Medicaid Section 1115 Waivers: Flexible Approach to Approving
Demonstrations Could Increase Federal Costs (Chapter Report, 11/08/95,
GAO/HEHS-96-44).
Pursuant to a congressional request, GAO examined the financing
arrangements for four approved section 1115 Medicaid demonstration
waivers.
GAO found that: (1) the approved spending limits for demonstration
waivers in Oregon, Hawaii, and Florida are not budget neutral and could
increase federal Medicaid expenditures; (2) Tennessee's 1115 waiver
agreement should cost less than the continuation of its prewaiver
program and result in savings; (3) although additional federal funding
is relatively lower than demonstration spending under federal
expenditure caps, federal Medicaid expenditures could increase rapidly
if similar flexibility in reviewing state 1115 financing strategies is
allowed; (4) five waivers have been approved since late 1994, and the
backlog of pending waivers includes three states with large Medicaid
programs; (5) additional federal funding is available to protect against
the uncertainties states face in implementing demonstrations; and (6) it
remains unclear whether the states implementing demonstrations will
exceed their 1115 waiver funding caps.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: HEHS-96-44
TITLE: Medicaid Section 1115 Waivers: Flexible Approach to
Approving Demonstrations Could Increase Federal Costs
DATE: 11/08/95
SUBJECT: Medicaid programs
Health maintenance organizations
State-administered programs
Waivers
Health care cost control
Federal/state relations
Medical economic analysis
Financial management
Health resources utilization
IDENTIFIER: Tennessee
Florida
Oregon
Hawaii
Minnesota
Illinois
New York
Kentucky
Medicaid Disproportionate Share Hospital Program
Texas
Oklahoma
Arizona
MinnesotaCare Program
Medicaid Medically Needy Program
Rhode Island
Ohio
**************************************************************************
* This file contains an ASCII representation of the text of a GAO *
* report. Delineations within the text indicating chapter titles, *
* headings, and bullets are preserved. Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are *
* identified by double and single lines. The numbers on the right end *
* of these lines indicate the position of each of the subsections in the *
* document outline. These numbers do NOT correspond with the page *
* numbers of the printed product. *
* *
* No attempt has been made to display graphic images, although figure *
* captions are reproduced. Tables are included, but may not resemble *
* those in the printed version. *
* *
* A printed copy of this report may be obtained from the GAO Document *
* Distribution Facility by calling (202) 512-6000, by faxing your *
* request to (301) 258-4066, or by writing to P.O. Box 6015, *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time. *
**************************************************************************
Cover
================================================================ COVER
Report to the Ranking Minority Member, Committee on Finance, U.S.
Senate
November 1995
MEDICAID SECTION 1115 WAIVERS -
FLEXIBLE APPROACH TO APPROVING
DEMONSTRATIONS COULD INCREASE
FEDERAL COSTS
GAO/HEHS-96-44
Medicaid 1115 Demonstrations
(108200)
Abbreviations
=============================================================== ABBREV
CPI - Consumer Price Index
DSH - Disproportionate Share Hospital Program
ERISA - Employee Retirement Income Security Act of 1974
HCFA - Health Care Financing Administration
HHS - Department of Health and Human Services
HMO - health maintenance organization
OMB - Office of Management and Budget
PCCM - primary care case management
PPO - preferred provider organization
PPPM - per person per month
SFY - state fiscal year
Letter
=============================================================== LETTER
B-257396
November 8, 1995
The Honorable Daniel Patrick Moynihan
Ranking Minority Member
Committe on Finance
United States Senate
Dear Senator Moynihan:
This report, prepared at your request, examines the financing
arrangements for four approved section 1115 Medicaid demonstration
waivers.
As arranged with your office, unless you publicly announce its
contents earlier, we will make no further distribution of this report
until 30 days after its issue date. At that time, we will send
copies to the Secretary of Health and Human Services; the
Administrator, Health Care Financing Administration; the Director,
Office of Management and Budget; and other congressional committees.
Copies of this report will also be made available to others upon
request.
If you or your staff have any questions, please call me at (202)
512-7123. Major contributors to this report are listed in appendix
VI.
Sincerely yours,
William J. Scanlon
Associate Director
Health Financing Issues
EXECUTIVE SUMMARY
============================================================ Chapter 0
PURPOSE
---------------------------------------------------------- Chapter 0:1
In what amounts to an ambitious experiment to determine whether the
Medicaid program can add several million new beneficiaries without
increasing federal costs, the administration has approved requests
for 11 demonstration waivers authorized under section 1115 of the
Social Security Act. The cost of extending Medicaid coverage to
uninsured individuals under these demonstrations is said to be offset
by the reinvestment of managed care savings. The administration has
asserted that approval for such a major restructuring hinges on
budget neutrality, that is, federal costs over 5 years must be no
more than if the state had opted to continue its smaller, prewaiver
Medicaid program. (A glossary at the end of this report defines
budget concepts and other program terminology.) In fact, some states
suggest that the net result will be lower costs--even though managed
care savings are being reinvested. Eleven more states have pending
waivers and other applications are anticipated.
Because a significant number of states may eventually restructure and
operate their Medicaid programs under "comprehensive" 1115
demonstration waivers,\1 the Ranking Minority Member of the Senate
Committee on Finance asked GAO to examine the financing arrangements
for applications approved since 1992, concentrating on the potential
net impact on federal Medicaid expenditures. GAO focused its review
on section 1115 waivers in Tennessee, Florida, Oregon, and Hawaii,
but also closely monitored other approved and pending waivers.
--------------------
\1 The administration refers to section 1115 waivers that seek to
restructure state Medicaid programs as "comprehensive" to distinguish
them from other, less ambitious 1115 demonstrations.
BACKGROUND
---------------------------------------------------------- Chapter 0:2
To constrain rising health care costs, states are increasingly
turning to mandatory enrollment of some or all Medicaid beneficiaries
in managed care delivery plans--arrangements that limit a
beneficiary's choice of physicians and hospitals. In many cases,
these managed care plans are prepaid a fixed amount per enrollee.
This financing arrangement--known as "capitation"--has demonstrated
the ability to lower service utilization, which in turn can hold down
costs. In order to implement mandatory managed care programs, states
must obtain a waiver of certain Medicaid requirements. The waiver
authority that gives states the greatest flexibility--it allows the
executive branch to waive most federal Medicaid requirements for
demonstration projects likely to assist in promoting program
objectives--is section 1115 of the Social Security Act.
Since 1992, 22 states have asked to use this demonstration authority
to restructure their Medicaid programs. As of September 1995, 5 of
the 11 section 1115 waivers approved by the administration were being
implemented--those for Tennessee, Hawaii, Oregon, Rhode Island, and
Minnesota. (App. III summarizes the status of all approved and
pending waivers.) The common thread present in what have been coined
comprehensive 1115 waivers--complex, sometimes controversial
proposals often tied to broader health care reforms--is (1) the
switch from a fee-for-service to a managed care approach to
delivering health benefits to traditional beneficiaries and (2) the
use of anticipated savings plus other funding streams to expand
coverage to groups previously ineligible for Medicaid. In return for
extending coverage, the federal government has increased states'
flexibility to use managed care as a cost-containment strategy. Only
a few states have applied for 1115 waivers without also proposing to
expand coverage.
The transformation of Medicaid through 1115 waivers is being
facilitated by a new federal flexibility in assessing the budget
neutrality of such demonstrations, particularly the administration's
openness to new methods to estimate what the continuation of a
state's existing Medicaid program would have cost. By redefining the
terms of budget neutrality, the administration has made it easier for
states to demonstrate that the waiver will cost less than the
existing Medicaid program. Each state with an approved waiver has a
5-year expenditure cap on the federal share of its demonstration. As
spelled out in the waiver agreement, yearly expenditure targets may
be exceeded, but total spending over the 5-year period of the
demonstration must be at or under the overall cap.
According to waiver applications, states are relying on a mix of
funding sources to pay for extending Medicaid to some portion of
their uninsured. Analytically, these funding sources can be grouped
into two major categories:
redirected resources obtained by eliminating or reducing existing
programs, such as Medicaid's Disproportionate Share Hospital
Program (DSH),\2 public health programs, or optional Medicaid
coverage expansions, and
expected savings, primarily from different forms of managed care
delivery but also from other Medicaid reimbursement reforms or
reductions in benefits.
The extent of reliance on a particular funding source varies from
state to state and, contrary to the conventional wisdom, managed care
savings predominate in only one of the four waiver applications we
examined. In addition, many states collect user fees--the premiums
charged to new program enrollees. Finally, in a few cases, state
funds used to subsidize insurance programs for low-income residents
are being folded into the waiver.
--------------------
\2 DSH compensates hospitals that serve the uninsured and a large
Medicaid clientele.
RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3
Contrary to the administration's assertion, the approved spending
limits for demonstration waivers in Oregon, Hawaii, and Florida are
not budget neutral and could increase federal Medicaid expenditures.
Only Tennessee's 1115 waiver agreement should cost no more than the
continuation of its smaller, prewaiver program and, in fact, should
result in savings. Overall, the net additional federal funding is
small in relation to demonstration spending allowed under federal
expenditure caps--likely less than 3 percent. However, federal
Medicaid expenditures could grow significantly if the administration
continues to show a similar flexibility in reviewing state 1115
financing strategies. Five waivers have been approved since
Florida's in late 1994, and the large backlog of pending waivers
includes three states with large Medicaid programs--New York,
Illinois, and Texas.
Additional federal dollars are available along with other funding
sources identified in state waiver applications. GAO believes that
the potential for additional federal funding serves as a hedge
against the many challenges and uncertainties states face in
implementing these ambitious demonstrations--including changing
economic conditions, the accuracy of cost-containment assumptions,
the availability of anticipated funding identified in waiver
applications, and the lack of reliable cost data on the uninsured.
The three states in GAO's sample currently implementing
demonstrations are already reacting to these challenges. How they
adjust to differences between expectations and actual experience in a
changing political landscape will determine whether they live within
or exceed their 1115 waiver funding caps.
PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4
WAIVER AGREEMENT SPENDING
LIMITS COULD INCREASE
FEDERAL COSTS
-------------------------------------------------------- Chapter 0:4.1
To determine whether the four approved waivers were budget neutral,
GAO compared (1) the estimated cost of the traditional Medicaid
program--absent the demonstration--with (2) each state's capped
budget under its waiver funding agreement.
Projecting growth in each state's traditional Medicaid program was
challenging because of the lack of consistently generated,
state-specific forecasts. Lacking such data, GAO relied on the only
available forecasts--the Office of Management and Budget's national
projections, on a current services basis, of how the traditional
Medicaid program would grow over the 5-year duration of 1115
waivers.\3 GAO then examined state-specific information to identify
any reasons why growth in future expenditures for the traditional
program should exceed expectations of average growth nationwide.
Using this methodology, the waiver expenditure caps negotiated by the
administration exceeded the expected costs of continuing the
traditional Medicaid program in three states--Oregon, Hawaii, and
Florida. GAO saw no support in state-specific data to justify the
differences. Only Tennessee's 1115 demonstration is budget neutral.
Administration officials told GAO that, since some states' Medicaid
expenditures had been growing faster than the national average, the
budget neutrality of each proposed waiver should be evaluated
independently in order to capture these variations. However, neither
the negotiating record nor other state-specific data analyzed by GAO
justified higher than average future funding for Florida, Oregon, or
Hawaii. Even in the case of Florida, whose program had been growing
faster than the national average, the state's own estimates showed
that key factors contributing to past growth were not expected to be
sustained.
--------------------
\3 A current services forecast provides a policy-neutral benchmark
that only reflects medical inflation, normal growth in the eligible
population, and changes in utilization for the entire Medicaid
program.
ACCESS TO ADDITIONAL FUNDING
HAS NOT ELIMINATED COST
PRESSURES
-------------------------------------------------------- Chapter 0:4.2
Rather than precise funding road maps, state financing strategies to
pay for expanded coverage under an 1115 waiver are outlines subject
to the test of reality. The uncertainties states face in
implementing these ambitious waivers make it difficult to predict the
extent to which the "cushion" provided by access to additional
federal dollars will be needed to supplement the redirected funds and
program savings outlined in the waiver applications. During the
first year of implementation, these uncertainties have played out in
different ways for the three states in GAO's review that are
currently implementing waivers.\4 For diverse reasons, all three
states are reacting to pressures to reduce waiver expenditures.
Oregon: Unlike Tennessee, Hawaii, and Florida, coverage expansion
plans in Oregon rely predominately on managed care savings.
Increased managed care savings resulting from significantly higher
enrollment in fully capitated plans allowed the state to spend less
than the federal expenditure cap--despite the enrollment of a
greater-than-estimated number of uninsured residents. These
optimistic first-year results, however, have not dampened the concern
of state officials about future waiver costs. Under the waiver
agreement, federal matching funds for newly eligible individuals are
linked to the number of traditional eligibles, a major component of
which experienced a numerical decline in 1994. Should this decline
continue, fewer new eligibles would qualify for federal matching
funds. State concerns also stem from (1) higher-than-expected costs
of new eligibles; (2) a number of new, non-health-related, spending
priorities; and (3) the impact of a 1991 tax limitation initiative.
Oregon began implementing a number of cost-containment measures such
as new eligibility rules in October 1995. The state is still waiting
for the administration to approve a reduction in benefits and
premiums for newly eligible enrollees.
Hawaii: Only Hawaii exceeded its federal waiver funding cap for the
first year--costs that must be offset in later years. Any advantage
from the current slowdown in medical inflation was offset by the
surge in newly eligible individuals who signed up for coverage under
the waiver. The state attributes this higher-than-expected
enrollment to Hawaii's current economic slowdown and to individuals
who dropped private insurance in favor of less expensive coverage
through the waiver. Hawaii is taking a number of steps--including
increasing premium contributions from new enrollees and using
stricter eligibility standards--to lower its waiver costs.
Tennessee: After maintaining open enrollment for a full year and
achieving about 80 percent of its coverage expansion goal, Tennessee
abruptly cut off enrollment as a result of a budget crisis that it
attributes, in part, to the 1115 waiver. In contrast to Hawaii,
Tennessee's first year demonstration costs were well below its
federal spending cap and on a par with the cost of its significantly
smaller prewaiver program--even though the demonstration added
several hundred thousand previously uninsured individuals. Moreover,
the state has lowered its waiver spending estimates for future years.
According to administration officials, lower-than-expected
expenditures may be linked to problems in identifying state matching
funds. They told GAO that Tennessee's elimination of its DSH program
could have served as a major source of funding for coverage
expansion; however, the state also discontinued the hospital tax that
served as the source of state match with the onset of waiver
implementation. Primarily because of the steep price discount that
the state required from participating managed care plans, Tennessee's
waiver expenditure cap may have provided more money than the state
actually needed to expand coverage.
Section 1115 Medicaid waivers were approved during a period of
economic recovery and a slowdown in medical inflation. The recession
of the early 1990s coupled with rapid medical price increases serves
as a reminder of the risks posed by fixed-cost agreements in
variable-cost environments. While states have benefited from recent
economic trends, the potential for a resurgence in medical inflation,
a recession, and a large increase in the number of traditional and/or
newly eligible beneficiaries could create problems for three states
currently implementing waivers--Tennessee, Oregon, and Hawaii.
Moreover, the political environment has already changed considerably
since many of these waivers were first conceived and approved.
National health care reform no longer appears imminent, and even
reform-minded states have retreated from earlier, broader goals. A
new cost-consciousness--unrelated to current economic conditions--has
emerged. Clearly, 1115 waivers will continue to evolve, perhaps in
unanticipated ways, as they are implemented.
--------------------
\4 In 1994 and again earlier this year, Florida's legislature refused
to authorize implementation of the waiver.
RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:5
GAO is making no recommendations in this report.
AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:6
In commenting on a draft of this report, the Department of Health and
Human Services and the Office of Management and Budget disagreed with
GAO's conclusion that the waiver funding caps for Oregon, Hawaii, and
Florida are not budget neutral. GAO continues to believe that the
administration's waiver funding caps for these states may result in
increased federal spending. A detailed discussion of the
administration's comments and GAO's response appears in chapter 5.
The administration's comments are reprinted in appendix V.
INTRODUCTION
============================================================ Chapter 1
Financed jointly by the federal government and states, Medicaid is
the nation's health care lifeline for two statutorily defined groups
of low-income residents--families, primarily women and children; and
the aged, blind, and disabled.\5 In reality, Medicaid is not 1, but
rather 56 separate programs that differ dramatically across states.\6
While federal statute mandates who is eligible for coverage and the
broad categories of services that must be provided, each
participating state designs and administers its own program by (1)
setting certain income and asset eligibility requirements, (2)
selecting which optional groups and services to cover, and (3)
determining the scope of mandatory and optional services. As a
result of this flexibility, Medicaid is not available to everyone who
is poor. In 1993, Medicaid provided health care coverage to less
than half of those with incomes below the poverty level.\7 The Health
Care Financing Administration (HCFA) within the Department of Health
and Human Services (HHS) monitors each state program for compliance
with federal regulations. (App. II contains more background
information on the Medicaid program.)
The $130 billion Medicaid program is at a crossroads. Between 1985
and 1993, Medicaid costs tripled and the number of beneficiaries
increased by over 50 percent. Current projections suggest that
program costs will double over the next 5 to 7 years. To constrain
rising health care costs, states are increasingly turning to
mandatory enrollment of some or all Medicaid beneficiaries in managed
care delivery plans--arrangements that limit a beneficiary's choice
of physicians and hospitals. In many cases, these managed care plans
are prepaid a fixed amount per enrollee. This financing arrangement
has demonstrated the ability to lower service utilization, which in
turn can hold down costs.
In order to implement mandatory managed care programs, states must
obtain a waiver of certain Medicaid requirements. The waiver
authority that gives states the greatest flexibility is section 1115
of the Social Security Act. Section 1115 allows the executive branch
to waive most federal Medicaid requirements for demonstration
projects likely to assist in promoting program objectives.\8 Since
1992, 22 states have asked to use this demonstration authority to
restructure their Medicaid programs. The common thread present in
what have been termed "comprehensive" 1115 demonstration waivers is
(1) the switch from a fee-for-service to a managed care approach to
delivering health benefits and (2) the use of anticipated savings
plus other funding streams to expand coverage to groups previously
ineligible for Medicaid.
--------------------
\5 The federal government matches state expenditures according to a
prescribed formula based on state per capita income, providing, on
average, 58 cents for every dollar spent. Federal matching for
individual states varies from a minimum of 50 percent to a maximum of
83 percent of total Medicaid costs.
