Medicare Subvention Demonstration: Greater Access Improved	 
Enrollee Satisfaction but Raised DOD Costs (31-OCT-01,		 
GAO-02-68).							 
								 
In the Balanced Budget Act of 1997, Congress established a 3-year
demonstration, called Medicare subvention, to improve access to  
Medicare-eligible military retirees for care at military	 
treatment facilities (MTF). The demonstration allowed retirees to
get their health care largely at MTF by enrolling in a Department
of Defense (DOD) Medicare managed care organization known as	 
TRICARE Senior Prime. During the subvention demonstration, access
to health care for many retirees who enrolled in Senior Prime	 
improved, while access to MTF care for some of those who did not 
enroll declined. Many enrollees said that, compared with their	 
experience before Senior Prime, they were better able to get care
when they needed it. In addition, they reported that access to	 
doctors in general as well as care at MTF's improved. The results
of enrollees' improved access and high utilization were mixed.	 
Enrollees generally were more satisfied with their care than	 
before the demonstration. However, the demonstration did not	 
improve enrollees' self-reported health status. In addition,	 
compared to nonenrollees, enrollees did not have better health	 
outcomes, as measured by their mortality rates and rates of	 
"preventable" hospitalizations. Moreover, DOD's costs of care	 
were high, reflecting enrollees' heavy use of hospitals and	 
doctors. These costs were significantly higher than Medicare's	 
costs for comparable fee-for-service beneficiaries.		 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-02-68						        
    ACCNO:   A02387						        
  TITLE:     Medicare Subvention Demonstration: Greater Access	      
Improved Enrollee Satisfaction but Raised DOD Costs		 
     DATE:   10/31/2001 
  SUBJECT:   Health care costs					 
	     Health care programs				 
	     Managed health care				 
	     Retirement benefits				 
	     Retired military personnel 			 
	     DOD TRICARE Extra Program				 
	     DOD TRICARE Prime Program				 
	     DOD TRICARE Program				 
	     DOD TRICARE Senior Prime Program			 
	     DOD TRICARE Standard Program			 
	     Medicare Choice Program				 
	     Medicare Program					 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-02-68
     
A

Report to Congressional Committees

October 2001 MEDICARE SUBVENTION DEMONSTRATION

Greater Access Improved Enrollee Satisfaction but Raised DOD Costs

GAO- 02- 68

Letter 1 Results in Brief 2 Background 3 Senior Prime Enrollees Got More
Care While Some Nonenrollees

Were Crowded Out of MTFs 8 Improved Enrollee Access Resulted In Higher
Patient Satisfaction

but Costs for DOD Were High 17 Concluding Observations 20 Agency Comments 20

Appendixes

Appendix I: Health Care Survey of Subvention Demonstration Beneficiaries 24
Panel Study 24 Cross Section Study 26 Questionnaire Design 28

Appendix II: Measures of Access to Care And Satisfaction With Care 33
Measures of Change in Access and Satisfaction 33 Measures of Access and
Satisfaction at the End of the

Demonstration 42

Appendix III: Models of Utilization 64 Data 64 Models of Risk- Adjusted
Utilization 65 Acute Hospitalization Model 66 Outpatient Physician Visit
Model 67 Profile of Sites? Utilization and Risk Factors 68

Appendix IV: Crowd- Out of Nonenrollees 72 Change in Satisfaction With
Access to Military Care Among

Crowd- Outs 72 Trend in Utilization Among Crowd- Outs 73

Appendix V: Health Outcomes Analysis 74 Mortality Analysis 74 Health Status
Analysis 76 Preventable Hospitalizations 79

Appendix VI: Comments From the Department of Defense 80

Appendix VII: Comments From the Centers for Medicare and Medicaid Services
82

Appendix VIII: GAO Contacts and Staff Acknowledgments 83 GAO Contacts 83

Related GAO Products 84 Tables Table 1: Enrollment at Medicare Subvention
Demonstration Sites 7

Table 2: Enrollees? Change in Access to Health Care 8 Table 3: Change in
Enrollees? Views of the Convenience of Their

Doctors? Hours and Doctors? Timeliness 10 Table 4: Usual Waiting Time to See
Doctor Varied by Type of Visit for Senior Prime Enrollees 11

Table 5: Change in Satisfaction With Doctors for Enrollees 18 Table 6:
Sample and Population Sizes in the Panel 25 Table 7: Reasons for Nonresponse
in the Panel Sample 25 Table 8: Sample and Population Sizes in the Cross
Section 27 Table 9: Reasons for Nonresponse in the Cross Section Sample 27
Table 10: Survey Questions Used in This Report 29 Table 11: Net Improvement
in Access and Quality 33 Table 12: Level of Access and Quality at the End of
the

Demonstration 44 Table 13: Comparison of Senior Prime Enrollees With
Medicare Fee- for- Service Beneficiaries in the Demonstration Areas in 1999
65

Table 14: Estimated Effects of Selected Factors on Acute Hospitalization of
Medicare Fee- for- Service Beneficiaries 67 Table 15: Estimated Effects of
Certain Factors on Medicare

Fee- for- Service Outpatient Utilization 68 Table 16: Site Profiles of
Senior Prime and Medicare Fee- for- Service

Utilization in 1999 70 Table 17: Change in MTF Utilization Among
Nonenrollees 72

Table 18: Change in Self- Reported Access to MTF Care and Satisfaction With
MTF Care Among Crowd- Outs 73 Table 19: Changes in Utilization of
Nonenrollees Who Were Crowded

Out of MTFs 73 Table 20: Profile of 2- Year Mortality Rate 74 Table 21:
Factors Affecting 2- Year Mortality Rate 75 Table 22: Profile of SF- 12
Scores 77 Table 23: Factors Affecting Change in SF- 12 Score 77

Figures Figure 1: Enrollees Who Did Not Use MTFs Before Demonstration
Reported Greatest Improvement in Access to MTFs 9

Figure 2: Three- fifths of Retirees Who Had Been Heavy MTF Users Before the
Demonstration Enrolled in Senior Prime 14 Figure 3: Many Nonenrollees Who
Were Heavy MTF Users Before

the Demonstration Were Crowded Out 15 Figure 4: Few Nonenrollees Were Heavy
MTF Users Before the

Demonstration 16 Figure 5: Medicare Fee- for- Service Care Did Not Offset
Drop in MTF Care for Nonenrollees Who Were Crowded Out 17

Abbreviations

BBA Balanced Budget Act of 1997 CMS Centers for Medicare and Medicaid
Services COPD Chronic Obstructive Pulmonary Disease DOD Department of
Defense ESRD end- stage renal disease FFS fee- for- service HCC Hierarchical
Coexisting Conditions ICD- 9 International Classification of Diseases, Ninth
Revision MTF military treatment facility

Lett er

October 31, 2001 Congressional Committees In the Balanced Budget Act (BBA)
of 1997, 1 Congress established a 3- year demonstration, called Medicare
subvention, designed to improve the access of Medicare- eligible military
retirees to care at military treatment facilities (MTF). Historically,
military retirees age 65 and over have had only limited access to military
health care. Until they turned 65, they could enroll in TRICARE Prime, the
Department of Defense?s (DOD) managed care plan, which gave them priority
access to MTFs. Alternatively, they could use one of DOD?s other plans that
pays part of the cost of civilian health care. However, when they turned 65
and became eligible for Medicare, retirees lost their right to military
health care and could obtain care at MTFs only if space were available after
higher priority beneficiaries were treated. 2

The demonstration allowed retirees 3 to get their care largely at MTFs by
enrolling in a DOD- run Medicare managed care organization known as TRICARE
Senior Prime. Enrollees in Senior Prime could receive the full range of
Medicare services, as well as some additional TRICARE services, and they
would incur minimal out- of- pocket costs. For enrollees, the MTF became the
focal point of their medical care- the source of all their primary care and
much of their specialty care, as well as referrals to civilian network
providers. Those retirees who chose not to enroll could still use MTFs on a
space- available basis. However, given the MTFs? new responsibilities for
treating Senior Prime enrollees, nonenrollees might no

longer be able to get care at MTFs. 1 P. L. 105- 33, sec. 4015. 2 The Floyd
D. Spence National Defense Authorization Act for Fiscal Year 2001 (P. L.
106- 398, sec. 712) established a new program, known as TRICARE For Life,
which started October 1, 2001. Under this program, TRICARE is a secondary
payer for Medicare, paying nearly all

beneficiary cost- sharing for Medicare- covered services obtained from
civilian providers. 3 Throughout this report, we use the term ?retirees? to
refer to military retirees and their dependents and survivors aged 65 and
over.

In establishing the subvention demonstration, the BBA also directed us to
evaluate the demonstration over its initial 3- year period, which ended in
December 2000. 4 Our evaluation covers three major areas: the effect of the
demonstration on retirees? access to care and the quality of care received
by enrollees; the cost of the demonstration to beneficiaries, Medicare, and
DOD; and management and implementation issues encountered during the
demonstration.

This report, focusing on access to care, is one of several addressing these
issues. 5 Our objectives were to examine (1) the effect of the demonstration
on enrollees? and nonenrollees? access to health care and (2) the
consequences of changes in access to care for retirees? satisfaction, health
outcomes, and DOD costs. To address these issues, we surveyed about

20, 000 enrollees and eligible nonenrolled retirees by mail at the beginning
of the demonstration and at the end of 2000. (See app. I.) We supplemented
the surveys with Medicare and DOD administrative data, but did not
independently verify these data. Although the survey data covered the period
of the initial demonstration, the administrative data related primarily to
1999, due to reporting lags. We performed our work according to generally
accepted government auditing standards from June 1998 through September
2001.

Results in Brief During the subvention demonstration, access to health care
for many retirees who enrolled in Senior Prime improved, while access to MTF
care

for some of those who did not enroll declined. Many enrollees said that,
compared with their experience before Senior Prime, they were better able to
get care when they needed it. In addition, they reported that access to

doctors in general as well as to care at MTFs improved. DOD?s and Medicare?s
own data confirmed enrollees? self- reports: They had more hospital stays
and more visits to doctors than before the demonstration. Enrollees also
used more health care than their counterparts in Medicare

fee- for- service. Although the demonstration did not affect most
nonenrollees, access to military health care declined sharply during the 4
The BBA authorized the demonstration to start in January 1998 and run
through December 2000. However, the first site did not become operational
until September 1998. By January 1999, all other sites were operational.
Congress extended the demonstration for an additional year through December
2001. As directed by law, our evaluation covers the period from start- up
through December 2000. 5 A list of related GAO products is included at the
end of the report.

demonstration for the minority of nonenrollees who had relied on MTF care.
The results of enrollees? improved access and high utilization were mixed.

Enrollees generally were more satisfied with their care than before the
demonstration. However, the demonstration did not improve enrollees? self-
reported health status. In addition, compared to nonenrollees,

enrollees did not have better health outcomes, as measured by their
mortality rates and rates of ?preventable? hospitalizations. Moreover, DOD?s
costs of care were high, reflecting enrollees? heavy use of hospitals and
doctors. These costs were significantly higher than Medicare?s costs

for comparable fee- for- service beneficiaries. In commenting on a draft of
this report, DOD stated that the report was accurate and thorough, while
noting some issues concerning the utilization data. The Centers for Medicare
and Medicaid Services (CMS) agreed with the findings of the report.

Background DOD?s health system, TRICARE, currently offers health care
coverage to approximately 6.6 million active duty and retired military
personnel under

age 65 and their dependents and survivors. An additional 1.5 million
retirees aged 65 and over can obtain care when space is available. TRICARE
offers three health plans: TRICARE Standard, a fee- for- service plan;
TRICARE Extra, a preferred provider plan; and TRICARE Prime, a managed care
plan. In addition, TRICARE offers prescription drugs at no

cost from MTF pharmacies and, with co- payments, from retail pharmacies and
DOD?s National Mail Order Pharmacy.

Retirees have access to all of TRICARE?s health plans and benefits until
they turn 65 and become eligible for Medicare. Subsequently, they can only
use military health care on a space- available basis, that is, when MTFs
have unused capacity after caring for higher priority beneficiaries. 6
However, MTF capacity varies from a full range of services at major medical
centers to limited outpatient care at small clinics. Moreover, the amount of
space

available in the military health system has decreased during the last decade
with the end of the Cold War and subsequent downsizing of military bases and
MTFs. Recent moves to contain costs by relying more on military care and
less on civilian providers under contract to DOD have also contributed to
the decrease in space- available care.

Although some retirees age 65 and over rely heavily on military facilities
for their health care, most do not, and over 60 percent do not use military
health care facilities at all. In addition to using DOD resources, retirees
may receive care paid for by Medicare and other public or private insurance
for which they are eligible. However, they cannot use their Medicare
benefits at MTFs, and Medicare is generally prohibited by law from paying
DOD for health care.

Medicare Medicare is a federally financed health insurance program for
persons age 65 and over, some people with disabilities, and people with end-
stage kidney disease. Eligible beneficiaries are automatically covered by
part A,

which covers inpatient hospital, skilled nursing facility, and hospice care,
as well as home health care that follows a stay in a hospital or skilled
nursing facility. They also can pay a monthly premium to join part B, which
covers physician and outpatient services as well as those home health
services not covered under part A. Traditional Medicare allows beneficiaries
to choose any provider that accepts Medicare payment and

requires beneficiaries to pay for part of their care. Most beneficiaries
have supplemental coverage that reimburses them for many costs not covered
by Medicare. Major sources of this coverage include employer- sponsored 6 It
is not yet clear how the new TRICARE For Life program will affect access to
MTF care. In addition to offering military retirees coverage that wraps
around Medicare coverage,

DOD has established a new program, TRICARE Plus, that allows retirees to get
their primary care at MTFs. However, TRICARE Plus coverage of specialist
care is limited. If MTF specialists are not available, TRICARE Plus
enrollees are referred to civilian providers, with Medicare as the primary
payer and TRICARE covering out- of- pocket costs. MTFs can cap enrollment in
TRICARE Plus, and there will not necessarily be space for all who wish to
enroll. Retirees are not allowed to enroll in TRICARE Prime.

health insurance; ?Medigap? policies, sold by private insurers to
individuals; and Medicaid, a joint federal- state program that finances
health care for low- income people. The alternative to traditional Medicare,
Medicare+ Choice, offers beneficiaries the option of enrolling in managed
care or other private health plans. All Medicare+ Choice plans cover basic
Medicare benefits,

and many also cover additional benefits such as prescription drugs.
Typically, these plans have limited cost sharing but restrict members?
choice of providers and may require an additional monthly premium. The
Subvention

Under the Medicare subvention demonstration, DOD established and
Demonstration

operated Medicare+ Choice managed care plans, called TRICARE Senior Prime,
at six sites. 7 Enrollment in Senior Prime was open to military retirees
enrolled in Medicare part A and part B who resided within the plan?s service
area. About 125,000 dual eligibles (military retirees who were also eligible
for Medicare) lived in the 40- mile service areas of the six sites- about
one- fifth of all dual eligibles nationwide living within an MTF?s service
area. DOD capped enrollment at about 28, 000 for the demonstration as a
whole. Over 26, 000 enrolled- about 94 percent of the cap. In addition,
retirees enrolled in TRICARE Prime could ?age in? to Senior Prime upon
reaching age 65, even if the cap had been reached, and

about 6,800 did so. Beneficiaries enrolled in the program paid the Medicare
part B premium, but no additional premium to DOD. Under Senior Prime, all
primary care was provided at MTFs, although DOD purchased some hospital and
specialty care from its network of civilian providers. Senior Prime
enrollees received the same priority for care at the MTFs as younger
retirees enrolled in TRICARE Prime. Care at the MTFs was free of charge for
enrollees, but they had to pay any applicable cost- sharing amounts for care
in the civilian network (for example, $12 for an office visit). 7 See table
1 for a list of the six sites.

