Private Health Plans in Medicare -- Cost Trends and Ideological BattlesJeff Lemieux
revised May 31, 2004
for a minor editorial correction in the "references and links" section -- originally published April 21, 2004
Since 1997, Medicare's payment rates to private health plans have growth at roughly the same rate as spending in the traditional fee-for-service program. However, payment growth has varied by geographic area. Reimbursements to plans in "minimum update" (urban and high-cost) areas have grown about 11 percentage points slower than nationwide fee-for-service costs. Meanwhile, Congress has raised payments to plans in "floor" (rural and low-cost) areas almost 14 percentage points faster than comparable fee-for-service costs since 1997. (Until recently, most plans operated in minimum update areas.)
Trend analysis doesn't prove whether private plans are more or less costly or efficient than the fee-for-service program. It simply shows that per-enrollee spending growth has been roughly equivalent in both sectors, at least relative to the base year of 1997. There are increasingly strong reasons to believe that private plans in Medicare do not have healthier-than-average enrollees. Therefore, it's wrong to say that Medicare's private plans have "cost the Medicare program money" in recent years, at least relative to their cost in 1997.
Medicare's private plans are caught up in a larger political struggle. Liberals are opposed to "privatization" in Medicare, even though private plans have helped many seniors with low incomes. Liberals do not want to dilute or diminish the importance of the fee-for-service program's price-setting system, which they believe has the potential to be the foundation of a national health insurance program.
By contrast, Medicare reformers and some conservatives believe Medicare's budget problems could be eased via direct competition between private plans and the fee-for-service program. However, conservatives sometimes seem distracted from this goal by their zeal for high-deductible health insurance and health savings accounts.
On balance, a healthy rivalry between private health plans and the fee-for-service program has sparked cost reductions and value improvements in Medicare. It has the potential to improve Medicare's overall value in the future, even if direct competition between the sectors is not established.
Outline:
Private Health Plans and Medicare Reform
Enrollment and Payment Rates Since 1997
Comparing the Growth of Spending in Private Plans and Fee-For-Service
Trends in Relative Cost Ratios
The Larger Battle -- For and Against Private Health Insurance
Liberal Arguments Against Private Plans
Do Private Plans Get Healthier Enrollees?
Conservative Schizophrenia on Private Plans
The Best Hope for Long-Term Value -- A Healthy Rivalry
Appendix: Step-By-Step Guide to the Calculations in this Report
Selected References and Links
Private Health Plans and Medicare Reform: Medicare's alternative HMO program became controversial in the mid- and late-1990s. At that time, private health plans participating in Medicare seemed to have a significant efficiency advantage over the traditional fee-for-service program. That perception, in turn, helped trigger strong political reactions intended to slow the growth of Medicare spending and improve its benefits.
In a sense, this is precisely why Medicare has a private plan option. Private plans allow Medicare to test new ways of providing health insurance and observe their clinical effectiveness, cost savings, and political acceptability. Medicare's private plans can signal a need for change, or the possibility of improvement.
However, the impact of the Congressional response in the late-1990s was not well understood at the time. Congress did not foresee that tight restrictions on HMO reimbursements enacted in 1997 would dramatically diminish private plans' ability to offer Medicare coverage.
Last year, Congress attempted to resuscitate Medicare's private plan offerings by increasing payments back up to their earlier levels. This has sparked a new round of controversy.
This report attempts to put the debate over Medicare's private plans into perspective, using new calculations of cost and spending trends and a characterization of the political forces battling over the prospect of direct public-private competition in Medicare.
One goal of Medicare reform is to find a way to slow the growth in the program's costs without cutting benefits, raising taxes, or reducing access to health care.
Another goal is to ensure Medicare benefits are the highest possible value.
Judging value is harder than assessing costs. Improvements in health status are part of the measure. Health care quality is also important, as is the reduction of medical errors. Finally, the level of beneficiaries' cost sharing and the necessity of supplemental insurance, especially for seniors with low incomes, are key factors in Medicare's overall value.
Ideally, Medicare reform is a win-win outcome. Taxpayers would win if Medicare outlays could be restrained. Senior citizens would win if care could be delivered more effectively or efficiently.
The main vision for reform involves competition between the traditional fee-for-service Medicare program and a variety of private health plans, in either of two ways:
1. Direct competition patterned after the Federal Employees Health Benefits Program (FEHBP). Expensive plans -- either private plans or the fee-for-service program -- would have higher premiums, based on formulas set in advance. Medicare beneficiaries could save money by choosing a less expensive plan, and the government would share the savings. This sort of competition could slow the growth of Medicare spending, especially if the fee-for-service program had sufficient administrative flexibility to adjust its operations without explicit acts of Congress.