\6 All 50 states plus the District of Columbia, American Samoa, Guam,
the Northern Mariana Islands, Puerto Rico, and the Virgin Islands
have Medicaid programs.
\7 Throughout this report, "poverty level" refers to the federal
poverty guidelines, which are used to establish eligibility for
certain federal assistance programs. The guidelines are updated
annually to reflect changes in the cost of living and vary according
to family size. Guidelines are uniform across the continental United
States and slightly higher for Alaska and Hawaii. For example, in
1994 the poverty level for a family of four in the continental United
States was $14,800, while in Hawaii it was $17,020.
\8 Demonstration projects have made significant contributions to the
development of Medicaid policy, including early experiments with
prospective payment for inpatient care and school-based services for
young children.
MEDICAID PROGRAMS RESTRUCTURED
UNDER REDEFINED BUDGET
NEUTRALITY POLICY
---------------------------------------------------------- Chapter 1:1
Eleven states with 1115 waivers approved since 1993 have undertaken
an ambitious experiment to demonstrate that the Medicaid program can
actually save money while simultaneously expanding coverage.\9
The administration has entered into 5-year budget commitments that
allow each state to reinvest managed care savings and redirect other
funds in order to expand coverage to currently uninsured individuals.
Compared to expenditure trends for the predemonstration program,
states suggest that the net result of waivers will be lower
costs--even though managed care savings are being reinvested. Eleven
more waivers are pending, and all but a few applicants are pursuing a
similar managed care cost-containment/coverage expansion strategy.
Only Illinois, Oklahoma, and Kentucky (the state's 1995 demonstration
application) propose using 1115 waiver authority to reduce both state
and federal Medicaid expenditures without expanding coverage.\10
The use of 1115 waivers to restructure state Medicaid programs has
been facilitated by a new federal flexibility in assessing the budget
neutrality of such demonstrations--particularly, the administration's
avowed openness to "new methodologies" to estimate what the
continuation of a state's existing Medicaid program would have cost.
By redefining the terms of budget neutrality, the administration has
made it easier for states to demonstrate that waivers will cost less
than their existing Medicaid programs. The administration asserts,
nonetheless, that all approved 1115 waivers are budget neutral and
that the demonstration authority is not being used to expand
entitlement spending.
--------------------
\9 In 1982, Arizona was granted an 1115 waiver to initiate a
statewide managed care program. Previously, the state had not
participated in Medicaid. For a description of Arizona's Medicaid
program, see Arizona Medicaid: Competition Among Managed Care Plans
Lowers Program Costs (GAO/HEHS-96-2, Oct. 4, 1995).
\10 Kentucky's 1993 demonstration application was one of the first to
be approved. Because of the state legislature's concern about the
adequacy of funding to extend eligibility to certain uninsured
residents, the state submitted a new 1115 waiver application in 1995
that omits any coverage expansion.
OREGON FIRST TO REQUEST 1115
MEDICAID DEMONSTRATION
WAIVER
-------------------------------------------------------- Chapter 1:1.1
Oregon's 1991 application--a major component of a broader reform of
its health care system--was the first request since 1982 to operate a
comprehensive, statewide demonstration program. After a year of
review, the Bush administration rejected the application on the
grounds that the state's proposed benefits package might violate the
Americans With Disabilities Act. A report prepared for the National
Governors' Association criticized the handling of Oregon's
application, noting that the review process did not give due weight
to the state role in Medicaid, had a "chilling" effect on innovative
proposals, and imposed budget neutrality requirements that were too
narrow.\11 The Clinton administration subsequently announced plans to
streamline and expedite the review process for 1115 waivers.
Oregon revised and resubmitted its waiver application in late 1992
and the following March became the first state in over 10 years to
obtain approval for a comprehensive 1115 demonstration. Since then,
10 additional waivers have been approved. Another 11 states have
applications pending, and others have expressed interest in
submitting requests for 1115 waivers. States with approved and
pending waivers account for about 49 percent of the nation's Medicaid
beneficiaries and 52 percent of total Medicaid expenditures.\12 As of
July 1995, however, only five states--Oregon, Hawaii, Tennessee,
Rhode Island, and Minnesota--had begun implementation. Appendix III
summarizes the status of all approved and pending comprehensive 1115
demonstrations submitted since 1992.
--------------------
\11 John Luehrs, Flexibility and Waiver Authority for Health Care
Reform: A Primer for States, Health Policy Studies, Center for
Policy Research, National Governors' Association (Washington, D.C.:
1992).
\12 Throughout this report, Medicaid beneficiary/recipient statistics
include all enrolled individuals in the 50 states and the District of
Columbia regardless of whether they received services during the
year.
1115 WAIVERS REMOVE
ROADBLOCKS TO MANAGED CARE
-------------------------------------------------------- Chapter 1:1.2
Requirements intended to preserve quality by protecting a Medicaid
beneficiary's freedom to choose a provider have limited states'
ability to mandate enrollment in HMO-style managed care.\13
HMO-style health plans are prepaid a fixed amount based on the number
of enrollees rather than reimbursed after each service is rendered.
Such prepayment on a per capita basis is often referred to as
"capitation." A convergence of trends--spiraling Medicaid
expenditures coupled with a stronger interest on the part of
mainstream HMOs in serving the Medicaid population--has given impetus
to as well as facilitated the adoption of a managed care
cost-containment strategy by states. And rather than turning to
physician gatekeeper arrangements that dominated past managed care
experiments, many states are using the flexibility gained under 1115
Medicaid waivers to adopt capitated alternatives.
Reacting to quality-of-care, marketing, and other problems that
surfaced in a number of Medicaid managed care programs, the Congress
enacted provisions in 1976 with the general goal of encouraging HMOs
to provide public clients a quality of care comparable to that
available to private clients.\14 At the time, HMOs were the prevalent
form of managed care. One provision discouraged the creation of HMOs
serving only Medicaid beneficiaries by requiring that at least a
certain percentage of the patients be privately insured. The
participation of private-paying patients, who presumably have a
choice of health plans, was instituted as a proxy for quality. A
second provision added in 1981 allows recipients to terminate
enrollment in an HMO at any time. From the beneficiary's
perspective, these provisions offer protection against enrollment in
an HMO seeking excessive profit at the expense of quality. From the
HMO's perspective, however, unrestricted freedom to disenroll makes
it difficult to plan financially and therefore renders the enrollment
of Medicaid recipients less attractive. This provision added to the
hesitancy of many mainstream HMOs to participate in Medicaid, further
restricting states' ability to experiment with fully capitated health
plans.\15
In 1993, about 4.8 million beneficiaries, or about 12 percent of the
Medicaid population, were enrolled in some type of managed care.\16
Though a number of voluntary and mandatory managed care options are
available to states, the enrollment of Medicaid beneficiaries has
lagged behind national trends. Without an 1115 waiver, states
essentially have three managed care options:
voluntary enrollment in an HMO, in which case the beneficiary must
also have a choice of obtaining services on a fee-for-service
basis and be allowed to disenroll at will;
mandatory enrollment in an HMO, provided the beneficiary can choose
from among a number of competing HMOs, with disenrollment
allowed on a monthly basis (or every 6 months if an HMO meets
certain federal requirements); and
voluntary or mandatory enrollment in a physician gatekeeper system
in which either the physician's charges are partially capitated
or the physician is reimbursed under fee-for-service.
The last two options require waivers of Medicaid provisions that have
been widely provided under section 1915(b) of the Social Security
Act. As of July 1994, 37 states operated 1915(b) waiver programs.
These programs were primarily substate (that is, in a limited
geographic area), voluntary, and involved physician gatekeepers
rather than HMOs. Until 1993, only two states--Arizona and
Minnesota--operated mandatory HMO managed care programs under the
authority of an 1115 waiver.\17 Table 1.1 delineates the additional
flexibility available under an 1115 waiver compared with 1915(b)
authority.
Table 1.1
Comparison of Managed Care Flexibility
Available Under 1915(b) Versus 1115
Waivers
Section 1915(b) program Section 1115 demonstration
waivers\a waivers
---------------------------- ----------------------------
HMOs must still meet federal HMOs may enroll Medicaid
requirement for more than 25 patients exclusively
percent private enrollment
Full range of mandatory Benefit package may be
services must be offered modified\b
Enrollment "lock-in" limited Enrollment "lock-in" may be
to 1 month\c extended to
12 months
No restrictions on access to Access to family planning
family planning providers providers may be restricted
----------------------------------------------------------
\a When originally enacted in 1981, section 1915(b) expressly allowed
waiver of section 1903(m) HMO provisions in the Social Security Act.
The following year, the law was amended to eliminate the authority to
waive these provisions.
\b To date, only Oregon has been permitted to modify the benefit
package for traditional Medicaid eligibles. Other states have been
permitted to offer a modified package only to those newly eligible
under the demonstration.
\c The lock-in is 6 months for an HMO meeting certain federal
qualifications.
--------------------
\13 Though no commonly accepted definition exists for the term
"managed care," a number of features are typically associated with
it: (1) provider networks with explicit criteria for selection, (2)
alternative payment methods and rates that often shift some financial
risk to providers, and (3) utilization controls over hospital and
specialist physician services. Despite the confusing nomenclature
used to distinguish a variety of managed care plans--HMO, PPO,
PCCM--most include one or more of these common cost-control features.
Health maintenance organizations (HMO), the most tightly controlled
type of managed care plans, require patients to use affiliated
physicians who may be salaried, paid on a per capita basis (often
referred to as "capitated"), or reimbursed for each service.
Typically, a patient's care, especially referrals to specialists and
hospitalization, is coordinated by a primary care physician--often
called a "gatekeeper." Preferred provider organizations (PPO) provide
enrollees with a financial incentive--lower cost-sharing
(co-payments)--to receive care from a network of providers that are
normally reimbursed at a discounted rate. Finally, many state
Medicaid programs have conducted experiments using a primary care
case management (PCCM) approach in which physician gatekeepers must
authorize a patient to see a specialist or obtain hospital care.
Gatekeeper physicians may be partially capitated or paid for each
service delivered. For a description of the evolution and use of the
term managed care, see Managed Health Care: Effect on Employers'
Costs Difficult to Measure (GAO/HRD-94-3, Oct. 19, 1993).
\14 These provisions were incorporated in subsection (m) of the
Social Security Act, which was added to section 1903 (42 U.S.C.
1396b(m)).
\15 Other factors that dissuaded HMOs from participating in the
Medicaid program included (1) inability to offer incentives, such as
lower out-of-pocket costs or better benefits, to encourage
beneficiaries to enroll; (2) lack of experience in marketing to
individuals as opposed to large employers; and (3) a widespread
perception that Medicaid beneficiaries were costly and had difficulty
adjusting to HMO restrictions.
\16 This percentage is calculated using the number of individuals
enrolled in Medicaid. Occasionally, a slightly higher percentage is
reported that is based only on beneficiaries who actually receive
services during the year.
\17 As noted earlier, Arizona's program operated statewide. In
contrast, Minnesota's 1115 waiver primarily covered the urban
counties surrounding Minneapolis.
1115 COVERAGE EXPANSION
GOALS AND STATE MEDICAID
BASELINES VARY
-------------------------------------------------------- Chapter 1:1.3
Though all approved 1115 waivers expand eligibility, the nature and
extent of coverage expansion varies. Waivers use income standards to
define and, in effect, limit who is eligible, but states further
restrict the number of individuals that can actually obtain coverage
through explicit enrollment caps, other barriers, and even premiums.
Because of greater than anticipated enrollment, several states are
taking additional steps to limit the number applying for coverage.
Most states are adding groups who were previously ineligible for
Medicaid--single adults and childless couples. However, it is
important to recognize that some categories of newly eligible
individuals in one state may already be enrolled in Medicaid in
another because of differences in the qualifying income levels
established by states for families and optional coverage expansions
allowed by statute.\18 Moreover, states may have previously provided
health coverage to some new eligibles outside of Medicaid--that is,
through state-funded programs. For example, Kentucky's planned
expansion under its now suspended 1993 waiver includes both
previously ineligible individuals and individuals who could have been
included in Medicaid at state option. In Rhode Island, Hawaii, and
Minnesota the expansions either are limited to or include individuals
who could have been Medicaid- eligible at state option. Both Hawaii
and Minnesota offered state-funded coverage to many included in their
waivers.\19
Table 1.2 summarizes the coverage expansion goals of seven states
with approved 1115 waivers. Generally, expansion goals are stated in
terms of the number of newly eligible individuals expected to be
covered under the waiver at the end of the 5-year period of each
demonstration. Enrollment in year 1 may be less than the 5-year
enrollment goal, since states typically anticipate reaching "full
enrollment" gradually. State goals should be viewed as targets
subject to a number of constraints, especially financing. Some
waivers indicate that the state will limit enrollment if funding
proves to be inadequate. At least one state, Tennessee, has already
done so.
Table 1.2
Coverage Expansion Goals and Cost-
Sharing Arrangements for Seven Approved
1115 Demonstration Waivers
Tennesse Rhode Minnesot
e Oregon Hawaii Island a Kentucky Florida
------------------- -------- -------- -------- -------- -------- -------- --------
Poverty level No 100% of 300% of 250% of 275% of 100% of 250% of
criterion (no upper poverty poverty poverty poverty poverty poverty
individual/family limit level level level level level level
with income above (childre
this limit n only)
qualifies for
coverage)\a
Estimated 500,000 96,400\b 77,000 7,000 135,000 201,000 1,100,00
enrollment, year 5 0
Enrollment, year 1 400,425 69,666 63,500 716 47,000 Suspende 145,558
(actual, (average (actual, (actual, (actual, d by (estimat
7/7/95) , 12/ 7/95) 5/31/ 7/95) state ed)
31/94) 95)
Percentage of 100% of \d 100% of 185% of 70% of No Enrollee
poverty level at poverty poverty poverty poverty premiums with any
which premium level\c level\e level\f level income
payments begin must pay
a
premium
Income level at 400% of \d 201% of Over Over No Over
which enrollee pays poverty poverty 250% of 275% of premiums 250% of
full premium level\c level\e poverty poverty poverty
level level level
Average monthly $27.35 \d $140\e No $36 No Approxim
premium for a premium premiums ately
family of 4 at 150% at this $21\g
of poverty level poverty
level\f
-----------------------------------------------------------------------------------------
\a For example, in Oregon a family of four whose income exceeds
$14,800 would not be eligible for coverage under the waiver, while in
Minnesota the income standard is $40,700.
\b Because of the employer mandate, the state expected the number of
new eligibles to decline to this lower level by the fifth year of the
waiver as some low-income individuals shift to private insurance.
\c Participants are also expected to contribute co-payments for most
services and annual deductibles up to a maximum.
\d In June 1995, Oregon requested HCFA approval to begin charging
premiums as of October. As of late October, HCFA officials had not
responded to this request. Under the proposed changes, all new
eligibles are expected to pay a premium. The maximum premium for a
family of four or more at 100 percent of the poverty level would be
$28.
\e Hawaii implemented these cost-sharing arrangements on August 1,
1995. During the preceding 12 months, out-of-pocket costs for new
eligibles were lower: (1) only individuals at and above 133 percent
of the poverty level were required to pay premiums, (2) individuals
at or above 296 percent of the poverty level were required to pay the
full premium, and (3) the monthly premium payment for a family of
four was only $51.40.
\f Rhode Island offers two cost-sharing options--a premium and a
point-of-service co-payment. To date, most enrollees have selected
the point-of-service co-payment option. According to a state
official, the two options are designed to be actuarially equivalent,
that is, the net cost to enrollees is the same. Premium shares for
demonstration eligibles are stratified by age and represent 3 percent
of the total premium amount. A family of four consisting of a
mother, one infant, and two children between 185 and 250 percent of
the poverty level would pay a monthly premium of about $14.50.
\g The total premium is actually $42, but Florida assumes that
enrollees and their employers will split the cost. An enrollee would
be responsible for any portion of the total premium not paid by the
employer.
At one end of the expansion continuum, Tennessee's waiver program has
no eligibility income limit and requires only those enrollees above
400 percent of the poverty level to pay the full premium. Tennessee
set an original enrollment cap of 500,000 on new eligibles and
excluded individuals who were insured before a certain cutoff date to
forestall a migration from private insurance.\20 The state lowered
its cap in December 1994, in effect limiting the expansion to the
approximately 400,000 previously uninsured individuals already
enrolled.\21 In contrast, Oregon's expansion goals are less ambitious
because, in part, the employer mandate was expected to cover
low-income workers. The waiver expands eligibility up to 100 percent
of the poverty level, requires no premiums from participants, and has
no explicit enrollment cap. In order to address budgetary
constraints, the legislature enacted several changes that the state
began implementing on October 1, 1995. However, HCFA has yet to
approve Oregon's request to require premiums of some newly eligible
participants.
Some 1115 waivers, such as Hawaii's, target individuals who were
already covered by state-funded expansion programs outside of
Medicaid.\22 Consequently, Hawaii did not anticipate much additional
enrollment as a result of its waiver. Unlike Hawaii, Minnesota's
waiver application did not include single adults and childless
couples covered under state-funded MinnesotaCare. State officials
told us that requesting a federal match for this group would have
created a budget neutrality problem because of the cost. Instead,
the approved waiver shifts children covered under MinnesotaCare into
Medicaid--a group that could have been included in the program at
state option. Finally, Rhode Island's expansion group is limited to
pregnant women and children who could also have been covered at state
option without an 1115 waiver.
--------------------
\18 One such option is section 1902(r)(2) of the Social Security Act,
42 U.S.C. 1396(b)(r)(2), which enables states to liberalize
financial criteria for certain individuals not receiving cash
assistance. Women and children and the aged, blind, and disabled may
qualify for Medicaid coverage in this way. No precise upper limit on
income exists under this option. Thus, at least one state has
established financial criteria sufficiently liberal to cover
individuals with incomes up to 300 percent of the federal poverty
level.
\19 These programs are the General Assistance and State Health
Insurance Programs in Hawaii and MinnesotaCare in Minnesota.
\20 Tennessee established an overall cap on program
participation--1.3 million the first year (1 million traditional
Medicaid recipients and 300,000 previously uninsured individuals) and
1.5 million in subsequent years (1 million traditional recipients and
500,000 previously uninsured individuals). In effect, however, the
cap only applied to the enrollment of the previously uninsured, since
traditional Medicaid recipients and those whose medical conditions
make them uninsurable would continue to be enrolled if the overall
cap is reached.
\21 On December 31, 1994, Tennessee closed enrollment to previously
uninsured individuals who do not qualify for Medicaid or are not
considered uninsurable. Applications pending at that time were
processed.