The demonstration authorized Medicare to pay DOD for Medicare- covered
health care services provided to retirees at an MTF or through private
providers under contract to DOD. As established in the BBA, capitation
rates- fixed monthly payments for each enrollee- for the demonstration were
discounted from what Medicare would pay private managed care plans in the
same areas. However, to receive payment, DOD had to spend at least as much
of its own funds in serving this dual- eligible population as it had in the
recent past. 8 The six demonstration sites are each in a different TRICARE
region and

include 10 MTFs 9 that vary in size and types of services offered. (See
table 1.) The five MTFs that are medical centers offer a wide range of
inpatient services and specialty care as well as primary care. They
accounted for over 75 percent of all enrollees in the demonstration, and the
two San Antonio medical centers had 38 percent of all enrollees. MTFs that
are community hospitals are smaller, have more limited capabilities, and
could accommodate fewer Senior Prime enrollees. At these smaller facilities,
the civilian network provides much of the specialty care. At Dover, the MTF
is a clinic that offers only outpatient services, thus requiring all
inpatient and

specialty care to be obtained at another MTF or purchased from the civilian
network. 8 For more information on the payment mechanism, see Medicare
Subvention Demonstration: DOD Data Limitations May Require Adjustments and
Raise Broader Concerns (GAO/ HEHS- 99- 39, May 28, 1999). 9 Two sites have
more than one MTF.

Table 1: Enrollment at Medicare Subvention Demonstration Sites Percentage of

Percentage of Demonstration site, location of

Eligible Total

eligible retirees demonstration

military treatment facility Facility type retirees a enrollment b enrolled

wide enrollment Colorado Springs Fort Carson

Community 6,530 2, 371 36 7

Colorado Springs, Colo. hospital U. S. Air Force Academy Community

8,458 1, 750 21 5 Colorado Springs, Colo. hospital Dover c

Dover Air Force Base Clinic 3,894 1, 062 27 3

Dover, Del. Keesler Keesler Air Force Base Medical center 8,309 3, 507 42 11
Biloxi, Miss. Madigan

Fort Lewis Medical center 21, 072 4, 674 22 14

Tacoma, Wash. San Antonio

San Antonio Area Fort Sam Houston

Medical center 21, 354 5, 928 28 18 San Antonio, Tex. Lackland Air Force
Base

Medical center 15, 153 6, 523 43 20 San Antonio, Tex. Texoma Area Sheppard
Air Force Base

Community 2,820 1, 074 38 3

Wichita Falls, Tex. hospital Fort Sill Community

4,873 1, 467 30 4 Lawton, Okla. hospital San Diego San Diego, Calif. Medical
center 34, 485 4, 751 14 14

Total 126, 948 33, 107 26 100 d

Note: Although the law specifies six sites, for the purpose of analysis we
treat the San Antonio area and the Texoma area, which are roughly 300 miles
apart, as separate sites. a As of December 31, 2000.

b As of December 31, 2000. Total enrollment includes age- ins. c Dover dual-
eligibles as of June 1998. d Percentages do not add to 100 due to rounding.

Source: TRICARE Senior Prime Plan Operations Report (Washington, D. C.: DOD,
Dec 31, 2000). The number of eligible retirees (by site and total) is drawn
from DOD?s Defense Enrollment Eligibility Reporting System (DEERS).

Senior Prime Enrollees Compared with their access to care before the
demonstration, many Got More Care While enrollees reported that their access
to care overall- their ability to get care

when they needed it- had improved. They reported better access to MTFs Some
Nonenrollees as well as to doctors. Although at the start of the
demonstration enrollees

Were Crowded Out of had reported poorer access to care than nonenrollees, by
the end of the MTFs

demonstration about 90 percent of both groups said that they could get care
when they needed it. Enrollees? own views are supported by administrative
data: they got more care than they had received from Medicare and DOD
combined before the demonstration. However, most nonenrollees who had relied
on MTFs before the demonstration were no

longer able to rely on military health care. Enrollees Obtained More

Most enrollees reported that their ability to get care when they needed it
Health Care Than Before the was not changed by the demonstration, but those
who did report a change Demonstration were more likely to say that their
access to care- whether at MTFs or from the civilian network- had improved.
10 (See table 2.)

Table 2: Enrollees? Change in Access to Health Care Change in access
Percentage

Improved 32 Stayed the same 54 Declined 14

Tot al 100

Source: GAO Survey of Medicare- Eligible Military Retirees and Family
Members.

When asked specifically about their access to MTF care, those who had not
used MTFs in the past reported the greatest improvement. (See figure 1.)

10 Appendix II contains information by site on access to care and
satisfaction with care.

Figure 1: Enrollees Who Did Not Use MTFs Before Demonstration Reported
Greatest Improvement in Access to MTFs 80 Percentage improvement in MTF
access

70 60 50 40 30 20 10

0 None Some Most All Amount of care at MTFs before demonstration

Note: Improvement in access to care was determined by comparing enrollees?
reports on ability to get needed care at the MTF before the demonstration
and at the end of its initial period.

Source: GAO Survey of Medicare- Eligible Military Retirees and Family
Members.

About one- third of all enrollees said that their access to physicians had
improved, and a significantly smaller fraction said that it had declined.
For example, 32 percent of enrollees said that, under the demonstration,
their primary care doctor?s office hours were more convenient, while 20
percent said they were less so. Similarly, enrollees said that they did not
have to wait too long to get an appointment with a doctor and, once they
reached the office, their doctor saw them more promptly. (See table 3.)

Table 3: Change in Enrollees? Views of the Convenience of Their Doctors?
Hours and Doctors? Timeliness Numbers in percent

Improved Unchanged Declined

Primary care doctor?s 32 48 20

hours convenient Did not have to wait too long for an 35 39 26 appointment
with the primary care doctor Primary care doctor saw me 34 49 17

promptly Source: GAO Survey of Medicare- Eligible Military Retirees and
Family Members.

For two aspects of access, however, Senior Prime enrollees? experience was
mixed. TRICARE has established standards for the maximum amount of time that
should elapse in different situations between making an appointment and
seeing a doctor: 1 month for a well- patient visit, 1 day for an urgent care
visit, and 1 week for routine visits. 11 According to TRICARE

policy, MTFs should meet these standards 90 percent of the time. While
Senior Prime met the standards for the time it took to get an appointment
and see a doctor for well- patient visits (like a physical), it fell
slightly short of the standard for urgent care visits (such as for an acute
injury or illness like a broken arm or shortness of breath) and, more
markedly, for routine

visits (such as for minor injuries or illnesses like a cold or sore throat).
12 (See table 4.) 11 We modified these standards slightly when making a
comparison of DOD?s standards to the responses to our questionnaire, using
30 days for a well- patient visit and less than 3 days for urgent care.

12 We have previously reported that TRICARE Prime also had difficulty in
meeting these goals for active duty and other Prime enrollees. See Defense
Health Care: Appointment Timeliness Goals Not Met; Measurement Tools Need
Improvement (GAO/ HEHS- 99- 168, Sept. 30, 1999).

Table 4: Usual Waiting Time to See Doctor Varied by Type of Visit for Senior
Prime Enrollees

Numbers in percent

Enrollees receiving care Standard within standard

Well- patient visits: less than 30 days 91 Urgent care visits: less than 3
days 87 Routine visits: less than 1 week 69 Source: GAO Survey of Medicare-
Eligible Military Retirees and Family Members.

When asked about their ability to choose their own primary care doctors,
enrollees were somewhat more likely to say that it was more difficult than
before the demonstration. This is not surprising, in view of the fact that
Senior Prime assigned a primary care doctor (or nurse) to each enrollee.

However, regarding specialists, enrollees said that their choice of doctors
had improved.

Enrollees reported fewer financial barriers to access under Senior Prime.
They said that their out- of- pocket spending decreased and was more
reasonable than before. By the demonstration?s end, nearly two- thirds said

that they had no out- of- pocket costs. Even at the smaller demonstration
sites, where care from the civilian network, which required co- payments,
was more common, about half of enrollees said they had no out- of- pocket
costs.

These enrollee reports of better access under Senior Prime are largely
supported by DOD and Medicare administrative data. Enrollees received more
services from Senior Prime than they had obtained before the demonstration
from MTFs and Medicare combined. Specifically, their use of physicians
increased from an average 12 physician visits per year before enrolling in
Senior Prime to 16 visits per year after enrollment, and the

number of hospital stays per person also increased by 19 percent.

Enrollees? use of services not only increased under Senior Prime- as did
other measures of access to care- but exceeded the average level in the
broader community. Enrollees used significantly more care than their
Medicare fee- for- service counterparts. These differences cannot be
explained by either age or health- enrollees were generally younger and
healthier. Adjusted for demographics and health conditions, physician visits
were 58 percent more frequent for Senior Prime enrollees than for their
Medicare counterparts, and hospital stays were 41 percent more frequent.
Nonetheless, enrollees? hospital stays- adjusted for

demographics and health conditions- were about 4 percent shorter. 13 We
found three probable explanations for enrollees? greater use of hospital and
outpatient care:

 Lower cost- sharing. Research confirms the commonsense view that patients
use more care if it is free. 14 Whereas in traditional Medicare the
beneficiary must pay part of the cost of care- for example, 20 percent of
the cost of an outpatient visit 15 -in Senior Prime all primary care and
most specialty care is free.

 Lack of strong incentives to limit utilization. Although MTFs generally
tried to restrain inappropriate utilization, they did not have strong
financial incentives to do so. MTFs cannot spend more than their budget, but
space- available care acts as a safety valve: that is, when

costs appear likely to exceed funding, space- available care can be reduced
while care to Senior Prime enrollees remains unaffected. MTFs also had no
direct incentive to limit the use of purchased care, which is funded
centrally, and the managed care contractors also lacked an

incentive, since they were not at financial risk for Senior Prime. 13 See
appendix III for our analysis of utilization. 14 See Joseph P. Newhouse,
Free for All: Lessons From the RAND Health Insurance Experiment (Cambridge,
MA: Harvard University Press, 1993). More recent evidence shows that
beneficiaries with supplemental insurance covering most or all of their
Medicare costsharing requirements have higher Medicare utilization and
spending than otherwise similar people with Medicare coverage only. See
Sandra Christensen and Judy Shinogle, ?Effects of Supplemental Coverage on
Use of Services by Medicare Enrollees,? Health Care Financing Review, Fall
1997, and Physician Payment Review Commission, Annual Report to Congress,
1997, Chapter 15.

15 Most Medicare beneficiaries have supplemental insurance, such as Medigap
or employer coverage, that moderates the effect of Medicare cost- sharing
requirements.

 Practice styles. Military physicians? training and experience, as well as
the practice styles of their colleagues, also affect their readiness to
hospitalize patients as well as their recommendations to patients about
follow- up visits and referrals to specialists. 16 Studies have shown that
the military health system has higher utilization than the private sector.
17 Given that military physicians tend to spend their careers in the
military with relatively little exposure to civilian health care?s
incentives and

practices, it is not surprising that these patterns of high use would
persist.

Some Nonenrollees Could Although nonenrollees generally were not affected by
the demonstration,

No Longer Use MTFs the minority who had been using space- available MTF care
were affected

because space- available care declined. This decline is shown in our survey
results, and is confirmed by DOD?s estimate of the cost of space- available
care, which decreased from $183 million in 1996 18 to $72 million in 1999,
the first full year of the demonstration. However, for most nonenrollees,
this decline was not an issue, because they did not use MTFs either before
or

during the demonstration. Furthermore, of those who depended on MTFs for all
or most of their care before the demonstration, most enrolled in Senior
Prime, thereby assuring their continued access to care. (See figure 2.)

16 In civilian health care, much of the variation in use of health care
among states and counties is attributed to the clinical practice styles of
their physicians. See W. P. Welch and others, ?Geographic Variation in
Expenditures for Physician Services in the United States,? New England
Journal of Medicine, Vol. 328, No. 621 (Mar. 4, 1993); John E. Wennberg and
Alan Gittelsohn, ?Small Area Variations in Health Care Delivery,? Science
Vol. 182, No. 4117 (Dec. 1973); and The Quality of Medical Care in the
United States: A Report on the Medicare Program (American Hospital
Association, 1999). 17 See Susan D. Hosek and others, The Demand for
Military Health Care: Supporting Research for a Comprehensive Study of the
Military Health Care System (Santa Monica,

Calif.: RAND, MR- 407- PA& E, Jan. 1994) and The Institute for Defense
Analysis and Center for Naval Analysis Corporation, Evaluation of the
TRICARE Program: FY1998 Report to Congress (Washington, D. C.: 1998). 18 In
1999 dollars.

Figure 2: Three- fifths of Retirees Who Had Been Heavy MTF Users Before the
Demonstration Enrolled in Senior Prime

38% Nonenrollees

62% Enrollees

Note: Data are based on survey respondents who reported receiving all or
most of their care at the MTF before the demonstration.

Sources: GAO Survey of Medicare- Eligible Military Retirees and Family
Members and Iowa Foundation for Medical Care enrollment file.

Since there was less space- available care than in the past, many of those
who had previously used MTFs and did not enroll in Senior Prime were
?crowded- out.? Crowd- out varied considerably, depending both on the

types of services that nonenrollees needed and the types of physicians and
space available at MTFs. Nonenrollees who required certain services were
crowded out while others at the same MTF continued to receive care. We focus
on nonenrollees who experienced a sharp decline in MTF care: those

who said they had received most or all of their care at MTFs before the
demonstration but got no care or only some care at MTFs during the
demonstration. Of those nonenrollees who had previously depended on MTFs for
their care, over 60 percent (about 4, 600 people) were crowded out. (See
figure 3.)

Figure 3: Many Nonenrollees Who Were Heavy MTF Users Before the
Demonstration Were Crowded Out

36%

Crowded out

64% Heavy MTF use before demonstration; little or none during. Heavy MTF use
before and during demonstration.

Note: Data are based on nonenrollees who received all or most of their care
at the MTF before the demonstration.

Source: GAO Survey of Medicare- Eligible Military Retirees and Family
Members.

The small number of nonenrollees- 10 percent of the total- that had depended
on MTFs for their care before the demonstration limited crowdout. (See
figure 4.) Consequently, only a small proportion of all nonenrollees- about
6 percent- was crowded out. 19 Somewhat surprisingly, a small number of
nonenrollees who had not previously used MTFs began obtaining all or most of
their care at MTFs.

19 Using a stricter definition- those who had previously received all of
their care at the MTF before the demonstration and got no MTF care during
the last year of the demonstration- about 1, 500 or 2 percent of all
nonenrollees were crowded out. This represents just over one- third of those
who had previously depended on MTFs for all their care. For a further
discussion of the range of estimates of crowd- out, see appendix IV.