2. Indirect competition or rivalry, with savings possible through ad-hoc political adjustments to Medicare payments. Beneficiaries could save money by choosing cheaper health plans, but the government would not share the savings automatically. Likewise, the competition would not directly raise beneficiaries' premiums in any plan. However, the rivalry could highlight the efficiencies of either the private plans or the fee-for-service program. The government could save money over time by adjusting benefits and payment rates within the legislative process.
The first method is better. It would allow direct competitive adjustments without the need for legislative changes, which can be awkward, disruptive, or just wrongly calculated.
Nevertheless, the indirect "political adjustment" method was used successfully in the 1990s, when Congress cut Medicare payments and tweaked its benefits in a reaction to the model of private health plans.
There are two subsidiary visions for Medicare reform. First, the fee-for-service program needs the administrative flexibility to adjust its benefits and payments for local market conditions and beneficiaries' needs. The fee-for-service program should hold health providers more accountable for improved results.
Another imperative is to switch Medicare's payments away from the "acute care" model of health insurance, which is focused on paying for health care "disasters," toward helping patients with long-term or "chronic" health conditions manage their health and avoid sudden flare-ups or health disasters in the first place.
In the 1970s and 1980s, most private health insurance plans had benefits designed very much like Medicare's fee-for-service program, with few local initiatives or accountability systems and an acute care focus.
However, that started to change in the 1980s with the rise of HMOs. The pace of change accelerated in the 1990s as Preferred Provider Organizations (PPOs) created a hybrid of traditional insurance and HMO-style managed care. The changes in private health insurance were tumultuous, with lowered premiums, expanded benefits, controversies over restrictions on access to care, and a sharply negative reaction from health care providers. But the changes generally moved private plans toward local flexibility and improved chronic care.
For the last decade, politicians have been at each other's throats over how to inject some of these new characteristics of private health plans into the Medicare program. The hottest, most vitriolic topic has been the role of private health plans as an alternative to the fee-for-service program.
Liberals desperately want private health plans to fail in Medicare, period. The reasons for this are complex and sometimes hard to understand, but it is the reality.
Conservatives generally want private plans to succeed in Medicare, although their commitment is sometimes soft, and they do not always seem sure which types of private health plans they'd prefer.
The latest flare-up was sparked by a routine report from MedPAC, the Medicare Payment Advisory Commission. MedPAC followed-up its March 2004 report with a technical memo estimating that payments to Medicare private health plans were 107 percent of comparable spending in the traditional fee-for-service program. This figure was seized by liberals as evidence private health plans cost too much.
Some context is in order. The following sections track the payments received by private plans since 1997, and compare those figures with the comparable trends in the traditional fee-for-service program. Extra effort was made to ensure that the results could easily be reproduced and evaluated.
Enrollment and Payment Rates Since 1997: In the mid-1990s, enrollment in Medicare's HMO program surged (see Figure 1). Simultaneously, Congress began to rein in the growth of Medicare spending. First, Congress and Medicare administrators attempted to reduce fraud in the fee-for-service program. Later, payments to fee-for-service health providers and Medicare HMOs were reduced. The main payment reductions were enacted in the Balanced Budget Act of 1997 (BBA).
As a result, Medicare spending dropped precipitously in the late-1990s, and Medicare HMO enrollment started to fall.
Figure 1.
(Centrists.Org published a discussion of the political impact of private plans on Medicare spending in the 1990s in The Curious, Counter-Intuitive Relationship Between Medicare Costs and HMO Enrollment, February 8, 2004.)
For most Medicare HMOs, the BBA limited the annual increase in Medicare payments to 2 percent, beginning in 1998. This was briefly increased in subsequent years by so-called "BBA giveback" bills designed to smooth over excessive payment reductions, both for HMOs and other health providers.
However, the distribution of the BBA payment reductions and the subsequent "givebacks" was not spread evenly across the country.
Beginning in 1998, Congress raised payments in some areas by establishing a new payment "floor." This initially had a very small impact on overall Medicare payments because floor payments were restricted to rural areas where few HMOs operated.
However, a subsequent payment floor for large urban areas with low health costs -- like Portland, Seattle and Minneapolis -- caused a spike in payment rates to HMOs in those areas in 2001.
By 2004, there were four methods of "updating" (that is, calculating the annual increase in) payments to Medicare's private plans: the minimum update method, the "floor" method, the "blend" method, and the "100 percent of fee-for-service" method.
These formulas gave many health plans their first substantial payment increase since 1997 (see Table 1). On average, the 2004 rate updates have brought reimbursements to private plans back up to their 1997 level relative to fee-for-service spending.
