By Kip Sullivan, J.D.
Minnesota Physician, January 2022
But the introduction of HMOs (and after 1997, all types of insurance companies) into Medicare did not cut Medicareās costs. Throughout the half century during which insurance companies have participated in Medicare, they have raised Medicareās costs. They have been paid more, at times much more, to insure their enrollees than if those enrollees had remained in the original Medicare program.
The Medicare Advantage program
Not surprisingly, the large overpayments have facilitated enormous growth in the Medicare Advantage program, the current name of the Medicare program in which insurance companies participate. Today, 46% of all Medicare beneficiaries in the country, and 56% of beneficiaries in Minnesota, are enrolled in one of the Medicare Advantage insurance companies. Minnesotaās rate is second only to the 57% rate for Hawaii. Ten giant insurance companies, led by UHC Healthcare, Humana and CVS Health Corporation, account for four-fifths of the enrollment in Medicare Advantage plans.
How did this happen? Medicare was enacted, after all, because the insurance industry did not want to insure the elderly. And HMOs were allowed into Medicare only because they were expected to lower Medicareās costs. Why has Medicare overpaid insurance companies for half a century?
We can eliminate one explanation immediately: that Congress didnāt know about the overpayments. Beginning in 1980, analysts inside and outside the federal government regularly published research demonstrating that Medicare was overpaying HMOs. Note that the phrase fee-for-service (FFS) refers to the traditional Medicare program in which doctors and hospitals are paid a fee only after they render a service. In contrast, Medicare Advantage insurance companies are paid a lump sum per enrollee in advance. In a 1995 report to Congress entitled āGrowing Enrollment Adds Urgency to Fixing HMO Payment Problem,ā the US General Accounting Office (GAO) stated, āMedicare has paid HMOs more than it would have paid for the same patientsā care by fee-for-service providers.ā In a 2005 report entitled āPayments Exceed Cost of Fee-for-Service Benefits, Adding Billions to Spending,ā the GAO stated, āIt is largely . . . excess payments, not managed care efficiencies, that enable plans to attract beneficiaries by offering a benefit package that is more comprehensive than the one available to FFS beneficiaries, while charging modest or no premiums. Nearly all of the 210 plans in our study received payments in 1998 that exceeded expected FFS costsā¦.ā
The explanation for the overpayments is two-fold: the insurance companies have always enrolled healthier beneficiaries (a problem known as favorable selection), but they have been paid as if they attracted beneficiaries of average health. For example, if the average annual cost of insuring a beneficiary in the traditional Medicare program was $10,000 and United Healthcare (UHC) enrolled healthier beneficiaries whose average cost was $9,000, it could be said UHC enjoyed favorable selection. If the Centers for Medicare and Medicaid Services (CMS, the current name of the agency that runs Medicare) were to pay United $10,000 for each enrollee, the overpayment would equal $1,000 per enrollee. The obvious solution would be for CMS to pay UHC a premium commensurate with the health status of its enrollees ā in this case, $9,000. But CMS has never been able to reduce its payments to insurance companies to match the lower cost of the healthier beneficiaries they enroll.
There is no solution to either problem. The favorable selection problem will persist as long as a Medicare Advantage program exists alongside the traditional program. CMSās effort to improve the accuracy of its payments to Medicare Advantage plans only made the problem worse.
Favorable selection
Medicare beneficiaries and Medicare Advantage plans both contribute to the favorable selection problem.
Medicare beneficiaries contribute because they vary in their willingness to give up choice of provider (doctor and hospital) in exchange for the extra services the Medicare Advantage plans can afford to offer (because they are overpaid). Sicker beneficiaries are less likely to want to join an HMO or any other type of insurance company that limits choice and attempts to influence physician-patient decisions.
The insurance companies aggravate favorable selection by their advertising and recruiting tactics (cherry-picking) and their treatment of the sicker enrollees they cannot avoid (lemon-dropping). Medicare Advantage plans are not allowed to refuse to enroll applicants, but they can encourage healthier beneficiaries to apply by their marketing tactics (where they market, what they say in their advertisements and what they cover at little or no expense to the enrollee), and they can encourage sicker enrollees to disenroll and return to traditional Medicare by denying or delaying services.
Dozens of studies published over the last half century have shown that insurance companies that participate in Medicare have enjoyed favorable selection. Here are five examples spanning the last 40 years.
A study published by Paul Egger in 1980 concluded that beneficiaries who joined Group Health Cooperative of Puget Sound (one of the nationās oldest HMOs) āhad a rate of hospital inpatient use over 50% below the comparison groupā consisting of beneficiaries in traditional Medicare.
A 1988 report to the Health Care Financing Administration (the former name of the agency that runs Medicare) stated that the beneficiaries who enrolled in Medicare HMOs cost 21% less than those who stayed in traditional Medicare.