\22 Some individuals covered by Hawaii's state-funded expansion had
been ineligible for Medicaid, while others could have been included
at state option.
TREATMENT OF LOW-INCOME
POPULATIONS UNDER WAIVERS IS
CONTROVERSIAL
-------------------------------------------------------- Chapter 1:1.4
Embedded in many comprehensive 1115 demonstrations is a controversial
philosophical shift in the way publicly supported health care is
provided to the poor--a shift (1) inherent in the adoption of managed
care delivery systems and (2) visible in the treatment of those newly
eligible under the demonstration. The cumulative effect of these
changes is to place more responsibility on the individual
beneficiary. Advocacy groups for the poor\23 and waiver states
disagree on the extent to which low-income beneficiaries can shoulder
these responsibilities.
Advocacy groups typically see Medicaid beneficiaries, as well as many
of the working poor to whom coverage is being extended, as a
vulnerable, high-risk, and sicker segment of the population. They
are concerned about the ability of the poor to access services in a
managed care system. As noted earlier, some critics of managed care
argue that it creates a "perverse incentive" to deliver fewer or less
costly services than may be needed. Advocacy groups also worry that
managed care will (1) further reduce the historically low Medicaid
reimbursement levels, (2) diminish quality of care, and (3) curtail
access to providers that have traditionally served the poor.
Finally, they see a disincentive for the poor to participate in
waiver programs when premiums, though subsidized, are high relative
to income. States, on the other hand, point out that the waivers
require them to implement significant quality assurance programs such
as collecting and analyzing encounter data and conducting annual
satisfaction surveys. They contend that access and quality were
never optimal under a fee-for-service delivery system in which choice
was guaranteed but not necessarily available and high emergency room
use was an underlying symptom of access problems. Finally, states
generally see employment as a proxy for health and evidence for
distinguishing between the newly eligible working poor and the more
vulnerable beneficiaries typically enrolled in Medicaid.
The "mainstreaming" of new eligibles--that is, the attempt to treat
them as if they were purchasers of private insurance--is perhaps the
hallmark of most state 1115 demonstrations.\24 For the newly insured,
states have attempted to break Medicaid's psychological link with
welfare by establishing eligibility criteria and enrollment
mechanisms distinct from those that apply to traditional
beneficiaries, many of whom qualify for Medicaid by virtue of
receiving cash assistance. For example, most waivers eliminate the
asset test, often criticized as intrusive and expensive to
administer, and rely instead on a gross income test.\25 Moreover,
rather than the frequent redeterminations of eligibility associated
with Medicaid, many newly eligible individuals are enrolled for
periods ranging from 6 to 12 months. In some states, enrollment of
new eligibles is not handled by the agency that administers Medicaid.
In fact, Florida plans to hand this task over to insurance
agents--further underscoring the similarity to a private insurance
product. In general, application forms are simpler and in some
states can even be mailed. Newly eligible individuals usually
receive the full acute care benefit package available to traditional
Medicaid recipients. Only in the case of Florida is the benefit
package more restrictive. Finally, many states require individuals
with incomes above the poverty level to contribute toward the cost of
health care coverage by charging premiums, co-payments, and
deductibles.
For the traditional Medicaid population, the major change associated
with 1115 waivers is the wholesale movement from fee-for-service to
some type of managed care--even in rural areas and, often, even for
those who are aged, blind, or disabled.\26
For some, such as low-income families in Oregon and Minnesota who
were already enrolled in mandatory managed care programs, the change
may be imperceptible. Under most 1115 waivers, eligibility
requirements are unchanged and benefits remain the same or are more
generous. For example, Tennessee lifted service restrictions on its
Medicaid benefit package. Only Oregon altered benefits to help
finance coverage expansion. The redefined package, commonly known as
the "prioritized list," eliminates some costly health services while
adding a broad array of preventive care.\27
--------------------
\23 In June 1994, the National Association of Community Health
Centers filed a lawsuit to stop the implementation of comprehensive
1115 demonstrations. The Association noted that it was acting to
ensure that the rights of vulnerable populations are protected--not
to thwart state efforts to improve health care delivery. In a press
release announcing the lawsuit, the Association argued that
beneficiaries were being "forced to enroll in managed care plans that
are under no compulsion to meet federal safeguards designed to ensure
access, services, and quality control."
\24 Since 1986, states have had greater leeway in adopting a
streamlined eligibility process for Medicaid recipients who do not
receive cash assistance. Most states have adopted one or more of the
available alternatives.
\25 Certain of these changes result in the loss of coverage for a
small number of individuals who would qualify for Medicaid were the
demonstration not in place. For a more extensive discussion of the
impact of eliminating the asset test, see Sara Rosenbaum and Julie
Darnell, Medicaid Section 1115 Demonstration Waivers: Approved and
Proposed Activities as of November 1994, Center for Health Policy
Research, The George Washington University (Washington, D.C.: Nov.
1994), p. 3.
\26 Tennessee and Florida include the aged, blind, and disabled in
their transition from fee-for-service to managed care. HCFA approved
Oregon's request to add this population to the state's waiver program
beginning in January 1995. Hawaii anticipates seeking HCFA approval
to include these populations at a later date.
\27 The process for establishing the revised benefit package was
controversial because the state developed a method to rank, or
prioritize, services in descending order from most to least useful.
Independent actuaries set cost estimates for individual services on
the list. Each legislative session, state legislators "draw the
line" according to the amount of funds appropriated for the program,
thereby defining the list of available services. State officials
believe the use of the prioritized list will ensure continuous
coverage for demonstration enrollees despite funding fluctuations,
which states historically remedy by eliminating coverage for optional
groups and services. Examples of services not included are
aggressive cancer treatment that will not result in a 5-percent
probability of a 5-year survival (hospice care, comfort care, and
treatment of symptoms are still covered); medical treatment for a
sore throat or diaper rash; and cosmetic services.
WAIVER DESIGNS OFTEN REFLECT
STATE CONTEXT AND HEALTH REFORM
GOALS
---------------------------------------------------------- Chapter 1:2
A notable feature of the growing number of 1115 waivers is their
divergent techniques for using Medicaid as a springboard to achieve
some degree of reform in state health care systems. Numerous states
that have submitted waiver applications are recognized leaders
representing diverse approaches to health care reform. Though many
waivers were conceived when national reform appeared imminent, the
recent retrenchment from broader reform goals suggests that Medicaid
waivers have become a more important component of state health care
reform. Appendix IV compares several states across a range of
indices that are relevant to understanding the diversity evident in
waiver designs.
The 1115 waiver is health care reform in Tennessee. The state's
emphasis on managed care promises to increase penetration by that
delivery system in a region long resistant to such a change while at
the same time significantly reducing the number of uninsured. The
success of this policy hinges on the adoption of a stringent
cost-containment strategy with regard to health care financed through
the waiver.
In contrast to Tennessee's nascent reform program, Florida's as yet
unimplemented waiver represents a logical progression from earlier
small market reforms intended to provide access to affordable
insurance for the working poor. The state hopes to use its waiver to
achieve a dramatic enrollment expansion in the state's voluntary,
small business-oriented purchasing cooperatives and a significant
reduction in its estimated 24-percent uninsured rate. Florida stands
alone in the extent to which it distinguishes between traditional and
newly eligible Medicaid recipients. While the former are required to
choose between different forms of managed care, newly eligible
recipients may select any health plan offered by state-supported
purchasing cooperatives with the sole proviso that the enrollee is
responsible for any difference between the subsidy and the plan
premium. Florida's 1992 health reform legislation established a goal
of universal coverage, and the legislature promised to revisit the
choice of a voluntary over a mandatory approach unless there was a
significant reduction in the number of uninsured.
In Minnesota, Oregon, and Hawaii, the waiver is only one element of a
much more ambitious reform agenda--an agenda that some state
officials believe has been brought into question since the 1994
health care debate. All three states have universal coverage as a
goal. Hawaii already has a limited exemption from the requirements
of the Employee Retirement Income Security Act of 1974 (ERISA)\28
that allows it to require employers to offer health insurance to
their workers. Oregon's Medicaid waiver is also built around an
ERISA exemption to permit enactment of an employer mandate.
Obtaining that exemption is now considered unlikely. In contrast,
Minnesota had been attempting to finance universal coverage with an
individual mandate, though state officials told us that this approach
is no longer considered a possibility. Both Hawaii and Minnesota
have relatively small uninsured populations and, prior to the 1115
Medicaid waiver, had already taken steps to address this problem
through the establishment of state-funded coverage expansions. All
three states have significant managed care penetration. Moreover,
Oregon and Minnesota have a decade of experience with Medicaid
managed care.
--------------------
\28 Public Law 93-406, 88 Stat. 829 (classified as amended at 29
U.S.C. 1001 et seq.). ERISA prevents states from mandating that
employers provide health benefits. Concurrent with the original
passage of ERISA, Hawaii enacted comprehensive health reform
requiring all employers to provide full-time employees a standard
health package and pay for a substantial portion of the premium.
Some dependents, part-time workers, and certain other individuals
were not covered by the mandate. After several large employers
successfully argued that ERISA precluded such a mandate, the Congress
grandfathered Hawaii's statute by granting an exception to ERISA
requirements.
FINANCING OF WAIVERS IS MOST
CONTENTIOUS ISSUE
---------------------------------------------------------- Chapter 1:3
An important element of the administration's commitment to streamline
and expedite the review of 1115 waivers was the promise to maintain
the principle of waiver budget neutrality "more flexibly than has
been the case in the past." Despite this commitment, a number of
factors have contributed to a lengthening of the process, including
extensive negotiations over financing.\29 In fact, both
administration and state officials told us that budget neutrality is
often the most contentious issue.
Though a policy of budget neutrality has been in effect since the
early 1980s, previous 1115 Medicaid demonstrations were usually
small-scale experiments targeted at specific populations (for
example, pregnant drug-users on Medicaid) or implemented in a limited
geographic area.\30 As a result, the potential impact on state and
federal expenditures was more circumscribed and the task of devising
a cap to ensure cost neutrality was less challenging.\31 Oregon's
1991 waiver application was the first of a new breed of
"comprehensive" 1115 demonstrations--complex proposals that were
sometimes controversial and often tied to broader health reform
agendas. HCFA officials told us that the number, scope, and
complexity of such comprehensive demonstrations in effect elevated
the importance of budget neutrality while making it more difficult to
evaluate and enforce.
In a memo preceding promulgation of a more flexible approach to
budget neutrality, HHS officials recognized the incentive for states
to shift costs to the federal government and the need to constrain
such behavior. Nonetheless, they outlined several arguments for a
less strict approach. For example, they pointed out that the federal
government might want to (1) share in the risks and costs of testing
innovations that were ultimately in its own interest; (2) set the
stage for health reform by supporting changes that should not wait
even if they are somewhat more costly; and, finally, (3) provide some
fiscal relief to states overburdened by the rising number of
uninsured, increasing charity care requirements, and federally
mandated expansions. Most importantly, the memo recognized that
whatever policy was adopted needed to be clear and consistently
applied.
There are two key aspects to the administration's revised budget
neutrality policy. First, states are allowed to demonstrate budget
neutrality over the life of the waiver rather than on a yearly basis,
allowing more time to recoup any associated start-up costs. Second,
recognizing the difficulty in estimating the costs of continuing the
prewaiver program over the period of the demonstration and the
inherent element of judgment in undertaking such an estimate, the
administration announced that it was open to state suggestions on the
development of a new baseline methodology. According to HHS and
Office of Management and Budget (OMB) officials who share
responsibility for implementing this revised policy, there are three
critical steps in determining baseline costs: (1) selecting a base
year, (2) developing a trend factor for growth from the base year to
the first year of implementation, and (3) developing a trend factor
for baseline costs over the period of the waiver. The method used to
develop baseline costs is important because it is the benchmark
against which the administration assesses waiver costs. The higher
the baseline, the easier it is for a state to demonstrate cost
neutrality.
--------------------
\29 Five waivers were submitted between November 1992 and mid-1993
and each was approved before the end of 1993. The shortest approval
took 3 months and the longest 7 months. In 1994, however, only one
of nine waivers pending was approved for implementation. Additional
factors that have lengthened the review process include (1) the
complexity of and variation found in state applications, (2) concerns
raised over the rapid approval and implementation of Tennessee's
waiver, and (3) the growing backlog of waivers submitted.
\30 Past insistence on budget neutrality had effectively limited the
scope of 1115 waivers to small, substate demonstrations. The only
prior use comparable to recent statewide waiver applications was the
1982 initiation of a managed care program in Arizona, a state that
previously had not participated in Medicaid.
\31 Minnesota obtained an 1115 waiver in 1982 to move low-income
families and elderly recipients in the Minneapolis area and one rural
county into HMOs. Budget neutrality was addressed by requiring that
the capitation rate be structured at 90 to 95 percent of
fee-for-service costs for these beneficiaries.
MONITORING AND ENFORCING
BUDGET NEUTRALITY
-------------------------------------------------------- Chapter 1:3.1
The final waiver agreement consists of a set of terms and conditions
that, among other things, spells out how budget neutrality will be
monitored and enforced. Although there are important state-specific
variations, the administration has taken two basic approaches to
enforcement. Tennessee and Florida have an aggregate cap on the
amount of federal matching funds available for their demonstrations,
while all other states have a per capita limit. The federal
government will not match state expenditures above the specified
caps. Table 1.3 highlights the important differences in what the
expenditure caps cover and how they work in the four states whose
budget neutrality agreements we assessed.
Table 1.3
Methodology for Enforcing Budget
Neutrality Expenditure Limits
State Methodology
------------------ --------------------------------------
Aggregate cap
----------------------------------------------------------
Tennessee Expenditure limit is on total Medicaid
program, including long-term care,
which is not part of the
demonstration; expenditures for
traditional and new eligibles are
matched up to the cap.
Florida Expenditure limit only applies to
acute care costs and DSH; expenditures
for traditional and new eligibles are
matched up to the cap.
Per capita limit
----------------------------------------------------------
Hawaii Expenditure limit applies to acute
care costs of traditional Medicaid
population; expenditures on new
eligibles are not matched, but program
savings can be applied to cover their
costs.
Oregon Expenditure limit applies to acute
care costs of traditional Medicaid
population and new eligibles, but a
formula limits the number of those
newly eligible for which federal
matching funds are available.
----------------------------------------------------------
Aggregate caps are the most straightforward and uniform of the two
approaches. In Tennessee and Florida, the federal government agreed
to an explicit expenditure limit on demonstration costs. With the
exception of certain recipient growth in Florida, the federal
government will not match any costs above this cap. For states that
use the per capita approach, the federal government agreed to a
cost-per-recipient limit. For Hawaii, this cost limit is based on
per capita fee-for-service costs from 1993 trended forward to the
first year of the demonstration. If, as the state anticipates, the
switch to managed care produces savings over fee-for-service rates,
they can be applied to the costs of those newly eligible. In Oregon,
the agreement specifies per capita cost limits for both traditional
and newly eligible enrollees. However, the number of new eligibles
is limited to an agreed-upon percentage of traditional Medicaid
enrollment. With the exception of Tennessee, no state is held at
risk for growth in the Medicaid population caused by an economic
downturn. Florida has an escape valve from its aggregate cap if
growth in the Medicaid population exceeds projections by 3 percent or
more.
OBJECTIVES, SCOPE, AND
METHODOLOGY
---------------------------------------------------------- Chapter 1:4
We reviewed the financing arrangements for approved 1115 Medicaid
demonstration waivers in several states, with a focus on (1) the
relationship between the waiver and other state health reform
initiatives, (2) the planned sources of funding available to finance
expanded coverage, (3) the potential net impact of these waivers on
federal Medicaid expenditures, and (4) the actual waiver expenditures
of states with the most implementation experience. Although our
study focused on 1115 waivers in Tennessee, Florida, Oregon, and
Hawaii, we closely monitored other pending waivers, which we use as
examples throughout this report. For a detailed description of our
methodology, see appendix I. Our review was conducted from August
1994 through August 1995 in accordance with generally accepted
government auditing standards.
THEORY BEHIND EXPANDING MEDICAID
COVERAGE WITHOUT INCREASING
PROGRAM EXPENDITURES
============================================================ Chapter 2
Although the administration has adopted a more flexible approach
toward budget neutrality, it contends that all approved comprehensive
1115 demonstrations are in fact budget neutral. Before addressing
this issue in the next chapter, this chapter describes and
categorizes waiver funding strategies--strategies that states say
will result in coverage expansion without increasing expenditures
beyond what their smaller, prewaiver Medicaid programs would have
cost. In fact, compared to the cost of continuing the existing
Medicaid program, many states project that the demonstrations could
actually save money. State officials estimate that the four 1115
demonstrations whose financing we examined in detail could add up to
2 million previously uninsured individuals while yielding savings of
about $6 billion over 5 years.
Table 2.1
State Projections of 5-Year Savings
Under 1115 Waivers
(Dollars in billions)
Tennes Florid
see a Oregon Hawaii
-------------------------------------- ------ ------ ------ ------
State share $1.6 $.845 $.017 $.429
Federal share 3.2 0.0 .029 .005
======================================================================
Total $4.8 $.845 $.046 $.434
----------------------------------------------------------------------
PLANNED SOURCES OF FUNDING FOR
1115 MEDICAID WAIVERS
---------------------------------------------------------- Chapter 2:1
States rely on a similar mix of funding sources that, analytically,
can be grouped into two major categories: (1) Medicaid resources
redirected from existing programs, such as the Disproportionate Share
Hospital Program (DSH),\32 and (2) expected savings, primarily from
different forms of managed care delivery. In addition, many states
collect user fees--the premiums charged to new program enrollees.
Finally, in a few cases, state funds used to subsidize insurance
programs for low-income residents are being folded into the waiver.
Based on a review of waiver applications and discussions with HCFA
and state officials, table 2.2 (1) summarizes the planned funding
sources for expanded coverage in four states over the 5-year terms of
the demonstrations and (2) highlights the relative importance of the
various categories of funding.
Table 2.2
Planned Sources of Funding for Expanded
Coverage Over the 5-Year Terms of 1115
Demonstrations
Tennessee Florida Oregon Hawaii
-------------- -------------- -------------- -------------- ----------------
Redirected Major $$$ Important $$ No role Major $$$
DSH DSH DSH
Public health Medically
programs Needy Program
Program Important $$ Major $$$ Major $$$ Important $$
savings
Managed care Managed care Managed care Managed care
Reimbursement Priority list
reform
Employer
mandate
Other Less important Less important Less important Less important $
$ $ $
Premiums Premiums State-funded Premiums
programs
State-funded
programs
--------------------------------------------------------------------------------
Note: This categorization is based on expected funding sources
before states started implementing their waivers. Chapter 4
addresses the extent to which state expectations have been realized.