Table 1.
| Post-BBA Payment Rate Changes for Medicare Private Plans |
|
|
|
(annual percent change) |
|
|
|
|
|
|
|
| |
1998 |
1999 |
2000 |
2001 |
2002 |
2003 |
2004 |
| |
|
|
|
|
|
|
|
| Payment Update Type: |
|
|
|
|
|
|
|
| "Minimum Update" Areas |
2.0 |
2.0 |
2.4 |
3.0 |
2.0 |
2.0 |
6.3 |
| "Floor" Areas |
3.4 |
2.1 |
6.6 |
11.9 |
4.8 |
2.0 |
8.8 |
| "Blend" Areas |
2.0 |
2.0 |
6.7 |
3.0 |
2.1 |
2.0 |
9.9 |
| "Fee-For-Service" Areas |
2.0 |
2.0 |
2.8 |
3.5 |
2.4 |
2.0 |
14.8 |
| |
|
|
|
|
|
|
|
| All Private Plans |
2.4 |
2.1 |
4.1 |
5.8 |
3.0 |
2.0 |
10.8 |
| |
|
|
|
|
|
|
|
| Memo. |
|
|
|
|
|
|
|
| Fee-For-Service Costs |
-0.6 |
0.7 |
4.0 |
8.8 |
5.6 |
4.5 |
6.3 |
| |
|
|
|
|
|
|
|
| Source: Centrists.Org |
|
|
|
|
|
|
|
Note: Categories refer to the update method used in 2004. Most plans received the "minimum" update in preceding years, regardless of the update received in 2004. Weighting is by December 2002 enrollment. These payment updates are for so-called "Medicare +Choice" plans or "Medicare Advantage" plans -- those private health plans which receive capitation payments from Medicare. Data do not include the PPO demonstration program or HMOs reimbursed on a cost basis.
Figure 2 shows the distribution of private plans by type of update in 2004. Of course, most plans received the minimum update in years prior to 2004.
Three technical notes about these premium rate calculations:
First, Medicare publishes the actual payment rates for private plans in each county in the U.S. in its so-called "ratebook" tables. The actual rates and the annual percent change since 1997 are available via attached files: Rate of growth in county-by-county payment rates to health plans. Actual monthly rates by county.
Second, the averaged rate trends in Table 1 and Table 2 (below) can easily be replicated from the estimates of nationwide fee-for-service spending and the actual county-by-county payment rates for private plans, weighted by enrollment in each county. (There are instructions on how to replicate these calculations in an appendix at the end of this report.)
Third, to make them easy to replicate from readily available Medicare data, these calculations use enrollment data from December 2002. This may cause the enrollment-weighted trend figures in Tables 1 and 2 to be slightly different than figures calculated from more up-to-date enrollment data. However, the differences are not analytically meaningful.
Figure 2.
Comparing the Growth of Spending in Private Plans and Fee-For-Service: There is an on-going dispute over whether Medicare's private plans are paid too much or too little.
There are two components of the debate:
1. Does Medicare pay private plans too much or too little relative to fee-for-service costs?
2. Do private Medicare plans have healthier or sicker enrollees than the fee-for-service program, and are payments properly adjusted for health status or risk?
This section addresses the first question. (The discussion of health status and risk composition of Medicare enrollees in private plans is considered below.)
Before 1998, Medicare HMOs were paid at 95 percent of fee-for-service Medicare spending. The BBA changed that formula to a combination of minimum updates, blends, and floor payments.
However, the BBA also changed the way Medicare's Graduate Medicare Education (GME) payments are counted for the purposes of comparing payments to private plans with those in fee-for-service.
In 1997, GME payments were considered part of fee-for-service spending, and HMO rates were calculated as 95 percent of that amount. However, the BBA subtracted GME payments from fee-for-service spending for rate-setting purposes.
The Medicare Payment Advisory Commission (MedPAC) computes that Medicare payments were actually 98 percent of fee-for-service spending in 1997, based on the earlier definition. According to MedPAC, the 98 percent figure -- instead of the desired 95 percent -- was simply a calculation error.
Now, MedPAC calculates that payments to private plans average 107 percent of fee-for-service spending, based on the new definition that excludes GME payments.
The HMO industry vigorously rejects those calculations, saying that in a proper apples-to-apples comparison, payment rates are not much better now than they were in 1997, despite the large rate increases in 2004.
Medicare fee-for-service spending has many facets, of course, and the choice of the "denominator" determines how these sorts of measures turn out. Choose a small measure of fee-for-service spending, and private plans will appear overpaid -- choose a large measure and private plans will seem underpaid.
Fortunately, there is a relatively unbiased way to examine what has happened to payment rates for private plans since 1997.
