How Not to Reform Medicare: Lessons From the Medicare+Choice Experiment

Managed Care's Unexpected Response

Medicare HMOs have been available to some seniors for decades. Kaiser Permanente and United Healthcare, for example, started enrolling Medicare patients in 1977 and 1982, respectively.¹¹ Nationwide, Medicare managed care plans enrolled as many as 85,000 new enrollees a month.¹² Five years ago, 7 percent of Medicare's population had enrolled in a managed care plan; by 1998, that number had grown to 17 percent, or 6.5 million of the 39 million beneficiaries.¹³ HMOs are popular in states such as California and Minnesota, where commercial membership in managed care plans is high and members of the commercial plans who retire can stay in the plans.¹⁴ This option has enabled many of these retirees to escape the discontinuity of care experienced by working Americans with employer-based health care.

Surveys of senior HMO enrollees demonstrate high levels of satisfaction with their managed care plans. In fact, 96 percent of HMO Medicare enrollees rated their overall health care as "good" to "excellent."¹⁵ According to one recent study, chronically ill Medicare HMO enrollees were as satisfied as or more satisfied than their counterparts in traditional Medicare or Medicare supplemental plans--a finding that should dispel the simplistic perception that HMO Medicare enrollees are more satisfied only because they have fewer medical problems.¹⁶

Despite such anecdotal evidence, Washington's approach to health care for seniors is based on a presumption that Medicare patients need the federal government to interpret the medical world and make their medical choices for them. But Medicare patients, like most seniors, are proving to be smart shoppers, as demonstrated by the numbers who stayed with the plans they had during their working years. Moreover, the cost-benefit differential--or "medigap"--between managed care plans and supplemental insurance influences their decisions regarding Medicare HMO enrollment. Studies indicate that Medicare patients most often choose managed care options when that differential is the greatest.¹⁷

As of October 1997, more than two-thirds of Medicare's managed care plans did not charge a supplemental monthly premium.¹⁸ Medicare managed care plans also are able to provide services not available in traditional Medicare. For example, over two-thirds cover prescription drugs, 83 percent pay for routine eye exams, 72 percent cover ear exams, and almost all cover routine physicals.¹⁹ These added benefits are usually provided at little or no extra cost to beneficiaries. On average, Medicare HMO enrollees receive more benefits for far less in out-of-pocket expenses than if they had purchased a Medicare supplemental (medigap) package.²⁰

Until recently, Medicare HMO enrollees could quit an HMO and join another plan at any time (commercial arrangements may require a one-year lock-in period). Disenrollments were relatively infrequent, amounting to 14.2 percent in one study.²¹ The same study found that HMO disenrollees returned to traditional fee-for-service Medicare in only 5.4 percent of cases.²² Some disenrollees from fee-for-service plans (25 percent in one study²³ ) had moved out of their HMO's service area and thus were no longer eligible for plan membership. Other research shows large variations in disenrollment rates among HMOs in large urban areas, where competition among plans is more likely.²⁴

Although HMOs may be criticized for disenrollment rates, a better interpretation of the disenrollment data is that market forces are working. Seniors leave their HMOs to find better health care bargains. The HMOs do not make money on disenrollments, which in fact can be quite costly. Depending on the market, "attracting an enrollee can cost $700 to $1,300 so high attrition is costly to health plans."²⁵

By most measures, then, Medicare HMOs have been a success. It is therefore not surprising that the expected congressional expansion of Medicare options through new managed care plans in Medicare+Choice was greeted with optimism by analysts in industry and government, as well as among health care professionals. For example, the Congressional Budget Office (CBO) and officials at HCFA predicted a 20 percent growth rate for Medicare risk-based plans.²⁶ The American Hospital Association reported that 33 percent of its members (1,500 hospitals) planned to initiate point-of-service (POS) plans. With such an initial response, HCFA predicted 50 POS applications in 1999.²⁷

But the repercussions of the Balanced Budget Act of 1997 did not bear out this initial optimism. Whatever Congress's intent in enacting the health policy changes, the effect was to enable HCFA to reverse Medicare's steady progress in managed care.

