew about the demonstration, given
                            information about the person’s traits—for example, over age 75, had
                            employer-sponsored health insurance, and so on. The coefficient on each
                            variable measures its effect on beneficiaries’ knowledge.2 These
                            coefficients pertain to the entire demonstration population, not just those
                            beneficiaries in our survey sample. To make the estimates generalizable to


                            1
                             Individual sites were represented by binary or dummy variables; for example, Humboldt
                            County, California had a value of one when a beneficiary lived in that site, and a value of
                            zero when the beneficiary lived in another site.
                            2
                             To avoid statistical problems with analyzing the probability directly, logistic regression
                            analyzes a related dependent variable—a function of the probability, P, divided by (1-P).
                            However, the estimated probability, P, can be calculated from the logistic regression. In our
                            analysis, P refers to each retiree’s probability of knowing about the demonstration.




                            Page 35                                       GAO-03-547 Military Retiree Health Benefits
Appendix II: Data, Methods, and Models Used
in Analyzing Factors Affecting DOD-FEHBP
Demonstration Enrollment




the entire eligible population, we applied sample weights to all
observations.

In view of the large difference in enrollment between the mainland sites
and Puerto Rico, we tested whether the same set of coefficient estimates
was appropriate for the mainland sites and the Puerto Rico site. Our
results showed that the coefficient estimates for the mainland and for
Puerto Rico were not significantly different (at the 5 percent level), so it
was appropriate to estimate a single logistic regression model for all sites.

Table 7 shows for each variable its estimated effect on knowledge, as
measured by the variable’s coefficient and odds ratio. The odds ratio
expresses how much more likely—or less likely—it is that a person with a
particular characteristic knows about the demonstration, compared to a
person without that characteristic. The odds ratio is based on the
coefficient, which indicates each explanatory variable’s estimated effect
on the dependent variable, holding other variables constant. For the
mainland sites, retirees were more likely to know about the demonstration
if they were male, were married, were officers, were covered by employer-
sponsored health insurance, lived less than an hour from a military
treatment facility (MTF), or belonged to military retiree organizations.
Retirees were less likely to know about the demonstration if they were
African American; were older than age 75; or lived in Camp Pendleton,
California, Dallas, Texas, or Fort Knox, Kentucky.




Page 36                                       GAO-03-547 Military Retiree Health Benefits
                                                Appendix II: Data, Methods, and Models Used
                                                in Analyzing Factors Affecting DOD-FEHBP
                                                Demonstration Enrollment




Table 7: Estimated Effects of Selected Factors on Whether Eligible Retirees Knew about the DOD-FEHBP Demonstration

                                                                                                                    Odds ratio             Coefficient
 Demographic and economic factors           African American                                                              0.67                  -0.40a
                                            Higher income (over $40,000)                                                  1.29                   0.26a
                                            Hispanic                                                                      0.61                  -0.49b
                                            Male                                                                          1.38                   0.32c
                                            Married                                                                       1.43                   0.36c
                                            Officer                                                                       1.49                   0.40c
                                            Older than age 75                                                             0.71                  -0.35c
 Health status                              Self or spouse in fair or poor health                                         0.85                   -0.16
 Health insurance coverage                  Covered by a Medigap policy                                                   1.10                    0.09
                                            Covered by employer-sponsored health insurance                                1.39                   0.33c
                                            Enrolled in a Medicare+Choice plan in 1999                                    0.97                   -0.03
                                            Enrolled in Medicare part B on January 1, 2000                                1.12                    0.11
 Military-related factors                   Less than 1 hour from an MTF                                                  1.46                   0.38c
                                            Member of military retiree organization                                       1.70                   0.53c
                                            Used VA care during fiscal years 1998 or 1999                                 0.81                   -0.21
 Site effectsd                              Camp Pendleton, Calif.                                                        0.58                  -0.55c
                                            Dallas, Tex.                                                                  0.65                  -0.43c
                                            Dover, Del.                                                                   0.72                   -0.33
                                            Fort Knox, Ky.                                                                0.59                  -0.52a
                                            Greensboro, N.C.                                                              1.18                    0.16
                                            Humboldt County, Calif.                                                       0.93                   -0.07
                                            Puerto Rico                                                                   0.77                   -0.26
                                            Constant                                                                                            -0.73c
                                            Observations                                                                                        3,504

Sources: GAO-DOD-OPM survey, CMS, and VA.

                                                Note: GAO analysis of GAO-DOD-OPM survey data, CMS enrollment data and VA enrollment data.
                                                The odds ratio expresses how much more likely—or less likely—it is that a person with a particular
                                                characteristic knows about the demonstration, compared to a person without that characteristic. The
                                                coefficient indicates each explanatory variable’s estimated effect on the dependent variable, holding
                                                other variables constant.
                                                a
                                                    Significant at the 5 percent level.
                                                b
                                                    Significant at the 10 percent level.
                                                c
                                                    Significant at the 1 percent level.
                                                d
                                                 The site effects consisted of a dummy variable for each site; the comparison site is New Orleans,
                                                La., selected at random from the eight sites. The mainland site effects were jointly significant at the 5
                                                percent level.