The best way to examine relative payment levels is to set payments in 1997 equal to an index value of 1.00, and examine how the average payment rates to private plans (weighted by enrollment) have changed since that date.
This method abstracts away from the dispute over whether payment rates were 95 percent or 98 percent of fee-for-service spending in 1997. The index method does not care whether or not education payments should have been included in the computation in a given year, or not.
Instead, these calculations simply set a base year of 1997, one year prior to when the payment formula was changed. This analysis doesn't speak to whether private plans were overpaid or underpaid in 1997 -- it just sets that year as a starting point and tracks what has happened since.
On this index basis, payment rates for private plans have grown at almost exactly the same rate as comparable fee-for-service spending between 1997 and 2004 (see Figure 3).
Figure 3.
Since 1997, fee-for-service spending has increased by 33 percent. Therefore, the index value has increased from 1.00 in 1997 to 1.33 in 2004, an average annual growth rate of just over 4.1 percent.
In the meantime, average payment rates for all private plans increased by 34 percent over the same period. Therefore the index level for all private plans shown in Figure 3 rose from 1.00 in 1997 to 1.34. That is an average annual growth rate of just under 4.3 percent.
Of course, plans in different areas received dramatically different payment rate increases, depending on which type of updates they received in their particular service areas.
For example, payment rates for minimum update plans have grown by much less than nationwide fee-for-service costs (see Figure 4.)
Figure 4.
Payment rates for plans receiving the minimum update in 2004 grew by only 21 percent between 1997 and 2004, over 11 percentage points less than nationwide fee-for-service costs during that period. (The growth rates don't sum to precisely 33 percent -- the fee-for-service growth rate -- due to rounding.)
On the other hand, payment rates for plans that qualified for "floor" payments in 2004 have risen dramatically since 1997. Although there were few plans in this category in 1998, by 2001 a significant number of Medicare plans received the special floor payment designed for urban areas with low fee-for-service costs.
As a group, plans that received the "floor" update in 2004 have seen total payment rate increases of over 46 percent since 1997, almost 14 percentage points more than fee-for-service spending (see Figure 5).
Figure 5.
The other two categories, "blend" counties and "100 percent fee-for-service" counties, ended up in 2004 about where they started in 1997, with payment rates that grew at almost exactly the same rate as fee-for-service spending during the 1997-2004 period.
For example, Figure 6 shows how payment rates for plans that received the 100 percent of fee-for-service update in 2004 snapped back up to par with growth in fee-for-service spending.
Figure 6.
Trends in Relative Cost Ratios. By dividing the index values for private plans by the index for fee-for-service spending, we can compute the percentage by which payment rates for private plans exceeded (or fell below) the growth of fee-for-service costs during the 1998-2004 period, relative to the starting point in 1997.
On average, the ratio of private plan reimbursements to fee-for-service spending grew from 1.00 in 1997 to 1.01 in 2004, an increase of one percentage point (see Figure 7).
Figure 7.
Table 2 shows the details of these figures by the type of payment update a plan received in 2004.
Reimbursements in "minimum update" areas fell by over 8 percentage points relative to fee-for-service between 1997 and 2004. By contrast, relative payments in "floor" areas increased by over 10 percentage points.
The ratio of payment rates for "blend" plans to fee-for-service spending fell by a little more than 1 percentage point. Payments to "100 percent fee-for-service" areas increased in 2004 to par with trends in overall fee-for-service spending.
Table 2.
| Private Plan Reimbursements Relative to Fee-For-Service |
|
|
|
| (1997=1) |
|
|
|
|
|
|
|
|
| |
1997 |
1998 |
1999 |
2000 |
2001 |
2002 |
2003 |
2004 |
| |
|
|
|
|
|
|
|
|
| Payment Update Type: |
|
|
|
|
|
|
|
|
| "Minimum Update" Areas |
1.000 |
1.026 |
1.039 |
1.023 |
0.969 |
0.936 |
0.914 |
0.914 |
| "Floor" Areas |
1.000 |
1.040 |
1.054 |
1.081 |
1.112 |
1.104 |
1.078 |
1.103 |
| "Blend" Areas |
1.000 |
1.026 |
1.039 |
1.066 |
1.010 |
0.977 |
0.953 |
0.986 |
| "Fee-For-Service" Areas |
1.000 |
1.026 |
1.039 |
1.027 |
0.978 |
0.948 |
0.926 |
1.000 |
| |
|
|
|
|
|
|
|
|
| All Private Plans |
1.000 |
1.030 |
1.044 |
1.046 |
1.017 |
0.992 |
0.969 |
1.010 |
| |
|
|
|
|
|
|
|
|
| Source: Centrists.Org |
|
|
|
|
|
|
|
|
The Larger Battle -- For and Against Private Health Insurance: Liberal health care advocates have spent the last six or seven years trashing HMOs.