Comparing the major funding sources across states can be tricky. For
example,
Hawaii's waiver application never quantified expected managed care
savings,
Tennessee's waiver application identified premiums but not DSH as a
financing source, and
Florida quantified its funding sources but omitted premiums because
they will be used to offset state costs.
Despite these obstacles, a few generalizations can be made about the
magnitude and relative importance of core, coverage expansion
funding. The conventional wisdom that 1115 expansions are financed
largely by managed care savings is misleading. In at least two
states with approved waivers, funds redirected from DSH and other
programs play a more significant role. On the other hand, states
with relatively small DSH allotments rely to a greater extent on
managed care or other forms of savings. Finally, premiums are a less
important and more uncertain source of funding.
--------------------
\32 DSH compensates hospitals that serve the uninsured and a large
Medicaid clientele.
FUNDS FROM DSH AND OTHER
PROGRAMS PLAY MAJOR ROLE
---------------------------------------------------------- Chapter 2:2
Based on a review of waiver applications and discussions with HCFA
and state officials, DSH and other redirected funds in Tennessee's
and Hawaii's demonstrations appear to be a more important coverage
expansion financing source than either expected managed care savings
or any capitation discount obtained from managed care
organizations.\33 As shown in table 2.3, both states had relatively
large DSH programs at the time their waivers were approved. The
theory behind eliminating or greatly reducing DSH payments to
hospitals is that fewer uninsured will translate into less
uncompensated care.\34 Compared to other financing sources, DSH
appears to be the most tangible and assured source of financing.
Table 2.3
DSH Allotment as a Percentage of
Medicaid Expenditures in Fiscal Year
1993 for Four Waiver States
(Dollars in millions)
DSH
allotment Percentage
------------------------------ ------------ ------------
Tennessee $430 16.1
Hawaii 44 11.5
Florida 240 4.8
Oregon 21 2.2
----------------------------------------------------------
Source: The Urban Institute.
Tennessee's waiver also proposes to redirect funds from two
additional sources--public health programs and DSH-like payments,
referred to as local government charity care. As with DSH, state
officials believe that routine access to health care by those
currently uninsured should decrease the funding needed for programs
such as those for communicable disease control and maternal and child
health.
Florida's 1115 waiver caps enrollment at about 1.1 million previously
uninsured individuals--less than half of the state's uninsured
population. Since hospitals would continue to face significant
levels of uncompensated care, the state was reluctant to redirect all
of its DSH funds. Consequently, Florida's finance plan only shifts
growth in its DSH resources toward coverage expansion. Because of
the relative modesty of its DSH payments, Florida was forced to
search for an alternative funding source. The state decided to
eliminate Medically Needy Program coverage and to reallocate those
funds to help subsidize the purchase of private health insurance. A
state is not required to offer a Medically Needy Program under
Medicaid. Because the Medically Needy Program pays for health
services only after individuals have already incurred large
liabilities, Florida considers the program to be similar to DSH.
That is, it reimburses hospitals for bills that otherwise might go
unpaid. Together with DSH, redirected funds are only about one-third
of the financing identified in Florida's waiver.
--------------------
\33 Our recent report on Tennessee's 1115 waiver program, Medicaid:
Tennessee's Program Broadens Coverage, But Faces Uncertain Future
(GAO/HEHS-95-186, Sept. 1, 1995), focuses primarily on the financial
aspects of the demonstration. It challenges the state's estimate of
the discount obtained from participating managed care organizations,
noting that Tennessee's rate-setting methodology understated
historical Medicaid costs by approximately 25 percent. See chapter 4
for a discussion of Tennessee's waiver expenditures for the first
year of the demonstration and the influence of DSH on the level of
state spending.
\34 Under Tennessee's waiver agreement, DSH was technically
eliminated, since the state was relieved of the obligation of making
payments to hospitals. However, Tennessee's level of DSH funding in
fiscal year 1993 is built into the waiver expenditure cap and
actually grows in future years. For further details, see chapter 3.
SAVINGS ANTICIPATED, PRIMARILY
FROM MANAGED CARE
---------------------------------------------------------- Chapter 2:3
States with relatively smaller DSH allotments like Oregon and Florida
rely to a much greater extent than Tennessee and Hawaii on program
savings to finance coverage expansion. Those expected savings,
however, represent more than just the transition to managed care.
Though all 1115 waivers anticipate managed care savings, there
appears to be little unanimity on how quickly savings can be achieved
or on what type of managed care delivery system is the most
efficient.
Without substantial DSH funding, Oregon's expansion goals, as
outlined in the waiver, rely almost exclusively on program savings.
In addition to managed care efficiencies, the state also anticipates
lower Medicaid costs under the waiver as a result of adopting its
redefined benefit package, known as the prioritized list, and an
employer mandate. The mandate would reduce state costs by requiring
employers to provide health insurance coverage to low-income workers
initially covered under the waiver. State estimates suggest that
about 85 percent of the financing for expanded coverage is
attributable to the combination of switching to managed care delivery
arrangements and using the prioritized list. Although the state
attributes a specific amount of savings to the list, its officials
told us that, in fact, it is difficult to distinguish such savings
from managed care efficiencies.
About two-thirds of Florida's financing also relies on program
savings. However, almost half of those savings would result from
proposed reimbursement reforms. Unlike Medicaid physician fees,
other medical services in Florida have had a built-in inflation
adjustment. Under the reimbursement reforms, price increases for
services rendered by HMOs, pharmacies, and clinics, and on an
outpatient basis at hospitals, will be lowered by limiting them to
increases in the Consumer Price Index plus a declining number of
percentage points with each subsequent year.
LACK OF CONSENSUS ON MANAGED
CARE SAVINGS
-------------------------------------------------------- Chapter 2:3.1
Despite the common thread of reliance on anticipated managed care
savings as a funding strategy, state definitions of just what
constitutes managed care and their approaches toward achieving those
savings differ. Thus, Tennessee uses a reimbursement strategy in the
form of a capitation discount to achieve immediate savings. In
contrast, Oregon offers relatively generous capitation payments but
expects control over utilization of services to reduce the rate of
future cost increases. These differing strategies reflect each
state's decision on how best to balance the need for savings against
(1) the extent and maturity of the state's managed care
infrastructure and (2) concerns about enrollee access, quality, and
choice.
Demonstrations typically rely on a mix of different types of managed
care delivery, though one is often predominant: HMO-style systems in
Oregon, PPOs in Tennessee, and physician gatekeeper arrangements in
Florida. Table 2.4 shows the actual enrollment of beneficiaries in
Tennessee and Oregon, and Florida's projection of enrollment if the
state implements its approved waiver. The high penetration of HMOs
in Oregon, with enrollment of almost one-third of the state's
population, facilitated the state's decision to rely on this type of
managed care delivery system. In Tennessee and Florida, enrollment
in HMOs is significantly lower--6 percent and 18 percent of each
state's residents, respectively. HMOs and PPOs participating in the
1115 demonstration in all three states are reimbursed on a per capita
basis, referred to as a capitation payment. Physician gatekeepers,
on the other hand, are often paid on a fee-for-service basis, though
some are partially capitated. What differentiates managed care from
fee-for-service is not only the method of reimbursement but the
attempt to control the utilization of services. Although gatekeepers
in Florida would be paid for each service delivered, enrollees must
get prior authorization to see a specialist. PPOs in Tennessee, on
the other hand, have 3 years to employ physician gatekeepers to help
control the length of inpatient hospital stays and the utilization of
other services.\35
Table 2.4
Percentage Enrollment in Different Types
of Managed Care Delivery Systems
Partially
HMO- Physician capitated
style PPO gatekeeper plans\a
------------------ -------- ---- ---------- ----------
Tennessee 33% 67% 0% 0%
Oregon 91% 0% 4% 5%
Florida 33% 0% 67% 0%
----------------------------------------------------------
Note: Enrollment data for Tennessee and Oregon are actual data as of
May and August 1995, respectively. Florida, which has not yet
implemented its waiver, provided enrollment projections in its waiver
application.
\a A plan that is capitated for a range of services that does not
include inpatient hospitalization.
Tennessee's approach to managed care reflects the strategy used by
large employers. The state asked for and received a substantial
capitation discount from participating HMOs and PPOs.\36 Tennessee
acknowledges that its shift from fee-for-service to capitation is
unlikely to result in significant utilization savings at the outset,
since most traditional and newly eligible recipients are enrolled in
PPOs that lack gatekeepers rather than in more structured HMO-style
managed care arrangements. Despite the belief that utilization
savings will be lower during this initial phase-in period,
substantial capitation discount savings still accrue to the state.
Tennessee officials noted that PPOs in turn often obtain significant
pricing discounts from their providers. Our recent report on the
Tennessee demonstration notes that a primary concern about the future
of the demonstration is the poor financial performance of
participating managed care plans and the willingness of physicians to
contract with those plans.\37 The demonstration's viability, we
concluded, may hinge on the continued willingness of the health care
community to participate in the program in spite of the low
reimbursement levels. Although analysis of access to and quality of
health care under the waiver has been limited because of problems in
collecting data on enrollee visits to providers, beneficiary surveys
and advocacy groups both indicate that access is a problem.
In contrast to Tennessee's steep, up-front, capitation discounts,
Oregon took a longer range approach that emphasizes access and
quality. Concerned about the adverse impact of its already low
Medicaid fee-for-service rates on the delivery of services, Oregon's
initial capitation rates represent an increase over comparable
fee-for-service rates prior to the 1115 demonstration. According to
state officials, this increase contributed to the decision of a large
number of HMOs to participate in the demonstration. As a result,
about 91 percent of the recipients covered under Oregon's 1115 waiver
are enrolled in some type of fully capitated HMO--over three times
more than the state's original estimate.\38 Oregon assumes that more
highly structured managed care will better control the utilization of
services and that over time health care costs will rise at a slower
rate than under the old fee-for-service reimbursement system.
Finally, there appears to be a wide spectrum of opinion among 1115
waiver states about the extent of savings from alternative managed
care structures. Florida assumes that a physician gatekeeper
arrangement, which preserves recipient choice, will produce the
greatest managed care savings and anticipates that the majority of
its Medicaid population will select this option. Only one-third of
the state's Medicaid population is expected to enroll in fully
capitated HMOs, the system that Oregon considers to be the most
cost-effective form of managed care delivery. The gradual erosion of
savings in Kentucky, which employs physician gatekeepers under a
1915(b) waiver, suggests that state enforcement and oversight of such
managed care arrangements are critical. According to state
officials, emergency room use has risen again after an initial
decline. Kentucky's 1115 waiver application envisioned an eventual
transition to more highly structured forms of managed care.
--------------------
\35 As of February 1995, about one-half of enrollees had been
assigned to primary care physicians. While one Tennessee PPO,
accounting for almost 50 percent of total enrollment, allows
specialist visits without prior authorization, most others require
preauthorization.
\36 According to the state's waiver application, a key element in
establishing rates for managed care organizations in Tennessee is the
assumption that per capita costs can be reduced from historic levels
because more people will have health care coverage. The reduction
recognizes that prior to the waiver, some of the cost of
uncompensated care in the state's health care system was shifted to
other payers. Skeptical that a reduction in uncompensated care would
eliminate a cost shift that had become "institutionalized," the state
discounted the initial capitation rate under the waiver by about 25
percent, which it characterized as an attempt to capture about
one-half of uncompensated care charges statewide. The dollars
"saved" through the discount help to finance expanded coverage.
\37 GAO/HEHS-95-186, Sept. 1, 1995.
\38 This estimate includes all new eligibles as well as both low-
income families and the aged, blind, and disabled. The state's
original estimate of enrollment in fully capitated health plans
excluded the latter group since it was not initially scheduled to be
part of the demonstration.
NEWLY INSURED ASKED TO
CONTRIBUTE TOWARD COSTS
---------------------------------------------------------- Chapter 2:4
Although the waiver proposals we reviewed rely primarily on
redirected funds and expected program savings to finance coverage
expansions, many state financing strategies incorporate new money
raised by charging premiums to certain enrollees. Such premiums
appear to make a modest contribution toward overall financing,
ranging from a high of about 15 percent to as little as 1 percent of
the core funding strategies we discussed. Moreover, at least one
state's application recognized that not all premiums are likely to be
collected.
The 1115 waivers in Tennessee, Florida, and Hawaii require most
recipients with incomes above the poverty level to pay premiums on a
sliding scale. In addition, Florida expects a minimal premium
contribution from individuals below the poverty level if they have
any income. Expected premiums in Tennessee are about 8 percent of
the combined total of redirected funds and the capitation discount.
In Florida, premiums account for about 15 percent of the funds the
state says it needs to provide insurance to a target group of about
1.1 million. Officials in Hawaii told us that premiums expected from
newly eligible beneficiaries represent only about 1 percent of
coverage expansion funding. Under the approved waiver agreements, a
substantial portion of the premiums collected in Tennessee and
Florida can be counted as state match, with no reduction in federal
expenditures. In Hawaii, the state and federal governments share
equally in the cost offset represented by individual premiums. Table
2.5 provides hypothetical examples of the different types of
arrangements used to allocate premiums.
Table 2.5
Hypothetical Examples of Premium
Allocation Agreements Under Approved
Waivers
Enrollee's
Cost of Federal Federal State premium
coverage match rate share share contribution
---------------------- ------------ ------------ ----------- ---------- ------------
State A (no enrollee $100 50% $50 $50 0
premium)
State B (only state 100 50% 50 40 $10
benefits from
premiums)
State C (premium 100 50% 45 45 10
revenue divided
between federal and
state governments)
-----------------------------------------------------------------------------------------
TWO STATES FOLD SUBSIDIZED
INSURANCE PROGRAMS FOR LOW-
INCOME RESIDENTS INTO WAIVERS
---------------------------------------------------------- Chapter 2:5
Hawaii and Minnesota are folding existing state subsidized insurance
programs for low-income residents into their waiver programs and, in
the process, bringing along the state dollars that financed them.\39
These program dollars now qualify for federal match.\40 While some
individuals in these state-sponsored expansions were eligible for
Medicaid under optional programs for pregnant women and children
authorized in the late 1980s, others were not.
Previously, states chose to expand coverage outside of Medicaid for a
number of reasons. First, men, single adults, and childless couples
were generally ineligible for Medicaid unless they were elderly or
disabled. Second, such programs made it easier for states to provide
coverage to entire families. Third, freed from Medicaid rules,
states were able to offer more modest benefits and to require
participants to pay premiums, co-payments, and deductibles. At least
one state, Minnesota, cited another rationale for its self-funded
program, MinnesotaCare. State consultants concluded that the lack of
an employer mandate would result in the migration of children from
private insurance to Medicaid. Eligibility rules in MinnesotaCare
were designed to prevent such a migration. Thus, MinnesotaCare
enrollees must have been uninsured for the 4 months immediately
preceding enrollment and may not have had access to
employer-subsidized health insurance for the previous 18 months.
HCFA allowed Minnesota to maintain these barriers in its approved
1115 waiver.
--------------------
\39 While Tennessee had no means-tested, subsidized insurance
programs, the state did fold its high-risk pool for otherwise
uninsurable residents into its 1115 demonstration. Hawaii had no
high-risk pool, and Minnesota and Florida elected to maintain
separate programs for the uninsurables outside of their waivers.
\40 Minnesota officials told us that they planned to continue
extending MinnesotaCare coverage of single adults and childless
couples by using state funds freed up through the 1115 waiver's shift
of children in MinnesotaCare into Medicaid.
THREE OF FOUR DEMONSTRATION
WAIVERS ARE POTENTIALLY NOT BUDGET
NEUTRAL
============================================================ Chapter 3
Contrary to the administration's assertion that approved Medicaid
1115 waivers are budget neutral, net federal spending in the four
states we examined could potentially exceed projected without-waiver
program costs over the 5-year duration of the demonstrations. The
net additional federal funding available in these four states is
small in relation to allowable demonstration spending. However,
overall federal Medicaid expenditures could grow significantly if the
administration shows a similar flexibility in reviewing the large
backlog of pending waivers.
Administration officials told us that, since some states' Medicaid
expenditures were growing faster than the national average in the
past, the budget neutrality of each proposed waiver should be
evaluated independently in order to capture these variations. Such
an approach is difficult because of the lack of consistently
generated, state-specific forecasts. Lacking such data, we relied on
the only available forecasts--national projections of how the current
Medicaid program would grow over the 5-year duration of waiver
programs.\41 At the same time, we reviewed waiver applications and
talked with state officials to identify factors suggesting whether a
state's Medicaid expenditures would indeed exceed the national norm.
We found no evidence to support the high budget caps agreed to by the
administration.
--------------------
\41 Since we were evaluating the administration's budget neutrality
assessments, we used OMB's national Medicaid projections rather than
those produced by the Congressional Budget Office.
SPENDING LIMITS INCONSISTENT
UNDER FOUR APPROVED WAIVERS
---------------------------------------------------------- Chapter 3:1
Medicaid 1115 waiver programs are popular because they allow the
administration to grant states significant program flexibility.
Since the 1980s, OMB has used its budget neutrality policy to ensure
that states were not given access to additional federal funding at
the same time they were provided with greater program flexibility.
Rather than applying a uniform methodology to measure the budget
neutrality of waiver applications approved since 1993, the
administration has allowed considerable variation in growth of
baseline costs from state to state.
According to the administration, each of the waiver programs is
budget neutral, even though the individual growth rates vary
significantly. The results of the administration's flexible,
state-specific approach are shown in table 3.1, which summarizes the
rates of increase allowed in the four agreements--Florida, Tennessee,
Oregon, and Hawaii. Also included in the table are the
administration's projected rates of growth for Medicaid on a
nationwide, current services basis over roughly the same time period.
Table 3.1
Allowed Growth Rates Under Four Approved
1115 Waiver Agreements
(Numbers in percent)
Year 1 Year 2 Year 3 Year 4 Year 5 Average
----------------------------- -------- -------- -------- -------- -------- --------
OMB FY94 projection\a 14.5 15.4 13.0 12.2 11.2 13.3
Tennessee 16.4 8.3 7.5 5.7 5.1 8.5
Oregon\b 11.2 36.9 20.6 6.3 11.5 16.8
Hawaii\b 30.5 15.2 13.3 14.4 14.4 17.4
OMB FY95 projection\a 10.6 12.2 12.3 12.2 11.7 11.8
Florida 16.3 15.6 15.4 14.8 14.2 15.3
-----------------------------------------------------------------------------------------
\a Each year the president's budget contains OMB's projection of
expected growth in the Medicaid program on a nationwide, current
services basis. States are grouped under the OMB projection for the
fiscal year in which their waiver was approved.