                                                Page 37                                             GAO-03-547 Military Retiree Health Benefits
Appendix II: Data, Methods, and Models Used
in Analyzing Factors Affecting DOD-FEHBP
Demonstration Enrollment




Decision to enroll in the demonstration. To account for a retiree’s
decision to enroll or not to enroll, we considered four categories of
individual-level variables similar to those in the “knowledge of the
demonstration” regressions, and a site-level variable for Puerto Rico. We
also introduced a set of health insurance factors pertaining to the area in
which the retiree lived—the premium for a Medigap policy and the
proportion of Medicare beneficiaries in a retiree’s county of residence
enrolled in a Medicare+Choice plan.

In our logistic regression analysis of enrollment, we included only those
people who knew about the demonstration. Despite the large enrollment
differences between the mainland sites and Puerto Rico, our statistical
tests determined that the mainland sites and the Puerto Rico site could be
combined into a single logistic regression of enrollment. We included a
variable for persons in the Puerto Rico site. (See table 8.)




Page 38                                       GAO-03-547 Military Retiree Health Benefits
                                                             Appendix II: Data, Methods, and Models Used
                                                             in Analyzing Factors Affecting DOD-FEHBP
                                                             Demonstration Enrollment




Table 8: Estimated Effects of Selected Factors on Whether Eligible Retirees Enrolled in an FEHBP Plan

                                                                                                                                   Odds ratio            Coefficient
 Demographic and economic factors                                        African American                                                0.51                 -0.68a
                                                                         Hispanic                                                        1.19                   0.17
                                                                         Higher income (over $40,000)                                    1.35                  0.30b
                                                                         Male                                                            0.74                 -0.31c
                                                                         Married                                                         5.06                  1.62c
                                                                         Officer                                                         1.46                  0.38a
                                                                         Older than age 75                                               1.32                   0.28
 Health status                                                           Self or spouse in fair or poor health                           0.93                  -0.07
 Health insurance coverage                                               Covered by a Medigap policy                                     1.32                  0.28b
                                                                         Covered by employer-sponsored health insurance                  0.40                 -0.92a
                                                                         Enrolled in a Medicare+Choice plan in 1999                      0.53                 -0.64a
                                                                         Enrolled in Medicare part B on January 1, 2000                  2.01                  0.70a
 Military-related factors                                                Less than 1 hour from an MTF                                    0.36                 -1.01a
                                                                         Member of a military retiree organization                       1.49                  0.40a
                                                                         Used VA care during fiscal years 1998 or 1999                   1.00                   0.00
                                                                                                                                              e
 Geographic effects                                                      Medicare+Choice enrollment in countyd                             --                 -0.01a
                                                                         Medigap price for county and age category                         --e                -0.38
                                                                                                                                                                   a


                                                                         Puerto Rico site                                                2.96                  1.09a
                                                                         Constant                                                                             -2.69a
                                                                         Observations                                                                         1,913

Sources: GAO-DOD-OPM survey, Quotesmith Inc., CMS, and VA.

                                                             Note: GAO analysis of GAO-DOD-OPM survey, DOD enrollment data, CMS enrollment data, VA
                                                             enrollment data, and Quotesmith Inc. Medigap premium data.
                                                             a
                                                                 Significant at the 1 percent level.
                                                             b
                                                                 Significant at the 5 percent level.
                                                             c
                                                                 Significant at the 10 percent level.
                                                             d
                                                             The proportion of Medicare beneficiaries in a retiree’s county of residence enrolled in a
                                                             Medicare+Choice plan.
                                                             e
                                                              Odds ratios are not reported for continuous variables, such as the number of enrollees and the price
                                                             in dollars, because, unlike binary variables, the choice of values to make a comparison is arbitrary.


                                                             We found that retirees were less likely to enroll in the demonstration if
                                                             they were African American, enrolled in Medicare+Choice plans, had
                                                             employer-sponsored health insurance, lived in areas with a high
                                                             proportion of Medicare beneficiaries enrolled in a Medicare+Choice plan,
                                                             lived in areas where Medigap was more expensive, or lived less than an
                                                             hour from an MTF. Retirees who had higher incomes, were officers, were
                                                             members of a military retiree organization, were enrolled in Medicare part
                                                             B, lived in Puerto Rico, or were covered by a Medigap policy were more
                                                             likely to enroll.