Medicare HMOs serving more than 6 million patients were repackaged under Medicare+Choice. HCFA's overly optimistic and erroneous assumptions led many analysts to expect that the new program would allow beneficiaries to participate in popular plans already available to privately insured working Americans--such as preferred provider organizations (PPOs), provider sponsored organizations (PSOs), and new HMOs with point-of-service options. Over half of the 70 million Americans enrolled in Blue Cross/Blue Shield's managed care plans chose to participate in PPOs rather than traditional HMOs.²⁸ Medicare patients generally remained locked in more traditional HMO models, with Medicare Parts A and B covering most inpatient and many outpatient services. As of January 1, 1999, Medicare+Choice was scheduled to make popular private-sector options available as well--including medical savings accounts (MSAs), PSOs, POS options, and new fee-for-service insurance plans.

Unfortunately, the initial response was disheartening. Just before the August 31, 1998, deadline for private plan participation in the new program, HCFA announced that only three private plans had submitted applications.²⁹ Nor did Medicare HMOs jump on board. In late October 1998, almost 100 HMOs had exited the program in some part.

The system quickly unraveled. Instead of holiday shopping at the end of 1998, hundreds of thousands of seniors faced a last-minute search to find alternative health plan coverage. Fifty thousand had no choice but to return to traditional Medicare. Not only did existing HMOs withdraw from the program in the first year, but no new options such as PSOs, fee-for-service plans, or MSAs became available.

In general, the BBA's Medicare mandates resulted in:

• Fewer choices.
  By the end of 1998, many HMOs that had serviced Medicare patients for years had withdrawn altogether or partially from their service regions. An HMO plan was more likely to vacate areas it had served since 1992.³⁰ The number of withdrawals was also unprecedented (in previous years, only two or three plans quit the system).³¹ But the expected cuts in payments were not the only reason for their flight. Although the average per capita Medicare payment was projected to be $413 across all U.S. counties at the time, the average payment rate in the 72 counties with plan withdrawals was $434. Health plans even left counties that had payment rates as high as $721 per member per month.³²

• Fewer participating plans.
  In addition to HMO withdrawals, 20 other health plans, including 15 health care prepayment plans (Part B only), did not renew their contracts with HCFA.33 Instead of finding themselves with expanded choice, nearly 500,000 Medicare patients in 29 states found themselves with fewer options.

• More counties with no Medicare managed care option.
  Senior citizens in 71 percent of all counties now have no Medicare managed care option, compared with 68 percent in 1998.³⁴

• More seniors needing medigap coverage.
  About 50,000 Medicare HMO patients were forced to return to traditional Medicare and find affordable medigap insurance.³⁵

• Fewer applications from PSOs.
  Rather than the 50 applications HCFA expected from provider sponsored organizations, only four applications³⁶ came in by October 1998, and HCFA approved only one.³⁷

• No private fee-for-service plans or medical savings accounts (MSAs).
  There were no takers for medical savings account plans or other fee-for-service insurance.

• Benefit cuts.
  Faced with rate cuts and the compliance costs, the remaining HMOs dropped important benefits. For example, 21 percent of Medicare enrollees lost coverage for glasses, and 12 percent lost coverage for hearing aids. Fewer plans offered prescription drug coverage in 1999.³⁸

The repercussions of the Balanced Budget Act of 1997 surprised half a million seniors who thought they had solid private health care coverage. Perhaps HCFA officials assumed that HMOs' complaints about the effects of the cuts and regulations were exaggerated. But the deadly combination of HCFA micromanagement and overly broad congressional mandates demonstrated once again how Washington was "out of sync" with the best practices of private-sector health care delivery.