Figure 4: Few Nonenrollees Were Heavy MTF Users Before the Demonstration

90% 10% Received some or no care at MTF. Received most or all care at MTF.

Note: Data are based on all nonenrollees? reports on the amount of care
received at the MTF before the demonstration.

Source: GAO Survey of Medicare- Eligible Military Retirees and Family
Members.

Although Medicare fee- for- service care increased for those who were
crowded out of MTF care, the increase in Medicare outpatient care was not
nearly large enough to compensate for the loss of MTF care. (See figure 5.)
Retirees who were crowded out had somewhat lower incomes than other
nonenrollees and were also less likely to have supplemental insurance,

suggesting that some of them may have found it difficult to cover Medicare
out- of- pocket costs. By the end of the initial demonstration period, less
than half of all nonenrollees said they were able to get care at MTFs when
they needed it, a modest decline from before the demonstration.

Figure 5: Medicare Fee- for- Service Care Did Not Offset Drop in MTF Care
for Nonenrollees Who Were Crowded Out Number of outpatient physician visits
12

11 10

9 8 7 6 5 4 3 2 1 0

Before During demonstration

demonstration

Medicare fee- for- service care MTF care Sources: GAO survey of Medicare-
Eligible Military Retirees and Family Members, and DOD and Medicare fee-
for- service encounter data (1997- 1999).

Improved Enrollee Enrollees? improved access to care had both positive and
negative

Access Resulted In consequences. Many enrollees in Senior Prime reported
that they were more satisfied with nearly all aspects of their care. Some
results were Higher Patient neutral: enrollees? self- reported health status
did not change and health

Satisfaction but Costs outcomes, such as mortality and preventable
hospitalizations, were no for DOD Were High

better than those achieved by nonenrolled military retirees. However,
enrollees? heavy use of health services resulted in high per- person costs
for DOD compared to costs of other Medicare beneficiaries. Enrollee
Satisfaction Satisfaction with almost all aspects of care increased for
enrollees. Improved Moreover, by the end of the demonstration, their
satisfaction was generally as high as that of nonenrollees.

Patients? sense of satisfaction or dissatisfaction with their physicians
reflects in part their perceptions of their physicians? clinical and
communication skills. Under Senior Prime, many enrollees reported

greater satisfaction with both their primary care physicians and
specialists. Specifically, enrollees reported greater satisfaction with
their physicians? competence and ability to communicate- to listen, explain,
and answer questions, and to coordinate with other physicians about
patients? care. 20 (See table 5.) Table 5: Change in Satisfaction With
Doctors for Enrollees

Numbers in percent

Improved Unchanged Declined Quality- primary care doctor

Received excellent care 30 56 14 Thorough examination 33 48 19 Careful in
taking medical

33 52 16 history Spent enough time with me 34 48 18

Skillful and competent 30 52 18

Communication- primary care doctor

Explained things clearly 33 47 20 Really listened 31 51 18

Quality- specialist

Skillful and competent 25 57 18

Communication- specialist

Told me about my treatment 29 50 21 Answered all my questions 27 53 20
Doctors communicated with 32 47 22 one another Source: GAO Survey of
Medicare- Eligible Military Retirees and Family Members.

20 For the complete set of all patient satisfaction measures, see appendix
II.

Demonstration Did Not Senior Prime did not appear to influence three key
measures of health

Affect Health Outcomes outcomes- the mortality rate, self- reported health
status, and preventable

hospitalizations.

 Mortality rate. Although there were slightly more deaths among
nonenrollees, the difference between enrollees and nonenrollees disappears
when we adjust for retirees? age and their health conditions at the start of
the demonstration.

 Health status. We also found that Senior Prime did not produce any
improvement in enrollees? self- reported health status. We base this on
enrollees? answers to our questions about different aspects of their health,
including their ratings of their health in general and of specific areas,
such as their ability to climb several flights of stairs. This finding is
not surprising, given the relatively short time interval- an average of

19 months- between our two surveys. We also found that, like enrollees,
nonenrollees did not experience a significant change in health status.

 Preventable hospitalizations. The demonstration did not have a clear
effect on preventable hospitalizations- those hospitalizations that experts
say can often be avoided by appropriate outpatient care. Among patients who
had been hospitalized for any reason, the rate of preventable
hospitalizations was slightly higher for Senior Prime enrollees than for
their Medicare fee- for- service counterparts. However, when all those with
chronic diseases- whether hospitalized or not-

were examined, the rate among Senior Prime enrollees was lower. 21 Access
and High Utilization

A less desirable consequence of enrollees? access to care was its high cost
Resulted in High Costs for for DOD. Under Senior Prime, DOD?s costs were
significantly higher than

DOD Medicare fee- for- service costs for comparable patients and comparable
benefits. 22 These higher costs did not result from Senior Prime enrollees

being sicker or older than Medicare beneficiaries. Instead, they resulted
from heavier use of hospitals and, especially, greater use of doctors and 21
See appendix V for a discussion of our analyses of health outcomes. 22 On
average, providing Senior Prime enrollees with the Medicare benefits package
(which excludes prescription drugs) cost DOD about $6,400 per person
annually- about 30 percent more than Medicare fee- for- service costs for
comparable people in the demonstration areas. See Medicare Subvention
Demonstration: DOD Costs and Medicare Spending (GAO- 02- 67, Oct. 31, 2001).

other outpatient services. In other words, the increased ability of Senior
Prime enrollees to see physicians and receive care translated directly into
high DOD costs for the demonstration. Concluding From the perspective of
enrollees, Senior Prime was highly successful. Observations

Their satisfaction with nearly all aspects of their care increased, and by
the end of the demonstration enrollees were in general as satisfied as
nonenrollees, who largely used civilian care. However, enrollees?
utilization and the cost of their care to DOD were both higher. Although
subvention is not expected to continue, the demonstration raises a larger
issue for DOD: can it achieve the same high levels of patient satisfaction

that it reached in Senior Prime while bringing its utilization and costs
closer to the private sector?s?

Agency Comments We provided DOD and CMS an opportunity to comment on a draft
of this report, and both agencies provided written comments. DOD said that
the report was accurate. It noted that the report did not compare Senior
Prime enrollees? utilization rates with those of Medicare+ Choice plans and
suggested that our comparison with fee- for- service might be misleading,

because it did not take account of the richer benefit package offered by
Senior Prime. DOD further stated that the utilization data should cover the
full 3 years of the demonstration experience and that utilization might be
higher during the initial phase of a new plan. Finally, DOD stated that

access and satisfaction for TRICARE Prime enrollees were adversely affected
by the demonstration. CMS agreed with the report?s findings and suggested
that higher quality of care might be an explanation for Senior Prime
enrollees? higher use of services. (DOD and CMS comments appear in
appendixes VI and VII.)

In comparing utilization rates with Medicare fee- for- service in the same
areas, we chose a comparison group that would be expected to have higher
utilization than Senior Prime or any other managed care plan. Fee-
forservice beneficiaries can obtain care from any provider without
restriction, whereas Medicare+ Choice plans typically have some limitations
on access.

Consequently, the fact that Senior Prime utilization was substantially
higher than fee- for- service utilization is striking. As mandated by law,
our evaluation covers the initial demonstration period (through December

2000). We therefore did not attempt to obtain information on utilization
during 2001 and, in any case, the lag in data reporting would have prevented
our doing so. However, during the first 2 full years of the demonstration
utilization declined slightly: outpatient visits in 2000 were 2 percent
lower than in 1999. As we have reported elsewhere, site officials found
little evidence that the demonstration affected TRICARE Prime enrollees?
satisfaction or access to care. 23 Regarding the possible impact of quality
of care on use of services, we examined several health outcome indicators
and found no evidence of such an effect.

We are sending copies of this report to the Secretary of Defense and the
Administrator of the Centers for Medicare and Medicaid Services. We will
make copies available to others upon request. If you or your staffs have
questions about this report, please contact me at (202) 512- 7114. Other GAO
contacts and staff acknowledgments are listed in appendix VIII.

William J. Scanlon Director, Health Care Issues

23 See Medicare Subvention Demonstration: DOD?s Pilot HMO Appealed to
Seniors, Underscored Management Complexities (GAO- 01- 671, June 14, 2001).

List of Addressees

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

The Honorable Max Baucus Chairman The Honorable Charles E. Grassley Ranking
Minority Member Committee on Finance United States Senate

The Honorable Bob Stump Chairman The Honorable Ike Skelton Ranking Minority
Member Committee on Armed Services House of Representatives

The Honorable W. J. ?Billy? Tauzin Chairman The Honorable John D. Dingell
Ranking Minority Member Committee on Energy and Commerce House of
Representatives

The Honorable William M. Thomas Chairman The Honorable Charles B. Rangel
Ranking Minority Member Committee on Ways and Means House of Representatives

Appendi xes Health Care Survey of Subvention

Appendi x I

Demonstration Beneficiaries To address the questions Congress asked about
Medicare subvention, we fielded a mail survey of military retirees and their
family members who were eligible for the subvention demonstration. The
survey had two interlocking components: a panel of enrollees and
nonenrollees, who were surveyed both at the beginning and the end of the
demonstration, and two

cross sections or snapshots of enrollees and nonenrollees- one taken at the
beginning of the demonstration and the other at the end. Panel Study To
assess those questions that involved change over time, we sampled and

surveyed by mail enrollees and nonenrollees, stratified by site, at the
beginning of the demonstration. These same respondents were resurveyed from
September through December 2000, shortly before the demonstration?s initial
period ended. 1 Because a prior report describes our initial survey, this
appendix focuses on our second survey. 2

Sample Design To conduct the second round of data collection, we began with
15,223 respondents from the first round of surveys. To be included in the
panel, three criteria had to be met: (1) the person must still be alive, (2)
the person

must still reside in an official demonstration area, and (3) the person must
have maintained the same enrollment status, that is, enrolled or not
enrolled. Based on these criteria we mailed 13, 332 surveys to our panel
sample of enrollees and nonenrollees. Response Rates Starting with a sample
of 13, 332 retirees and their family members, we

obtained usable questionnaires from 11,986 people, an overall response rate
of 91 percent. (See table 6, which also shows the adjustments to the initial
sample and to the estimated population size. See table 7 for the reasons for
nonresponse.)

1 The demonstration was initially authorized for 3 years, ending December
2000, but it was extended for 1 additional year. 2 For a full discussion of
the first survey, see Medicare Subvention Demonstration: Enrollment in DOD
Pilot Reflects Retiree Experiences and Local Markets (GAO/ HEHS- 0035, Jan.
31, 2000).

Table 6: Sample and Population Sizes in the Panel Sample Estimated
population

Initial unadjusted size 13,332 92, 669 Exclusions Died before sampling 89
591

Final size 13,243 92, 078 Respondents 11,986 Nonrespondents 1, 257

Response rate (percentage) 91 Table 7: Reasons for Nonresponse in the Panel
Sample

Number of people excluded from final Reason

response

No information received 923 Moved out of the demonstration area or 138
undeliverable address Refused 97

Too sick to respond 27 Died after sampling 41 Others 31

Total nonresponse 1, 257

Cross Section Study To enable comparisons between enrollees and nonenrollees
at the end of the demonstration, the second survey was augmented to include
persons who had enrolled since the first survey as well as additional
nonenrollees. The overall composition of the Senior Prime enrollee
population had changed from the time of our first survey. When we drew our
second sample in July 2000, 36 percent of all enrollees were new- that is,
they had enrolled since our first survey- and over two- fifths of them were
age- ins who had turned 65 since the demonstration started. From the time of
our first survey to the time of our second survey, only 861 people had

disenrolled from Senior Prime. Therefore, we surveyed all voluntary 3
disenrollees. Data from all respondents- those we surveyed for the first
time as well as those in the panel- were weighted, to yield a representative
sample of the demonstration population at the end of the program. Sample
Design The sample for the cross section study included the panel sample as
well as the augmented populations. We defined our population as all
Medicareeligible military retirees living in the demonstration sites and
eligible for Senior Prime. The sample of new enrollees was drawn from all
those

enrolled in the demonstration according to the Iowa Foundation?s 4
enrollment files. The supplemental sample of nonenrollees was drawn from all
retirees age 65 and over in the Defense Enrollment Eligibility Reporting
System who (1) had both Medicare part A and part B coverage, (2) lived
within the official demonstration zip codes, (3) were not enrolled in Senior
Prime, and (4) were not part of our first sample. We stratified our sample
of new enrollees and new nonenrollees by site and by whether they aged in.
We oversampled each stratum to have a large enough number to conduct
analyses of subpopulations. 5 The total sample for all sites was 23, 967,
drawn from a population of 117,618.

3 Voluntary disenrollees are persons who chose to disenroll from Senior
Prime. Those who died, moved out of the service area, or lost their Medicare
part A or part B coverage are excluded.

4 The Iowa Foundation for Medical Care is a DOD contractor that handled
enrollment. 5 We specified a sample size sufficient to detect a minimum
difference of 5 percent between enrollees and nonenrollees at each site,
using a 95- percent confidence interval, with a power of 0.8 (the
probability of rejecting the null hypothesis when it is false).

Response Rates Starting with a sample of 23, 967 retirees and their family
members, we obtained complete and usable questionnaires from 20,870 people,
an overall response rate of 88 percent. (See table 8, which also shows the
adjustments to the initial sample and to the estimated population size. See
table 9, which shows the reasons for nonresponse.) Response rates varied

across sites and subpopulations. Rates ranged from 95.3 percent among aged-
in new enrollees to 66.7 percent among disenrollees.

Table 8: Sample and Population Sizes in the Cross Section Sample Estimated
population

Initial unadjusted size 23,967 117, 618 Exclusions Died before sampling 120
660

Moved out of official 14 37

subvention zip code before sampling

Final size 23,833 116, 921 Respondents 20,870 Nonrespondents 2, 963

Response rate (percentage) 88 Table 9: Reasons for Nonresponse in the Cross
Section Sample

Number of people excluded Reason from final response

No information received 2, 275 Moved/ undeliverable address 310 Refused 240
Too sick to respond 42 Died after sampling 61 Others 35

Total nonresponse 2, 963

Questionnaire Design The original questionnaire that was sent to our panel
sample was created based on a review of the literature and five existing
survey instruments. In addition, we pretested the instrument with several
retiree groups. For the

second round of data collection, we created four different versions of the
questionnaire, based on the original questionnaire. The four versions were
nearly the same, with some differences in the sections on Senior Prime and

health insurance coverage. (See table 10 for a complete list of all the
survey questions used in our analyses.)

For the panel sample, our objective was to collect the same data at two
points in time. Therefore, in constructing the questionnaires for the panel
enrollees and panel nonenrollees we essentially used the same instrument as
the original survey to answer questions about the effect of the
demonstration on access to care, quality of care, health care use, and
outof- pocket costs. However, we modified our questions about plan
satisfaction and health insurance coverage.