                                                             Page 39                                            GAO-03-547 Military Retiree Health Benefits
                            Appendix II: Data, Methods, and Models Used
                            in Analyzing Factors Affecting DOD-FEHBP
                            Demonstration Enrollment




Calculating the Impact on   We estimated what the demonstration’s enrollment rate would have been
Enrollment if Those         in 2000 if everyone eligible for the demonstration had known about it. For
Eligible Had Known about    the 54 percent of retirees who did not know about the demonstration, we
                            calculated their individual probabilities of enrollment, using their
the Demonstration           characteristics (such as age) and the coefficient estimates from the
                            enrollment regression.3 Aggregating these individual estimated enrollment
                            probabilities, we found that if all eligible retirees had known about the
                            demonstration, enrollment in 2000 would have been 7.2 percent of eligible
                            beneficiaries, compared with actual enrollment of 3.6 percent.4


Estimating Health Status    To measure the health status of retired enrollees and nonenrollees, as well
Based on PIP-DCG Scores     as of civilian FEHBP enrollees, we calculated scores for individuals using
                            the Principal Inpatient Diagnostic Cost Group (PIP-DCG) method. This
                            method—used by the Centers for Medicare & Medicaid Services (CMS) in
                            adjusting Medicare+Choice payment rates—yielded a proxy for the
                            healthiness of military and civilian retirees as of 1999, the year before the
                            demonstration. The method relates individuals’ diagnoses to their annual
                            Medicare expenditures. For example, a PIP-DCG score of 1.20 indicates
                            that the individual is 20 percent more costly than the average Medicare
                            beneficiary. In our analysis, we used Medicare claims and other
                            administrative data from 1999 to calculate PIP-DCG scores for eligible
                            military retirees and their counterparts in the civilian FEHBP in the
                            demonstration sites.

                            Using Medicare part A claims for 1999, we calculated PIP-DCG scores for
                            Medicare beneficiaries who were eligible for the demonstration. We used a




                            3
                             In these calculations, we used only the characteristics from the model to simulate
                            enrollment, which means we assumed the people who did not know about the
                            demonstration would have behaved the same with respect to their decision to enroll, given
                            their characteristics, as those who knew. We also adjusted for the difference between the
                            enrollment rate in the demonstration as a whole and the enrollment rate of those included
                            in the logistic regression analysis for whom there were no missing data.
                            4
                             Retirees who reported that they did not know about the demonstration before the survey
                            may have included some retirees who had known about it at one time. About 9 months
                            elapsed between DOD’s final mailing to beneficiaries about the demonstration and the end
                            of our survey. Our logistic regression for enrollment considered only people who
                            responded in the survey that they knew about the demonstration. We excluded people from
                            the enrollment regression who were enrolled but responded that they did not know about
                            the demonstration. This did not affect our results because nearly all (more than 99 percent)
                            of those who said they did not know about the demonstration did not enroll.




                            Page 40                                       GAO-03-547 Military Retiree Health Benefits
Appendix II: Data, Methods, and Models Used
in Analyzing Factors Affecting DOD-FEHBP
Demonstration Enrollment




DOD database to identify enrollees as well as those who were eligible for
the demonstration but did not enroll.

We also calculated PIP-DCG scores based on 1999 Medicare claims for
each Medicare-eligible person enrolled in the civilian FEHBP. We obtained
from OPM data on enrollees in the civilian FEHBP and on the plans in
which they were enrolled. We restricted our analysis to those Medicare-
eligible civilian FEHBP enrollees who lived in a demonstration site.

Results of PIP-DCG calculations. We compared the PIP-DCG scores of
demonstration enrollees with those of eligible retirees who did not enroll.
In every site, the average PIP-DCG score was significantly less5 for
demonstration enrollees than for those who did not enroll. We also
compared the PIP-DCG scores of those enrolled in the demonstration with
those enrolled in the civilian FEHBP: For every site, these scores were
significantly less for demonstration enrollees than for their counterparts in
the civilian FEHBP.6 (See table 9.)




5
The scores were significantly less at the 5 percent level.
6
The scores were significantly less at the 5 percent level.




Page 41                                       GAO-03-547 Military Retiree Health Benefits
                   Appendix II: Data, Methods, and Models Used
                   in Analyzing Factors Affecting DOD-FEHBP
                   Demonstration Enrollment




                   Table 9: Health Status Comparisons of DOD-FEHBP Demonstration Enrollees with
                   Eligible Retirees Who Did Not Enroll and with Civilian FEHBP Retirees, Based on
                   PIP-DCG Scores

                                                                   Ratio of PIP-DCG scores of enrollees in a
                                                                              demonstration plan
                                                                   Compared to eligible
                                                                   military retirees who Compared to civilian
                       Site                                                did not enroll   retirees in FEHBP
                       All sites                                                     0.87                  0.81
                       Camp Pendleton, Calif.                                        0.88                  0.83
                       Dallas, Tex.                                                  0.82                  0.75
                       Dover, Del.                                                   0.76                  0.71
                       Humboldt County, Calif.                                       0.91                  0.86
                       Fort Knox, Ky.                                                0.79                  0.73
                       Greensboro, N.C.                                              0.84                  0.77
                       New Orleans, La.                                              0.78                  0.73
                       Puerto Rico                                                   0.94                  0.93

                   Source: CMS, DOD, and OPM.