And the bad news continues. HCFA recently announced that, for the year 2000, 99 Medicare contracts are not being renewed or else the companies are reducing service areas covered.³⁹ According to a recent report in American Medical News, over a quarter of a million more Medicare patients will lose their health plan coverage in 2000 because of program plan withdrawals, and almost all other enrollees will experience a reduction in benefits, an increase in payments, or both.⁴⁰

Cutting Payments

Before passage of the BBA, Congress set per capita payments to Medicare HMOs at 95 percent of a county's average Medicare Part A and Part B per capita fee-for-service cost. Payments were adjusted for demographic factors, such as patients' age and sex and whether they were on Medicaid or institutionalized. At the time, most Medicare HMOs were providing many more services than traditional Medicare provided. For example, 67 percent of HMOs covered outpatient medication, 83 percent covered eye exams, 72 percent paid for ear exams, and 97 percent covered routine physicals.⁴⁴

Yet providing such additional services at a 5 percent discount did not prevent federal officials from criticizing Medicare's HMOs for receiving too much money. In 1997, the GAO concluded that the HCFA would have paid less, on average, for the care of HMO enrollees if those patients had remained in the fee-for-service program,⁴⁵ because Medicare HMO enrollees generally were healthier than their fee-for-service counterparts, even those with the same age and demographic characteristics. While HMO enrollments have doubled,⁴⁶ the trend in the so-called overpayments to HMOs had decreased from the late 1980s to the early 1990s, according to the GAO.⁴⁷ A recent Sachs Group report published in Business Wire, which was based on more timely data, also found that most of these health plans have numbers of chronically ill and aged patients that are similar to those found in the Medicare fee-for-service program.⁴⁸

Congress's objectives in passing the BBA included saving $116 billion on Medicare. For the HMOs, this translated into the six "provider" reimbursement cuts that are listed above and an escalating difference between spending in fee-for-service Medicare and HMO reimbursements. Even the CBO acknowledged that if current payment policies remain in effect, by 2004 Medicare will pay HMOs only 75 percent of what they pay their fee-for-service providers.⁴⁹ A PriceWaterhouse study finds a similar difference; it shows that payments in the traditional Medicare program will outpace payments to some Medicare managed care health plans by 20 percent by the year 2003.⁵⁰ These projections, moreover, do not take into account the impact of an "evolving risk adjuster," which will decrease HMO payments by an additional 7.5 percent.

In most cases, the HMOs--burdened with rapidly rising medical costs, negative bottom lines, and costly regulatory mandates--were expecting a 2 percent rate hike, which did not reflect everything else occurring in the health care sector of the economy. In the private sector, rates were projected to be increasing from 8 percent to 10 percent. Managed care drug costs were increasing because the costs for pharmaceuticals in general had increased 14 percent in 1997 and from 12 percent to 17 percent in 1998.⁵¹ Although traditional Medicare does not cover pharmaceuticals, over two-thirds of Medicare HMOs do. Americans over age 65, who make up 12 percent of the U.S. population, consume 30 percent of all prescription drugs; ambulatory elderly patients take an average of 4.5 prescribed drugs each day.⁵²

The HMOs, facing rate increases that were out of touch with private-sector health care costs, to say nothing of costly regulatory mandates, therefore confronted a Catch-22: They could reduce benefits and raise supplemental payments, or they could discontinue contracting with HCFA altogether.

For understandable reasons, most notably the country's broader budgetary and fiscal problems, Congress historically has focused on Medicare spending. As a result, Medicare policy too often has been subsumed by budget policy. A far more fruitful use of Congress's time would be to improve the way in which quality care is delivered to Medicare patients.

Congress delegates broad discretionary authority to HCFA to run Medicare; and under Congress's broad mandates, HCFA makes detailed decisions about the plans, the level of care, and the kind of physician and specialty payments that will be reimbursable under Medicare. Because of this micromanagement, in the normal course of business, doctors, hospitals, and private insurance officials are locked in an inevitable conflict over administrative and reimbursement issues. Although Congress may intervene periodically as a reluctant referee among these competing interests, Medicare patients are caught in the middle. Whatever Members of Congress conclude about the cost-effectiveness of HMOs, the central issue should be the reorientation of congressional policy to give Medicare beneficiaries more control.