In constructing the questionnaires for the new enrollees, we generally
adopted the same questions in the panel enrollee instrument to measure
access to care, quality of care, health care use, and out- of- pocket costs.
However, we also asked the new enrollees about their health care experiences
in the 12 months before they joined Senior Prime. For new

nonenrollees, we were able to use the same instrument as we had used for the
panel nonenrollees, because their health care experiences were not related
to tenure in Senior Prime. Finally, the disenrollee questionnaire, like the
other versions, did not change from the original instrument in the measures
on access to care, quality of care, health care use, and out- ofpocket
costs. However, we added questions on the reasons for disenrollment.

Table 10: Survey Questions Used in This Report Question Possible answers
Access to care

I could get health care when I needed it. Strongly agree, Agree, Neither
agree nor disagree, Disagree, Strongly disagree, Not applicable I could not
get medical information by phone when I needed it. Strongly agree, Agree,
Neither agree nor disagree, Disagree, Strongly disagree, Not applicable I
could not get care when I needed it at night or on weekends. Strongly agree,
Agree, Neither agree nor disagree, Disagree, Strongly disagree, Not
applicable When you went to a civilian or military place during the past 12
Same Day, 1- 3 Days, 4- 7 Days, 8- 14 Days, 15- 30 Days, 31- 60 months, how
long did you USUALLY wait between the time you Days, More than 60 days, Does
not apply made an appointment for care and the day you actually saw a doctor
or other health care professional? When going for:

Well- patient visit (like a physical) Routine visit for minor illness or
injury (like a cold or sore throat) Urgent care visit for an acute injury or
illness (like a broken arm or shortness of breath)

Access to military care

During the past 12 months, NOT including getting prescriptions None, Some,
Most, All

filled, about how much of your health care was at military health care
facilities?

I was able to get care at military health care facilities when I needed
Strongly agree, Agree, Neither agree nor disagree, Disagree, it during the
past 12 months. Strongly disagree, Not applicable

I prefer to get my health care at military health care facilities. Strongly
agree, Agree, Neither agree nor disagree, Disagree, Strongly disagree, Not
applicable It was difficult for me to schedule appointments at military
health

Strongly agree, Agree, Neither agree nor disagree, Disagree, care facilities
during the past 12 months. Strongly disagree, Not applicable

Access to a primary care doctor

My primary care doctor?s office was conveniently located. Strongly agree,
Agree, Neither agree nor disagree, Disagree, Strongly disagree The hours
when my primary care doctor?s office was open were not

Strongly agree, Agree, Neither agree nor disagree, Disagree, convenient for
me. Strongly disagree I had to wait too long between making an appointment
and seeing

Strongly agree, Agree, Neither agree nor disagree, Disagree, the doctor.
Strongly disagree I was able to choose my own primary care doctor. Strongly
agree, Agree, Neither agree nor disagree, Disagree, Strongly disagree Once I
got to the office, my doctor saw me promptly. Strongly agree, Agree, Neither
agree nor disagree, Disagree, Strongly disagree

Access to a specialist doctor

I was satisfied with the choice of specialists available to me. Strongly
agree, Agree, Neither agree nor disagree, Disagree, Strongly disagree

(Continued From Previous Page)

Question Possible answers

I did not have to wait a long time between making an appointment Strongly
agree, Agree, Neither agree nor disagree, Disagree, and seeing the
specialist. Strongly disagree

Quality of care

I am satisfied with the health care that I received. Strongly agree, Agree,
Neither agree nor disagree, Disagree, Strongly disagree

Quality of military care

I am satisfied with the health care I received at military health care
Strongly Agree, agree, Neither agree nor disagree, Disagree, facilities
during the past 12 months. Strongly disagree

Doctors and staff at military health care facilities did not treat me
Strongly agree, Agree, Neither agree nor disagree, Disagree, with courtesy
and respect during the past 12 months. Strongly disagree I would not
recommend military health care to my family or friends Strongly agree,
Agree, Neither agree nor disagree, Disagree, who need care. Strongly
disagree

Quality of primary care doctor

I received excellent care from my primary care doctor. Strongly agree,
Agree, Neither agree nor disagree, Disagree, Strongly disagree My primary
care doctor examined me thoroughly. Strongly agree, Agree, Neither agree nor
disagree, Disagree, Strongly disagree My primary care doctor was very
careful in taking and

Strongly agree, Agree, Neither agree nor disagree, Disagree, understanding
my medical history. Strongly disagree My primary care doctor did not explain
things clearly. Strongly agree, Agree, Neither agree nor disagree, Disagree,
Strongly disagree My primary care doctor really did not listen to me.
Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly
disagree My primary care doctor did not spend enough time with me during
Strongly agree, Agree, Neither agree nor disagree, Disagree, visits.
Strongly disagree My primary care doctor was not skillful and competent.
Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly
disagree

Quality of specialist doctors

My specialists were skillful and competent. Strongly agree, Agree, Neither
agree nor disagree, Disagree, Strongly disagree My specialists did not tell
me all I wanted to know about my Strongly agree, Agree, Neither agree nor
disagree, Disagree, condition or treatment. Strongly disagree My specialists
did not answer all of my questions. Strongly agree, Agree, Neither agree nor
disagree, Disagree, Strongly disagree My doctors did not communicate with
each other about my care. Strongly agree, Agree, Neither agree nor disagree,
Disagree, Strongly disagree

Health status

In general, would you say your health status is: Excellent, Very Good, Good,
Fair, Poor

(Continued From Previous Page)

Question Possible answers

The following questions are about activities you might do during a Yes,
limited a lot; Yes, limited a little; No, not limited at all.

typical day. Does your health now limit you in these activities? If so how
much? a. Moderate activities, such as moving a table, pushing a vacuum
cleaner, bowling, or playing golf? b. Climbing several flights of stairs
During the past 4 weeks, have you had any of the following Ye s , N o
problems with your work or other regular daily activities as a result of
your physical health?

a. Accomplished less than you would like b. Were limited in the kind of work
or other activities During the past 4 weeks, have you had any of the
following Ye s , N o problems with your work or other regular daily
activities as a result of any emotional problems (such as feeling depressed
or anxious)?

a. Accomplished less than you would like b. Di dn?t do work or other
activities as carefully as usual During the past 4 weeks, how much did pain
interfere with your

Not at all, A little bit, Moderately, Quite a bit, Extremely normal work
(including both work outside the home and housework)?

How much of the time during the past 4 weeks? All of the time, Most of the
time, A good bit of the time, Some of the

time, A little of the time, None of the time a. Have you felt calm and
peaceful? b. Did you have a lot of energy? c. Have you felt downhearted and
blue? During the past 4 weeks, how much of the time has your physical

All of the time, Most of the time, A good bit of the time, Some of the
health or emotional problems interfered with your social activities time, A
little of the time, None of the time (like visiting with friends, relatives,
etc.)?

Compared to one year ago, how would you rate your health in Much better than
one year ago, Somewhat better than one year general now? ago, About the
same, Somewhat worse now than one year ago, Much worse now than one year ago

Because of your health, do you need help from another person with Ye s , N o
activities such as eating, bathing, dressing, or getting around the house?

(Continued From Previous Page)

Question Possible answers Insurance and income

Do you currently have Medicare supplemental insurance? Yes, No, Don?t know
What was your family?s TOTAL income last year BEFORE taxes? Less than $20,
000, $20,000-$ 39, 999, $40,000-$ 59, 999, $60,000$ (Include wages before
taxes; dividends; interest; social security;

79,999, $80,000 and over pensions; alimony; net business or farm income; and
any other money income received by members of the family who are 15 years of
age or older.)

Note: Our questionnaire included the SF- 12 ? Health Survey. Reproduced with
permission of the Medical Outcomes Trust. Copyright ï¿½ 1994 the Health
Institute, New England Medical Center.

Measures of Access to Care And Satisfaction

Appendi x II

With Care Measures of Change in To detect the effects the demonstration had
on both enrollees? and Access and nonenrollees? access to care and
satisfaction with care, we compared the

differences between survey responses at both points in time and among
Satisfaction

each demonstration site. For most questions, retirees were asked both before
the demonstration and at the end of the demonstration how much they agreed
or disagreed with each statement. They were given five possible answers:
strongly agree, agree, neither agree nor disagree,

disagree, and strongly disagree. To calculate change, responses were
assigned a numeric value on a five- point scale, with five being the highest
and one being the lowest. To properly quantify the response, some scales had
to be reversed. Where necessary, questions were rescaled so that ?agree?
represents a positive answer and ?disagree? a negative answer. To obtain a
measure of change, the value of the response from the first survey

was subtracted from the value of the response from the second survey. A
positive value indicates improvement, a negative value indicates decline.
The net improvement is calculated as the difference between the proportion
of respondents within each sample population who improved and the proportion
of those who declined.

Four separate significance tests were performed. (See table 11.) The first
test was for net improvement (the difference between improved and declined)
among enrollees. The second test was for net improvement among nonenrollees.
The third test was for the difference of net improvement between enrollees
and nonenrollees. Finally, we tested

whether the net improvement for each site is significantly different from
the net improvement of the other sites. (See tables 11 and 12.) Table 11:
Net Improvement in Access and Quality

Numbers in percent

Enrollees Nonenrollees Difference in net

improvement Net

Net between enrollees

Question and site Improved Declined change Improved Declined change and
nonenrollees

Access to care

I was able to get care when I needed it. 32 14 18 a 18 22 -4 a 22 a Madigan
24 17 7 a 18 25 -7 a 14 a San Antonio 38 12 26 a 19 22 -3 29 a San Diego 26
13 13 a 16 22 -6 a 19 a

(Continued From Previous Page)

Numbers in percent

Enrollees Nonenrollees Difference in net

improvement Net

Net between enrollees

Question and site Improved Declined change Improved Declined change and
nonenrollees

Keesler 27 17 10 a 20 20 0 10 Texoma 30 14 16 a 17 20 -3 19 a Colorado 34 15
19 a 20 20 0 19 a Dover 30 20 10 a 18 16 2 8 a

I was able to get info by phone when I 37 26 11 a 30 33 -3 12 a needed it. b
Madigan 31 26 5 30 28 2 3

San Antonio 37 27 10 30 33 -3 13 San Diego 38 22 16 a 29 35 -6 22 a Keesler
37 271032 284 6 Texoma 38 28 10 28 32 -4 14 Colorado 43 26 17 a 26 35 -9 26
a Dover 43 25 18 a 34 30 4 14

I was able to get care on nights and 34 23 11 a 29 30 -1 12 a

weekends. b Madigan 31 20 11 a 31 28 3 8 San Antonio 34 27 7 27 34 -7 14 San
D iego 31 191229 290 12 Keesler 33 221137 2611 0 Texoma 35 23 12 28 27 1 11
Colorado 41 18 23 a 26 32 -6 29 a Dover 46 11 35 a 33 24 9 26 a

Military care

I am satisfied with care at military facilities. 26 14 12 a 20 29 -9 a 21 a
Madigan 21 17 4 a 17 27 -10 a 14 a San Antonio 29 13 16 a 20 33 -13 a 29 a
San Diego 19 11 8 a 22 24 -2 10 Keesler 24 18 6 a 25 27 -2 8 Texoma 26 13 13
a 22 27 -5 18 a Colorado 27 16 11 a 22 29 -7 18 a Dover 27 13 14 a 19 31 -12
26 a

(Continued From Previous Page)

Numbers in percent

Enrollees Nonenrollees Difference in net

improvement Net

Net between enrollees

Question and site Improved Declined change Improved Declined change and
nonenrollees

I was able to get care at military facilities 35 12 23 a 19 33 -14 a 37 a
when I needed it. Madigan 26 15 11 a 21 31 -10 a 21 a

San Antonio 40 13 27 a 20 38 -18 a 45 a San Diego 25 9 14 a 14 31 -17 a 31 a
Keesler 36 15 21 a 18 29 -11 32 a Tex oma 38 8 30 a 23 22 1 29 a Colorado 42
12 30 a 23 26 -3 33 a Dover 29 11 18 a 22 33 -11 29 a

I prefer to get my care at military facilities. 18 13 4 a 24 25 -1 5 a
Madigan 16 14 2 23 22 1 1 San Antonio 20 13 7 a 25 28 -3 10 a San Diego 16 9
7 a 22 22 0 7 Keesler 12 17 5 a 24 29 -5 10 Texoma 19 14 5 a 20 32 -12 a 17
a Colorado 17 13 4 a 25 23 2 6 Dover 21 13 8 a 22 25 -3 11

It was not difficult to schedule 47 17 30 a 26 35 -9 a 39 a

appointments at military facilities. b Madigan 35 22 13 a 23 36 -13 a 26 a
San Antonio 56 13 43 a 28 33 -5 48 a San Diego 27 19 8 a 20 48 -28 a 36 a
Keesler 47 18 29 a 23 25 -2 31 a Texoma 49 16 33 a 28 29 -1 34 a Colorado 46
21 25 a 39 23 16 a 9 Dover 40 22 18 a 17 41 -24 a 42 a

Doctors and staff treated me with respect 21 16 5 a 18 32 -14 a 19 a at
military facilities. b Madigan 18 16 2 13 34 -21 a 23 a

San Antonio 24 15 9 a 20 29 -9 18 a

(Continued From Previous Page)

Numbers in percent

Enrollees Nonenrollees Difference in net

improvement Net

Net between enrollees

Question and site Improved Declined change Improved Declined change and
nonenrollees

San Diego 15 13 2 19 33 -14 16 Keesler 16 20- 424 26- 2 -2 Texoma 22 17 5 21
37 -16 21 a Colorado 22 21 1 17 39 -22 a 23 a Dover 19 17 2 14 38 -24 a 26 a

I would recommend military care. b 18 16 2 a 25 27 -2 4 Madigan 16 16 0 26
24 2 -2 San Antonio 19 14 5 a 26 27 -1 6 San D iego 14 15- 123 27- 4 3
Keesler 17 20- 323 36- 13 a 10 Texoma 19 16 3 31 29 -2 5 Colorado 20 19 1 27
26 1 0 Dover 23 16 7 22 37 -15 a 22 a

Satisfaction with care

I am satisfied with the care I received. 29 14 15 a 18 22 -4 a 19 a Madigan
21 17 4 19 24 -5 a 9 a San Antonio 33 13 20 a 19 22 -3 23 a San Diego 25 11
14 a 17 22 -5 a 19 a Keesler 23 19 4 a 21 22 -1 5 Texoma 28 12 16 a 16 20 -4
20 a Colorado 38 13 25 a 18 21 -3 28 a Dover 29 17 12 a 18 17 1 11 a

Satisfaction with primary care providers

I received excellent care. 30 14 16 a 17 20 -3 a 19 a Madigan 26 16 10 a 20
23 -3 13 a San Antonio 32 13 19 a 18 19 -1 20 a San Diego 26 12 14 a 14 20
-6 a 20 a Keesler 22 19 3 20 16 4 -1 Texoma 29 11 18 a 19 17 2 16 a

(Continued From Previous Page)

Numbers in percent

Enrollees Nonenrollees Difference in net

improvement Net

Net between enrollees

Question and site Improved Declined change Improved Declined change and
nonenrollees

Colorado 39 13 26 a 16 22 -6 a 32 a Dover 23 14 9 a 18 17 1 8 a

The doctor?s office was conveniently 28 16 12 a 17 23 -6 a 18 a

located. Madigan 23 17 6 a 19 23 -4 10 a San Antonio 32 16 16 a 17 24 -7 a
23 a San Diego 23 15 8 a 14 23 -9 a 17 a Keesler 23 19 4 23 17 6 -2 Texoma
28 15 13 a 21 17 4 9 a Colorado 32 18 14 a 16 24 -8 a 22 a Dover 23 13 10 a
19 18 1 9 a

