Payment to Medicare+Choice
Organizations
Table of Contents:
The Balanced Budget Act of 19971
(BBA) established the Medicare+Choice program (M+C) that expanded private
coverage options in the Medicare program. It also included a new payment
methodology that replaced the one that had been used to pay M+C precursor
organizations (i.e., Medicare risk HMOs).
Before the BBA was enacted, Medicare's payments to HMOs were based on
95 percent of average county-level fee-for-service costs,2
also known as the adjusted average per capita cost (AAPCC). An HMO was
paid 95 percent of an estimate of what it would have cost Medicare to
cover a beneficiary had the beneficiary remained in the traditional
Medicare program.3
These payments were updated annually based on the estimated county growth
in per beneficiary fee-for-service spending.
The Adjusted Community Rate
The AAPCC was the maximum amount Medicare could pay an HMO. This amount
was compared to an HMO's adjusted community rate (ACR), or the amount it
would cost the plan to provide services to Medicare enrollees while
earning the same rate of return as for a commercial enrollee. If the
plan's average payment from Medicare was greater than its ACR, the HMO
could use the difference in one of the following ways: (1) return all or
part of the difference to the Health Care Financing Administration (HCFA);
(2) apply it to a reduction in beneficiary cost sharing, premiums, or in
providing additional benefits; or (3) apply a portion of the difference to
a benefit stabilization fund.
Under the old payment methodology, any savings achieved by the HMO were
not shared with Medicare. Usually, HMOs elected to use the difference
between the government payment and the ACR to provide additional benefits
or reduce beneficiary cost sharing or premiums. Thus, for example, health
plans typically provided preventive services, physical examinations, and
some outpatient prescription drug coverage.
In 1997, the difference between Medicare HMOs' estimated cost of
providing the Medicare benefit package and the projected Medicare payment
averaged about 13 percent of payments. Therefore, on average, these plans
had to provide additional benefits worth $60 per member per month.4
Flaws in the AAPCC Methodology
The AAPCC methodology was flawed for several reasons.
(1) As county fee-for-service costs varied widely, so too did payments
to HMOs. Therefore, in those parts of the country where fee-for-service
costs were high (hence HMO reimbursement was high), plans were able to
offer richer benefit packages or lower premiums than in areas where
fee-for-service costs were low. Moreover, in certain geographic areas,
particularly rural counties, payments were so low that no HMOs entered the
market.
(2) Rates fluctuated widely from year-to- year, especially in sparsely
populated areas where just a few relatively expensive cases in a year
could drive up Medicare fee-for-service expenditures; conversely, a
"healthy year" could drive expenditures down. Since HMO payment rates were
tied to fee-for-service expenditures, HMO payments from year-to-year were
not stable. This lack of predictability was another factor that
discouraged market entry by HMOs.
(3) Finally, the AAPCC was a poor predictor of HMO costs because the
risk adjusters included in the AAPCC methodology did not adequately
predict a beneficiary's expected health care use.5
Therefore, on average, Medicare was overcompensating HMOs. Estimates of
the overpayment range from 7 percent to 37 percent.6
In the BBA, Congress tried to address the flaws in the AAPCC payment
methodology by amending the way Medicare paid HMOs and other risk-based
M+C plans.8
These changes were intended to attract more private, risk-based plans to
rural areas (which had historically seen little or no HMO enrollment),
promote savings for the Medicare program, and reduce geographic variation
in payment.
In general, under the new methodology, M+C plans are paid a monthly
capitation payment calculated for the payment area (usually a county) in
which a beneficiary lives. As with the AAPCC methodology, the payment is
adjusted for certain beneficiary characteristics (age, sex, welfare,
institutional status, and work status). Beginning in 2000, a health
status risk adjuster is being phased in. (See Risk Adjustment below.)
The 1997 rates are the starting point for determining M+C payment rates.
The payment to M+C plans is the greatest of a: (1) blended capitation
rate consisting of national (adjusted for input prices) and local rates;
(2) minimum (or floor) payment amount; or (3) minimum percentage increase
set at 2 percent over the prior year's rate. The blended rate is subject
to a budget neutrality adjustment, which means that total M+C spending
must equal what HCFA would have paid to M+C plans if only local rates had
been used. The area-specific portion of the blended rate and the minimum
payment amount are updated each year by the national average per capita
Medicare (fee-for-service ) growth rate less a specified statutory
reduction.