Why would liberals do this? Sure, HMOs could be a nuisance. Getting referrals and being restricted to a network of health providers was annoying for people used to generous, unrestricted health benefits.
But there are millions of lower-income working people who don’t have generous health coverage through a union shop or academia, or through public-sector employment. HMOs were a godsend for them. They could get decent, mainstream health care for a wide variety of items -- routine and preventive services, mental health, prescription drugs -- without unaffordable deductibles and coinsurance rates.
In Medicare it was much the same. HMOs helped several million seniors avoid heavy out-of-pocket costs. Moreover, the quality of care received by HMO enrollees is widely considered to be at least as good as the health care received by enrollees in Medicare’s traditional fee-for-service program. (A full discussion of the literature on quality comparisons between private plans and the fee-for-service program is beyond the scope of this report. However, it is clear that HMOs are leading the move toward innovative treatments, such as disease management and coordinated care for patients with chronic health conditions, which are now being mimicked by the fee-for-service program.)
Clearly, liberals' distaste for HMOs and other managed care plans is based on more than practical concerns -- it is an ideological phenomenon.
The meta-struggle over private health plan options in Medicare is rooted in a clash of ideologies that goes far beyond Medicare itself, or the particular needs of its beneficiaries.
At its deepest level, it is a battle for and against the reach of government.
Old-style liberals fear a "privatization" of government services. At their core, they believe that only government control can rationalize the chaos and inequities of the U.S. health insurance system. Liberals worry that any successes of private plans would weaken the traditional fee-for-service program -- their model of price controls and claims-paying efficiency -- and that the essential government infrastructure for national health insurance could be lost.
Conservatives generally approve of private plans in Medicare as a means of thwarting liberals’ push for centralized national health insurance and limiting the portion of the health sector subject to government price setting. But they are at best ambivalent toward HMOs and other private plans with generous, comprehensive benefits. This leads to a sort of schizophrenia in conservative thinking.
Centrists and classic conservatives push for private options in Medicare for two reasons: (1) because they believe competitive systems and consumer choice lead to innovation and efficiency, and (2) to demonstrate to seniors that it is the service itself -- the health benefits -- that they appreciate, not the government per se. If more benefits could be provided by private plans, direct government control over the health system would be partly reduced, even if federal spending and taxation remained unchanged.
But some conservatives dislike HMOs and other private health plans with generous benefits, because they believe such benefits lead to overuse of health services and overly high health spending.
Instead, they prefer high-deductible "catastrophic" coverage coupled with personal spending accounts, which would give patients an incentive to spend wisely for routine or non-catastrophic services. (Conservatives assume patients would bargain for reasonable prices when they were spending “their own money,” and believe that networks of reasonably priced health providers would spring up to provide services to those with spending accounts.)
Liberal Arguments Against Private Plans: Liberals argue against private plan options in Medicare, variously claiming that plans attract healthier-than-average enrollees or that the fee-for-service program's claims-paying efficiency and "bargaining clout" (that is, price-setting capability) renders it inherently more efficient.
The bargaining clout or price control argument is inherently subjective. Conservatives and most centrists believe price controls usually distort the marketplace and diminish innovation, and could reduce access to health care. Liberals believe well-calibrated price controls could reduce health spending without negative effects.
There is no conclusive evidence. Analysts can look to the experience of other countries that employ price controls to varying degrees, and guess what would happen if they were extended in the U.S. But there is no way to know for sure.
On the other hand, there is a growing consensus that we can measure or infer that Medicare's private plan enrollees are not in fact disproportionately healthy or low-risk.
Do Private Plans Get Healthier Enrollees? It's becoming increasingly likely that private health plans in Medicare don't have healthier-than-average enrollees.
Importantly, the Congressional Budget Office and Medicare's Office of the Actuary -- the groups that provide the official cost estimates of Medicare proposals -- do not generally assume that Medicare's private plans now attract healthier enrollees when they make cost estimates.
But this was not always the assumption.
In the 1990s, the main support for the theory that Medicare's private plans had healthier-than-average enrollees came from so-called "joiner" studies conducted by the General Accounting Office (GAO) and the Physician Payment Review Commission (which has since become MedPAC).
The joiner studies looked at Medicare beneficiaries who switched from the fee-for-service program to HMOs. They concluded, not surprisingly, that people who switched had lower-than-average costs while they were in the fee-for-service program the year before.
GAO assumed that the reason people who switched to HMOs had lower prior-year costs in fee-for-service was because they were healthier than average in that year.