Problems with Pricing, User Fees, and Regulation

HCFA Medicare payment schemes are more complex than the IRS tax code. Yet HCFA officials are not solely to blame for this regulatory mess. Congress is at fault as well for enabling several of HCFA's complicated formulas that often base payments to HMOs on miscalculations, misperceptions, and a poor understanding of managed care. Consider a few of the problems facing managed care organizations today:

• Miscalculating adjustments.
  One intent of Congress in passing the BBA was to improve payments to Medicare HMOs serving seniors in rural counties, compared with payments to HMOs in urban areas. These higher payments, Congress reasoned, would make serving Medicare patients in rural areas more attractive. Provider sponsored organizations would be quick to jump into the market, and existing HMOs would expand their service areas to include more seniors.

In fact, the opposite occurred. HCFA's promised "adjustments" were based on spending in traditional Medicare, and because of cuts in that program, its projections were off the mark.⁵³ The Medicare Payment Advisory Commission reported to Congress in March 1999 that there "has been no reduction in the disparity between high and low rates above the floor"⁵⁴ set by the BBA. For Medicare HMOs, the expected 2 percent rate increase was much lower than the rate of inflation in private-sector health care--and it would be even less when "user fees" and the enormous costs associated with regulatory compliance were factored into the equation. Consequently, fewer HMOs joined the program.

• Provider taxes for an elusive Medicare education campaign.
  The BBA authorized a provider tax as a "user fee" to fund a campaign to educate seniors. Medicare HMOs have been assessed nearly $190 million ($95 million in 1998 and another $95 million in 1999) to fund a patient information program for all 39 million Medicare beneficiaries. The HMOs were expected to pay the entire cost, despite enrolling only 17 percent of Medicare patients.

For this multimillion-dollar investment in its centrally controlled campaign, HCFA produced a small 35-page information handbook for Medicare beneficiaries in five test areas, initiated a Medicare "helpline," and set up a Web site on the Internet. Despite HCFA's original predictions that the "helpline" would receive 20,000 calls a day (or about 7.9 million calls in the first two months, for which HCFA planned to hire 3,000 additional personnel), the number of calls averaged about 250.⁵⁵ Many callers are referred back to their own health plans or to local community groups for the information they seek.⁵⁶

On its Medicare Web site, HCFA lists several HMO quality measures to help consumers choose from among the available plans. Noticeably absent is a comparable rate for each type of service under traditional Medicare, even when such comparisons may be available. For example, one study reports a 62 percent mammography rate for women enrolled in managed care health plans versus a 39 percent rate in traditional Medicare.⁵⁷ A more recent study suggests an even greater disparity: "The latest edition of the Dartmouth Atlas of Health Care...reports that less than one third of women between the ages of 65 and 69 obtain mammograms every two years."⁵⁸ Published rates for other important preventive care services are off by similar amounts, or worse.⁵⁹ And it is not clear that the information produced by HCFA would be of any interest to seniors even if it were available.⁶⁰

By comparison, federal workers and retirees enrolled in the popular FEHBP system receive easy-to-use guides on private-sector plans that include information on prices, benefits, quality, and service. The Washington Consumers' Checkbook publishes such a guide annually to supplement the comparative information produced for FEHBP enrollees by the Office of Personnel Management (OPM), the federal agency that administers the FEHBP. As Walton Francis, former Department of Health and Human Services official and current editor of the Washington Consumers' Checkbook Guide to Health Insurance for Federal Employees, notes:

It is unclear why HCFA did not copy the economical and effective OPM system...10 years or more ago. I was once told by a HCFA official that the reason why no usable information was provided was because the plans didn't want to make comparisons easy, and HCFA felt obliged to defer to the plans' wishes. This theory is so scandalous that it is hard to believe. I am more inclined to believe in tight budgets and weak imagination. Another theory is that a well-run Choice program would drive Medicare (a grossly inferior insurance product) into the ground and that agency staff are unwilling to foster fair competition. Regardless, the record on Medicare HMO information is atrocious.⁶¹

• Projections of rates sought before HMOs knew the rules of the game.
  HCFA asked the Medicare HMOs to submit information about anticipated costs, premiums, and benefits packages by May 1, 1998. But these projections were based on HCFA's early miscalculation; the deadline meant HMO plan actuaries would have to forecast their plan expenditures for the coming year based on three months of information from the previous year.⁶² Many Medicare HMOs took that leap of faith after assessing the BBA mandates.