                   Note: GAO analysis of CMS claims data, DOD enrollment data, and OPM enrollment data.
                   Comparisons used 1999 claims data and measured enrollment status as of September 2000. The
                   difference between the PIP-DCG scores for the enrollees in the demonstration and the scores of
                   military retirees who did not enroll was statistically significant at the 5 percent level for each
                   demonstration site. The difference between the PIP-DCG scores for the enrollees in the
                   demonstration and the scores of civilian retirees in FEHBP was statistically significant at the 5 percent
                   level for each demonstration site.


Medigap Premiums   We compiled data from Quotesmith Inc. to obtain a premium price for
                   Medigap plan F in each of the counties in the eight demonstration sites.7
                   We collected the lowest premium quote for a Medigap plan F policy for
                   each sex at 5-year intervals: ages 65, 70, 75, 80, 85, and over 89. A person
                   age 65 to 69 was assigned the 65-year-old’s premium, a person age 70 to 74
                   was assigned the 70-year-old’s premium, and so on. Using these data, we
                   assigned a Medigap plan F premium to each survey respondent age 65 and
                   over, according to the person’s age, sex, and location.




                   7
                    Quotesmith.com, Inc. Instant Medicare Supplemental Insurance Quotes (Darien, Ill.: June
                   2000), http://www.quotesmith.com/index.html#medsup (downloaded on June 27, 2000).




                   Page 42                                             GAO-03-547 Military Retiree Health Benefits
                                            Appendix III: Enrollment in the DOD-FEHBP
Appendix III: Enrollment in the DOD-FEHBP   Demonstration



Demonstration

                                            Tables 10, 11, and 12 show enrollment rates by site and for the U.S.
                                            mainland sites as a whole for each year of the demonstration, 2000
                                            through 2002.

Table 10: Enrollment in the DOD-FEHBP Demonstration, 2000

                                                                          Enrollees      Eligible beneficiaries   Percentage enrolled
 Mainland sites             Camp Pendleton, Calif.                              303                     24,907                    1.2
                            Dallas, Tex.                                        520                     13,607                    3.8
                            Dover, Del.                                          35                      4,384                    0.8
                            Fort Knox, Ky.                                      134                      7,757                    1.7
                            Greensboro, N.C.                                    285                      3,278                    8.7
                            Humboldt County, Calif.                             221                      2,919                    7.6
                            New Orleans, La.                                     96                      5,083                    1.9
 Total for mainland sites                                                     1,594                     61,935                    2.6
 Other site                 Puerto Rico                                         913                      6,907                   13.2
 Total                                                                        2,507                     68,842                    3.6

Source: DOD.

                                            Note: Data are as of January 1, 2000.



Table 11: Enrollment in the DOD-FEHBP Demonstration, 2001

                                                                     Enrollees        Eligible beneficiaries      Percentage enrolled
 Mainland sites               Adair County, Iowa                         1,564                       29,584                       5.3
                              Camp Pendleton, Calif.                       421                       27,328                       1.5
                              Coffee County, Ga.                           867                       27,329                       3.2
                              Dallas, Tex.                                 949                       16,159                       5.9
                              Dover, Del.                                   64                        4,868                       1.3
                              Fort Knox, Ky.                               188                        9,121                       2.1
                              Greensboro, N.C.                             334                        4,033                       8.3
                              Humboldt County, Calif.                      305                        3,461                       8.8
                              New Orleans, La.                             142                        6,095                       2.3
 Total for mainland sites                                                4,834                     127,978                        3.8
 Other site                   Puerto Rico                                2,687                        9,401                      28.6
 Total                                                                   7,521                     137,379                        5.5

Source: DOD.

                                            Note: Data are as of March 14, 2001.




                                            Page 43                                         GAO-03-547 Military Retiree Health Benefits
                                          Appendix III: Enrollment in the DOD-FEHBP
                                          Demonstration




Table 12: Enrollment in the DOD-FEHBP Demonstration, 2002

                                                                      Enrollees     Eligible beneficiaries    Percentage enrolled
 Mainland sites             Adair County, Iowa                              484                    29,530                     1.6
                            Camp Pendleton, Calif.                          145                    27,287                     0.5
                            Coffee County, Ga.                              212                    27,284                     0.8
                            Dallas, Tex.                                    354                    16,133                     2.2
                            Dover, Del.                                      36                     4,867                     0.7
                            Fort Knox, Ky.                                   70                     9,113                     0.8
                            Greensboro, N.C.                                 85                     4,024                     2.1
                            Humboldt County, Calif.                          65                     3,454                     1.9
                            New Orleans, La.                                 74                     6,085                     1.2
 Total for mainland sites                                                 1,525                   127,777                     1.2
 Other site                 Puerto Rico                                   2,842                     9,453                    30.1
 Total                                                                    4,367                   137,230                     3.2

Source: DOD.

                                          Note: Data are as of February 21, 2002.