The doctor?s hours were convenient. b 32 20 12 a 24 27 -3 15 a Madigan 29 25
4 24 29 -5 9 a San Antonio 33 20 13 a 27 24 3 10 a San Diego 32 19 13 a 22
29 -7 a 20 a Keesler 28 20 8 a 28 25 3 5 Texoma 39 16 23 a 25 23 2 21 a
Colorado 30 19 11 a 22 25 -3 14 a Dover 29 22 7 28 20 8 a -1

I did not have to wait long between making 35 26 9 a 25 28 -3 12 a an
appointment and seeing the doctor. Madigan 30 31- 125 27- 2 1

San Antonio 41 24 17 a 29 27 2 15 a San Diego 34 20 14 a 22 30 -8 a 22 a
Keesler 27 30- 329 227 -10 a Texoma 33 22 11 a 27 24 3 8 Colorado 34 30 4 25
27 -2 6 Dover 30 29 1 25 24 1 0

The doctor saw me promptly. 34 17 17 a 22 24 -2 19 a Madigan 26 19 7 a 24 24
0 7 a

(Continued From Previous Page)

Numbers in percent

Enrollees Nonenrollees Difference in net

improvement Net

Net between enrollees

Question and site Improved Declined change Improved Declined change and
nonenrollees

San Antonio 39 15 24 a 24 24 0 24 a San Diego 34 16 18 a 20 25 -5 23 a
Keesler 27 22 5 25 22 3 2 Texoma 31 16 15 a 22 22 0 15 a Colorado 38 16 22 a
18 27 -9 a 31 a Dover 25 19 6 23 21 2 4

The doctor did a thorough examination. 33 19 14 a 20 25 -5 a 19 a Madigan 27
23 4 21 26 -5 a 9 a San Antonio 38 16 22 a 20 23 -3 a 25 a San Diego 31 18
13 a 19 26 -7 a 20 a Keesler 23 26- 321 201 -4 Texoma 32 16 16 a 20 23 -3 19
a Colorado 38 19 9 a 20 27 -7 a 16 a Dover 25 19 6 22 23 -1 7

The doctor was careful in recording my 33 16 17 a 19 23 -4 a 21 a medical
history. Madigan 28 18 10 a 21 25 -4 14 a

San Antonio 36 13 23 a 19 23 -4 27 a San Diego 32 16 16 a 17 23 -6 a 22 a
Keesler 24 23 1 22 17 5 -4 Texoma 30 12 18 a 21 19 2 16 a Colorado 38 18 20
a 19 25 -6 a 26 a Dover 25 17 8 a 21 21 0 8

I was able to choose my own doctor. 32 36 -4 a 20 21 -1 -3 Madigan 27 37 -10
a 19 25 -6 a -4 San Antonio 37 32 5 22 21 1 4 San D iego 32 34- 218 19- 1 -1
Keesler 24 44 -20 a 26 18 8 a -28 a Texoma 29 35 -6 19 15 4 -10 a Colorado
25 47 -22 a 18 25 -7 a -15 a

(Continued From Previous Page)

Numbers in percent

Enrollees Nonenrollees Difference in net

improvement Net

Net between enrollees

Question and site Improved Declined change Improved Declined change and
nonenrollees

Dover 34 27 7 23 20 3 4 The doctor explained things clearly. b 33 20 13 a 22
24 -2 a 15 a Madigan 30 22 8 a 22 25 -3 11 a San Antonio 36 19 17 a 23 26 -3
20 a San Diego 32 17 15 a 19 24 -5 20 a Keesler 22 27- 526 215 -10 a Texoma
34 16 18 a 24 23 1 17 a Colorado 39 17 22 a 22 24 -2 24 a Dover 29 19 10 a
21 22 -1 11

The doctor really listened. b 31 18 13 a 22 23 -1 15 a Madigan 28 19 9 a 21
27 -6 15 a San Antonio 33 17 16 a 23 24 -1 17 a San Diego 30 16 14 a 22 21 1
13 a Keesler 20 23- 326 215 -8 Texoma 33 14 19 a 22 24 -2 21 a Colorado 38
19 19 a 20 23 -3 22 a Dover 24 20 4 21 20 1 3

The doctor spent enough time with me. b 34 18 16 a 22 27 -5 a 19 a Madigan
31 23 8 a 23 29 -6 14 a San Antonio 36 16 20 a 24 27 -3 23 a San Diego 35 15
20 a 21 26 -5 25 a Keesler 23 27- 425 223 -7 Texoma 37 12 25 a 23 27 -4 29 a
Colorado 40 18 22 a 19 28 -9 a 31 a Dover 28 19 9 a 23 23 0 9

The doctor was skillful and competent. b 30 18 12 a 20 22 -2 14 a Madigan 29
19 10 a 19 24 -5 15 a San Antonio 31 18 13 a 22 24 -2 15 a San Diego 29 15
14 a 18 21 -3 17 a

(Continued From Previous Page)

Numbers in percent

Enrollees Nonenrollees Difference in net

improvement Net

Net between enrollees

Question and site Improved Declined change Improved Declined change and
nonenrollees

Keesler 22 24- 223 203 -5 Texoma 33 14 19 a 20 26 -6 25 a Colorado 36 16 20
a 19 21 -2 22 a Dover 24 21 3 20 16 4 -1

Satisfaction with specialists

I am satisfied with my ability to choose 27 17 10 a 18 21 -3 a 13 a
specialists. Madigan 22 18 4 22 22 0 4

San Antonio 31 16 15 a 20 19 1 14 a San Diego 25 15 10 a 16 24 -8 a 18 a
Keesler 19 23- 417 152 -6 Texoma 26 17 9 a 18 20 -2 11 a Colorado 32 16 16 a
18 22 -4 20 a Dover 20 20 0 19 20 -1 1

I didn?t wait too long for my appointment. 32 22 10 a 23 27 -4 a 14 a
Madigan 27 26 1 26 26 0 1 San Antonio 36 20 16 a 27 26 1 15 a San Diego 33
18 15 a 19 30 -11 a 26 a Keesler 26 24 2 23 21 2 0 Texoma 25 27 -2 20 22 -2
0 Colorado 33 22 11 a 21 26 -5 16 a Dover 29 22 7 22 22 0 7

The doctor was skillful and competent. 25 18 7 a 18 20 -2 9 a Madigan 22 19
3 21 22 -1 4 San Antonio 27 17 10 a 20 20 0 10 a San Diego 25 14 11 a 15 21
-6 a 17 a Keesler 19 23- 417 18- 1 -3 Texoma 26 20 6 17 19 -2 8 Colorado 25
18 7 a 18 21 -3 10 a Dover 24 15 9 a 19 18 1 8

(Continued From Previous Page)

Numbers in percent

Enrollees Nonenrollees Difference in net

improvement Net

Net between enrollees

Question and site Improved Declined change Improved Declined change and
nonenrollees

The doctor told me all I wanted to know 29 21 8 a 24 25 -1 9 a about my
treatment. b Madigan 27 23 4 26 26 0 4

San Antonio 30 21 9 a 23 27 -4 13 a San Diego 32 19 13 a 25 24 1 12 a
Keesler 19 23- 421 25- 4 0 Texoma 29 20 9 a 26 24 2 7 Colorado 34 19 15 a 23
23 0 15 a Dover 26 24 2 26 19 7 a -5

The doctor answered all my questions. b 27 20 7 a 22 23 -1 8 a Madigan 24 23
1 25 24 1 0 San Antonio 28 19 9 a 21 25 -4 13 a San Diego 30 16 14 a 22 23
-1 15 a Keesler 20 22- 221 25- 4 2 Texoma 24 18 6 22 23 -1 7 Colorado 34 17
17 a 22 22 0 17 a Dover 21 25- 425 187 a -11

The doctors communicated with one 32 22 10 a 27 27 0 10 a

another. b Madigan 31 25 6 32 24 8 a -2 San Antonio 33 19 12 a 27 29 -2 14 a
San Diego 33 20 13 a 25 28 -3 16 a Keesler 25 26- 126 260 -1 Texoma 30 20 10
a 25 24 1 9 Colorado 35 24 11 a 23 27 -4 15 a Dover 25 25 0 26 23 3 -3

Note: Bold indicates that the demonstration site percentage is significantly
different from the overall percentage. a Significant at .05 level.

b Questions reversed, in order to calculate improvement or decline. The
exact wording of all questions used in this analysis can be found in
appendix I, table 10. Source: GAO Survey of Medicare- Eligible Military
Retires and Family Members

Measures of Access In addition to the change of access and quality among
enrollees and and Satisfaction at the

nonenrollees, we also examined the level of access and quality at the time
of the second survey among the cross section sample. (See table 12.) End of
the Demonstration

Three separate significance tests were performed. The first test of
significance was between enrollees and nonenrollees who said they strongly
agreed with each statement. The second test of significance was between
enrollees and nonenrollees who said they either strongly agreed or agreed
with each statement. The final test was whether the site percentage differs
significantly from the overall percentage.

Table 12: Level of Access and Quality at the End of the Demonstration

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

Access to care

I was able to get care when I 48 42 5 3 1

needed it. Madigan 48 43 4 3 1 San Antonio 51 40 5 3 1 San Diego 51 40 5 4 1
Keesler 44 45 6 4 1 Texoma 45 44 6 3 1 Colorado 44 45 7 3 1 Dover 38 51 6 3
1

I was able to get info by phone 26 31 15 17 11 when I needed it. b Madigan
283217 14 10

San Antonio 26 30 15 17 12 San D iego 273019 15 10 Keesler 243412 18 12
Texoma 22 30 14 22 11 Colorado 27 29 15 18 10 Dover 263214 18 10

I was able to get care on nights 36 33 13 11 7

and weekends. b Madigan 39 38 10 7 5 San Antonio 39 32 12 10 7 San Diego 35
30 16 12 7 Keesler 32 37 11 13 8 Texoma 28 33 14 15 9 Colorado 32 35 14 12 8
Dover 273116 15 11

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

42 49 5 3 1 6 a 90 91 -1 41 49 5 3 1 7 a 92 91 1 40 48 6 4 2 11 a 91 89 3 a
46 47 5 2 1 5 a 91 93 -2 39 53 4 2 2 5 89 93 -3 a 42 51 5 2 0 3 89 93 -3 a
37 54 5 4 1 7 a 89 91 -2 45 50 4 1 1 -6 a 90 94 -4 a

23 31 17 19 10 3 a 57 54 2 25 33 15 19 8 3 60 58 2 23 28 16 21 11 3 56 52 5
25 30 18 16 11 1 56 55 2 18 33 17 18 14 6 a 57 51 7 a 17 32 16 21 14 5 a 53
49 4 20 34 17 21 8 7 a 56 53 2 24 34 15 15 12 2 57 58 -1

30 32 16 13 10 6 a 69 61 8 a 34 29 16 12 9 6 77 63 15 a 28 31 16 15 10 11 a
71 58 13 a 33 31 15 11 10 2 65 64 0 21 38 17 13 11 11 a 69 59 10 a 23 34 18
13 11 5 61 58 3 22 38 15 16 9 10 a 66 60 6 29 29 18 13 11 -2 58 58 0

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

Military care

I am satisfied with care at 56 36 4 2 1

military facilities. Madigan 57 37 4 2 1 San Antonio 58 35 4 2 1 San Diego
58 36 3 2 1 Keesler 54 38 4 2 2 Texoma 53 37 6 2 2 Colorado 53 38 5 2 1
Dover 47 45 4 3 1

I was able to get care at military 52 39 4 4 2

facilities when I needed it. Madigan 52 42 4 2 1 San Antonio 53 38 4 4 1 San
Diego 55 37 4 3 2 Keesler 48 38 5 6 4 Texoma 49 38 4 4 4 Colorado 49 38 5 5
3 Dover 44 45 5 3 2

I prefer to get my care at military 68 28 3 1 0 facilities. Madigan 67 29 3
1 0

San Antonio 71 26 2 0 0 San Diego 69 27 3 0 0 Keesler 68 28 3 1 0 Texoma 61
30 6 2 1 Colorado 64 31 4 0 1 Dover 58 33 7 1 0

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

24 38 18 8 12 32 a 93 63 30 a 26 38 18 9 10 31 a 93 63 30 a 29 36 14 8 13 29
a 93 65 28 a 24 43 20 6 7 34 a 94 67 27 a 18 30 14 9 30 37 a 92 47 45 a 15
31 19 12 23 38 a 90 46 44 a 16 43 20 8 13 37 a 91 60 32 a 18 29 23 11 18 29
a 92 47 45 a

15 29 17 13 26 36 a 90 44 46 a 17 28 16 13 25 34 a 93 45 48 a 18 30 14 14 24
36 a 92 48 43 a 16 34 24 9 17 39 a 92 50 42 a 8 1410 11 57 39 a 86 22 63 a
11 20 14 16 39 38 a 88 31 57 a 10 26 15 15 34 38 a 87 36 51 a 11 21 24 11 33
33 a 89 32 57 a

27 22 22 15 14 42 a 96 48 48 a 32 21 21 13 13 35 a 96 53 43 a 29 25 20 14 13
42 a 97 54 44 a 21 18 28 19 15 48 a 96 39 58 a 28 21 17 14 19 40 a 96 50 46
a 19 21 20 18 23 42 a 91 39 52 a 30 26 20 13 11 34 a 95 56 39 a 16 21 26 16
22 43 a 92 36 55 a

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

It was not difficult to schedule 38 35 10 11 6 appointments at military
facilities. b

Madigan 39 35 11 11 4 San Antonio 39 33 10 12 5 San Diego 42 35 10 8 4
Keesler 34 37 9 12 8 Texoma 38 35 8 10 9 Colorado 36 35 10 12 8 Dover 33 40
11 10 5

Doctors and staff treated me 61 32 4 2 2 with respect at military
facilities. b

Madigan 63 30 4 1 2 San Antonio 62 31 4 1 2 San Diego 62 31 3 1 2 Keesler 59
34 4 2 2 Texoma 56 33 6 3 2 Colorado 58 34 5 2 2 Dover 59 35 4 1 1

I would recommend military 67 25 4 2 2 care. b Madigan 68 23 5 2 1

San Antonio 68 24 4 2 2 San Diego 71 22 3 2 2 Keesler 64 27 5 2 2 Texoma 61
28 6 3 3 Colorado 64 28 5 1 2 Dover 60 30 6 3 1

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

11 15 22 23 30 27 a 73 25 47 a 12 14 23 23 28 27 a 74 26 48 a 12 16 16 23 33
27 a 72 28 44 a 10 15 33 23 19 32 a 77 25 52 a 9 8 12 19 53 26 a 71 16 55 a
10 8 19 23 39 28 a 73 18 55 a 11 17 21 21 31 25 a 71 28 43 a 9 1022 26 32 24
a 73 19 55 a