This makes sense, because people who are in the middle of an acute health care episode are unlikely to want to switch health plans and providers, unless they are in desperate financial straights. People who aren't having health problems might have weaker attachments to particular health providers, and would therefore be more willing to switch.
However, some low-income people who switched to HMOs probably did so precisely because they couldn't afford the deductibles and coinsurance required by the fee-for-service program. Their prior-year spending might have been low because they couldn't afford the care they wanted.
Moreover, the joiner methodology used by GAO made an extremely important -- and debatable -- assumption in the way it handled "regression to the mean" in its estimates.
GAO estimated that HMO joiners had prior-year costs that were 67 percent of average fee-for-service spending. To impute what those switchers in their first year of HMO enrollment would have cost had they stayed in fee-for-service, GAO extrapolated from this figure using estimates of "regression to the mean" -- that is, healthier people get sicker over time -- based on the experience of groups of fee-for-service enrollees with similar costs.
A typical regression to the mean estimate is shown in Figure 8. GAO assumed that people who started out "healthier" than average gradually got sicker. But their costs were never assumed to exceed the average.
This is reasonable, but it is not the only plausible assumption.
What about seniors who joined HMOs because they needed health services, but couldn't afford the deductibles and coinsurance in fee-for-service? Might not their health spending spike upward in their first year of enrollment in an HMO? Is the fee-for-service pattern of regression toward the mean really the best estimate for people who actually joined an HMO?
Figure 8 shows some other plausible assumptions for regression toward the mean. It's not unreasonable to assume that the "mean" cost for some HMO enrollees could be higher than the fee-for-service average.
We really don't know. The GAO assumption is reasonable, but there are other reasonable possibilities GAO didn't include in its calculations.
Using a different assumption about the cost characteristics of HMO enrollees would have led to a different conclusion.
Figure 8.
There are three additional reasons to believe that HMO enrollees are not healthier than average:
1. Alternative survey data. In late 2000, Gregory Pope and his colleagues from the firm Health Economics Research published a study of the self-reported health status of enrollees in the fee-for-service program and Medicare's private health plans. Instead of imputing the cost or healthiness of HMO enrollees, this survey asked directly about enrollees' health status.
Like the GAO study, the survey data were from 1998 and 1999, near the peak of Medicare HMO enrollment. The data indicated that the gaps in health status between enrollees in private plans and the fee-for-service program were neither statistically nor clinically significant, once they were controlled for age, sex, Medicaid eligibility, and previous disability status. (These are some of the factors used to adjust Medicare's payments to private plans).
2. Low-income and minority enrollment. In 2002, economist Ken Thorpe of Emory University studied the demographic characteristics of Medicare enrollees who were not eligible for Medicaid and did not have employer-based retiree or supplemental coverage.
Thorpe and his coauthors concluded that Hispanic and African-American retirees were much more likely to use Medicare HMOs for extra coverage than to supplement their fee-for-service coverage with a "Medigap" plan. Likewise, Medicare beneficiaries with low incomes were slightly more likely to depend on Medicare's private comprehensive plans.
A disproportionate low-income or minority enrollment is not in itself evidence that Medicare's private plans have a less healthy mix of enrollees. But it is suggestive. Based on surveys conducted in the 1980s, health economists thought that low-income people had lower overall health costs, despite poorer-than-average health status. This was presumably because low-income people had limited access to high-tech or intensive health services, either because high-tech services were not available where they lived, or because the cost-sharing requirements or cultural factors restricted their use.
However, low-income Medicare beneficiaries probably don't have significant barriers to the most intensive (and expensive) health services at this point, especially in urban areas where most HMO enrollment occurs. And the health status of people with low incomes is still thought to be poorer than average. At the least, higher low-income and minority enrollment is probably an indicator that Medicare's private plans are not "skimming the cream" of low-cost seniors.
3. Analysis of trends in enrollment and spending. Medicare's cost experience over the last decade was the opposite of what would have happened if Medicare's HMOs had enrolled a dramatically healthier-than-average population of seniors.
Figure 9 (below) shows the trend in Medicare's overall costs and its HMO enrollment. Medicare's overall spending growth fell after its HMO enrollment surged in the mid-1990s. Likewise, Medicare spending grew more rapidly after HMO enrollment started to fall in the late-1990s.
Clearly, anti-fraud measures and payment rate reductions in both the fee-for-service program and in HMO reimbursements were the main reasons spending growth fell, and "giveback" provisions contributed to the subsequent rise in spending.