In 1998 and 1999, payments to M+C plans were set either at the 2
percent increase or the guaranteed minimum rate. No plan was paid based on
the blended rates because the combination of low national growth rates in
the traditional Medicare program and the 2 percent minimum increase
precluded payment rates based on the blended formula.9
However, in 2000 payments to M+C plans reflect a 5.76 percent increase
in the national growth in fee-for-service per capita spending, a rate of
increase that permits the implementation of the blended rate formula. In
2000, 29 percent of counties qualify for the minimum payment.10
|
M+C Payment Increases in 2000 |
|
Percent of Counties |
Range Percent of Increase |
|
25 |
2-5 |
|
40 |
5.0 -10 |
|
5 |
10 -15 |
| Source: American Association of Health Plans analysis
of HCFA data, October 1999. |
On March 1, 2000, HCFA announced the annual M+C capitation rate for
each M+C payment area for 2001. Since the final estimate of the increase
in the National Per Capita M+C Growth Percentage is -1.28 percent, 69
percent of the county rates will reflect the minimum percentage increase
of 2 percent, and 31 percent will reflect the "floor" amount of $415.01
for aged beneficiaries in 2001. In general, those counties that received
blended rates in 2000 will receive the 2 percent minimum increase.
The BBA changes achieved savings for the Medicare program by modifying
payment to M+C plans in three key areas.
(1) The new payment approach ties M+C annual updates to the national
rate of spending in the traditional Medicare program and breaks the direct
link between county fee-for-service spending and M+C payments. This is
expected to gradually reduce the impact of local costs on payment to M+C
organizations. The blended rates will reduce payments for counties that
have higher fee-for-service costs and increase payments to counties where
fee-for-service costs have been lower than the national average. Between
1998-2002, M+C annual increases are to be set at Medicare's estimate of
the current year's national growth rate for Medicare fee-for-service
spending minus a statutory reduction of 0.8 percentage points in 1998 and
0.5 percentage points annually from 1999 through 2001. In 2002, the
statutory reduction will be 0.3 percentage points.11
(2) Reimbursement for graduate medical education is being gradually
removed from M+C rates over a five-year phase-out period.
In 2000, a health status adjuster will be applied that will be
phased-in to M+C payments gradually.
Risk Adjustment
The BBA required the Medicare program to begin implementing a new risk
adjustment methodology by January 2000 that would adjust M+C capitation
payments based on the health status and demographic characteristics of
each Medicare enrollee in the plan. Risk adjustment is intended to
recognize the health status of a plan's enrollees and more accurately
reflect their expected medical costs in M+C payment rates. The Medicare
Payment Commission (MedPAC) predicts that the new risk adjustment system
will likely reduce the extent to which HCFA overpays M+C plans in the
aggregate.
The risk adjustment approach selected by HCFA, called the Principal
Inpatient Diagnostic Cost Group (PIP- DCG) is prospective. It will look at
diagnoses in the base year to adjust payment in the payment year.
Initially, a beneficiary's inpatient diagnoses (as well as age,
sex, Medicaid eligibility, and disability status12)
in one year will be used to predict health expenditures for that
individual the next year. The predicted expenditures will be converted to
relative factors that will be applied to county rates. Inpatient diagnoses
are being used, initially, because the data are easier to obtain from
plans than comprehensive encounter data for all other types of patient
visits.
To avoid sharp payment changes to M+C plans and changes in plan
benefits and cost sharing that could arise if the new risk adjustment
model were implemented all at once, HCFA decided to phase in the new risk
adjustment approach between 2000 and 2004 by applying a blended approach
consisting of both the demographic system (i.e., no adjustment for health
status) and the PIP-DCG risk adjustment methodology. HCFA intended the
blended approach to provide a transition period during which time plans
could adjust to payment rates that were increasingly adjusted for enrollee
health status.13
In the Balanced Budget Refinement Act of 199914
(BBRA), Congress directed HCFA to modify the phase-in period of risk
adjustment based on health status. Accordingly, in 2000 and 2001 only 10
percent of an M+C organization's payment will be adjusted by enrollee
health status, and no more than 20 percent of the payment will be based on
health status in 2002. In addition, Congress directed MedPAC and HCFA to
conduct several studies concerning HCFA's risk adjustment methodology as
well as the effects, costs, and feasibility of requiring Medicare
fee-for-service providers to comply with quality standards and related
reporting requirements comparable to those for M+C plans. Presumably, HCFA
has the discretion to complete the phase-in of the PIP-DCG risk adjustment
methodology after 2002, unless Congress enacts further legislation on this
subject.
Although there are numerous reasons why many M+C health plans have
recently terminated their contracts with Medicare or reduced their service
areas, one of the key reasons cited is M+C payments.15
The net effect of M+C withdrawals for contract years 1999 and 2000 has
been a reduction in the number of M+C plans and higher out-of-pocket costs
for Medicare beneficiaries.