\b Oregon's and Hawaii's rates of growth depend on an estimated
number of individuals enrolled as defined in the waiver agreements.
Actual rates of growth will vary depending on program enrollment.
See appendix I for a detailed explanation of our methodology for
arriving at these figures.
Table 3.1 shows that in three out of four states, the waiver
agreements permit growth above what OMB projected for the Medicaid
program as a whole at the time the waivers were approved. The growth
patterns among states also vary. The growth rates of Tennessee,
Florida, and Hawaii are the highest in the first year of their
waiver. While Florida's growth rate declines gradually in each
subsequent year, both Tennessee's and Hawaii's drop dramatically in
the second year and then decline more slowly in the remaining years
of their demonstrations. In Oregon, however, the highest rate of
growth is in the second and third years, with dramatically lower
increases in the last two.\42
Given the unique state setting of each Medicaid program, some
variation in the rate of Medicaid growth among states is to be
expected. However, even though OMB was predicting overall lower
growth in Medicaid, state waiver applications did not identify future
trends to justify their higher-than-average growth rates over the
course of the demonstrations. Instead, states used a variety of
arguments primarily based on history and options available under
current statute to convince the administration that their particular
situation warranted a high rate of growth.
HCFA and state officials admit that, in some states, continued
Medicaid growth at historical rates is unsustainable because of the
great strain it places on state budgets. Moreover, several of the
primary contributors to the growth of state Medicaid budgets over the
past 5 years are no longer present. For example, some states' use of
targeted provider taxes and donations contributed to the rapid rise
in DSH funding between 1989 and 1993, but recent legislation strictly
limits--and in some cases caps--such growth. It also appears
unlikely that states will be asked to absorb major new federally
mandated expansions of populations and benefits--a practice that
contributed to high growth rates in the past.\43
As shown in figure 3.1, the assumption that higher historical rates
of Medicaid growth will continue runs contrary to the
administration's own projections of nationwide Medicaid growth on a
current services basis. Each successive projection since 1993 shows
a decline in the rate of growth in Medicaid. In addition to pointing
to the history of recent rapid growth in Medicaid expenditures,
states used the so-called "hypotheticals" argument to justify higher
baselines. They argued that groups who were hypothetically eligible
for Medicaid coverage under existing law, but had not been included
in a state's Medicaid plan, should be considered part of the state's
baseline population for the purpose of determining budget neutrality.
Including hypotheticals raises baseline costs, making budget
neutrality easier to achieve.
Figure 3.1: OMB Projections of
Medicaid Current Services
Outlays
(See figure in printed
edition.)
The Hawaii, Kentucky, Minnesota, Rhode Island, and Ohio waiver
agreements allow hypothetical populations to be included in the
baselines. To date, the inclusion of hypotheticals has been limited
to those individuals who would actually be covered by the
demonstration and who are optionally eligible for Medicaid under
section 1902(r)(2) of the Social Security Act.\44 Hawaii and
Minnesota have covered some of this population outside the Medicaid
program in the past through state-only funded programs.\45 A Hawaii
Medicaid official estimated that including the 1902(r)(2) population
added approximately $56 million to the state's waiver baseline over
the 5-year life of the program--about 4 percent of total waiver
agreement funding. In Kentucky, Rhode Island, and Ohio, however,
hypotheticals were not covered by any state-funded program.
In each waiver we reviewed that included hypotheticals in the
baseline, state officials mentioned cost containment as a primary
consideration in seeking 1115 demonstration authority. It is
questionable, therefore, that these states would have added optional
eligibility groups to their Medicaid programs without the waiver.
--------------------
\42 Oregon's estimates of expenditure growth assume high enrollment
in the early years, as individuals who were previously ineligible for
coverage move into the program. The estimates also assume that
during the last 2 years the state will implement its employer
mandate, resulting in a decrease in the number of individuals
enrolled in the program and lower costs. In addition, by the fourth
year of the waiver, managed care is expected to restrain medical
costs.
\43 No major federal mandates have been added since 1991.
\44 Under section 1902(r)(2) of the Social Security Act states were
essentially given the option to expand coverage of women, children,
the elderly, and the disabled by employing less restrictive
eligibility methodologies.
\45 Some individuals eligible for Medicaid under section 1902(r)(2)
have been covered under Hawaii's General Assistance or State Health
Insurance Program and under MinnesotaCare in Minnesota.
NATIONWIDE MEDICAID EXPENDITURE
FORECASTS USED TO ASSESS BUDGET
NEUTRALITY
---------------------------------------------------------- Chapter 3:2
The complexity of and variation in individual state programs makes it
difficult to assess budget neutrality without a consistent frame of
reference.\46 Lacking state-specific Medicaid expenditure forecasts,
we used OMB's current services projections of growth in Medicaid for
the nation as a whole. A current services projection is policy
neutral and only reflects medical inflation, normal growth in the
eligible population, and changes in utilization for the entire
Medicaid program.
To determine if the four approved waivers were budget neutral, we
first estimated the cost of continuing the traditional Medicaid
program--absent the demonstration--in each state. This
estimate--referred to as "without-waiver spending"--was developed by
adjusting for inflation in the following manner: We adjusted the
cost of providing Medicaid in the year prior to waiver implementation
at the rate specified by OMB in its forecast of future Medicaid
current services outlays. We compared this without-waiver spending
estimate to total projected costs under the waiver expenditure caps
negotiated by each state and the administration. The difference
between our without-waiver projection and the waiver expenditure cap
in each state is the basis for our conclusion of whether an agreement
is budget neutral. When the difference was positive, we examined
state- specific information to determine if there was any
identifiable reason why the waiver expenditure cap should exceed our
without-waiver spending projection and still be regarded as budget
neutral.
We applied this methodology to the two types of waiver spending caps
agreed to by the administration, aggregate and per capita expenditure
limits. For states that use the aggregate cap--Tennessee and
Florida--the waiver funding limits are specified in the terms and
conditions approving the demonstration. For states with per capita
limits--Hawaii and Oregon--the spending cap depends on the actual
number of enrollees in the waiver program. To assess the potential
budget neutrality of these per capita agreements, we used enrollment
projections developed by each state and submitted to HCFA in
conjunction with approval of the waiver. If actual enrollment proves
to be higher than these initial projections, then the waiver
agreement funding limit will generally be higher and the state will
have access to more funds than we projected. Conversely, if
enrollment falls below these projections, fewer additional resources
would be available.\47
--------------------
\46 Reliable HCFA Medicaid expenditure data that are sufficiently
flexible to analyze waiver programs are limited. For example, one
HCFA report contains auditable expenditure data by service category;
however, expenditures on these services cannot be related to eligible
individuals. On the other hand, a different report on expenditures
by eligibility category omits items such as DSH and capitation
payments. Moreover, these data are unreliable because components may
not sum to report totals or may rely on estimated amounts. Finally,
auditable expenditure reports do not allow for the separation of
acute care services for recipients covered by the waiver programs
from those who are not part of the demonstration. In a number of
states, waivers exclude participation by aged, blind, and disabled
beneficiaries. Since 1115 waivers often segment the Medicaid
population and services in ways or time periods not reflected in HCFA
reports, budget neutrality analysis often relies on other state
databases.
\47 In both Oregon and Hawaii, actual enrollment for the first year
of the waiver was higher than estimated.
TENNESSEE
-------------------------------------------------------- Chapter 3:2.1
For the Tennessee demonstration, the administration specified an
aggregate federal funding cap, with any spending above that cap
ineligible for federal match. The relatively low rates of growth
under this spending limit mimic the growth caps used in the
President's 1993 health care reform proposal. The waiver agreement
expenditure cap covers all aspects of Tennessee's Medicaid program,
whether or not the associated populations are being moved into
managed care. Figure 3.2 shows a comparison of Tennessee's waiver
agreement spending cap and our without-waiver projection of spending.
Our without-waiver estimate was derived by increasing actual
expenditures in the year prior to implementation of the demonstration
at the national current services growth rate projected by OMB. When
compared with our without-waiver spending projection, Tennessee's
waiver expenditure cap is budget neutral. Savings in subsequent
years make up for initial demonstration costs that exceed projected
without-waiver spending.
Figure 3.2: Tennessee:
Comparison of Waiver Agreement
Spending Cap and Projected
Spending Without Waiver
(See figure in printed
edition.)
While the Tennessee waiver agreement is budget neutral using the
current services methodology, the administration's treatment of DSH
funds raises another issue. After the first year of the waiver, DSH
funding disappears as a budget item since it is built into the
baseline that increases at the agreed-to rates of growth on overall
Medicaid spending.\48 However, DSH funding in Tennessee and a number
of other states is capped by law because it is more than 12 percent
of the state's total Medicaid expenditures. In such states, DSH
funding is only permitted to grow when it falls below this 12-percent
cap.
Under the waiver agreement, Tennessee's DSH funding is allowed to
grow after the first year of the waiver, even though it exceeds the
12-percent limit.\49 Consequently, Tennessee is eligible for
approximately $250 million in DSH growth that would not have been
allowed without the waiver. Without this additional DSH funding, net
savings to the state and the federal government under the Tennessee
waiver agreement would have been higher.
Since the Tennessee agreement, the administration has separated DSH
funding from other aspects of waiver program funding. This approach
allows the cap on DSH growth to be enforced. Moreover, HCFA
officials told us that if DSH funding is growing at a slower rate
than the other program elements covered by the waiver, then that
lower growth rate is applied to any DSH growth.
--------------------
\48 According to HCFA, Tennessee's DSH payments were not allowed to
grow in the first year of the waiver program--state fiscal year (SFY)
1994. Holding DSH payments constant results in an overall Medicaid
funding increase of 19.6 percent for the remaining elements of the
state's Medicaid program in SFY 1994. On the other hand, simply
comparing SFY 1993 Medicaid expenditures, including DSH, with the
level permitted under the waiver agreement in SFY 1994 results in an
overall increase of 16.4 percent.
\49 In contrast, our assessment of the budget neutrality of
Tennessee's waiver agreement held DSH payments constant until they
would have constituted less than 12 percent of total Medicaid
expenditures. Thus, our approach is consistent with statutory DSH
limits.
FLORIDA
-------------------------------------------------------- Chapter 3:2.2
As in Tennessee, the Florida waiver agreement has an aggregate cap on
demonstration expenditures.\50 The cap only applies to the acute care
and DSH portions of the state's Medicaid program. Florida relied
heavily on an historical argument to justify its higher-than-average
rates of growth under the waiver, even though some state officials
later told us that it was unlikely that such growth could be
sustained.\51 Figure 3.3 compares estimated waiver agreement spending
and spending without the waiver. Our without-waiver estimate was
derived by increasing actual expenditures in the year prior to
implementation of the demonstration at the national current services
growth rate projected by OMB. The comparison shows that the waiver
spending cap exceeds our without-waiver estimate, with the difference
equaling $4.5 billion in state and federal funding.
Figure 3.3: Florida:
Comparison of Estimated Waiver
Agreement Spending Cap and
Projected Spending Without
Waiver
(See figure in printed
edition.)
We also analyzed the extent to which Florida's coverage expansion
goals depend on this $4.5 billion in excess funding. Our analysis
shows that Florida's enrollment plans would have to be scaled back
without the excess funds provided under the waiver agreement. If the
funding limits for Florida's waiver agreement had been based on
national projections of growth in the Medicaid program, both DSH and
the state's Medically Needy Program would have grown at slower rates.
As a result, almost $1 billion less than the amount needed to meet
the state's expansion goals would have been available. As shown in
table 3.2, the $4.5 billion in excess funding potentially available
under the waiver more than covers that shortfall. We believe that
the difference between the excess funds available and the
shortfall--about $3.5 billion--provides a backup if state assumptions
about managed care savings or other funding sources prove faulty.
Table 3.2
Comparison of Florida's Expansion Cost
Estimates and Budget Neutral Growth in
Without-Waiver Program
(Dollars in billions)
Budget projection Amount
------------------------------------------------ --------
Expansion costs $5.864
Funds available to expand coverage assuming 4.869
budget neutral growth in DSH and Medically
Needy Program
Shortfall .995
Funds in excess of our without-waiver spending 4.527\a
projection
Backup: difference between excess funds and 3.532
shortfall
----------------------------------------------------------
\a The federal share is about $2.5 billion (55 percent), and the
remainder consists of state matching funds.
--------------------
\50 The waiver agreement includes a circuit-breaker provision that
allows Florida to eclipse the waiver agreement cost cap in the event
that the state's traditional Medicaid population exceeds projections
by 3 percent or more.
\51 A Florida official told us that the state's 1115 waiver was
presented to the legislature at about the same time the application
was submitted to HCFA. This official was skeptical that the
legislature would sustain the past rates of growth in Medicaid in
order to finance coverage expansion under the waiver. A more likely
scenario, he said, is that the legislature would choose to let
Medicaid grow at a slower rate than in the past. The Florida
legislature adjourned in both 1994 and 1995 without agreeing to
implement the waiver program. During the 1995 session, the
Governor's waiver bill was never reported out of subcommittee. In
its place, the legislature considered alternatives that differed
considerably from the waiver approved by the administration in
September 1994. No consensus emerged, and the governor planned to
call a special legislative session during the summer of 1995.
According to a Florida official, the state budget approved by the
legislature already moves money from social services programs to
prison construction--including managed care savings accrued under the
state's more limited 1915(b) and voluntary HMO programs. Managed
care savings are an important source of funding for expanded
coverage.
OREGON
-------------------------------------------------------- Chapter 3:2.3
It is more difficult to apply our methodology to states with per
capita waiver agreements, like Oregon. As implied by the term "per
capita," estimating both the waiver agreement spending cap and
without-waiver expenditures requires assumptions about enrollment.
Moreover, in Oregon, the mix of benefits changed in the transition
from traditional fee-for-service Medicaid to the demonstration,
making it more difficult to arrive at a base-year cost. In
addressing these methodological challenges, we used the projected
enrollment in the waiver agreement to estimate spending, and we
derived a base-year cost from state reports. We discussed our
methodology with Oregon officials, who agreed that it was
appropriate. Figure 3.4 compares estimated waiver agreement spending
and spending without the waiver. Our without-waiver projection was
derived by increasing base-year estimated expenditures at the
national current services growth rate projected by OMB. The
comparison shows that the Oregon waiver spending ceiling exceeds our
without-waiver projection.
Figure 3.4: Oregon:
Comparison of Estimated Waiver
Agreement Spending Cap and
Projected Spending Without
Waiver
(See figure in printed
edition.)
HAWAII
-------------------------------------------------------- Chapter 3:2.4
The Hawaii waiver expenditure cap is also based on per capita costs.
As with the Oregon spending limit, all the pieces needed to calculate
the projected waiver costs were not included in the waiver agreement
documents. HCFA and Hawaii have agreed to use 1993 as the base year
to calculate budget neutrality, but have not yet agreed on per capita
costs for the various eligible populations. Our analysis was further
complicated by the fact that Hawaii has not yet completed its final
enrollment count for the base year. To make our calculation of
program costs, we used a preliminary state average per capita cost
and an estimate of the base-year enrollment from waiver documents.
According to state officials, these were the best figures available.
Figure 3.5 compares estimated waiver agreement spending and spending
without the waiver. Our without-waiver projection was derived by
increasing base-year estimated expenditures at the national current
services growth rate forecast by OMB. The comparison shows that the
waiver expenditure cap exceeds our without-waiver spending
projection.
Figure 3.5: Hawaii:
Comparison of Estimated Waiver
Agreement Spending Cap and
Projected Spending Without
Waiver
(See figure in printed
edition.)
Of the four waiver agreements we analyzed, only Hawaii's included a
hypothetical population in its baseline. As illustrated by figure
3.6, our analysis shows that the inclusion of this hypothetical
population--made up of children that had previously been covered by
state-only funded programs and who were eligible for Medicaid under
section 1902(r)(2)--did not significantly affect the cost of
implementing the waiver over its 5-year life.\52
Figure 3.6: Impact of
Including "Hypothetical"
Populations in Hawaii's Program
Baseline
(See figure in printed
edition.)
--------------------
\52 Kentucky also included the optional 1902(r)(2) population in its
waiver agreement baseline. Because this population is so much larger
in Kentucky than it is in Hawaii, it would have had a significant
impact on the cost of Kentucky's demonstration program. Kentucky's
legislature refused to permit implementation of the state's approved
waiver.
STATE VARIATION DOES NOT
JUSTIFY WAIVER AGREEMENT
EXPENDITURE CAPS
---------------------------------------------------------- Chapter 3:3
In responding to our work on budget neutrality, the administration
said that the characteristics of individual states--primarily
historical trends--justified waiver growth limits higher than
projected national average program growth. However, none of the four
waiver applications we analyzed in detail offered a rationale for the
expected higher-than-average rates of growth in expenditures,
enrollment, or medical inflation. OMB officials also told us that
the four state Medicaid programs we analyzed in detail had been
growing faster than the national average. Table 3.3 compares the
national average with growth in Medicaid spending from 1988 to 1993
for Florida, Tennessee, Oregon, and Hawaii.
Table 3.3
Recent Medicaid Spending Trends in Four
Waiver States
(Numbers in percent)
1988-1993
trend
-------------------------------------------- ------------
Florida 25.8
Tennessee 20.9
Oregon 20.4
Hawaii 17.5
==========================================================
National average 19.4
----------------------------------------------------------
Note: Trends and growth in expenditures exclude growth in DSH.
This analysis shows that, with the exception of Florida, these states
were not growing significantly faster than the national average. In
fact, the trend in Medicaid expenditures in Hawaii suggests that it
might be appropriate for its waiver program to grow more slowly than
the national average.
Federal mandates have contributed significantly to variations in
state Medicaid growth rates over the last 8 years. According to a
1994 Urban Institute study, states with historically more restrictive
Medicaid programs grew very rapidly during the period 1988 to 1992,
with much of the growth attributable to newly eligible adults and
children coming into the program under federal mandates.\53
Such new mandates were at least partially responsible for escalating
costs in Florida, a state whose Medicaid growth rate from 1988 to
1993 was significantly higher than the national average. Thus,
previously ineligible adults, children, elderly, and disabled
individuals accounted for more than 30 percent of the increase in
Medicaid spending between 1989 and 1990. According to state
officials, the establishment of an optional Medically Needy
Program--which totaled $98 million by 1993--also contributed to state
expenditure growth. More significantly, it had the unintended
consequence of helping to greatly expand enrollment of low-income
families. Florida officials explained that an outreach program
designed to increase participation in the Medically Needy Program
uncovered many low-income families eligible for Medicaid. Enrollment
of low-income families rose nearly 24 percent per year from 1990 to
1993, with associated expenditures more than doubling from just under
$500 million to nearly $1.2 billion. Finally, state officials told
us that the recession in 1991 also contributed to growth in the
number of low-income families enrolled in Medicaid.