A study published in āMedical Research and Reviewā in 1997 concluded āin the six months prior to their enrollment, new HMO enrollees use on average 37% fewer services than do beneficiaries in traditional fee-for-service Medicare. Furthermore, HMO disenrollees use 60% more services in the six months after disenrollment than do fee-for-service beneficiaries.ā
A study published in the āNew England Journal of Medicineā in 1997 found HMO enrollees were much healthier than traditional Medicare beneficiaries when they signed up with HMOs, and those that later disenrolled from HMOs were much sicker. The authors reported, āBefore their enrollment in HMOs, the HMO enrollees had a rate of hospital admissions that was two-thirds the rate in the fee-for-service group.The members of the HMO-disenrollment group had a substantially higher rate of [hospital] admissions than the fee-for-service beneficiariesā¦ā The title of this article was, appropriately enough, āThe Medicare-HMO revolving door: The healthy go in and the sick go out.ā
A study published by the Kaiser Family Foundation in 2019 found that Medicare beneficiaries who enrolled in a Medicare Advantage plan were 13% less expensive than the average cost of a traditional Medicare enrollee. In 2016, beneficiaries who stayed in traditional Medicare incurred Part A (hospital) and B (physician) medical expenses of $9,362 on average, while those who signed up with a Medicare Advantage plan cost just $8,109.
Risk adjusting premiums accurately is impossible
If favorable selection and the overpayments have been obvious for four decades, why hasnāt CMS eliminated the overpayments? Why doesnāt CMS just reduce the premiums down to a level commensurate with the better healthāthe lower costāof the insurance companiesā enrollees? Answer: because CMS sets the premiums prior to the year for which the beneficiaries are enrolled, and predicting what each insurance companyās pool of enrollees will cost the next year is a primitive science.
Between 1973 and 2004, CMS (and its predecessor, the Health Care Financing Administration) tried to risk adjust premiums downward using four factors: age, sex, Medicaid status and nursing home status. Although each of these factors is correlated with medical spending (80-year-olds typically cost more than 65-year-olds, for example), the correlation between these demographic factors and spending is very weak. They only predict one percent of the variation in spending. To explain what that means, letās go back to the United Healthcare example. If UHC was overpaid by $1,000, CMSās demographic risk adjuster would only call for reducing UHC premium payments by one percent, or $10, from $10,000 to $9,990. UHC would still be overpaid by $990.
In the Balanced Budget Act of 1997, Congress instructed CMS to improve the risk adjuster by adding diagnoses to the four demographic factors. Itās easy to see the logic in that proposal: People diagnosed with diseases and conditions cost more than people with no diagnoses, and people with serious diagnoses cost more than those with mild or temporary diseases or conditions. Medicare beneficiaries with diabetes, for example, cost 1.5 times as much as the average beneficiary; beneficiaries with a cancer diagnosis cost 1.7 times as much. CMS spent the next seven years preparing a new risk adjuster called the Hierarchical Condition Categories (HCC) risk adjuster. To the four demographic factors, CMS added 3,000 codes for diagnoses.
The addition of all those codes greatly improved the accuracy of the risk adjuster. The new HCC could predict 11%, and within a few years, 12% of the variation in spending between individual beneficiaries. But 12% is still negligible. To return to our United Healthcare example, now CMS can reduce UHCās overpayment by 12% from $1,000 to $880. The overpayment is still immense. Twelve percent is obviously 12 times better than one percent, but it still gives Medicare Advantage plans an enormous incentive to continue cherry-picking and lemon-dropping.
Another way to illustrate the crudeness of the HCC risk adjuster is to demonstrate its disparate impact on the healthy and the sick. According to the Medicare Payment Advisory Commission (MedPAC, an agency created by the Balanced Budget Act of 1997 to advise Congress on Medicare), the HCC results in huge overpayments for the healthiest beneficiaries and enormous underpayments for the sickest. For example, looking at the healthiest 20% of beneficiaries; it overpays by 62% but for the sickest one percent, it underpays by 29%. MedPAC and researchers at CMS who were hired to develop the HCC have made it clear that for both technical and financial reasons it is not possible to improve the HCC by even a small amount.
The HCC was implemented between 2004 and 2007. It backfired immediately. It gave the insurance companies an incentive to make sure every disease and condition is listed in patientsā medical records (which is legal), as well as an incentive to add diagnoses to medical records that are inaccurate (they overstate the seriousness of the condition), inappropriate (the diagnosis was made by someone not authorized to make the diagnosis) or flat out fraudulent. Adding inappropriate or fraudulent codes is called upcoding. Both behaviorsāmore thorough listing of legitimate diagnoses and upcodingā increased CMSās payments to the Medicare Advantage plans more than the increased accuracy of the HCC reduced them.
This conclusion is supported by the Medicare Advantage plansā behavior following the implementation of the HCC. Medicare Advantage plans began to advertise richer benefits for lower premiums, and that in turn caused enrollment to soarāfrom 13% of all beneficiaries in 2005 to 46% in 2020. Congress has tried to address the upcoding issue by instructing CMS to reduce plan payments by a few percentage points each year, but that approach has not stopped upcoding or overpayments. Over the last several years, federal prosecutors have sued several of the larger Medicare Advantage plans for defrauding MedicareāUHC, Anthem and Kaiser Permanente (Kaiser)āfor example. A complaint filed by the US Department of Justice against Kaiser last August stated that each additional diagnosis was worth $3,000 to Kaiser.