                                          Page 44                                       GAO-03-547 Military Retiree Health Benefits
                Appendix IV: DOD’s Approach to Informing
Appendix IV: DOD’s Approach to Informing
                Beneficiaries about the DOD-FEHBP
                Demonstration


Beneficiaries about the DOD-FEHBP
Demonstration
                The program for informing and educating eligible beneficiaries about the
                demonstration was modeled on OPM’s approach to informing eligible
                civilian beneficiaries about FEHBP. Elements of OPM’s approach include
                making available a comparison of FEHBP plans and holding health fairs
                sponsored by individual federal agencies. DOD expanded upon the OPM
                approach–for example, by sending postcards to inform eligible
                beneficiaries about the demonstration because they, unlike civilian federal
                employees and retirees, were unlikely to have any prior knowledge of
                FEHBP. In addition, DOD established a bilingual toll-free number. During
                the first year’s enrollment period, DOD adjusted its information and
                education effort, for example, by changing the education format from
                health fairs to town meetings designed specifically for demonstration
                beneficiaries. In the second year of the demonstration, DOD continued
                with its revised approach. In the third year, after TRICARE For Life (TFL)
                began, DOD significantly reduced its information program but continued
                to mail information to all eligible beneficiaries. It limited town meetings to
                Puerto Rico, the only site where enrollment remained significant during
                the third year.


Mailings        DOD sent a series of mailings to all eligible beneficiaries. These included

            •   a postcard announcing the demonstration, mailed in August 1999,1 that
                alerted beneficiaries to the demonstration–the returned postcards allowed
                DOD to identify incorrect mailing addresses and to target follow-up
                mailings to beneficiaries with correct addresses;
            •   an OPM-produced booklet, The 2000 Guide to Federal Employees Health
                Benefits Plans Participating in the DOD/FEHBP Demonstration Project,
                received by all eligible retirees from November 3 through 5, 1999, that
                contained information on participating FEHBP plans, including coverage
                and consumer satisfaction;
            •   a trifold brochure describing the demonstration, which was mailed on
                September 1 and 4, 1999; and
            •   a list of Frequently Asked Questions (FAQ) explaining how Medicare and
                FEHBP work together.

                At the time of our survey, after the first year’s information campaign, over
                half of eligible beneficiaries were unaware of the demonstration. Among
                those who knew about it, more recalled receiving the postcard than



                1
                Dates for this and subsequent mailings refer to the first year of the demonstration.




                Page 45                                      GAO-03-547 Military Retiree Health Benefits
                        Appendix IV: DOD’s Approach to Informing
                        Beneficiaries about the DOD-FEHBP
                        Demonstration




                        recalled receiving any of the later materials—although the FAQ was cited
                        more often as being useful. (See table 13.)

                        Table 13: Beneficiaries Who Recalled Receiving DOD-FEHBP Demonstration
                        Mailings and Who Found Them Useful

                                                                                           Percentages
                                                                                    Beneficiaries    Beneficiaries
                                                                                    who recalled       who found
                                                                                       receiving materials usefulb
                                                                                      materialsa
                            Postcard announcing the DOD-FEHBP
                            demonstration                                                           31                 61
                            Booklet entitled, The 2000 Guide to Federal
                            Employees Health Benefits Plans Participating
                            in the DOD/FEHBP Demonstration Project                                  27                 67
                            Trifold brochure describing the demonstration                           17                 69
                            FAQ about coordination of Medicare and
                            FEHBP benefits                                                          17                 72

                        Source: GAO-DOD-OPM survey.

                        Note: These materials were mailed in 1999 for the 2000 enrollment period.
                        a
                         The question was asked only of those who knew that, as part of the new demonstration, they could
                        join an FEHBP health plan.
                        b
                        Entries are percentages of beneficiaries who recalled receiving them.




Health Fairs and Town   Initially, the health fairs that DOD sponsored for military bases’ civilian
Meetings                employees were its main effort—other than the mailings—to provide
                        information about the demonstration to eligible beneficiaries. At these
                        health fairs, plans set up tables at which their representatives distributed
                        brochures and answered questions. At one site, the military base refused
                        to allow the demonstration representatives to participate in its health fair
                        because of concern about an influx of large numbers of demonstration
                        beneficiaries. At another site, the turnout exceeded the capacity of the
                        plan representatives to deal with questions and DOD officials told us that
                        they accommodated more people by giving another presentation at a
                        different facility or at the same facility 1 month later.

                        A DOD official discovered, however, that it was difficult to convey
                        information about the demonstration to large numbers of individuals at
                        the health fairs. DOD officials determined that the health fairs were not
                        working well, so by January 2000, DOD replaced them with 2-hour
                        briefings, which officials called town meetings. In these meetings, a DOD
                        representative explained the demonstration during the first hour and then



                        Page 46                                          GAO-03-547 Military Retiree Health Benefits
                            Appendix IV: DOD’s Approach to Informing
                            Beneficiaries about the DOD-FEHBP
                            Demonstration




                            answered questions from the audience. A DOD official told us that these
                            town meetings were more effective than the health fairs.2

                            For the first year of the demonstration, just under 6 percent of those
                            eligible attended either a health fair or a town meeting. The number of
                            eligible beneficiaries who reported attending these meetings varied
                            considerably by site—from about 3 percent in New Orleans and Camp
                            Pendleton to 4 percent in Fort Knox and 18 percent in Humboldt County.
                            Roughly 11 percent of beneficiaries reported attending in Puerto Rico, the
                            site with the highest enrollment.