But if Medicare's private plans had a healthier-than-average enrollment mix, the opposite pressures would have been in place. When HMO enrollment grew rapidly, Medicare spending would have been pressured higher, as healthier people (who would have cost the fee-for-service program little) went to HMOs that were paid 95 percent (or even 100 percent) of fee-for-service costs.
For example, suppose a group of healthy Medicare beneficiaries, whose claims costs were only 50 percent of average, moved from the fee-for-service program to an HMO. If Medicare payments to the HMO were 100 percent of average fee-for-service costs, Medicare spending on their behalf would double.
However, Medicare spending fell after lots of seniors joined HMOs in the mid-1990s. This is the opposite of what the "healthier-than-average" theory would imply.
Likewise, after 1999, when HMO enrollment started to fall, the "healthier-than-average" theory implies that Medicare spending would decelerate, as healthier HMO enrollees returned to fee-for-service, where Medicare would spend less on their behalf. However, Medicare spending started to rise as HMO enrollment began to fall.
Again, spending and enrollment trends do not prove that the "healthier-than-average" hypothesis is inaccurate. But it is a relatively clear and obvious indicator that something is wrong with the theory. For non-experts and experts alike, the question becomes: "What are you going to believe -- the old theory or your own eyes?"
Figure 9.
Conservative Schizophrenia on Private Health Plans: Conservative health policy tends to oscillate between a focus on competition and choice of health plans -- the federal employees' model -- and a desire to encourage high-deductible health insurance coverage that is coupled with tax-favored savings accounts.
For the last decade, conservatives have touted the federal employees' model of health insurance in Medicare. The FEHBP allows its enrollees -- federal workers and their dependents -- a wide choice of plans, and the program has succeeded at keeping costs within reason and improving benefits. FEHBP plans are generally able to innovate and adjust their benefit and payments. There is some regulation, but the system overall is quite flexible from a plan's point of view.
In the early 1990s, many centrists embraced a less flexible vision for competing private health plans dubbed "managed competition." Managed competition included strict controls on the benefits plans could offer.
The Clinton health reform proposal, which was designed from managed competition principles, allowed state-based purchasing groups to set prices for health providers as well.
Conservatives had no problem opposing the Clinton plan. The benefit and price controls were far too difficult to swallow.
Since the downfall of the managed competition model, however, many centrists and some liberals have embraced the earlier, more flexible version of competition modeled by the federal employees system. For example, Senator Kerry has endorsed a federal employees-style coverage option for small businesses.
At this point, conservatives don't seem to know what to think. Some free-market think tanks still robustly support the FEHBP model, but others have switched their emphasis to "consumer-directed health care," which is code for high-deductible insurance with a savings account from which routine purchases of health care services would be made.
The recently passed Medicare prescription drug bill reflects this shift. Instead of insisting on an FEHBP-style reform to Medicare, many conservatives were persuaded to vote for the bill because of its new tax breaks for Health Savings Accounts (HSAs) -- which are available only to the non-Medicare population.
The Best Hope for Long-Term Value -- A Healthy Rivalry: Politicians of either party, health plan administrators, or think tank experts can't possibly know in advance what is the best kind of health insurance -- not now, and certainly not in the future.
Therefore, Medicare should offer a wide variety of public- and private-sector options: private HMOs and PPOs, variations on the government-run fee-for-service program, even high-deductible plans. The only way to discover the best forms of insurance is to try all of the options.
This means the Medicare should operate by the ethic of "any willing insurer," allowing health plans with varying ideas on how to cut costs and improve health service to make their case to Medicare beneficiaries.
To be sure, Medicare should continue its work on "risk adjustment," which ensures that payments to health plans reflect the health status of their enrollees. Combined with a sensible level of regulation and oversight, risk adjustment levels the playing field for all health plans, so that plans can't design coverage that might appeal only to the healthiest seniors and thereby profit from "risk selection," not efficiency.
But in general, worries that inexpensive health plans would garner healthier-than-average enrollees have probably been overblown, and have impeded experimentation with innovative coverage ideas in Medicare.
Ultimately, the best way to ensure that Medicare spending remains under control is to allow a wide variety of health plans into the system, and take the "best practices" -- both cost saving ideas and benefit improvement ideas -- from the successful plans.
Let both government-run and private health plans compete for seniors’ business, and let the competition directly and quickly pressure both types of coverage to find efficiencies, new and helpful benefits, and other value improvements for both seniors and taxpayers. Direct competition and choice would be a more efficient way for each sector to match the other’s improvements. Depending on the formulas used for beneficiaries' premiums, direct competition could automatically save the government money.
If direct competition is not politically possible, Congress will have to make the adjustments on an ad-hoc legislative basis. Legislative changes are slower, more cumbersome, and more error-prone.