Although these factors contributed to the sharp rise in Florida's
Medicaid expenditures in the early 1990s and resulted in waiver
growth rates significantly higher than the national average, even
Florida officials do not expect a continuation of past trends.
Actual experience appears to support their predictions. For example,
while the waiver agreement estimated that acute care expenditures
would grow at 17 percent in 1994, the base year, actual spending
increased by only 12 percent without implementation of the waiver.
Moreover, Florida officials told us that enrollment growth among
low-income families has leveled off at around 4 or 5 percent over the
past 18 months. The Florida waiver agreement estimates that
low-income family enrollment growth will fluctuate between 2 and 3
percent over the life of the waiver.
--------------------
\53 Theresa Coughlin and others, "States' Responses to the Medicaid
Spending Crisis: 1988 to 1992," Journal of Health Politics, Policy,
and Law, Vol. 19, No. 4 (Winter 1994).
THREE WAIVERS PROVIDE ACCESS TO
SIGNIFICANT ADDITIONAL FUNDING
---------------------------------------------------------- Chapter 3:4
While the administration contends that each of the waiver agreements
we reviewed is budget neutral, our analysis of both national and
state-specific data shows that most of the agreed-upon rates of
growth are too high. As a result, the agreements provide these
states with access to significant additional federal Medicaid
funding.\54 Table 3.4 compares the waiver agreement spending caps and
our without-waiver expenditure projections (based on national
Medicaid growth rates), aggregated over the 5-year duration of the
programs. The $1.9 billion in net additional federal funds should
not be interpreted as a precise prediction of the amount of
additional funds available under these four waivers; rather, it
reflects the significant magnitude of the differences between the two
projections.
Table 3.4
Comparison of Four Section 1115 Waiver
Agreement Expenditure Caps and Projected
Without-Waiver Spending
(Dollars in billions)
Estimated Projected
waiver without-
agreement waiver Additional Federal
spending spending funds share
-------- ------------ ------------ ---------- --------
Tennesse $20.9 $22.2 ($1.3) ($0.9)
e
Florida 37.2 32.7 4.5 2.5
Oregon 2.7 2.3 0.4 0.2
Hawaii 1.3 1.1 0.2 0.1
==========================================================
Net $3.8 $1.9
additio
nal
funds
----------------------------------------------------------
--------------------
\54 States are not obligated to spend up to the limits placed on them
by the agreements. Moreover, the federal government will only
provide funds to match actual state expenditures.
DESPITE LOWER INFLATION, STATE
DEMONSTRATIONS FACE COST PRESSURES
============================================================ Chapter 4
We analyzed preliminary financial results from 1115 demonstrations in
Tennessee, Oregon, and Hawaii--three states with significant waiver
implementation experience. The data suggest that waivers will
continue to evolve as states attempt to balance coverage expansion
goals against systemic cost- containment pressures. Changing
political/fiscal realities in all three states and the potential for
an acceleration in the rate of medical inflation underscore the
challenges in implementing fixed-cost agreements in a variable-cost
environment.
Enrollment of previously uninsured individuals in both Oregon and
Hawaii surpassed state estimates. Oregon was able to meet greater
than expected demand without exceeding its waiver agreement
expenditure cap. Hawaii, however, projects demonstration spending
will be about 23 percent higher than permitted by its waiver
agreement--costs it will have to offset in future years if it is to
live within the expenditure cap. After maintaining open enrollment
for a full year and achieving about 80 percent of its coverage
expansion goal, Tennessee abruptly cut off enrollment because of a
budget crisis it attributes to the demonstration. In contrast to
Hawaii, Tennessee's first-year demonstration costs were 14 percent
below its waiver agreement spending cap. While Tennessee's waiver
program covered several hundred thousand previously uninsured
individuals, total expenditures were on a par with its significantly
smaller prewaiver program.
Despite the slowdown in medical inflation, all three states face
pressures to contain future waiver costs. Oregon and Hawaii have
announced a number of initiatives to do so, including higher cost
sharing and new eligibility rules. Since resources up to a state's
waiver funding cap are available until the end of the demonstration,
expenditures to date may not be a reliable indication of
demonstration costs--particularly if medical inflation accelerates.
OREGON: FIRST-YEAR ENROLLMENT
AND COST EXPERIENCE
---------------------------------------------------------- Chapter 4:1
Greater than anticipated managed care savings allowed Oregon to offer
insurance to about 50 percent more new enrollees than anticipated
during 1994--without breaching the waiver funding agreement.\55
Though the exact amount of the cap is in dispute, HCFA data show that
waiver expenditures of $347 million were about $34 million less than
the administration's estimate of the waiver agreement ceiling.\56
As shown in table 4.1, actual 1994 expenditures were remarkably close
to the 1993 cost estimates that formed the basis of the federal-state
financing agreement--within about $600,000. However, as this table
also demonstrates, traditional Medicaid beneficiaries were less
expensive than the state estimated. Table 4.2 provides our analysis
of costs for traditional Medicaid beneficiaries and those newly
eligible under the waiver on a per-person-per-month (PPPM) basis.\57
On average, traditional eligibles cost 25 percent less than
anticipated. Those newly eligible, however, cost 36 percent more.
Oregon officials attribute the higher costs of new eligibles, in
part, to the fact that many are sick when they apply for coverage.
Thus, hospitals are signing up individuals for health benefits under
the waiver as soon as they are admitted. Currently, eligibility for
benefits commences with the submission date rather than with the
subsequent approval of the application. Until the newly eligible
individual is enrolled in a managed care plan, providers are
reimbursed on a fee-for-service basis, further increasing state
costs.
Higher costs for newly eligible individuals, however, were offset by
additional managed care savings of $38 PPPM for each traditional
Medicaid recipient. In part, state officials credit increased
savings to the fact that all but 8 of 36 counties are served by fully
capitated, HMO-style health plans. As a result, about 91 percent of
waiver enrollees, rather than the estimated 28 percent, are receiving
services from what the state believes is the most cost-effective form
of managed care.\58 We also believe that the current ebb in medical
inflation contributed to lower than expected costs.
Table 4.1
Comparison of 1994 Actual and Estimated
Costs and Enrollment Under Oregon's 1115
Waiver
(Dollars in millions and enrollment in
thousands)
Estima Differ Estima Differ
Actual ted ence Actual ted ence
---------------------- ------ ------ ------ ------ ------ ------
Traditional Medicaid $224.6 $286.8 ($62.2 181.9 173.5 8.4
)
New eligibles 122.3 60.7 61.6 69.7 46.8 22.9
======================================================================
Total $346.9 $347.5 ($.6) 251.6 220.3 31.3
----------------------------------------------------------------------
Table 4.2 Comparison of Estimated versus
Actual PPPM Costs for Traditional and
Newly Eligible Beneficiaries Under
Oregon's 1115 Waiver
Estima
Actual ted
costs costs
------------------------------------------------------ ------ ------
Traditional Medicaid $112 $150
New eligibles $160 $118
----------------------------------------------------------------------
--------------------
\55 Oregon's first year of waiver implementation--February to
December 1994--only covers 11 months.
\56 The waiver agreement specifies the maximum allowable costs for
each of three categories of both traditional and newly eligible
enrollees. It does not specify a methodology if, as was the case,
enrollment differs from projections, that is, when there are fewer
beneficiaries in a more expensive category or more in a less costly
classification.
\57 The aged, blind, and disabled were not part of the demonstration
until February 1995. Thus, enrollment/cost statistics for
traditional Medicaid recipients only reflect low-income families.
\58 This estimate includes the aged, blind, and disabled. Excluding
this group, the percentage of total enrollees in fully capitated
health plans is slightly higher--92 percent. Compared with new
eligibles and low-income Medicaid families (3 percent), a higher
percentage of the aged, blind, and disabled (10 percent) are
enrolling in the physician gatekeeper option and thus lowering the
percentage participating in fully capitated plans.
SOME COST CONCERNS SPRING
FROM DETAILS OF WAIVER
AGREEMENT
-------------------------------------------------------- Chapter 4:1.1
Despite greater-than-expected managed care savings in Oregon, state
officials are concerned about the financial implications of current
cost/enrollment trends. The concern stems from the fact that the
traditional Medicaid population determines the funding base for
covering new eligibles. Under the per capita cost agreement, federal
matching funds for new eligibles are tied to a fixed ratio of new to
current eligibles. Thus, for every four current eligibles, Oregon
can claim a federal match for one new eligible in the first year of
the waiver.
In the first year, unexpected enrollment by new eligibles and
higher-than-anticipated costs for this group was accompanied by a
drop in the number of low-income families--the major component of the
coverage expansion funding base.\59 Should these trends continue,
fewer federal dollars than needed would be available to meet future
waiver costs. In mid-1995, Oregon officials asked HCFA for approval
to implement a number of cost-reduction initiatives. As of October
1995, Oregon had received approval to change waiver eligibility
rules\60 and delay full implementation of mental health services.
HCFA has yet to approve the state's request to reduce benefits\61 and
require premiums/some co-payments for newly eligible individuals. In
addition, state officials told us that greater-than-anticipated
managed care efficiencies may allow them to reduce the capitation
rate.
--------------------
\59 Oregon officials attributed the large number of new eligibles to
higher-than-anticipated participation rates rather than to an
increase in the number of uninsured state residents. The waiver
intentionally simplified eligibility and enrollment in order to
encourage participation. The cost controls now under consideration
(asset tests and co-payments) are intended to reduce participation
rates.
\60 Oregon's proposed changes apply only to those newly eligible
under the waiver. Eligibility would be based on 3 months' rather
than 1 month's income, with liquid assets limited to $5,000. In
addition, full-time college students would no longer be able to
enroll.
\61 The funding line for covered services would be moved from line
606 to line 581 of the prioritized list of benefits. Examples of
services that would no longer be covered include (1) certain urinary
tract and yeast infections, such as thrush in infants; (2) painful
menstruation and pelvic discomfort; (3) chronic bronchitis; and (4)
surgical correction of deformities and injuries of the limbs or feet
that may result in greater mobility or function. The state
attributes the low priority attached to these services to the
availability of over-the-counter medicine, to the fact that
individuals get better on their own without medical intervention, or
to uncertainty over the benefits of medical treatment.
OTHER STATE FISCAL CONCERNS
-------------------------------------------------------- Chapter 4:1.2
Other, more general, fiscal concerns in Oregon stem from the impact
of a 1991 tax initiative, new state priorities that reflect the
outcome of the 1994 elections, and uncertainty about the fate of the
employer mandate. Under the tax initiative, any shortfall in
education funding that results from a mandated reduction in the
property tax rate must be offset by general revenues. Funding for
the 1115 waiver also comes from general revenues--rather than from a
dedicated tax paid by a specific group. Similarly, a number of new
priorities, such as prison construction, may further increase the
competition for state funds.
The employer mandate plays an important role in Oregon's waiver
finance plan. Oregon estimates that the mandate, originally
scheduled to be phased in during the last 2 years of the waiver,
would reduce both the number of traditional Medicaid beneficiaries
and newly eligible individuals covered under the waiver. In
addition, program costs will be reduced for low-wage workers who
obtain employer-provided coverage but have incomes below the poverty
level. For these individuals, Medicaid will only pay for costs not
covered by the employer-provided insurance.
Growing business opposition to the mandate coupled with a political
realignment in the state legislature creates considerable uncertainty
about the future of this funding source. In 1993, the legislature
postponed implementation of the mandate, potentially increasing state
costs in the process. Although state legislation requires that the
mandate be repealed unless the Congress grants Oregon an exemption to
ERISA by January 1996, the legislature recently sent the Governor a
bill that would have repealed the mandate outright. He vetoed the
bill in July 1995. According to state officials, the administration
has indicated that no adjustments will be made to the waiver
financing agreement if the employer mandate is not implemented. And
without the mandate, waiver costs will increase, forcing the state to
make up the difference or to develop additional cost- containment
strategies. In 1994, Oregon estimated that the employer mandate
accounted for about 16 percent of the funds necessary to finance
coverage expansion under the waiver.
HAWAII: FIRST-YEAR ENROLLMENT
AND COST EXPERIENCE
---------------------------------------------------------- Chapter 4:2
Hawaii believed that state-funded programs subsumed under the waiver
had already identified most of those newly eligible--the so-called
gap group that included those who were not eligible for Medicaid and
those who were either dependents or part-time workers not covered by
the state's limited employer mandate.\62
During the first year of operation, however, Hawaii enrolled about
36,000 newly eligible individuals who had not participated in the
former state-funded programs. According to state officials, a
significant number of these new recipients are hypothetically
eligible at state option under section 1902(r)(2) --pregnant women or
children--who had not enrolled in the previous state-funded program
but instead were covered by private insurance; the state believes
these individuals dropped private insurance in favor of less
expensive coverage through the waiver. These officials also
attributed the unexpected high enrollment to Hawaii's current
economic slowdown. Table 4.3 compares estimated and actual
enrollment under the waiver for both traditional and new eligibles.
Hawaii officials project that as a result of this
higher-than-expected enrollment, the waiver will exceed the federal
budget limit for 1994-95 by approximately 23 percent--$47 million.
Under the waiver agreement, higher costs in one year can be offset by
lower costs in another--as long as expenditures over the 5-year life
of the waiver do not exceed the cap. Hawaii officials told us that
the state expects waiver costs to be slightly under the cap for the
full 5 years of the program as a result of state efforts to reduce
program expenditures. The following changes in eligibility standards
and premiums were effective on August 1, 1995: (1) the point at
which enrollees will be charged the full premium will be reduced from
296 percent of the federal poverty level to 201 percent, (2)
individuals eligible for coverage under the employer mandate but who
meet demonstration income requirements will be disenrolled, (3)
self-employed individuals will be required to pay a minimum of 50
percent of the premium--regardless of their stated income, and (4)
parental income will be taken into consideration when determining the
eligibility of students under age 21. The state suspects that these
last two groups either understate income or do not appropriately
account for parental income. Officials in Hawaii told us that
premium collections are keeping pace with expectations.
Table 4.3
Comparison of 1994 Actual and Estimated
Enrollment Under Hawaii's 1115 Waiver
Estima Differ
Actual ted ence
---------------------------------------------- ------ ------ ------
Traditional Medicaid 71,899 61,000 10,899
New eligibles\a 63,490 27,100 36,390
======================================================================
Total 135,38 88,100 47,289
9
----------------------------------------------------------------------
\a Actual enrollment for new eligibles includes 22,675 individuals
hypothetically eligible at state option under section 1902(r)(2) who
were not previously covered by Medicaid and who will be counted as
traditional eligibles for purposes of budget neutrality.
--------------------
\62 Hawaii is the only state to require employers to provide health
insurance to their workers. Its expansion of health insurance
coverage through the 1974 Prepaid Health Care Act was built on a
tradition of employer-based health benefits. Under the act,
employers pay most of the health insurance premiums. Employees must
elect the insurance unless they have comparable coverage from another
source. Employers who provide an extensive benefits package that
meets standards described in the law are not required to cover
dependents. Employers offering a more limited, state-approved,
benefits package must then pay at least half the cost of dependent
coverage. See Health Care in Hawaii: Implications for National
Reform (GAO/HEHS-94-68, Feb. 11, 1994).
TENNESSEE: FIRST-YEAR
ENROLLMENT AND COST EXPERIENCE
---------------------------------------------------------- Chapter 4:3
During its first year of waiver implementation, Tennessee enrolled
about 418,000 previously uninsured or uninsurable individuals.
Although its waiver application proposed an open enrollment period
once a year, the state actually accepted and processed enrollment
requests throughout 1994. Moreover, Tennessee liberalized a
restriction that had disqualified participation by individuals with
access to insurance as of March 1993 by moving the effective date to
July 1994. In late December 1994, however, the state unexpectedly
announced an end to open enrollment for individuals not traditionally
eligible for Medicaid.\63 Enrollment--including both traditional
Medicaid and new eligibles was about 39,000 less than the state's
1994 enrollment cap of 1.3 million. Tennessee also informed HCFA
that the enrollment cap for the remainder of the demonstration would
be 1.3 million rather than 1.5 million beneficiaries. State
officials attributed the freeze in enrollment of new eligibles to a
budget crisis caused, in part, by demonstration costs.
HCFA reports indicate that the state spent about $443 million (14
percent) less than allowed under the waiver agreement cap.\64
Nonetheless, Tennessee covered several hundred thousand newly
eligible individuals while increasing expenditures by less than half
a percent from SFY 1993 to SFY 1994. Table 4.4 compares SFY 1993
enrollment and expenditures with those in SFY 1994, the first year in
which the waiver became effective. Moreover, as shown in table 4.5,
the state is now projecting lower waiver expenditures that could
increase federal savings over earlier estimates. In the first 3
years alone, lower expenditures could more than double the state's
previous estimate of savings due to the waiver.
Table 4.4
Comparison of Enrollment/Expenditures
Under Tennessee's Prewaiver Medicaid
Program and the 1115 Demonstration
(Dollars in billions)
Enrollment Expenditures
------------------------------ ------------ ------------
Prewaiver Medicaid Program--SFY 1993
----------------------------------------------------------
Traditional eligibles 777,431
New eligibles 0
==========================================================
Total\a 777,431 $2.702
1115 waiver--SFY 1994
----------------------------------------------------------
Traditional eligibles 758,192
New eligibles 361,264
==========================================================
Total\b 1,119,456 $2.703
----------------------------------------------------------
\a As of June 1993.
\b Enrollment as of August 10, 1994.
Source: HCFA and state reports.
Table 4.5
Comparison of Tennessee's Waiver
Agreement Spending Limits, First-Year
Costs, and Revised Expenditure
Projections
(Dollars in billions)
1994
(actual) 1995 1996 1997 1998
---------------------------- -------- ------ ------ ------ ------
Agreement $3.146 $3.407 $3.663 $3.872 $4.069
Expenditures 2.703\a 2.982 3.135
Difference 0.443 0.425 0.528
----------------------------------------------------------------------
Note: The state did not issue revised estimates for SFYs 1997 or
1998.
\a Data as of June 30, 1995.
Source: HCFA and state reports.