DOD’s Call Center and       DOD also established a call center and a Web site to inform eligible
Web Site                    beneficiaries about the demonstration. The call center, which was staffed
                            by Spanish and English speakers, answered questions and sent out printed
                            materials on request. In the GAO-DOD-OPM survey, about 18 percent of
                            those who knew about the demonstration reported calling the center’s toll-
                            free number. The proportion that called the toll-free number was much
                            higher among subsequent enrollees (77 percent) than among nonenrollees
                            who knew about the demonstration (13 percent). The Web site was
                            another source of information about the demonstration.


Beneficiaries’ Sources of   Although less than half of eligible beneficiaries knew about the
Information                 demonstration, most of those who did know said they obtained their
                            information from DOD’s mailings. Other important sources of information
                            included military retiree and military family organizations and FEHBP
                            plans. (See table 14.)




                            2
                             In Puerto Rico, the town hall meetings were conducted in Spanish, which, according to
                            one DOD official, was very effective in conveying the information to the eligible
                            beneficiaries at that site.




                            Page 47                                     GAO-03-547 Military Retiree Health Benefits
                                              Appendix IV: DOD’s Approach to Informing
                                              Beneficiaries about the DOD-FEHBP
                                              Demonstration




Table 14: Beneficiaries’ Sources of Information about the DOD-FEHBP Demonstration

                                                                                                         Percentages
 Source of information                                                                    All beneficiaries     Enrollees         Nonenrollees
 Received information mailed by DOD                                                                    81.8          78.1                 82.1
 Received information from a military retiree or family organization                                   33.1          43.3                 32.3
 Received information from one of the FEHBP plans                                                      25.0          37.3                 24.0
 Heard about demonstration from family or friends                                                       7.0          10.0                   6.8
 Attended a health fair or town meeting                                                                 5.9          25.6                   4.3
 Heard about it from office of Member of Congress                                                       2.1            5.5                  1.8
 Read article about the demonstration in the newspaper                                                  7.6            9.7                  7.4
 Saw newspaper advertisements by one or more FEHBP plans                                                1.9            2.2                  1.8
 Heard about demonstration on radio or television                                                       1.7            1.7                  1.7
 Othera                                                                                                 6.7          10.4                   6.3

Source: GAO-DOD-OPM survey.

                                              Note: The source of information is given only for those who knew before receiving the survey that, as
                                              a part of the new demonstration, they could join an FEHBP health plan. Percentages add to more
                                              than 100 because respondents could select more than one reason. Respondents reported
                                              information gained relating to 2000 enrollment.
                                              a
                                              “Other” refers to answers that could not be classified.


                                              Nearly all of enrollees (93 percent) and more than half of nonenrollees
                                              who said they considered enrolling in an FEHBP health plan (55 percent)
                                              reported that they had enough information about specific plans to make an
                                              informed decision about enrolling in one of them. More than three-fifths of
                                              these beneficiaries who enrolled or considered enrolling in an FEHBP plan
                                              said they used The 2000 Guide to FEHBP Plans Participating in the
                                              DOD/FEHBP Demonstration Project as a source of information. Other
                                              major sources of information were the plans’ brochures and DOD’s health
                                              fairs and town meetings. More than 18 percent of those who considered
                                              joining did not obtain information about any specific plan. (See table 15.)




                                              Page 48                                            GAO-03-547 Military Retiree Health Benefits
Appendix IV: DOD’s Approach to Informing
Beneficiaries about the DOD-FEHBP
Demonstration




Table 15: Sources of Information for Eligible Beneficiaries about Specific FEHBP
Plans

                                                                     Percentages
    Source of Information                                 Enrollees Nonenrollees       Total
    Reading The 2000 Guide to FEHBP Plans                      75.1           59.7       63.5
    Reading one or more plans’ brochures                       46.5           26.0       31.1
    Health fair or town meeting                                35.2           12.5       18.1
    Calling one or more plans                                  27.2            9.9       14.2
    Friends or family                                          14.3            8.6       10.0
    Internet                                                   10.3            3.8        5.4
    Advertising in a newspaper or other publication              1.6           3.0        2.7
    Othera                                                     10.1            6.8        7.6
    I did not get information about any specific
    FEHBP plans                                                  1.2           24.3       18.6

Source: GAO-DOD-OPM survey.

Note: Entries are percentages of respondents who considered joining an FEHBP plan. Percentages
add to more than 100 because respondents could select more than one reason. Respondents
reported information gained relating to 2000 enrollment.
a
“Other” refers to answers that could not be classified.