33 30 25 6 5 28 a 93 64 29 a 33 32 24 8 4 30 a 94 65 29 a 41 30 21 3 5 21 a
93 70 23 a 29 29 31 7 4 33 a 94 58 36 a 27 26 27 8 12 32 a 93 53 40 a 23 31
27 10 10 33 a 88 54 34 a 28 36 26 6 4 29 a 91 64 28 a 25 31 32 8 5 34 a 94
55 39 a

32 23 22 11 11 35 a 92 55 37 a 35 23 23 10 10 34 a 92 58 34 a 35 23 20 12 9
33 a 92 59 33 a 32 23 25 10 11 39 a 94 55 39 a 25 19 22 13 20 39 a 92 44 47
a 20 23 21 16 20 40 a 88 43 45 a 30 26 20 13 10 34 a 92 57 35 a 18 23 29 13
17 42 a 90 41 49 a

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

Satisfaction with care

I am satisfied with the care I 52 39 5 2 1

received. Madigan 52 40 5 2 0 San Antonio 55 37 5 2 1 San Diego 53 37 5 3 2
Keesler 49 42 6 3 1 Texoma 50 41 6 3 1 Colorado 49 41 7 3 1 Dover 44 48 6 2
0

Satisfaction with primary care providers

I received excellent care. 57 33 7 2 0 Madigan 56 35 7 1 0 San Antonio 59 32
7 1 0 San Diego 56 33 7 3 1 Keesler 57 33 7 2 0 Texoma 60 33 5 1 0 Colorado
55 33 9 3 1 Dover 55 39 6 1 1

The doctor's office was 51 41 5 2 1 conveniently located. Madigan 50 44 4 1
1

San Antonio 54 39 5 2 0 San Diego 49 42 6 2 1 Keesler 50 43 5 1 1 Texoma 52
40 4 2 1 Colorado 47 42 6 4 1 Dover 49 40 6 3 2

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

43 47 6 2 1 9 a 91 90 1 43 47 7 2 1 9 a 92 90 2 a 42 48 5 3 1 13 a 92 91 2
47 44 7 1 1 6 a 90 91 -1 40 50 7 3 1 9 a 91 90 1 42 49 6 2 1 8 a 91 91 -1 36
53 7 3 1 13 a 90 89 1 45 48 5 2 1 -1 91 92 -1

50 40 7 2 0 7 a 91 90 1 49 40 8 2 0 7 a 91 89 2 52 38 6 2 1 7 a 92 91 1 51
39 8 2 0 5 a 89 90 -1 51 42 5 2 0 6 a 91 93 -2 54 37 7 2 0 6 a 93 91 2 44 43
10 3 0 10 a 87 87 0 52 40 7 1 0 3 93 92 1

47 42 7 3 1 5 a 92 89 3 a 44 45 7 4 1 7 a 94 89 5 a 47 42 6 4 1 7 a 93 88 5
a 50 40 7 2 1 -1 91 90 1 48 44 6 2 1 2 93 92 2 50 42 5 2 1 2 93 92 1 39 47 7
5 1 8 a 90 86 3 a 52 40 4 2 1 -4 89 93 -4 a

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

The doctor's hours were 36 43 12 6 3 convenient. b Madigan 35 44 12 6 3

San Antonio 36 43 12 6 3 San Diego 37 41 12 7 3 Keesler 37 45 11 4 3 Texoma
36 42 12 6 4 Colorado 34 46 13 5 3 Dover 33 40 16 7 3

I did not have to wait long 29 36 18 13 5

between making appointment and seeing the doctor. Madigan 25 30 22 16 7 San
Antonio 30 36 17 12 5 San Diego 34 38 15 9 4 Keesler 29 36 17 15 4 Texoma 31
38 15 12 4 Colorado 28 34 19 14 5 Dover 25 39 19 13 4

The doctor saw me promptly. 32 48 13 6 1 Madigan 29 50 14 6 1 San Antonio 33
46 14 6 1 San Diego 31 49 14 6 0 Keesler 32 48 13 7 1 Texoma 35 49 10 6 1
Colorado 32 50 11 6 1 Dover 28 53 13 5 0

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

33 45 13 7 3 3 a 79 78 1 30 46 13 9 3 6 a 79 76 3 a 33 46 12 6 3 4 79 78 1
35 45 13 6 2 2 78 79 -2 34 45 11 7 3 3 82 80 2 35 41 13 7 3 1 78 77 2 31 46
14 7 3 2 80 77 3 37 45 10 5 3 -4 73 82 -8 a

31 41 17 9 3 -1 65 71 -6 a 29 43 17 9 2 -4 a 55 72 -17 a 31 41 17 8 3 -1 66
71 -5 a 29 40 17 10 4 5 a 72 69 3 37 40 16 5 3 -8 a 64 76 -12 a 36 41 15 5 3
-5 a 69 76 -7 a 29 41 18 10 2 -1 62 70 -8 a 36 44 12 6 3 -11 a 63 79 -16 a

26 51 14 8 1 6 a 80 77 3 a 29 51 12 7 1 0 79 80 -1 25 49 16 9 1 8 a 79 75 4
a 26 51 14 7 1 5 a 80 78 2 25 52 13 9 1 6 a 79 77 3 30 48 13 8 1 5 a 84 78 6
a 22 55 14 9 1 11 a 83 76 6 a 24 54 14 8 1 4 81 78 3

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

The doctor did a thorough 36 46 12 5 1 examination. Madigan 34 47 15 4 0

San Antonio 38 44 11 5 1 San Diego 35 46 13 5 1 Keesler 34 46 12 6 1 Texoma
39 46 10 4 1 Colorado 33 47 14 6 1 Dover 34 51 12 3 0

The doctor was careful in 42 44 10 3 1 recording my medical history. Madigan
40 45 11 3 0

San Antonio 45 41 10 3 1 San Diego 40 45 10 4 1 Keesler 41 45 10 3 1 Texoma
45 45 7 2 1 Colorado 37 45 13 4 1 Dover 37 52 8 3 0

I was able to choose my own 20 23 20 26 10

doctor. Madigan 141622 35 14 San Antonio 22 21 21 26 11 San Diego 23 29 20
20 8 Keesler 182120 29 12 Texoma 21 27 21 22 9 Colorado 21 29 17 22 10 Dover
23 33 20 21 4

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

31 49 13 5 1 4 a 82 81 1 31 49 12 6 2 2 81 81 0 32 51 9 6 1 6 a 83 84 -1 33
45 16 4 1 2 81 78 3 30 54 12 4 1 4 80 84 -3 35 49 10 5 1 4 85 84 1 26 51 15
6 1 7 a 79 77 2 32 53 12 3 1 2 84 84 0

37 49 10 3 1 5 a 86 86 0 37 48 10 4 1 4 a 86 85 1 36 50 9 4 1 9 a 87 86 1 39
47 11 3 1 2 85 85 0 37 52 7 3 0 3 86 89 -3 41 48 8 3 0 4 90 89 1 31 52 11 4
1 5 a 82 84 -2 38 51 9 2 0 -1 89 89 0

43 43 6 6 2 -22 a 43 86 -42 a 40 43 6 8 3 -26 a 29 83 -54 a 40 43 7 7 3 -18
a 43 82 -40 a 46 43 6 4 2 -23 a 52 87 -36 a 44 43 5 4 3 -26 a 39 87 -50 a 47
44 5 3 2 -26 a 48 91 -43 a 39 46 6 6 2 -18 a 50 86 -35 a 47 43 5 5 1 -24 a
56 90 -34 a

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

The doctor explained things 35 45 12 6 2 clearly. b Madigan 34 47 13 5 2

San Antonio 37 44 12 6 2 San Diego 35 44 12 7 2 Keesler 35 45 12 6 2 Texoma
39 42 11 5 2 Colorado 32 47 13 7 2 Dover 35 49 10 4 2

The doctor really listened. b 40 44 10 4 1 Madigan 39 46 10 4 1 San Antonio
42 44 9 4 1 San Diego 40 44 9 5 2 Keesler 39 46 9 4 2 Texoma 44 42 9 4 2
Colorado 36 44 12 6 2 Dover 38 48 9 4 2

The doctor spent enough time 37 42 12 7 2

with me. b Madigan 34 43 15 6 2 San Antonio 39 43 11 6 1 San Diego 37 41 12
7 2 Keesler 36 44 11 7 2 Texoma 42 41 11 5 2 Colorado 32 40 16 10 3 Dover 36
45 11 5 2

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

35 46 12 6 2 0 80 81 -1 33 47 12 6 2 0 80 80 0 35 47 10 7 2 2 81 82 -1 35 46
13 4 2 0 79 81 -2 37 45 11 6 1 -3 80 83 -3 38 44 11 6 2 2 82 81 1 33 46 15 6
1 -1 78 78 0 36 47 10 5 1 -1 84 83 0

40 44 10 5 1 0 85 84 1 37 45 11 5 2 2 85 82 2 42 43 8 6 1 0 86 85 1 42 43 10
4 1 -2 84 85 -1 43 44 9 3 1 -4 85 87 -2 40 43 11 5 2 4 86 82 3 36 46 11 5 1
0 80 82 -2 41 44 9 4 1 -3 86 86 0

34 41 15 8 2 3 a 79 75 5 a 31 43 15 8 3 2 77 74 2 35 40 15 7 2 4 a 82 75 7 a
34 39 16 9 2 3 78 73 5 a 39 42 11 7 2 -2 81 81 0 36 42 12 7 3 6 a 83 79 4 a
29 43 16 9 2 3 72 73 -1 36 45 11 5 2 0 82 81 0

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

The doctor was skillful and 48 38 10 3 2 competent. b Madigan 48 39 9 3 2

San Antonio 49 37 9 3 2 San Diego 48 38 9 3 2 Keesler 45 40 10 3 2 Texoma 50
36 9 3 1 Colorado 44 39 13 3 2 Dover 43 45 7 3 1

Satisfaction with specialists

I am satisfied with my ability to 49 43 5 3 1 choose specialists. Madigan 52
41 5 2 0

San Antonio 52 41 4 2 0 San Diego 50 41 5 3 1 Keesler 45 44 5 4 2 Texoma 43
46 8 3 1 Colorado 44 46 7 3 1 Dover 40 47 8 4 1

I didn't wait too long for my 33 45 11 9 1 appointment. Madigan 30 45 12 11
2

San Antonio 34 45 11 9 1 San Diego 35 43 12 8 2 Keesler 33 47 10 9 1 Texoma
30 48 10 9 3 Colorado 29 48 11 10 1 Dover 31 46 11 9 2

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

48 38 9 3 2 -1 86 87 -1 47 38 10 3 2 1 87 85 2 49 38 8 3 2 0 86 86 0 50 38 8
3 1 -2 86 88 -2 51 40 7 3 0 -6 a 85 90 -5 a 45 40 9 3 3 5 a 86 85 1 46 40 10
2 1 -2 83 86 -3 51 38 7 3 1 -6 a 89 88 1

43 49 5 2 1 6 a 92 92 0 43 47 6 3 1 8 a 93 90 3 a 47 47 4 2 1 5 a 93 93 0 41
50 6 2 1 9 a 91 91 0 42 52 4 2 0 3 89 94 -5 a 43 50 5 2 0 0 88 92 -4 a 40 53
5 1 1 4 90 93 -3 a 46 48 4 2 0 -5 87 93 -6 a

29 51 11 8 1 4 a 78 80 -2 a 27 49 12 9 2 2 75 77 1 34 48 11 7 1 1 79 82 -7
26 52 12 9 1 9 a 78 78 0 29 54 10 7 1 5 80 82 -2 32 53 8 6 1 -3 77 85 -8 a
27 53 10 9 1 2 77 80 -2 33 53 6 7 1 -2 78 86 -9 a

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

The doctor was skillful and 49 42 7 1 1

competent. Madigan 50 43 5 1 0 San Antonio 51 41 7 1 1 San Diego 52 38 7 2 1
Keesler 48 43 7 2 0 Texoma 44 45 9 1 1 Colorado 45 45 8 2 1 Dover 44 46 9 0
1

The doctor told me all I wanted 37 44 11 7 2

to know about my treatment. b Madigan 36 45 12 6 2 San Antonio 39 43 10 7 1
San Diego 38 40 11 8 3 Keesler 35 46 9 7 2 Texoma 33 42 12 10 3 Colorado 33
46 13 6 2 Dover 30 48 12 9 2

The doctor answered all my 39 45 9 5 1 questions. b Madigan 38 46 10 4 2

San Antonio 42 44 8 5 1 San Diego 41 42 10 6 2 Keesler 36 47 9 6 1 Texoma 34
45 12 7 2 Colorado 37 46 11 4 2 Dover 31 49 11 8 1

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

44 47 7 1 0 5 a 91 91 0 43 48 8 2 0 7 a 93 91 2 a 49 43 6 1 1 2 92 92 0 43
47 8 1 0 9 a 90 90 0 43 51 5 0 0 5 91 94 -3 a 43 50 6 1 0 1 89 93 -4 a 42 50
6 1 1 3 90 92 -2 46 47 5 1 0 -3 90 93 -3

33 45 12 7 2 3 a 81 79 2 a 30 48 10 10 2 6 a 81 78 3 35 46 10 7 2 4 82 81 1
34 44 13 7 3 4 78 78 0 34 46 11 7 2 1 81 80 1 31 48 12 7 2 2 75 80 -5 a 33
43 13 7 3 0 79 76 3 38 41 11 9 1 -8 a 78 79 -1

36 47 10 6 2 4 a 84 82 2 a 33 49 10 6 2 6 a 85 82 3 37 47 9 6 1 4 a 86 84 2
36 46 10 6 2 5 83 82 1 36 46 9 6 2 0 84 82 1 33 49 9 6 2 1 80 82 -3 35 45 12
7 1 1 83 80 3 40 43 9 7 1 -9 a 80 83 -3

Numbers in percent

Enrollees c Strongly

Strongly Question and site agree Agree Neither Disagree disagree

The doctors communicated with 35 36 21 6 2 one another. b Madigan 34 39 21 4
2

San Antonio 37 36 20 6 2 San Diego 37 35 20 6 3 Keesler 32 36 22 7 3 Texoma
32 37 22 6 2 Colorado 30 33 29 6 2 Dover 29 38 21 10 1

Nonenrollees c Difference

Difference between enrollees

Enrollees between enrollees

and nonenrollees in Strongly

Nonenrollees and nonenrollees in

Strongly Strongly

percentage strongly agree or

Strongly agree percentage strongly

agree Agree Neither Disagree disagree agreeing d

Agree or Agree

agreeing or agreeing d

31 37 21 7 3 3 a 70 68 2 a 31 42 19 6 2 3 72 73 0 33 36 21 8 3 3 73 69 4 30
36 25 7 3 7 a 71 65 6 a 32 36 19 10 3 0 68 68 0 33 38 18 8 3 -1 69 71 -2 30
36 22 9 3 0 62 66 -4 36 38 18 7 1 -7 a 67 74 -7 a

Note: Bold indicates that the demonstration site percentage is significantly
different from the overall percentage. a Significant at .05 level.

b Question reversed so that a response of strongly agree is always a
positive response. The exact wording of all questions in this table can be
found in appendix I, table I0. c Row percentages may not equal 100 due to
the effects of rounding. d Differences are based on numbers before rounding.
Source: GAO Survey of Medicare- Eligible Military Retirees and Family
Members.