In 2000, 69 percent of Medicare beneficiaries have access to one or
more M+C organizations. This is a decline from 1999 when 71 percent of
beneficiaries had such access. Beneficiaries living in nine states do not
have access to a M+C plan in 2000.
In 2000, 77 percent of Medicare beneficiaries enrolled in M+C plans are
enrolled in "zero" premium plans,16
a decline from 1999 when 85 percent of beneficiaries had no premiums. The
decline in the number of available zero premium plans was greater in rural
areas where, in 1999, 63 percent of beneficiaries with any M+C plan
available had access to such a plan; the corresponding proportion of
beneficiaries in 2000 is 40 percent.
The enrollment-weighted average premium for M+C plans increased from
$5.35 in 1999 to $15.84 in 2000. Those beneficiaries living in areas with
only one M+C plan are particularly affected by premium increases. Of the
207,000 beneficiaries in areas where the minimum monthly premium is $80 or
more, 94 percent have only one M+C plan available.
Finally, cost-sharing for physician office visits and covered
outpatient prescription drugs are also increasing in 2000. For all plans,
the average copayment for a physician office visit in 2000 is $8.33, up
from $6.90 in 1999. Although the number of Medicare beneficiaries with
access to an M+C plan offering drug coverage has not changed, more
enrollees will have to pay an additional premium for drug coverage, and
for the first time, all plans will charge copayments for outpatient
prescription drugs. In addition, the value of the drug benefit itself
declined in 2000 with most M+C plans offering more limited drug coverage
than in 1999 (e.g., more restrictive annual benefit limits).
It remains to be seen whether the changes in the BBRA that increase
payments to M+C plans will stem the number of plan terminations for 2001
and help to stabilize benefit package and premium charges.
Footnotes 1 Part C, Title
XVIII, Social Security Act
2
Countywide AAPCCs were calculated for Parts A and B for aged beneficiaries
and disabled beneficiaries. State-wide rates were calculated for Parts A
and B for beneficiaries with end-stage renal disease.
3 It was assumed that prepaid plans (e.g.,
HMOs) could operate more efficiently than the Medicare fee-for-service
program, hence the reduction of 5 percent from fee-for-service
costs.
4 U.S. General Accounting
Office. "Medicare+Choice: Reforms have Reduced, but Likely Not Eliminated,
Excess Plan Payments." (GAO/HEHS-99-144). June 1999. The GAO reports that
in 1997, for competitive or other reasons, plans voluntarily added an
additional $33 per member per month in extra benefits.
5 The AAPCC adjusted for age, sex,
institutional, welfare, and work status.
6 Riley, G. Tudor, C, Chiang, Y., Ingber,
M., "Health Status of Medicare Enrollees in HMOs and Fee-For-Service in
1994". Health Care Financing Review. Summer 1996/Volume 17, Number
4; "Risk Selection Remains a Problem in Medicare." PPRC Update.
July 1997. Number 21; Physician Payment Review Commission. "Risk Selection
and Risk Adjustment in Medicare". Annual Report to Congress,
chapter 15, 1996.
7 Section 1853,
Title XVIII
8 Preferred Provider
Organizations, Private Fee-For-Service Plans, HMOs with Point-of-Service
options.
9 This occurred because the
budget neutrality adjustment brought all rates to an amount below the
amount of the minimum 2 percent increase.
10 Source: Public Policy Institute analysis
of HCFA data, 1999.
11 The reduction
specified in the BBA for 2002 was 0.5 percentage points. However, the
Balanced Budget Refinement Act of 1999 reduced the statutory reduction,
which will somewhat mitigate the overall effects of the BBA reductions on
M+C payments.
12 The purpose of
including these demographic independent variables is to take into account
their unique cost implications not related to hospital admissions. (HCFA
Report to Congress, March 1999)
13
During the first year, it would have based only 10 percent of a plan's
payment on the PIP-DCG model and gradually increased the proportion of the
payment based on this model to 100 percent by 2004. In addition, in 2004,
HCFA planned to include (in addition to hospital data), expenditures for
physicians' offices, hospital outpatient departments, skilled nursing
facilities, and home health agencies to the risk adjustment formula.
14 P.L. 106-113
15 Besides payment, other reasons for
leaving the Medicare program are: low M+C enrollment levels; small share
of the M+C market; inability to maintain an adequate provider network; and
other business considerations.
16
This means that apart from the Medicare Part B premium, beneficiaries will
not pay an additional premium.
Written by Joyce Dubow, 
April 2000
©2000, AARP
May be copied only for
noncommercial purposes and with attribution; permission required for all
other purposes.
Public Policy Institute, AARP, 601 E Street, NW,
Washington, DC 20049
publication ID: DD47 publication date: April
2000 |