Page 49                                            GAO-03-547 Military Retiree Health Benefits
                                             Appendix V: Enrollees’ and Nonenrollees’
Appendix V: Enrollees’ and Nonenrollees’     Reasons for Joining or Not Joining a DOD-
                                             FEHBP Demonstration Plan


Reasons for Joining or Not Joining a DOD-
FEHBP Demonstration Plan
                                             Table 16 shows reasons cited by enrollees for enrolling in a DOD-FEHBP
                                             health plan in 2000, and table 17 shows reasons cited by nonenrollees for
                                             not enrolling.

Table 16: Survey Responses by Enrollees to the Question “Why Did You Join a DOD-FEHBP Health Plan?”

                                                                                                             Percentages
                                                                                                                           Location
                                                                                           All respondents             Mainland Puerto Rico
 The plan’s benefits package meets my needs (and those of my family)                                  66.7                 66.1        68.2
 I needed better coverage for prescriptions                                                           64.3                 60.0        74.7
 My current doctors are among those I can select under the plan                                       62.5                 63.0        61.3
 It costs less than other coverage that I could buy                                                   62.1                 58.8        69.9
 The plan’s benefits package is better than other coverage I could get                                50.8                 47.2        59.2
 It costs less than my previous coverage (insurance or health plan)                                   49.8                 48.9        51.8
 The plan has a good reputation for quality of care                                                   44.6                 39.6        56.6
 My spouse joined the plan, and it is more convenient if we’re both in the
 same plan                                                                                                34.6               28.6             48.7
 I can’t count on getting space-available care                                                            27.1               33.1             13.1
 It gives me a broader choice of doctors than I had before                                                26.5               21.8             37.5
 I don’t want to use military care                                                                        22.2               25.9             13.4
 Many civilian doctors don’t accept CHAMPUS/TRICAREa                                                      20.4               17.2             28.2
 My friends or relatives recommended that I join the plan                                                 14.2                9.4             25.3
        b
 Other                                                                                                    10.0               10.9              7.8
Source: GAO-DOD-OPM survey.

                                             Note: This question was asked only of people who knew about the demonstration at the time of the
                                             survey. Beneficiaries were given a list of possible answers as well as an “Other” option for which they
                                             could write their own answers.
                                             a
                                              CHAMPUS is the name given to the military health care program that preceded TRICARE.
                                             b
                                              Answers that could not be classified.




                                             Page 50                                            GAO-03-547 Military Retiree Health Benefits
                                              Appendix V: Enrollees’ and Nonenrollees’
                                              Reasons for Joining or Not Joining a DOD-
                                              FEHBP Demonstration Plan




Table 17: Survey Responses by Nonenrollees to the Question “Why Didn’t You Join a DOD-FEHBP Health Plan?”

                                                                                                             Percentages
                                                                                                                        Location
                                                                                       All respondents             Mainland     Puerto Rico
 I was satisfied with my current coverage                                                         64.1                 65.9            28.4
 It would cost too much                                                                           29.4                 29.9            17.6
 The program is new, and I’m waiting to see how it works                                          26.6                 26.4            30.2
 I wasn’t sure how it would work with Medicare                                                    26.2                 25.7            36.2
 I wouldn’t be able to use military pharmacies anymore                                            26.1                 26.4            20.8
 I couldn’t keep my current doctors                                                               25.5                 26.3             8.1
 The demonstration will end in 3 years                                                            22.0                 22.3            16.5
 I was afraid I wouldn’t be able to get my Medicare supplemental policy
 back after the demonstration ended                                                                   20.2                20.6                 12.1
 I can get care at military health care facilities when I need it                                     14.4                13.6                 30.7
 I heard about the demonstration, but did not have enough information to
 make a decision                                                                                      13.9                13.3                 27.5
 I was afraid I wouldn’t be able to get my retiree health insurance back after
 the demonstration ended                                                                              11.1                11.1                 11.0
 I can get care at the VA when I need it                                                              10.2                 9.1                 31.4
 I couldn’t decide which plan to join                                                                  9.5                 9.0                 20.8
 My spouse didn’t want to join so I decided not to                                                     5.7                 5.5                  9.2
 My friends and relatives recommend against it                                                         5.0                 5.2                  0.0
 I was not eligible                                                                                    4.5                 4.7                  1.1
 I didn’t know about the demonstration project                                                         3.5                 3.4                  4.4
 None of the plans available to me had a good reputation                                               3.0                 3.1                  1.1
 Othera                                                                                               15.8                15.9                 13.2

Source: GAO-DOD-OPM survey.

                                              Note: This question was asked only of people who knew about the demonstration at the time of the
                                              survey. Beneficiaries were given a list of possible answers as well as an “Other” option for which they
                                              could write their own answers. Answers relate to enrollment in 2000. Because beneficiaries could
                                              select multiple reasons, the percentages total more than 100.
                                              a
                                               Answers that could not be classified.