Appendi x I II

Models of Utilization In this appendix, we describe the DOD and Medicare
data that we used to analyze utilization. We also summarize the models that
we developed to risk adjust acute inpatient care and outpatient care and
give results both demonstration wide and by site. Data For these analyses,
we defined the Senior Prime enrollee population as those who had enrolled as
of December 31, 1999. We used DOD data for 1999 as the source of our counts
of hospital stays and outpatient visits to both MTF and civilian network
providers. 1 We limited our analysis to

hospital stays of 1 day or more to eliminate inconsistencies between
Medicare and TRICARE in the use of same- day discharges. Our counts of
outpatient utilization include (1) visits and ambulatory surgeries in MTF
outpatient clinics and (2) visits to network providers- doctors? offices,
ambulatory surgeries, hospital emergency rooms, and hospital outpatient
clinics. To identify our comparison group of fee- for- service beneficiaries
in the demonstration areas, we used CMS? 2 20- percent Medicare sample, and

extracted those beneficiaries residing in the subvention areas. We excluded
anyone who had been in a Medicare+ Choice plan for any part of the year. To
make the comparison fair, we also excluded certain groups not represented or
only minimally represented in Senior Prime: persons with end- stage renal
disease (ESRD), Medicaid beneficiaries, persons with disabilities (under age
65), and people who lost Medicare part A or part B entitlement for reasons
other than death. We derived our counts of Medicare fee- for- service
utilization for the sample from Medicare claims

files. For those who were in either Senior Prime or fee- for- service for
less than a full year, we estimated full- year utilization counts.

We identified a separate comparison group of persons eligible for the
demonstration who did not enroll. We collected both Medicare fee- forservice
claims and DOD encounter data for the sample of enrollees and nonenrollees
who answered both our first and second surveys.

1 See Medicare Subvention Demonstration CY 1999 Reconciliation Processing,
March 2001, SRA International, Inc. 2 Formerly the Health Care Financing
Administration.

Models of RiskAdjusted In order to compare the utilization of Senior Prime
enrollees to Medicare Utilization

fee- for- service beneficiaries in the demonstration areas, we developed
several models of fee- for- service utilization (for hospitalization, length
of stay, and outpatient care). We then applied each model to Senior Prime

enrollees- taking account of their demographic characteristics and health
status- to predict what their utilization would have been in Medicare
feefor- service. The ratio of their predicted utilization to their actual
Senior Prime utilization gives a measure of the amount by which Senior Prime
utilization exceeded or fell short of fee- for- service utilization for
people with the enrollees? characteristics. Table 13 compares the
characteristics of Senior Prime enrollees with Medicare fee- for- service
beneficiaries in the demonstration area.

Table 13: Comparison of Senior Prime Enrollees With Medicare Fee- for-
Service Beneficiaries in the Demonstration Areas in 1999

Senior Prime Fee- for- service

Characteristic enrollees beneficiaries a Size

Sample size 30, 216 b 84, 523 Estimated population size 30, 216 422, 615

Actual utilization (annualized)

Acute hospitalization rate 0.37 0.39 Average hospital stay (in days) 4. 75
5. 60 Outpatient physician visits 16. 71 10.47

Predicted utilization (annualized)

Acute hospitalization rate 0.26 c Average hospital stay (in days) 4. 96 c
Outpatient physician visits 10. 59 c

Ratio of actual to predicted utilization

Acute hospitalization rate 1.41 c Average hospital stay (in days) 0. 96 c
Outpatient physician visits 1.58 c

(Continued From Previous Page)

Senior Prime Fee- for- service

Characteristic enrollees beneficiaries a Health status

Average HCC score d 0.94 1.19 Number of unique diagnoses per individual 15.
34 16.13 Proportion deceased during 1999 0.02 0.05

Demographics

Average number of months in program (Senior 10. 11 11. 71

Prime or fee- for- service) in 1999 Average age 72.36 76. 43 Proportion male
0.53 0.43 a We used the Medicare 20- percent sample of fee- for- service
beneficiaries residing in the official

demonstration areas. We excluded Medicare+ Choice members, military
retirees, persons with ESRD, Medicaid beneficiaries, persons with
disabilities (under age 65), and people who lost Medicare part A or part B
entitlement for reasons other than death. b CMS identified 30,228 unique
enrollees when calculating the final payment to DOD. Our number differs
slightly because we used an earlier data file prepared by DOD?s contractor.
c Our model of fee- for- service utilization has the property that the
average predicted utilization equals

the average actual utilization. d A ratio derived from the Hierarchical
Coexisting Conditions (HCC) concurrent model, which reflects the costliness
of each person, based on clinical diagnoses and demographic traits, relative
to the

average Medicare fee- for- service beneficiary (who would have a score of
1.0). A lower score indicates lower- than- average costs.

Source: GAO analysis of DOD encounter and claims data and Medicare 20-
percent fee- for- service sample.

Acute Hospitalization Acute hospitalization is a relatively rare event: only
one out of five Model Medicare beneficiaries (in the counterpart 20- percent
fee- for- service sample) is hospitalized during the year, and about half of
those who are hospitalized are admitted again during the same year. We
therefore used

Poisson regression, which is designed to predict the number of occurrences
(counts) of a rare event during a fixed time frame, to estimate the number
of acute hospitalizations. Positive coefficients are interpreted as
reflecting factors that increase the hospitalization rate while negative
coefficients indicate a decrease in that rate. The strongest factor
affecting the number of hospitalizations is the HCC score, which measures
how ill

and how costly a person is. Its effect is not linear- both squared and cubed
terms enter the model. (See table 14.)

Table 14: Estimated Effects of Selected Factors on Acute Hospitalization of
Medicare Fee- for- Service Beneficiaries

95% confidence Characteristic Coefficient interval

HCC 1.081 1. 067 1.095 HCC 2 -0. 090 -0. 091 -0. 088 HCC 3 0.002 0. 002
0.002 Age (continuous) 0. 017 0. 016 0. 019 Gender - male -0. 047 -0. 069
-0. 025 Number of unique diagnoses 0. 008 0. 007 0. 009 Note: All
coefficients are significant at the .001 level. Source: GAO analysis of
Medicare 20- percent fee- for- service sample.

Outpatient Physician Unlike hospitalizations, outpatient physician visits
are relatively common Visit Model

events for most Medicare beneficiaries. Physician visits have a skewed
distribution, with a small number of people having a very large number of
visits. We categorized the number of visits into five groups and used an

ordered logit model, which predicts the odds of each person belonging to
each category, to estimate the number of outpatient visits. Positive
coefficients indicate higher odds of belonging to the highest utilization
category while negative coefficients indicate higher odds of belonging to
the lowest utilization category. Both the HCC score and ICD- 9 diagnostic
categories 3 are major factors in the model. (See table 15.)

3 Diagnostic groupings are based on the International Classification of
Diseases, 9th Revision, Clinical Modification (ICD- 9- CM).

Table 15: Estimated Effects of Certain Factors on Medicare Fee- for- Service
Outpatient Utilization 95% confidence Characteristic Coefficient interval

HCC 0. 471 0. 445 0. 497 HCC 2 -0.062 -0. 067 -0. 057 HCC 3 0.002 0. 002 0.
002 Male 0.639 0. 613 0. 665 Circulatory disease 0. 643 0. 614 0. 672
Respiratory disease 0. 572 0. 542 0. 601 Digestive disease 0. 504 0. 471 0.
537 Infection 0. 398 0. 355 0. 441 Neoplasm 0.842 0. 811 0. 874 Endocrine,
nutritional, and metabolic diseases and

0.421 0. 394 0. 449 immunity disorders Diseases of the nervous system and
sense organs 0. 875 0. 848 0. 902

Diseases of the musculoskeletal system and 0.859 0. 830 0. 888

connective tissue Injury and poisoning 0. 528 0. 494 0. 562 Supplementary
classification (V01- V82) 0. 805 0. 777 0. 833 Note: All coefficients are
significant at the .001 level. Source: GAO analysis of Medicare 20- percent
fee- for- service sample.

Profile of Sites? Using the same approach and models, we examined
utilization at each site. Utilization and Risk

(See table 16.) Adjusting for risk, both hospital stays and outpatient
visits were substantially greater in Senior Prime than in fee- for- service
at all sites. Factors However, the differences in length of stay were small,
with lengths of stay generally higher in fee- for- service.

Table 16: Site Profiles of Senior Prime and Medicare Fee- for- Service
Utilization in 1999

Acute hospitalization Actual/ predicted Actual Predicted

ratio Senior Prime enrollees (n= 30,216)

Madigan 0. 32 0. 27 1. 19 San Antonio 0. 40 0. 26 1. 54 San Diego 0. 43 0.
28 1. 54 Keesler 0. 37 0. 27 1. 37 Texoma 0. 37 0. 27 1.37 Colorado 0.26
0.21 1. 24 Dover 0. 28 0. 22 1. 27

Fee- for- service sample a (n= 84,523)

Madigan 0. 33 0. 35 0. 94 San Antonio 0. 43 0. 43 1. 00 San Diego 0. 34 0.
41 0. 83 Keesler 0. 51 0. 41 1. 24 Texoma 0. 47 0. 38 1.24 Colorado 0.38
0.37 1. 03 Dover 0. 42 0. 42 1. 00

Outpatient visits Average hospital stay Average Actual/

Actual/ Average

Proportion number of

predicted predicted

HCC deceased

unique Actual Predicted

ratio Actual Predicted ratio

Score during year

diagnoses

15. 87 10. 23 1. 55 4.78 5. 19 0. 92 0.97 0.03 15. 07 18. 56 11. 55 1. 61
4.76 4. 86 0. 98 0.95 0.03 16. 49 15. 40 9. 79 1. 57 4. 40 5. 00 0. 88 1. 00
0. 03 14.07 15. 45 10. 42 1. 48 5.26 5. 04 1. 04 0.97 0.03 14. 76 16. 22 10.
12 1. 60 5.07 5. 00 1. 01 0.98 0.02 15. 84 15. 24 9. 53 1. 60 4. 25 4. 89 0.
87 0. 80 0. 01 13.91 13. 76 9. 47 1. 45 5. 30 4. 84 1. 10 0. 84 0. 02 13.92

10. 46 10. 04 1. 04 4.97 5. 48 0. 91 1.07 0.05 15. 49 10. 19 10. 43 0. 98
6.34 5. 85 1. 08 1.27 0.06 15. 90 10. 87 10. 56 1. 03 5.37 5. 74 0. 94 1.19
0.05 16. 70 10. 13 10. 56 0. 96 5.98 5. 42 1. 10 1.22 0.05 16. 80 10. 31 10.
70 0. 96 5.60 5. 43 1. 03 1.20 0.06 16. 09 9. 40 10. 02 0. 94 5. 30 5. 56 0.
95 1. 18 0. 04 15.30 11. 46 11. 39 1. 01 5.78 5. 63 1. 03 1.30 0.05 17. 33

a Represents about 422,615 Medicare fee- for- service beneficiaries in the
demonstration area. Excludes members of Medicare+ Choice plans at any time
during 1999, military retirees, persons with ESRD, Medicaid beneficiaries,
persons with disabilities (under age 65), and those who lost Medicare part A
or part B entitlement for reasons other than death. Source: GAO analysis of
DOD encounter and claims data and Medicare 20- percent fee- for- service
sample.

Appendi x V I Crowd- Out of Nonenrollees ?Crowd- outs? were nonenrollees who
had used MTF care before the demonstration but were unable to do so after
the demonstration started. In this report, we define crowd- outs as those
4,594 nonenrollees (6 percent of all nonenrollees) who had, according to
their survey answers, received all or most of their care at an MTF before
the demonstration but received none or only some of their care at an MTF
after the demonstration started. 1

However, as table 17 shows, crowd- out can be defined either more narrowly
or more broadly. By the narrowest definition of crowd- out- those
nonenrollees who received all of their care at an MTF before the
demonstration but none of their care at an MTF after the demonstration
started- only 1,498 persons (2 percent of all nonenrollees) were crowded
out. However, if we count all those who received less care than before,

12, 133 (16 percent of nonenrollees) nonenrollees were crowded out.

Table 17: Change in MTF Utilization Among Nonenrollees MTF use during
demonstration None Some Most All Total MTF use before demonstration None 51,
261 2,413 167 693 54, 534

Some 7,113 4,424 403 313 12, 253 Most 1,064 1,231 699 222 3, 216 All 1, 498
l 801 426 1, 269 3, 994

Tot al 60, 936 8,869 1, 695 2, 497 73, 997 Note: Outlined box contains
nonenrollees who were crowded out. Italicized number refers to narrowest
definition of crowd- out. Shaded area represents broadest definition of
crowd- out.

Source: GAO Survey of Medicare- Eligible Military Retirees and Family
Members.

Change in Satisfaction As expected, many of the 4,594 nonenrollees whom we
characterized as With Access to Military

crowd- outs changed their attitudes toward military care during the
demonstration. As shown in table 18, they reported a decline in access to
Care Among CrowdOuts MTF care as well as lower satisfaction with care in
MTFs. However, they did not report significant changes in satisfaction on
issues not explicitly connected to MTFs.

1 A small number of nonenrollees (428) answered this question in only one of
the two surveys. In these cases, we used DOD and Medicare fee- for- service
administrative data to impute the missing answer.

Table 18: Change in Self- Reported Access to MTF Care and Satisfaction With
MTF Care Among Crowd- Outs

Numbers in percent

Net Improved Declined change

Satisfaction with access to military care

Able to get care at military facilities when I needed it. 10 53 -43 a
Difficult to schedule appointments at military facilities. 14 61 -47 a I
prefer to get my care at military facilities. 10 37 -27 a

Satisfaction with military care

Satisfied with care at military facilities. 12 42 -30 a Doctors and staff
did not treat me with respect at

14 40 -26 a military facilities. I would not recommend military care. 17 37
-20 a

a Significant at .05 level. Source: GAO Survey of Medicare- Eligible
Military Retirees and Family Members.

Trend in Utilization DOD?s MTF encounter data and network claims data
confirmed the selfreports Among Crowd- Outs

of crowd- outs. The crowd- outs? MTF outpatient care dropped dramatically
during the demonstration and the increase in fee- for- service (FFS)
outpatient visits was not sufficient to offset this decline. However, as
shown in table 19, there was no decline in acute hospitalizations.

Table 19: Changes in Utilization of Nonenrollees Who Were Crowded Out of
MTFs Before

During Type of utilization demonstration demonstration

MTF - Acute hospitalization a 0.14 0.14 MTF - Outpatient physician visits b
7.44 0.26 FFS - Acute hospitalization a 0.10 0.27 FFS - Outpatient physician
visits b 4.30 6.93 a Number of hospital stays per person. b Number of
outpatient physican visits per person.

Sources: GAO Survey of Medicare- Eligible Military Retirees and Family
Members and GAO analysis of DOD claims and encounter data and Medicare 20-
percent fee- for- service sample.

Appendi x V

Health Outcomes Analysis In this appendix, we describe our methods for
analyzing the effects of the subvention demonstration on three indicators of
health outcomes- mortality, health status, and preventable hospitalization.