Then, in June 1998, one month after the HMOs had committed themselves to specific benefit packages and premiums, HCFA published a rule requiring numerous changes in HMO administrative systems, contracts, quality improvement plans, and medical information systems. The HMOs then found that payment reimbursements would be much lower than anticipated. They asked HCFA whether they could submit new information about their benefits and premiums based on the new rules and rates, but HCFA denied these requests, prompting many plans to forego the opportunity to participate in Medicare+Choice. In any other context, Members of Congress would denounce such "bait and switch" activity.

• Risk adjustments with no relevance to managed care.
  In January 1999, HCFA imposed "risk adjusters" that would remain in effect until 2004. HCFA predicted that the application of these "temporary" adjusters would decrease HMO reimbursements another 7.5 percent.⁶³ The HMOs complained that these risk adjusters would "reduce payments to health plans even further than intended by Congress."⁶⁴

HCFA's risk adjustment method would use hospitalization data from the previous year to indicate that capitated payments for previously hospitalized patients should be higher. There are numerous problems with this approach.⁶⁵ It rewards plans that allow patients to become sick enough to require hospitalization, and because it does not count short stays in hospitals, it provides incentives for plans to keep people hospitalized longer than needed. Such risk adjustments contradict the very nature of managed care. HCFA's post-2004 risk adjustment scheme also is highly complex and will require the infusion of large amounts of personal information from Medicare beneficiaries.

Bureaucratic, Not Patient, Control

In Medicare Part C, the part of the Medicare law that governs the new Medicare+Choice program, Congress did not provide the statutory basis for a genuine system of patient choice. HCFA's implementation of Medicare Part C--effectively imposing new regulations, directives, data specifications, and operational policy letters to govern every detail of plan administration--mocks the concept of patient choice. Patient choice has meaning only if there are different plans and benefits from which to choose. Imagine, for example, if consumers could choose between Apple or IBM computers but both brands used the same operating systems, microprocessors, and modems. Would the consumer have a real "choice?"

True, health plans do have some flexibility, but Medicare Part C mandates represent a stranglehold on the way care is delivered. In the process of implementation, HCFA added over 100 pages of new regulations and dozens of operational policy letters ranging in length from a few pages to as many as 50 pages each. Thousands of additional pages of detailed instructions describe mandated data sets and collection processes, surveys, and quality improvement studies. HCFA even promulgates regulations governing plan-provider relationships that have no statutory basis.⁷⁰ The Medicare Payment Advisory Commission has described Medicare Part C regulations as involving "far more extensive compliance than under earlier risk programs."⁷¹ For this reason, health plans direct their priorities more at compliance than at quality, and compliance consumes so many resources that creativity and innovation are cast aside.

Congress's statutory construction of Medicare+Choice and HCFA's broad interpretation of the BBA together make HMOs little more than conduits for bureaucratically controlled health care of the elderly. Providers at every level will have to submit to additional chart reviews and assure that the patient encounter data they send to the HMOs conform to HCFA specifications. Such efforts run contrary to Congress's intent, which (in the words of the Medicare Payment Advisory Commission) is to "enable the Medicare program to utilize innovations that have helped the private market contain costs and expand health care delivery options."⁷²

Mandated HMO Data Collection

The various reports, data and information sets, analyses, and documents that HCFA requires HMOs to submit cannot be described in detail in this study. Although HCFA collected volumes of such information from Medicare HMOs for some time, little of it is available to the public.

As far back as October 1996, the GAO recommended that HCFA release useful information about Medicare HMOs to aid consumers in making choices.⁷³ HCFA, however, has not done so. It also provides little or no feedback to the HMOs on the information they submitted for more than three years, such as detailed information about patient encounters with providers that include diagnoses, treatments, and procedures. In addition, HCFA is requiring the HMOs to perform and possibly pay for at least five additional data submission projects. Ostensibly, the data will be used to set rates, perform risk adjustments, assess quality of care, provide information to the public, and hold managed care organizations accountable.