How much is enough?
Insurance companies cannot make a profit off Medicare beneficiaries if they are paid no more than what those beneficiaries would have cost had they remained in traditional Medicare. The reason is their high overhead.
UHC and all other Medicare Advantage plans incur administrative costs equal to about 15% of their total Parts A and B expenditures on average. Those costs include the cost of advertising, creating networks of providers that enrollees must use, influencing and overruling doctors, lobbying and profit. Profit constitutes a quarter to a third of that 15% figure. But insurance companies cannot cut spending on hospitals and clinics by 15%. Experts guesstimate they can only cut medical spending by about five percent. Research on how much Medicare Advantage plans reduce utilization of medical services is sparse and inaccurate because plans do not publish information on how they allocate their revenues. This means insurance companies must be overpaid by at least 10% in order to induce them to participate in Medicare Advantage. And since insurance companies need additional overpayments to finance the extra benefits, e.g., dental care, eye glasses, hearing aids, that lure beneficiaries out of the traditional program into Medicare Advantage, the overpayments must be even higher than 10%.
Bruce Vladeck, who administered CMS during Bill Clintonās presidency, discussed this issue in an article he co-authored two decades ago for the journal āHealth Affairs.ā
Managed care plans do not have to be just more efficient than FFS Medicare,ā he wrote, āthey have to be a lot more efficient. With administrative expenses in the range of 8 to 25%, [plans] have to incur medical expenditures 10 to 20 % less than FFSā¦just to break even. In most marketsā¦plans cannot attract enrollees unless they offer additional benefits, which also cost moneyā¦ā
The subtitle of Vladeckās article was āTheory meets reality, and reality wins.ā That accurately sums up the problem: the theory that insurance companies could save Medicare money was always inconsistent with reality. It was based on hype, not evidence. If it wasnāt inconsistent with reality, why did Congress need to enact Medicare in the first place? Congress enacted Medicare because the insurance industry did not want to insure the elderly (or the disabled, who were added to Medicare in 1972). Overpayment is the reason 46% of all Medicare beneficiaries are insured by health insurance companies today, 50 years after Congress first allowed HMOs into Medicare. If they were paid no more than CMS pays for traditional beneficiariesāa requirement Congress wrote into the original 1972 law allowing HMOs into Medicareāthey would not participate.
A deadline looms
Payments to Medicare Advantage plans (including the overpayments) are financed directly by taxpayers and indirectly by lower reimbursements to doctors and hospitals that treat Medicare patients. Taxpayers include: people of all ages who pay the Medicare payroll tax that funds Medicareās Part A trust fund, all Americans who pay federal taxes and all Medicare beneficiaries who pay what are called Medicare Part B premiums. Thus, virtually all health care professionals and all Americans have a stake in terminating the Medicare Advantage overpayments.
It is true that people who enroll in Medicare Advantage plans are getting some of the benefit of the overpayments in the form of coverage for services they couldnāt get if they remained in traditional Medicare. But those services are being purchased by the taxpayer at an unnecessarily high price. Money spent on insurance company administrative costs would be far better spent extending coverage for those same servicesādental, vision, hearing, etc.āto the 38 million beneficiaries in traditional Medicare. After all, those 38 million people paid the same taxes the Medicare Advantage enrollees paid to finance Medicare. Why shouldnāt they enjoy the same coverage? Alternatively, the money currently spent on insurance companiesā overhead could be used to raise reimbursements to doctors and hospitals or could be added back to Medicareās trust funds to lower costs for future taxpayers and Medicare beneficiaries.
Congress will soon have an opportunity to debate the overpayments. Medicareās trustees predict that the Part A trust fund (the one that funds hospital services) will run short of revenues beginning in 2026. Part A is financed by a payroll tax paid half by employers (1.45%) and half by employees (1.45%). Congress will be under great pressure to raise that tax as 2026 approaches. Medicare Advantage plans will probably not lobby Congress to raise the payroll tax (at least not openly), but we can be sure they will tell Congress not to reduce their overpayments.
Next steps
Taxpayers of all ages should let members of Congress know how they feel about the overpayments. Unfortunately, the American taxpayer is not represented by well-heeled organizations. But physicians and hospitals do have trade associations with a long history of representing their members in Congress. The American Medical Association, the American Hospital Association and the numerous other organizations that represent the professionals and institutions which treat patients insured by Medicare should speak out now in favor of terminating the overpayments to Medicare Advantage plans. They should also promote legislation that ensures that all Medicare enrollees have access to the extra benefits Medicare Advantage plans offer now, thanks to their overpayments. Doing so would undermine the counterargument we can expect from the insurance industry: That ending the overpayments means ending the extra services the overpayments finance. Change happens only when Congress feels pressure for change. The public and providers must let Congress know they want the overpayments stopped.
Kip Sullivan, J.D., is a member of the Health Care for All Minnesota Advisory Board. He was a member of Gov. Perpichās Health Plan Regulatory Reform Commission. His articles have appeared in the New England Journal of Medicine, Health Affairs, and other peer-reviewed journals.