However, with private plan options available, members of Congress will at least have a better idea of the possibilities for improved care or cost savings when then legislate.
Policymakers should keep that in mind when the controversy over private plans in Medicare flares up again.
Appendix: Step-By-Step Guide to the Calculations in this Report
1. To download the county-by-county payment rate and enrollment data from the Centers for Medicare and Medicaid Services, go to http://www.cms.hhs.gov/healthplans/rates/ and look for the Subheading "Medicare Advantage Ratebook." Just below, there is another subheading called "Rate calculation data." Click on the file calculationdata2004b.zip and save the .zip file. Then open the sub-file called "aged2004_revised.csv" in a spreadsheet.
2. This file contains the actual rates for 1997 through 2004, and includes enrollment data from December 2002, which is a roughly acceptable "weight" for enrollment by each county. Calculate the total "Risk HMO" enrollment by summing the column, then divide each county's enrollment by the total to get a rough enrollment weight for each county.
3. Take the annual percent change in rates for each county for each year, and weight by the enrollment share for that county. The sum of the enrollment-weighted percent changes is the weighted average overall rate growth for each year.
4. To calculate sub-categories, sort the file by the column labeled "Rate Category" and repeat the weighting process within each subcategory.
5. Create base year indexes with 1997 = 1 to duplicate the results in this report. The fee-for-service program spending is just the growth rates at the top left of the spreadsheet, with the reduction factors "added" back in. (The "adding" of the reduction factors is actually done multiplicatively, using growth rates.)
Note #1, the data for Puerto Rico were included in these calculations. This slightly exaggerates the estimated weighted average payment increases for private plans, because rates in Puerto Rico were increased dramatically in 1998 and in 2001, based on floor calculations. However, only a small percentage of Medicare beneficiaries in Puerto Rico are enrolled in private health plans, and the calculations should be similar with or without the inclusion of Puerto Rico in the data.
Note #2, no weighted average technique is perfect. This report uses as weights the December 2002 enrollment figures, because those are the figures provided with the 2004 rate book spreadsheet. Ideally, the growth rates would have been weighted by enrollment in each year, but that would have required linking several files by county, which is hard. Also, the "update" categories themselves are a little odd because they apply to plans based on the applicable update methodology in 2004, regardless of how plans in those counties received their updates in prior years. However, this method has the benefit of being easy to replicate, and it generally provides an interesting and (hopefully) unbiased result. However, people may wish to look at the detailed county-by-county growth rates and payment rate levels to see particular trends.
Selected References and Links:
Kenneth E. Thorpe and Adam Atherly Medicare+Choice: Current Role and Near-Term Prospects Health Affairs (July 17, 2002). The main conclusion of this report is that Medicare's private plans had a disproportionate minority and low-income enrollment.
Gregory C. Pope, Michelle Griggs and Nancy T. McCall "Comparison of the Health Status of Medicare Fee-For-Service and Managed Care Enrollees using the Health Outcomes Survey" Health Care Financing Administration Final Report, Master Order Contract No. 500-95-0058 Task Order No. 2 (November 16, 2000). This report finds that after controlling for demographic factors, private plan enrollees are not significantly healthier than comparable fee-for-service enrollees.
General Accounting Office Medicare+Choice Payments Exceed Cost of Fee-For-Service Benefits, Adding Billions to Spending (August 2000). This report uses the prior fee-for-service costs of beneficiaries who joined HMO plans and a "regression to the mean" technique to estimate that private plan enrollees were healthier than average.
Dana Goldman and Julie Zissimopoulos High Out-Of-Pocket Health Care Spending By The Elderly Health Affairs (May-June 2003). Using a survey from 1998, RAND researchers found that Medicare HMO enrollees (as well as beneficiaries with employer-based supplemental coverage or Medicaid) had significantly lower out-of-pocket spending than those with no supplemental coverage or with Medigap coverage, especially when enrollees with high health costs are considered. The study also notes no indications that Medicare HMO enrollees were healthier than average, although that question was not the main focus of the research.
Centers for Medicare and Medicaid Services, County-By-County Update Growth Rates for Medicare Private Plans
Centers for Medicare and Medicaid Services, County-By-County Monthly Payment Levels for Medicare Private Plans
Centrists.Org The Curious, Counter-Intuitive Relationship Between Medicare Costs and HMO Enrollment (February 8, 2004)
Centrists.Org Explaining Premium Support: How Medicare Reform Could Work (revised November 6, 2003)
Centrists.Org Testimony: Improving Chronic Care in Medicare (November 4, 2003)
Progressive Policy Institute McDonald's vs. Burger King: A "Nothing Burger" Debate on Medicare Reform (May 6, 2003)