Before the BBA, an HMO was required to initiate its own quality assurance program, which then was evaluated as part of its qualification process.⁷⁴ The general guidelines for this program were compatible with commonly accepted principles of quality improvement. The HMOs were required to submit detailed data on patient encounters, enrollments, and disenrollments, information on appeals and grievances, and benefits publications. More recent HCFA regulations added the Health Plan Employer Data and Information Set (HEDIS) and the Quality Improvement System for Managed Care (QISMC).

HEDIS.
The Health Plan Employer Data and Information Set is a quality measurement survey developed for employers, especially large companies, to assist them in choosing a health plan for their employees. It is the most widely used data instrument for assessing quality. But HEDIS does have some drawbacks:

1. It was not designed with seniors or the disabled in mind.

2. It is extremely costly to administer (some estimate that it could cost millions of dollars⁷⁵ ).

3. It includes few true measures of outcomes, even though outcomes are the usual standard for evaluating quality.

4. It relies on costly chart reviews that are disruptive to physicians and a threat to patient privacy.

5. It presents problems for preferred provider organizations and managed care options that deviate from a rigid HMO model.

6. Large employers depend only infrequently on HEDIS data in selecting a health plan.⁷⁶

7. Many employers do not see HEDIS as an accurate measure of quality.⁷⁷

As Steve Wetzell, executive director of policy and public affairs at the Buyers Health Care Action Group (BHCAG), points out, HEDIS does not meet the information needs of the health customer, the consumer.⁷⁸ Seniors, for example, want to know whether their doctors are in the plan, what their benefits under the plan would be, how much they will have to pay out of pocket, and whether their friends are happy with the plan.⁷⁹

HCFA Quality Standards.
HCFA sets standards for HMOs in the Quality Improvement System for Managed Care, a nearly 100-page highly proscriptive plan. Although the QISMC is based on quality improvement measures generally accepted by most experts in that field, it is a particularly burdensome measure. It sets 70 standards for HMOs, prescribes how many quality improvement studies are to be performed, picks some of the issues that HMOs must study (regardless of whether an issue is a concern for a particular HMO), and stipulates exactly how much improvement is to be achieved.

QISMC is nearly impossible for less restrictive managed care options such as PPOs to fulfill. Its prescriptive nature runs counter to the very principle of quality improvement that it promotes. Some HMOs scale back their ongoing quality improvement projects just to comply with its costly requirements.

Even worse, QISMC's proscriptive nature creates an inherent perverse incentive for HMOs to pick less difficult or challenging areas to measure. Plans are required to achieve a 10 percent reduction in adverse outcomes each year. If they fail to meet this target, they could lose Medicare contracts. The result, however, would be a greater number of seniors with fewer choices. By setting a 10 percent target (or any target, for that matter), HCFA removed the market incentive to do better than the chosen target. Because plans are unlikely to choose more difficult or challenging projects, QISMC is little better than a mandate for mediocrity instead of excellence in health plans.

Unrealistic Expectations.
The HMOs, the GAO, and the Medicare Payment Advisory Commission have pointed out several problematic and unreasonable HCFA mandates on Medicare HMOs.⁸⁰ An example is the requirement that the health plans certify that the encounter (claims) data they submit are 100 percent accurate--an unrealistic request. Ironically, even HCFA's Medicare Transaction System (the Medicare claims system) has been criticized for exceeding its budget. Although it was expected to cost $36.6 million, as of the spring of 1997 the program had spent almost $102 million and showed little improvement.⁸¹

HCFA requires plans to perform a costly certification process for each physician--which plans claim costs them at least $80 per participating physician and rarely yields additional useful information.⁸² Curiously, HCFA requires no similar system in the credentialing of providers under traditional Medicare.

HCFA also requires HMOs to contract with a designated subcontractor for the collection of detailed information about a patient's health--including sensitive information about mental health--on consumer satisfaction and health outcomes surveys. This expensive and burdensome data collection system had a lot to do with why the PPOs and PSOs did not enter the Medicare+Choice program. According to Bruce David, senior vice president of Blue Cross and Blue Shield in Florida, QISMC is designed for a "tightly managed HMO product. Broad access PPOs are simply not set up to monitor, measure and assure improvement in enrollee health status and physician outcomes."⁸³