Mortality Analysis Using our first survey, we calculated the mortality rate
from the date of the survey response to January 31, 2001. The source of
death information was

the Medicare Enrollment Database. We excluded Medicare+ Choice members
because we could not obtain their diagnoses, which we needed to calculate
risk factors. The unadjusted 2- year mortality rate was 0.06 for Senior
Prime enrollees and 0.08 for nonenrollees. 1 Although the difference is
significant, it disappears when we adjust for individual risk. The adjusted
2- year mortality rate is 0.06 for both enrollees and nonenrollees. (See
table 20.)

Table 20: Profile of 2- Year Mortality Rate Actual mortality rate Adjusted
mortality rate

Senior Prime enrollees 0.06 0.06 Nonenrollees 0.08 0.06 Source: GAO Survey
of Medicare- Eligible Military Retirees and Family Members and the Medicare
Enrollment Database

We used the Cox proportional hazard model to calculate individuals?
riskadjusted mortality rate. A hazard ratio greater than 1 indicates a
higher risk of death while a hazard ratio less than 1 indicates a lower
risk. 2 For

example, a hazard rate for males of 1.5 means that males are 50 percent more
likely to die than females, holding other factors constant. Similarly, a
hazard rate of 0. 5 for retirees with HCC scores in the lowest quartile
means that they are 50 percent less likely to die than those with HCC scores
in the middle two quartiles, holding other factors constant. Enrollment in
Senior 1 Using either the Kaplan- Meier method or the life- table method.
See J. D. Kalbfleisch and R. L. Prentice, The Statistical Analysis of
Failure Time Data, John Wiley & Sons, 1980,

pp. 10- 19. 2 See J. D. Kalbfleisch and R. L. Prentice, The Statistical
Analysis of Failure Time Data,

John Wiley & Sons, 1980, pp. 70- 118. For the computational method, see
STATA Statistical Software, Release 5, Estimate Cox proportional hazards
model, pp. 252- 271, Reference P- Z.

Prime did not have a significant effect on mortality. (See table 21 for a
description of the factors that entered our model and of their estimated
effects.)

Table 21: Factors Affecting 2- Year Mortality Rate 95% Hazard confidence
Significance Characteristic ratio

interval level a Demographics

Age: 65 - 69 0.604 0. 481 0.758 0. 000 d Age: 70 - 74 0.618 0. 489 0.781 0.
000 d Age: 75 - 79 0.816 0. 654 1.017 0. 070 Age: 85+ 1. 426 1.072 1. 898 0.
015 Male 1.496 1. 288 1.739 0. 000 d Marital status: Separated or divorced
1.351 1. 004 1.817 0. 047

Health status at time of responding to first survey

SF- 12 b physical score: 55+ (highest quartile) 0. 701 0.515 0. 954 0. 024
Self- evaluated health status: very good 0. 637 0.499 0. 814 0. 000 d Self-
evaluated health status: poor 2.106 1. 747 2.540 0. 000 d

Prior utilization during the past 12 months

Number of outpatient visits: none 1.599 1. 230 2.077 0. 000 d Number of
hospitalizations: 5 - 9 1.663 1. 145 2.413 0. 007

(Continued From Previous Page)

95% Hazard confidence Significance Characteristic ratio

interval level a Coexisting clinical conditions

HCC score in the lowest quartile: 0. 075 or 0.489 0. 331 0.722 0. 000 d
lower HCC score in the highest quartile: 1. 31 or 5.425 4. 494 6.547 0. 000
d

higher Current smoker 1.625 1. 329 1.986 0. 000 d Assistance required with
activities of daily

2.909 2. 409 3.512 0. 000 d living Neoplasm c 1.258 1. 083 1.461 0. 003

Mental disease c 1.395 1. 124 1.733 0. 003

Enrollment status at time of first survey

Senior Prime enrollee 0. 977 0.842 1. 134 0. 762 a The significance level
applies to the z- test.

b A standard scale for measuring self- reported health status. c ICD- 9
classification. d Significance level is less than 0.0005.

Sources: GAO Survey of Medicare- Eligible Military Retirees and Family
Members and the Medicare Enrollment Database and GAO analysis of DOD claims
and encounter data and Medicare 20- percent fee- for- service sample.

Health Status Analysis We used the SF- 12, a standard scale 3 for measuring
self- reported physical and mental health status. At the beginning of the
demonstration, the enrollees had slightly higher SF- 12 scores than
nonenrollees (that is, they reported that they were healthier), but the
difference between enrollees and nonenrollees was very small and not
significant. This was also true when we repeated the scale at the end of the
demonstration. (See table 22.)

3 See Ware, J. E., Kosinski, M., and Keller, S. D., SF- 12: How to Score the
SF- 12 Physical and Mental Health Summary Scales, The Health Institute, New
England Medical Center, Second Edition, pp. 12- 13.

Table 22: Profile of SF- 12 Scores Average

Average Adjusted

SF- 12 score in SF- 12 score in

Actual change change of

first survey second survey of SF- 12 score

SF- 12 score a

Senior Prime 42.73 43. 05 0. 32 0. 18

enrollees Nonenrollees 42.10 42. 23 0. 13 0. 19 Note: The difference in the
average SF- 12 score between Senior Prime enrollees and nonenrollees was not
significant (p > 0.05) in either the first or second survey. The change in
SF- 12 score did not differ between enrollees and nonenrollees.

a A linear regression was used to adjust for demographic and other factors.
Source: GAO Survey of Medicare- Eligible Military Retirees and Family
Members.

The change in the score between the two times was also insignificant. We
examined both the unadjusted score and the adjusted score, using a linear
regression model (see table 23), but neither was significant, and enrollment
in Senior Prime was not a significant factor in the model.

Table 23: Factors Affecting Change in SF- 12 Score 95% confidence
Significance Characteristic Coefficient interval level a

Demographics

Age > 75 -1. 836 -2. 322 -1.350 0. 000 b Female -0. 953 -1. 441 -0.465 0.
000 b

Health status at time of responding to first survey

Self- evaluated general health status: 1.694 1. 083 2. 307 0. 000 b
Excellent or very good Self- evaluated general health status: -1. 548 -2.
288 -0.807 0. 000 b

Fair or poor SF- 12 physical health score -0. 369 -0. 398 -0. 339 0.000 b
SF- 12 mental health score 0. 151 0. 119 0.183 0. 000 b

Health conditions HCC score -0. 443 -0. 649 -0. 237 0. 000 b Change in
number of chronic conditions -1. 421 -1. 764 -1. 078 0.000 b History of
heart disease -1. 170 -1.745 -0.594 0. 000 b

(Continued From Previous Page)

95% confidence Significance Characteristic Coefficient interval level a

History of diabetes -0. 995 -1.649 -0.342 0. 003 History of lung disease
(COPD) -2. 032 -2. 913 -1.152 0. 000 b Overweight -0. 741 -1. 334 -0. 149
0.014 Current smoker -0. 443 -0. 938 0. 053 0. 080

Others

Two or more outpatient visits during the -0. 721 -1. 280 -0.161 0. 012 past
12 months Would recommend military health care

0.822 0. 168 1. 474 0. 014 at time of first survey Senior Prime enrollee 0.
209 -0. 186 0. 605 0. 300

a The significance level applies to the t- test. b Significance level is
less than 0.0005.

Sources: GAO survey of Medicare- Eligible Retirees and Family Members and
GAO analysis of DOD claims and encounter data and Medicare 20- percent fee-
for- service sample.

Preventable We analyzed preventable hospitalizations- hospital stays that
can often be avoided by appropriate outpatient care- using several alternate
models. 4 Hospitalizations

Specifically, we estimated the effect of Senior Prime enrollment on the
likelihood of having a preventable hospitalization, adjusting for age, sex,
and health conditions. Measures of a person?s health conditions included the
HCC score, an index of comorbidities, 5 and the number of recent
hospitalizations. In addition, we controlled for the number of outpatient

clinic and physician visits, since outpatient care is considered a means of
preventing hospitalization. We analyzed data on Senior Prime enrollees and
on Medicare fee- forservice beneficiaries who were not military retirees and
who lived in the demonstration areas. Within this combined group of
enrollees and fee- forservice beneficiaries, we modeled preventable
hospitalizations for two populations: (1) those who had been hospitalized in
1999 and (2) those who had at least one chronic disease 6 in 1999- whether
they had been hospitalized or not.

Our analysis of the demonstration?s effect on preventable hospitalizations
yielded inconsistent results. For the first population (hospitalizations),
we found that Senior Prime enrollment was associated with more preventable
hospitalizations. By contrast, for the second population (the chronically
ill), Senior Prime enrollment was associated with fewer preventable

hospitalizations. 4 Our models were formulated as logistic regressions. 5
This index, known as the Deyo- Charlson Comorbidity Index, enables patients
to be classified from less ill to more ill. See Deyo, R. A., Cherkin, D. C.,
& Ciol, M. A., ?Adapting a Clinical Comorbidity Index for Use with ICD- 9-
CM Administrative Databases,? Journal of Clinical Epidemiology, 1992, 45: 6,
pp. 613- 619.

6 Chronic diseases that may result in preventable hospitalizations include
angina, chronic obstructive pulmonary disease, hypertension, congestive
heart failure, diabetes, and urinary tract infection.

Appendi x VI Comments From the Department of Defense

Comments From the Centers for Medicare and

Appendi x VII Medicaid Services

Appendi x VI II

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

Other GAO staff who made significant contributions to this work included
Acknowledgments

Jessica Farb, Maria Kronenburg and Dae Park. Robin Burke provided technical
advice and Martha Wood provided technical advice and assistance.

Related GAO Products

Medicare Subvention Demonstration: DOD Costs and Medicare Spending

(GAO- 02- 67, Oct. 31, 2001).

Medicare Subvention Demonstration: DOD?s Pilot Appealed to Seniors,
Underscored Management Complexities (GAO- 01- 671, June 14, 2001).

Medicare Subvention Demonstration: Enrollment in DOD Pilot Reflects Retiree
Experiences and Local Markets (GAO/ HEHS- 00- 35, Jan. 31, 2000).

Defense Health Care: Appointment Timeliness Goals Not Met; Measurement Tools
Need Improvement (GAO/ HEHS- 99- 168, Sept. 30, 1999).

Medicare Subvention Demonstration: DOD Start- up Overcame Obstacles, Yields
Lessons, and Raises Issues (GAO/ GGD/ HEHS- 99- 161, Sept. 28, 1999).

Medicare Subvention Demonstration: DOD Data Limitations May Require
Adjustments and Raise Broader Concerns (GAO/ HEHS- 99- 39, May 28, 1999).

(290035) Lett er

GAO?s Mission The General Accounting Office, the investigative arm of
Congress, exists to support Congress in meeting its constitutional
responsibilities and to help improve the performance and accountability of
the federal government for the American people. GAO examines the use of
public funds; evaluates federal programs and

policies; and provides analyses, recommendations, and other assistance to
help Congress make informed oversight, policy, and funding decisions. GAO?s
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of The fastest and easiest way to obtain copies of GAO
documents is through the Internet. GAO?s Web site (www. gao. gov) contains
abstracts and full- text files of GAO Reports and current reports and
testimony and an expanding archive of older products. The Testimony Web site
features a search engine to help you locate documents using key words and
phrases. You can print these documents in their entirety, including charts
and other graphics.

Each day, GAO issues a list of newly released reports, testimony, and
correspondence. GAO posts this list, known as ?Today?s Reports,? on its Web
site daily. The list contains links to the full- text document files. To
have GAO E- mail this list to you every afternoon, go to our home page and
complete the easy- to- use electronic order form found under ?To Order GAO
Products.?

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 P. O. Box 37050 Washington, D. C. 20013

To order by Phone: Voice: (202) 512- 6000 TDD: (301) 413- 0006 Fax: (202)
258- 4066

Visit GAO?s Document GAO Building Distribution Center

Room 1100, 700 4th Street, NW (corner of 4th and G Streets, NW) Washington,
D. C. 20013

To Report Fraud, Contact: Waste, and Abuse in

Web site: www. gao. gov/ fraudnet/ fraudnet. htm, E- mail: fraudnet@ gao.
gov, or

Federal Programs 1- 800- 424- 5454 (automated answering system).

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

a

GAO United States General Accounting Office

Page i GAO- 02- 68 Access to Care In DOD Medicare Subvention

Contents

Contents

Page ii GAO- 02- 68 Access to Care In DOD Medicare Subvention

Contents

Page iii GAO- 02- 68 Access to Care In DOD Medicare Subvention

Contents

Page iv GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 1 GAO- 02- 68 Access to Care In DOD Medicare Subvention United States
General Accounting Office

Washington, D. C. 20548 Page 1 GAO- 02- 68 Access to Care In DOD Medicare
Subvention

A

Page 2 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 3 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 4 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 5 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 6 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 7 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 8 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 9 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 10 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 11 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 12 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 13 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 14 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 15 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 16 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 17 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 18 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 19 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 20 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 21 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 22 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 23 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 24 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix I

Appendix I Health Care Survey of Subvention Demonstration Beneficiaries

Page 25 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix I Health Care Survey of Subvention Demonstration Beneficiaries

Page 26 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix I Health Care Survey of Subvention Demonstration Beneficiaries

Page 27 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix I Health Care Survey of Subvention Demonstration Beneficiaries

Page 28 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix I Health Care Survey of Subvention Demonstration Beneficiaries

Page 29 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix I Health Care Survey of Subvention Demonstration Beneficiaries

Page 30 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix I Health Care Survey of Subvention Demonstration Beneficiaries

Page 31 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix I Health Care Survey of Subvention Demonstration Beneficiaries

Page 32 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 33 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II

Appendix II Measures of Access to Care And Satisfaction With Care

Page 34 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 35 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 36 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 37 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 38 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 39 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 40 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 41 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 42 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 43 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 44 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 45 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 46 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 47 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 48 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 49 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 50 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 51 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 52 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 53 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 54 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 55 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 56 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 57 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 58 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 59 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 60 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 61 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 62 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix II Measures of Access to Care And Satisfaction With Care

Page 63 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 64 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix III

Appendix III Models of Utilization

Page 65 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix III Models of Utilization

Page 66 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix III Models of Utilization

Page 67 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix III Models of Utilization

Page 68 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix III Models of Utilization

Page 69 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix III Models of Utilization

Page 70 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix III Models of Utilization

Page 71 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 72 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix IV

Appendix IV Crowd- Out of Nonenrollees

Page 73 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 74 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix V

Appendix V Health Outcomes Analysis

Page 75 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix V Health Outcomes Analysis

Page 76 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix V Health Outcomes Analysis

Page 77 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix V Health Outcomes Analysis

Page 78 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix V Health Outcomes Analysis

Page 79 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 80 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix VI

Appendix VI Comments From the Department of Defense

Page 81 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Page 82 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix VII

Page 83 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Appendix VIII

Page 84 GAO- 02- 68 Access to Care In DOD Medicare Subvention

Related GAO Products Page 86 GAO- 02- 68 Access to Care In DOD Medicare
Subvention

United States General Accounting Office Washington, D. C. 20548- 0001

Official Business Penalty for Private Use $300

Address Correction Requested Presorted Standard

Postage & Fees Paid GAO Permit No. GI00
*** End of document. ***