Both the (1) gap between 1994 waiver expenditures and the federal cap
on spending and (2) projected reduction in waiver expenditures to
well below the amount permitted under the state's 1115 financing
agreement may be linked to problems in identifying state matching
funds for DSH. HCFA officials told us that, from the outset, they
anticipated Tennessee would have difficulty in drawing down federal
funds up to the maximum allowed under the waiver agreement.\65 They
pointed out that although the Tennessee DSH program was available to
provide a major source of funding for coverage expansion, the state
discontinued its hospital tax with the onset of waiver
implementation. This tax had been a source of state match for
federal DSH funds. Undoubtedly, the shortfall in premiums collected
from newly eligible enrollees and counted as part of state matching
funds also contributed to Tennessee's financing problems.
--------------------
\63 Tennessee officials indicated that applications received but not
approved as of the date of the announcement would be processed.
\64 In Tennessee, the first year's cap on federal expenditures
applied to the period July 1993 through June 1994 even though the
demonstration did not begin until midway through that time period
(Jan. 1994).
\65 In retrospect, a line item in Tennessee's waiver
budget--"additional state funds required"--may have been a clue to
the existence of a state funding shortfall.
TENNESSEE ENCOUNTERS
PROBLEMS IN COLLECTING
PREMIUMS FROM ENROLLEES
-------------------------------------------------------- Chapter 4:3.1
Tennessee has encountered serious problems in collecting enrollee
premiums. Initially, the state estimated that it would collect about
$21 million in premiums during the first 6 months of the waiver as
new eligibles gradually signed up for the program; premiums would
increase up to $117 million in the last year when full enrollment had
been achieved. However, for the first 6 months, Tennessee only
collected $2.4 million, forcing it to find other sources of state
matching funds.
Lower-than-expected premium revenues in Tennessee are due, in part,
to a series of administrative glitches. Even though enrollment of
the uninsured began in January 1994, initial premium notices were not
mailed until June 1994. The notice informed enrollees that premium
booklets would be mailed soon for monthly payments beginning with
July. Then, the state contractor failed to mail up to 80,000 premium
booklets, an error that was not discovered until November 1994. In
February 1995, the state sent letters to nearly 60,000 households
notifying them of past due premiums totaling $31 million.
Approximately 62,000 individuals--about 15 percent of new
eligibles--had been disenrolled from the program as of June 1995 and
upwards of 20,000 more were within the 30-day notification period for
termination. Another 17,000 families were placed on payment plans to
address overdue premiums.
IMPLEMENTING FIXED-COST
AGREEMENTS IN A VARIABLE-COST
ENVIRONMENT
---------------------------------------------------------- Chapter 4:4
Though the three waivers discussed in this chapter were approved
during a period of economic recovery and a slowdown in medical
inflation, the recession of the early 1990s coupled with rapid
medical price increases serve as a reminder of the risks posed by
fixed-cost agreements in variable-cost environments. While states
have benefited from recent economic trends, the potential for a
resurgence in medical inflation, a recession, and large numbers of
traditional and/or new eligibles could create problems for Tennessee,
Oregon, and Hawaii.
In Tennessee, the 1115 agreement provides the state with a fixed
budget to serve both traditional Medicaid and newly eligible
recipients. Though the state appears to have a tight lid on cost
increases, it is already under pressure to raise capitation rates
that most providers consider unrealistically low. A slowdown in
economic growth and the associated increase in Medicaid enrollment
due to rising unemployment could further exacerbate the state's
current budget crisis and provide additional ammunition to already
aggrieved providers.
Increased medical inflation and a recession could pose a somewhat
different dilemma for Oregon and Hawaii. Under the terms of their
waiver expenditure caps, these two states are not at risk for
changing economic conditions that could increase the number of
traditional Medicaid beneficiaries. Thus, the limit on demonstration
costs floats upward with enrollment, permitting increased federal and
state Medicaid expenditures. If increased state costs associated
with covering more traditional beneficiaries is accompanied by an
acceleration in medical price increases, however, the additional
budget resources required could threaten Oregon's and Hawaii's
coverage expansion plans.
CONCLUSIONS AND AGENCY COMMENTS
============================================================ Chapter 5
Comprehensive 1115 Medicaid demonstrations have given states
flexibility to test innovative approaches for the delivery of
publicly funded health care services. While the waivers were
intended to give states program flexibility, it is not clear whether
the administration's decision to simultaneously provide budgetary
flexibility is consistent with the current emphasis on reducing the
federal budget deficit. Under the four approved waivers we analyzed,
the federal government is potentially at risk for a net increase of
about $2 billion in Medicaid expenditures. While Tennessee's waiver
agreement meets the test of budget neutrality, those of Florida,
Oregon, and Hawaii do not. The agreements in these three states
represent the antithesis of the budgetary certainty that the Congress
appears to be moving toward in social program spending.
We believe the granting of additional section 1115 waivers merits
close scrutiny for several reasons. First, the potential budget
impact of 1115 waivers may increase if the administration continues
to show budgetary flexibility in its review of additional state
proposals. The administration has granted a number of additional
waivers since Florida's, the most recently approved waiver whose
budget neutrality agreement we examined in detail. Moreover, the
number of pending waivers continues to grow and now includes New
York, whose Medicaid expenditures represented about 16 percent of
national program costs in fiscal year 1993. Second, given the
priority attached to reducing the deficit, it may be appropriate to
consider whether or at what point taxpayers should benefit from
managed care savings that are currently being reinvested to expand
Medicaid coverage to millions of additional individuals.
Finally, though comprehensive 1115 Medicaid waivers were approved
during a period of economic recovery and a slowdown in medical
inflation, the recession of the early 1990s coupled with rapid
medical price increases serve as a reminder of the risks posed by
fixed agreements in variable-cost environments. The combination of
higher medical inflation, a recession, and large numbers of
traditional and newly eligible Medicaid enrollees could pose equally
unattractive alternatives for both the federal government and states:
(1) increasing funding or (2) reducing benefits/denying coverage to
hundreds of thousands of people newly enrolled under the waivers.
Consequently, we question whether demonstration waivers granted for a
limited period are the best approach to reducing states' uninsured
populations.
AGENCY COMMENTS
---------------------------------------------------------- Chapter 5:1
The Department of Health and Human Services and OMB disagreed with
our conclusion that the waiver funding caps for Oregon, Hawaii, and
Florida are not budget neutral. We continue to believe that the
administration's waiver funding caps for these states may result in
increased federal spending.
We do not believe that our methodology is the only appropriate method
to estimate budget neutrality baselines, and agree that using a
tailored, state-specific approach would be more appropriate. We did
tailor our methodology to the specific services covered by the
demonstrations and did reflect current DSH rules. We saw no
evidence, however, that the administration itself adopted this
approach. The only state-specific data evident in the negotiating
record were historical trends, which were clearly not expected to
continue. The acute and long-term care cost projections cited in the
administration's comments are not consistent with OMB's published
forecast of overall growth in the Medicaid program.
While OMB characterized its own approach to budget neutrality as "ad
hoc," we adopted a consistent and uniform methodology that challenges
the administration to support its contention that these
demonstrations should grow at such high rates. To date, the
administration's own methodology remains shrouded in generalities.
We believe that potential program cost increases of hundreds of
millions of dollars should be based on a more clearly specified
methodology.
Second, contrary to the administration's assertion, state variation
in Medicaid programs and expenditures was a central component of our
assessment of budget neutrality. After using OMB's national
forecasts to project without-waiver expenditure trends, we examined
the waiver negotiating record and asked state officials to identify
why future state Medicaid expenditures should exceed the national
norm. As noted, we found no state-specific evidence to support the
high budget caps agreed to by the administration. Even in the case
of Florida, whose Medicaid program had been growing faster than the
national average, the state's own estimates show that key factors
contributing to past growth were not expected to be sustained.
OMB maintains that "it is more appropriate to use a current law
rather than a current services baseline for adjudicating budget
neutrality." Yet, it points out that "the President's budget does not
differentiate between the two." We do not question OMB's authority to
estimate the baseline, including anticipated behavioral changes in
mandatory programs where such changes are allowable under current
law. We do question whether in this case such an adjustment is
appropriate, based on our review of state practices in these
programs. The only explicit use of current law evident in the waiver
approval process is OMB's decision to include those hypothetically
eligible for Medicaid under current law within their baseline. Of
the four states we reviewed, OMB's approach only affected the
baseline for Hawaii, as is reflected in figure 3.6. Since no attempt
was made in Hawaii's or in other states' waivers to suggest that they
would have expanded Medicaid eligibility to hypothetical groups if
their 1115 demonstrations had not been approved, we chose not to
include hypotheticals in the baseline.
Finally, the administration questioned the basis for our estimate of
expenditures under the waiver funding agreements for Oregon and
Hawaii, states with per capita funding limits. We asked and were
told by administration officials that no estimates had been made of
potential expenditures under those caps. Consequently, we worked
closely with state officials to develop such estimates. State
Medicaid officials reviewed and agreed with our methodology,
described in detail in appendix I.
OBJECTIVES, SCOPE, AND METHODOLOGY
=========================================================== Appendix I
To examine the financing arrangements for approved 1115 Medicaid
demonstration waivers in several states, we focused on (1) the
relationship between the waiver and other state health reform
initiatives, (2) the planned sources of funding available to finance
expanded coverage, (3) the potential net impact of these waivers on
federal Medicaid expenditures, and (4) the actual waiver expenditures
of states with sufficient implementation experience.
Although our study concentrated on four states, we closely monitored
other pending waivers, which we use as examples throughout this
report. Two criteria guided our sample selection: (1) whether the
state was engaged in other major health reform initiatives and (2)
whether it had begun implementing its approved waiver. At the time
we selected our sample, only three states--Tennessee, Oregon, and
Hawaii--had commenced implementation. Though its waiver had not yet
been ratified by the state legislature, we included Florida in our
sample since it was the first state with a large Medicaid program to
gain federal approval of an 1115 waiver.
During our review we (1) analyzed data contained in HCFA expenditure
reports, state 1115 waiver applications, correspondence between HCFA
and state officials concerning the demonstrations, and reports
prepared by state agencies and other interested parties, such as
advocacy groups; (2) interviewed state legislators and officials
responsible for the state Medicaid program and for other health
reform initiatives; (3) discussed the waivers with affected parties,
such as health plans, providers, and advocacy groups; and (4)
interviewed HCFA, HHS, and OMB officials responsible for reviewing
and approving waiver applications. We also reviewed the literature
on state health care reform and Medicaid managed care.
To determine the potential impact of Medicaid 1115 waivers on federal
Medicaid spending, we developed and applied a consistent framework
that compares spending limits approved in each waiver with a
benchmark based on current services budgeting concepts. The
framework consisted of (1) determining base-year costs, for the
existing fee-for-service Medicaid system in that state and inflating
these costs at national projected current services rates to give us
our without-waiver spending estimate; (2) determining the total
waiver program cost limits, based on the waiver agreements; and (3)
comparing the two costs.
COMMON ASPECTS OF OUR
MULTISTATE ANALYSIS
--------------------------------------------------------- Appendix I:1
The four waiver agreement cost limits we analyzed were different in
several respects. Nevertheless, we applied some common rules and
assumptions.
We used the OMB national current services growth rate in effect at
the time a waiver was approved to project without-waiver
spending. For example, HCFA approved the Oregon waiver in March
1993, so we analyzed the agreement using the fiscal year 1994
projection OMB made in February 1993.
We applied projections made for federal fiscal years to associated
state fiscal years or calendar years, depending on the starting
point of the programs.
In cases where the total waiver agreement cost depended on program
enrollment--the states that used a per capita cost limit--we
used the states' enrollment estimates included either in the
waiver application or in answers to subsequent HCFA questions
prior to approval.
The OMB projections of Medicaid current services outlays cover the
entire Medicaid program, not just the populations and services
associated with the waiver programs. By applying these growth
rates to the waiver programs, we are assuming that the rates of
growth for the waiver programs and the entire Medicaid program
as a whole would not be significantly different.
TENNESSEE
------------------------------------------------------- Appendix I:1.1
The Tennessee waiver agreement has an aggregate cap on total program
cost. The agreement documents specified the base-year cost and the
rate at which this cost would be allowed to grow under the waiver,
giving us the total potential program cost.
Because Tennessee has a high Disproportionate Share Hospital Program
(DSH) allotment and the base-year cost included DSH, we did not
simply adjust the base-year cost for inflation using the OMB current
services rates to develop our without-waiver spending projection.\66
To be consistent with statutory DSH limits, we held Tennessee's DSH
payments constant until they would have constituted less than 12
percent of total Medicaid expenditures. At that point, we allowed
DSH payments to grow at the same rate as the rest of the Medicaid
program. We adjusted the remainder of the program for inflation at
the rates specified in OMB's projection for fiscal year 1994 current
services outlays.
The net effect of holding DSH constant was that our without-waiver
projection grew more slowly than it would have if we had allowed the
whole program to grow at the OMB rate. Table I.1 shows the waiver
agreement funding limit, our without-waiver estimate, and the
difference.
Table I.1
Comparison of Tennessee Waiver Agreement
Spending Cap and Projected Without-
Waiver Spending--Year by Year
(Dollars in billions)
1994 1995 1996 1997 1998
------------------------------ ------ ------ ------ ------ ------
Waiver spending cap $3.146 $3.407 $3.663 $3.872 $4.069
Percent growth 16.4 8.3 7.5 5.7 5.1
Without-waiver spending $3.032 $3.433 $3.853 $4.325 $4.808
Percent growth 14.5 15.4 13.0 12.2 11.2
Difference (savings) $0.114 ($0.02 ($0.19 ($0.45 ($0.74
5) 1) 3) 0)
----------------------------------------------------------------------
--------------------
\66 DSH is limited by statute to 12 percent of total Medicaid program
funding. States with DSH allotments higher than 12 percent were
capped at their existing DSH levels and prohibited from any increase
in DSH until they fell below 12 percent.
FLORIDA
------------------------------------------------------- Appendix I:1.2
Like Tennessee, the Florida waiver agreement has an aggregate cap on
program costs.\67 However, this waiver only includes the acute care
and DSH segments of the Medicaid program. The waiver agreement
documents specified the base-year costs and the rate at which the
program would be allowed to grow, giving us the total potential
program cost.
To estimate without-waiver spending, we adjusted for inflation the
waiver agreement base-year cost at the rate OMB projected Medicaid
current services outlays would grow beginning in fiscal year 1995.\68
Because Florida has a low DSH allotment, we allowed DSH to grow at
the same rate as the rest of the program. Table I.2 shows the waiver
agreement, our without-waiver projection, and the difference.
Table I.2
Comparison of Florida Waiver Spending
Agreement Cap and Projected Without-
Waiver Spending--Year by Year
1995 1996 1997 1998 1999 2000
---------------------- ------ ------ ------ ------ ------ ------
Waiver spending cap $4.950 $5.723 $6.605 $7.582 $8.658 $9.887
2
Percent growth 16.3 15.6 15.4 14.8 14.2 14.2
Without-waiver $4.706 $5.282 $5.931 $6.653 $7.430 $8.297
spending
Percent growth 10.6 12.2 12.3 12.2 11.7 11.7
Difference- $0.061 $0.442 $0.675 $0.929 $1.228 $1.193
excess funds
----------------------------------------------------------------------
The Florida waiver program was scheduled to begin in the fourth
quarter of 1995 and end after the third quarter of 2000. To
calculate our without-waiver projection and the waiver agreement
spending cap, we assumed the program would run for a full year in
1995 and 2000, but only counted 25 percent of 1995 and 75 percent of
2000 for our calculation of additional funds. We did this to avoid
skewing the increases in the first and last years of the program.
--------------------
\67 The waiver agreement includes a circuit-breaker provision that
allows Florida to eclipse the waiver agreement cost cap in the event
that the state's traditional Medicaid population exceeds projections
by 3 percent or more.
\68 OMB's fiscal year 1995 projection did not include an estimate of
current services growth for the year 2000. Because there was little
variation in the projection from year to year, we used the 1999
figure for 2000.
OREGON
------------------------------------------------------- Appendix I:1.3
The Oregon waiver agreement includes per capita cost limits, rather
than an aggregate program cost cap. The agreement specifies per
capita cost limits for individual eligibility groups, with a limit on
the number of new eligibles.\69
Consequently, our analysis represents the waiver agreement limit
based on a fixed set of enrollment assumptions, including both the
number and distribution of health plan enrollees. We calculated the
program cost using the agreed-upon per capita costs for each eligible
population and the enrollment estimates provided by the state.\70
The waiver agreement documents for Oregon did not include a base-year
cost that we could adjust for inflation at the current services rates
to calculate without-waiver spending. Consequently, we constructed a
base-year cost that approximated the cost of providing traditional
Medicaid services to those individuals who would be eligible for the
Oregon waiver program using data supplied by the Oregon Office of
Medical Assistance Programs.\71 Then, we adjusted this base-year cost
for inflation using OMB's fiscal year 1994 current services
projections to arrive at our without-waiver projection for Oregon.
Table I.3 shows the results of these calculations.
Table I.3
Comparison of Oregon Waiver Spending
Agreement Cap and Projected Without-
Waiver Spending--Year by Year
(Dollars in billions)
1994 1995 1996 1997 1998
------------------------------ ------ ------ ------ ------ ------
Waiver spending cap $0.348 $0.476 $0.574 $0.610 $0.680
Percent growth 11.2 36.9 20.6 6.3 11.5
Without-waiver spending $0.328 $0.413 $0.467 $0.524 $0.582
Percent growth 14.5 15.4 13.0 12.2 11.2
Difference- $0.019 $0.063 $0.107 $0.086 $0.098
excess funds
----------------------------------------------------------------------
The Oregon waiver program was in operation for 11 months in 1994. To
account for this, we multiplied the waiver agreement and our
without-waiver projection by eleven-twelfths. As there was only a
1-month difference, we did not extend our analysis into a sixth year
(in this case, 1999) as we did with Florida and Hawaii.
--------------------
\69 The waiver agreement stipulates that the federal government will
match state funds spent to provide coverage for newly eligible
enrollees. However, the federal government will only provide
matching funds for a limited number of new eligibles--based on the
fixed ratio of current-to-new eligibles specified in the waiver
agreement.
\70 We could not calculate the maximum waiver agreement limit with
any other estimate of enrollment because, to date, HCFA and Oregon
have not agreed on a specific methodology for determining the maximum
allowable number of newly eligible enrollees in each category that
will qualify for federal matching funds.
\71 This exercise generally gave us the cost of providing Medicaid
services to the waiver population on a fee-for-service basis. A
segment of the Oregon Medicaid population was already involved in a
Medicaid managed care experiment; we assumed that this experiment
would continue unchanged and grow at the same rate as the rest of the
program.