                                              Page 51                                            GAO-03-547 Military Retiree Health Benefits
             Appendix VI: Comments from the Department
Appendix VI: Comments from the
             of Defense



Department of Defense




             Page 52                                     GAO-03-547 Military Retiree Health Benefits
                  Appendix VII: GAO Contacts and Staff
Appendix VII: GAO Contacts and Staff
                  Acknowledgments



Acknowledgments

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


                  Major contributors to this work were Michael Kendix, Robin Burke,
Acknowledgments   Jessica Farb, Martha Kelly, Dae Park, and Michael Rose.




                  Page 53                                GAO-03-547 Military Retiree Health Benefits
             Related GAO Products
Related GAO Products


             Defense Health Care: Oversight of the Adequacy of TRICARE’s Civilian
             Provider Network Has Weaknesses. GAO-03-592T. Washington, D.C.:
             March 27, 2003.

             Federal Employees’ Health Benefits: Effects of Using Pharmacy Benefit
             Managers on Health Plans, Enrollees, and Pharmacies. GAO-03-196.
             Washington, D.C.: January 10, 2003.

             Federal Employees’ Health Plans: Premium Growth and OPM’s Role in
             Negotiating Benefits. GAO-03-236. Washington, D.C.: December 31, 2002.

             Medicare+Choice: Selected Program Requirements and Other Entities’
             Standards for HMOs. GAO-03-180: Washington, D.C.: October 31, 2002.

             Medigap: Current Policies Contain Coverage Gaps, Undermine Cost
             Control Incentives. GAO-02-533T. Washington, D.C.: March 14, 2002.

             Medicare Subvention Demonstration: Pilot Satisfies Enrollees, Raises
             Cost and Management Issues for DOD Health Care. GAO-02-284.
             Washington, D.C.: February 11, 2002.

             Retiree Health Insurance: Gaps in Coverage and Availability. GAO-02-
             178T. Washington, D.C.: November 1, 2001.

             Medigap Insurance: Plans Are Widely Available but Have Limited
             Benefits and May Have High Costs. GAO-01-941. Washington, D.C.: July
             31, 2001.

             Health Insurance: Proposals for Expanding Private and Public
             Coverage. GAO-01-481T. Washington, D.C.: March 15, 2001.

             Defense Health Care: Pharmacy Copayments. GAO/HEHS-99-134R.
             Washington, D.C.: June 8, 1999.

             Federal Health Programs: Comparison of Medicare, the Federal
             Employees Health Benefits Program, Medicaid, Veterans’ Health
             Services, Department of Defense Health Services, and Indian Health
             Services. GAO/HEHS-98-231R. Washington, D.C.: August 7, 1998.

             Defense Health Care: Offering Federal Employees Health Benefits
             Program to DOD Beneficiaries. GAO/HEHS-98-68. Washington, D.C.:
             March 23, 1998.



(290026)
             Page 54                            GAO-03-547 Military Retiree Health Benefits
                         The General Accounting Office, the audit, evaluation and investigative arm of
GAO’s Mission            Congress, exists to support Congress in meeting its constitutional responsibilities
                         and to help improve the performance and accountability of the federal
                         government for the American people. GAO examines the use of public funds;
                         evaluates federal programs and policies; and provides analyses,
                         recommendations, and other assistance to help Congress make informed
                         oversight, policy, and funding decisions. GAO’s commitment to good government
                         is reflected in its core values of accountability, integrity, and reliability.


                         The fastest and easiest way to obtain copies of GAO documents at no cost is
Obtaining Copies of      through the Internet. GAO’s Web site (www.gao.gov) contains abstracts and full-
GAO Reports and          text files of current reports and testimony and an expanding archive of older
                         products. The Web site features a search engine to help you locate documents
Testimony                using key words and phrases. You can print these documents in their entirety,
                         including charts and other graphics.
                         Each day, GAO issues a list of newly released reports, testimony, and
                         correspondence. GAO posts this list, known as “Today’s Reports,” on its Web site
                         daily. The list contains links to the full-text document files. To have GAO e-mail
                         this list to you every afternoon, go to www.gao.gov and select “Subscribe to daily
                         E-mail alert for newly released products” under the GAO Reports heading.


Order by Mail or Phone   The first copy of each printed report is free. Additional copies are $2 each. A
                         check or money order should be made out to the Superintendent of Documents.
                         GAO also accepts VISA and Mastercard. Orders for 100 or more copies mailed to a
                         single address are discounted 25 percent. Orders should be sent to:
                         U.S. General Accounting Office
                         441 G Street NW, Room LM
                         Washington, D.C. 20548
                         To order by Phone:     Voice:    (202) 512-6000
                                                TDD:      (202) 512-2537
                                                Fax:      (202) 512-6061


                         Contact:
To Report Fraud,
                         Web site: www.gao.gov/fraudnet/fraudnet.htm
Waste, and Abuse in      E-mail: fraudnet@gao.gov
Federal Programs         Automated answering system: (800) 424-5454 or (202) 512-7470


                         Jeff Nelligan, Managing Director, NelliganJ@gao.gov (202) 512-4800
Public Affairs           U.S. General Accounting Office, 441 G Street NW, Room 7149
                         Washington, D.C. 20548