HAWAII
------------------------------------------------------- Appendix I:1.4
Like the Oregon agreement, the Hawaii waiver agreement is per capita
based, where the limit on total program spending is
flexible--depending on the number of traditional Medicaid eligibles
enrolled--but the cost per person in each eligibility group is fixed.
However, with Hawaii, HCFA cost sharing is based only on the number
of traditional eligibles--including hypotheticals; newly eligible
individuals are not counted in calculating the federal match.
In order to make our comparison, we used a preliminary estimate of
the base-year per capita cost and constructed a base-year enrollment
figure using actual enrollment figures for fiscal year 1992 and a
state estimate for fiscal year 1994. To calculate the total waiver
agreement cost, we adjusted the product of the base-year per capita
cost and the state's projected enrollment by the waiver agreement
inflation rate. The waiver agreement inflation rate is the Consumer
Price Index (CPI) for health care in Honolulu plus 4 percentage
points. However, because the Bureau of Labor Statistics does not
make CPI forecasts, we used an OMB nationwide projection as a proxy
for health care inflation in Honolulu.\72
To project without-waiver spending, we multiplied the base-year per
capita cost figure by the base-year enrollment estimate and adjusted
the total for inflation at the rate specified by OMB's fiscal year
1994 current services projections. Table I.4 shows the results of
these calculations.
Table I.4
Comparison of Hawaii Waiver Agreement
and Projected Without-Waiver Spending--
Year by Year
(Dollars in billions)
1994 1995 1996 1997 1998 1999
---------------------- ------ ------ ------ ------ ------ ------
Waiver spending cap $0.188 $0.216 $0.245 $0.280 $0.321 $0.367
Percent growth 30.5 15.2 13.3 14.4 14.4 14.4
Without-waiver $0.165 $0.190 $0.215 $0.241 $0.268 $0.298
spending
Percent growth 14.5 15.4 13.0 12.2 11.2 11.2
Difference-excess $0.012 $0.026 $0.027 $0.032 $0.040 $0.025
funds
----------------------------------------------------------------------
Hawaii's waiver program started in mid-1994, and the demonstration is
scheduled to be completed in mid-1999. In creating table I.4, we
annualized the 1994 and 1999 figures and, as we did with Florida, we
multiplied the amount of additional funds by the fraction of the year
the program was (or will be) in operation.
--------------------
\72 In a previous report, we found that per capita health care
expenditures in Hawaii, while lower than those in the nation as a
whole, were growing at roughly the same rate as the rest of the
nation. See Health Care in Hawaii: Implications for National Reform
(GAO/HEHS-94-68, Feb. 11, 1994).
OVERVIEW OF MEDICAID
========================================================== Appendix II
Financed jointly by the federal government and states, Medicaid is
the nation's health care lifeline for two statutorily defined groups
of low-income residents--families, primarily women and children; and
the aged, blind, and disabled. The federal government matches state
expenditures according to a prescribed formula, providing, on
average, 58 cents of every dollar spent. In 1993, Medicaid
expenditures for the 39 million beneficiaries enrolled in the 50
states and the District of Columbia totaled $130 billion, up
dramatically from just a decade ago.
In reality, Medicaid is not one, but 56 separate programs that differ
dramatically across states.\73 While federal statute mandates who is
eligible for coverage and the broad categories of services that must
be provided, each participating state designs and administers its own
program by (1) setting certain income and asset eligibility
requirements; (2) selecting which optional groups and services to
cover; and (3) determining the scope of mandatory and optional
services, for example, by limiting the number of covered hospital
days per year. As a result of this flexibility, Medicaid is not
available to everyone who is poor. In 1993, Medicaid provided health
care coverage to less than half of those with incomes below the
poverty level. HCFA, within HHS, monitors each state program for
compliance with federal regulations.
Medicaid costs have escalated sharply--tripling between 1985 and
1993, while the number of beneficiaries increased by over 50 percent.
Expenditure growth outpaced changes in the Consumer Price Index as
well as national health and Medicare spending. Currently, Medicaid
accounts for about 6 percent of all federal outlays and 19 percent of
state spending. Medicaid nearly equals state expenditures for
elementary and secondary education combined, generally the largest
segment of state budgets.
A number of factors contributed to this rapid cost growth. Between
1984 and 1990, the Congress mandated coverage for certain low-income
groups--primarily pregnant women, children, and Medicare
beneficiaries--and allowed coverage of others at state option.\74
Medical price inflation, higher provider reimbursements, utilization
growth, and an increase in the number of eligibles due to the
national recession also played a role. The relative importance of
each factor depends on the time period selected. For example, in
1993 the Kaiser Commission reported that enrollment, inflation, and
increased use of services each accounted for about one-third of the
increase in expenditures from 1988 through 1991.\75 The Urban
Institute has attributed part of the cost growth to increased
enrollment resulting from the sharp downturn in the economy during
that time period.\76
Figure II.1: Growth in
Medicaid Expenditures,
1985-1993
(See figure in printed
edition.)
As shown in figure II.1, growth in Medicaid expenditures soared in
1991 and 1992. The most important cost driver during this period was
creative financing techniques typically adopted by some states to
increase DSH payments. DSH provides supplemental payments to
hospitals serving a large number of Medicaid and other low-income
patients, thereby partially offsetting costs not covered by either
Medicaid, state charity care programs, or private insurance. In 2
years, DSH grew from just under $1 billion to over $17 billion and
represented about $1 out of every $7 Medicaid spent on medical
services. This rapid growth can be traced to several legal but
nonetheless questionable practices, including the use of rebated
provider taxes to gain federal match under DSH.\77 In some cases, a
portion of the federal matching funds was redirected to state general
revenues and spent on programs other than Medicaid. This swapping
and redirecting of revenues contributed greatly to Medicaid cost
escalation. In response to these practices, the Congress placed
restrictions on DSH in 1991 and 1993, effectively capping the program
and tying increases to the overall growth in state Medicaid programs.
--------------------
\73 All 50 states plus the District of Columbia, American Samoa,
Guam, the Northern Mariana Islands, Puerto Rico, and the Virgin
Islands have Medicaid programs.
\74 See Medicaid: Spending Pressures Drive States Toward Program
Reinvention (GAO/HEHS-95-122, Apr. 4, 1995), p. 27.
\75 The Medicaid Cost Explosion: Causes and Consequences (Baltimore,
Md.: The Kaiser Commission on the Future of Medicaid, Feb. 1993).
\76 Theresa A. Coughlin and others, "State Responses to the Medicaid
Spending Crisis: 1988 to 1992," Journal of Health Politics, Policy,
and Law, Vol. 19, No. 4 (Winter 1994).
\77 For more detail on these state financing practices, see Medicaid:
States Use Illusory Approaches to Shift Program Costs to the Federal
Government (GAO/HEHS-94-133, Aug. 1, 1994).
SUMMARY OF APPROVED AND PENDING
1115 MEDICAID DEMONSTRATION
WAIVERS SUBMITTED SINCE 1992
========================================================= Appendix III
Submission Approval Implementation/status
------------------------- -------------- -------------- ------------------------------
Approved demonstrations
-----------------------------------------------------------------------------------------
Oregon Nov. 1992 Mar. 19, 1993 Feb. 1994
Hawaii Apr. 1993 July 16, 1993 Aug. 1994
Kentucky May 1993 Dec. 9, 1993 Suspended--not approved by
state legislature; new waiver
proposal submitted June 22,
1995
Tennessee June 1993 Nov. 18, 1993 Jan. 1994
Rhode Island July 1993 Nov. 1, 1993 Aug. 1994
Florida Feb. 1994 Sept. 15, 1994 Not approved by state
legislature in session ending
May 1995
Ohio Mar. 1994 Jan. 17, 1995 State has decided not to
implement
Massachusetts Apr. 1994 Apr. 24, 1995 Awaiting state legislative
approval; expected
implementation in Jan. 1996
Minnesota July 1994 Apr. 27, 1995 July 1, 1995
Delaware July 1994 May 17, 1995 Approved by state legislature;
scheduled to begin Jan. 1996
Vermont Feb. 1995 July 31, 1995 Jan. 1, 1996
Provisionally approved
-----------------------------------------------------------------------------------------
South Carolina Mar. 1994 Nov. 18, 1994 In Apr. 1995, state abandoned
waiver and shifted to a
voluntary managed care
strategy
Pending
-----------------------------------------------------------------------------------------
New Hampshire June 1994 On hold, new proposal expected
Missouri June 1994 Amendment submitted Mar. 24,
1995
Illinois Sept. 1994 HCFA reviewing financing
issues
Louisiana Jan. 1995 Finance plan rejected
Oklahoma Jan. 1995 HCFA reviewing proposal
New York Mar. 1995 HCFA reviewing proposal
Kansas Mar. 1995 HCFA reviewing proposal
Kentucky June 1995 HCFA reviewing proposal
Utah July 1995 HCFA reviewing proposal
Alabama July 1995 HCFA reviewing proposal
Texas Sept. 1995 HCFA reviewing proposal
-----------------------------------------------------------------------------------------
Note: Information in table is as of October 6, 1995.
DIFFERING STATE HEALTH REFORM
CONTEXTS
========================================================== Appendix IV
Tennesse Minnesot
e Florida Kentucky Oregon Hawaii a U.S. total
-------- -------- -------- -------- -------- -------- -------- ----------
Access to insurance\a
--------------------------------------------------------------------------------
Uninsure 15.6% 24.1% 14.7% 17.2% 13.7%\b 12.7% 18.1%
d
(percent
of
nonelder
ly),
1993\c
Insured 58.7% 52.8% 58.7% 63.9% 68.8% 66.7% 60.8%
by
employer
(percent
of
nonelder
ly),
1993\c
Medicaid 16.0% 13.3% 18.4% 9.3% 9.4% 11.3% 12.8%
coverage
(percent
of
nonelder
ly),
1993\c
Individu 19.6% 17.8% 20.4% 11.8% 8.0% 11.6% 15.1%
als in
poverty
(percent
of total
populati
on),
1993
State health care reform agenda
--------------------------------------------------------------------------------
Universa No Yes No Yes Yes Yes
l
coverage
goal
Universa Not Voluntar Not Employer Employer Individu
l applicab y--to be applicab mandate mandate al
coverage le. reconsid le. to begin enacted mandate
financin ered if in 1997 in had been
g no if 1974.\d consider
mechanis evidence voluntar ed; no
m of y consensu
signific particip s on an
ant ation alternat
decline falls ive has
in short of emerged.
uninsure goals.
d.
Legislat No Agency Health Health No Health
ively for Care Services Care
establis Health Policy Commissi Commissi
hed Care Board on on
state Administ (1994) (1989)\e (1992)
health ration
reform (1992)
policy
organiza
tion
Small Yes Yes Yes Yes No Yes
group
insuranc
e
reforms
Publicly No Yes Yes No No Yes
sponsore
d health
purchasi
ng
cooperat
ive
High- Yes Yes No Yes No Yes
risk
insuranc
e pool
Coverage expansions pursued by state prior to 1115 waiver
--------------------------------------------------------------------------------
State No No No No Yes, Yes,
subsidiz individu families
ed als up with
insuranc to 300% children
e of up to
programs poverty 275% of
for level. poverty
lower level
income and
resident individu
s als up
to 125%
of
poverty
level.
Optional Up to Up to Up to No Up to Expanded
Medicaid 185% of 185% of 185% of 185% of beyond
coverage poverty poverty poverty poverty 185% of
to level. level. level. level. poverty
pregnant level;
women see next
and row.
infants
over
133% of
poverty
level
Optional No No No No No Yes,
Medicaid pregnant
coverage women up
expansio to 275%
n under of
1902(r)( poverty
2) level
and
children
up to
age 1.
Managed care penetration
--------------------------------------------------------------------------------
HMO 5.7% 17.6% 6.6% 31.5% 22.3% 30.1% 17.4%
enrollme
nt
(percent
of
populati
on),
1993
Previous Medicaid managed care experience
--------------------------------------------------------------------------------
Type of Substate Voluntar Statewid Substate Substate Voluntar
program managed y HMO e managed managed y HMO
care enrollme managed care care enrollme
waiver nt since care program waiver nt
under 1981, program under under starting
1915(b). and under 1915(b) 1915(b) in late
Renewal substate 1915(b) implemen implemen 1970s
denied physicia implemen ted in ted in and
in 1992. n ted in 1985. 1983. substate
gatekeep 1986. managed
er care
program program
under under
1915(b) 1115
implemen waiver
ted in implemen
1991.\f ted in
1985.
Managed 2.7% 17.3% 45.8% 22.7% 3.2% 18.1% 12.4%
care
enrollme
nt as
percent
of
Medicaid
populati
on, 1993
Managed HMO and HMO and Physicia HMO and HMO HMO
care physicia physicia n partiall
approach n n gatekeep y
gatekeep gatekeep er capitate
er er d health
plans
Categori Women Primaril Women Women Women Women,
es of and y women and and and children
traditio children and children children children , and
nal children elderly
Medicaid , but a
populati small
on number
enrolled of aged,
blind,
and
disabled
\f
1115 waiver managed care approach
--------------------------------------------------------------------------------
Managed 100% Not yet 1115 not Approxim 84% Approxim
care implemen implemen ately ately
enrollme ted ted, but 80%\g 33%
nt as 1915(b)
percent enrollme
of nt is
Medicaid 57%
populati
on, 1995
Categori Women, Women, Women Women, Women Women,
es of children children and children and children
traditio , aged, , aged, children , aged, children , and
nal blind, blind, blind, aged.
Medicaid and and and
populati disabled disabled disabled
on
enrolled
Managed Capitate Traditio Physicia Capitate Capitate Capitate
care d HMOs nal n d HMO- d HMOs d HMOs
approach and PPOs eligible gatekeep style and PPOs
s: er, plans,
choice phasing some
of into partiall
capitate capitate y
d HMO or d plans. capitate
physicia d plans,
n and some
gatekeep physicia
er. New n
elgibles gatekeep
: choice er
of programs
indemnit .
y and
HMO
plans
offered
by
state-
sponsore
d health
alliance
s.
--------------------------------------------------------------------------------
\a Numbers in bold indicate that amount is greater than the national
average.
\b Unlike most of the country, Hawaii is experiencing an economic
downturn and an associated rise in unemployment and the number of
uninsured. In 1992, Hawaii's percentage of uninsured was reported to
be 8.1 percent--the lowest in the nation.
\c Data are from Employee Benefit Research Institute, Sources of
Health Insurance and Characteristics of the Uninsured, Analysis of
the March 1994 Current Population Survey, Issue Brief Number 158
(Washington, D.C.: Feb. 1995).
\d Hawaii was the first state to attempt universal coverage with its
passage of the Prepaid Health Care Act in 1974, which implemented a
limited employer mandate. Because this act was passed before ERISA,
Hawaii is the only state granted an exemption under ERISA. In
addition, the state established publicly funded programs to insure
individuals not covered by the employer mandate.
\e Prepares prioritized list for legislative consideration, that is,
ranks health services from most to least important.
\f Florida has had authority to implement its 1915(b) waiver program
statewide since 1993, but as of August 1995 the program was still
limited to specific geographic areas of the state. It expects to
enroll all traditional medicaid recipients (with the exception of
those beneficiaries who are also eligible for Medicare) in either the
voluntary HMO program or its physician gatekeeper program by June
1996. A small number of aged, blind, and disabled beneficiaries have
been enrolled in the state's voluntary HMO program since 1981.
However, as of August 1995, the enrollment of this population in the
physician gatekeeper program was limited to two pilot areas.
\g Though Oregon's goal is to transition all recipients into managed
care, 20 percent remain in fee-for-service because there are no
appropriate providers in their area or the most appropriate provider
is not on the network of any participating managed care plan.
(See figure in printed edition.)Appendix V
COMMENTS FROM THE DEPARTMENT OF
HEALTH AND HUMAN SERVICES
========================================================== Appendix IV
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
MAJOR CONTRIBUTORS TO THIS REPORT
========================================================== Appendix VI
Michael Gutowski, Assistant Director, (202) 512-7128
Walter Ochinko, Senior Health Policy Analyst, (202) 512-7157
Cheryl Williams, Senior Health Policy Analyst
Mark Ulanowicz, Health Policy Analyst
GLOSSARY
============================================================ Chapter 1
BUDGET NEUTRALITY
-------------------------------------------------------- Chapter 1:0.1
An assessment of whether the cost of expanding coverage to additional
recipients under an 1115 Medicaid waiver is equal to or less than the
cost of continuing the original Medicaid program serving only those
traditionally eligible.
CURRENT SERVICES
-------------------------------------------------------- Chapter 1:0.2
A cost projection that assumes Medicaid spending will continue
unchanged except for increases resulting from medical inflation,
normal growth in the eligible population, and changes in utilization
of medical services.
TERMS AND CONDITIONS
-------------------------------------------------------- Chapter 1:0.3
The culmination of negotiations between a state and the
administration over an 1115 demonstration application in which the
administration specifies the provisions of the Social Security Act
that are being waived and any special conditions upon which
implementation of the waiver is contingent, including how budget
neutrality will be measured and enforced over the life of the
demonstration.
WAIVER FUNDING AGREEMENT
-------------------------------------------------------- Chapter 1:0.4
The portion of the terms and conditions governing an 1115
demonstration that spells out the limit (cap) on federal matching
funds available during the demonstration.
DISPROPORTIONATE SHARE
HOSPITAL PROGRAM (DSH)
-------------------------------------------------------- Chapter 1:0.5
A program that provides supplemental payments to hospitals serving a
large number of Medicaid and other low-income patients, thereby
offsetting costs not covered by either Medicaid, state charity care
programs, or private insurance. Eliminating or limiting payments
from this program is a major funding source for coverage expansion
under 1115 waivers in states with high DSH allotments.
MANAGED CARE
-------------------------------------------------------- Chapter 1:0.6
An umbrella term encompassing types of health insurance coverage with
some insurer control over the use of services and restrictions on the
choice of physicians and hospitals. In the context of Medicaid, it
includes arrangements in which the insurer is paid a single fee, in
advance, for each beneficiary rather than the state reimbursing
providers after the fact for each service provided.
NEW ELIGIBLES
-------------------------------------------------------- Chapter 1:0.7
Primarily lower income, uninsured individuals not traditionally
eligible for Medicaid who are being provided health care coverage
under 1115 waivers.
TRADITIONAL ELIGIBLES
-------------------------------------------------------- Chapter 1:0.8
The specified categories of persons eligible for Medicaid according
to statute: low-income families, primarily women and children; and
the aged, blind, and disabled, depending on their financial status.
*** End of document. ***