Blog Home


Medicare Advantage: Facts, Fallacies, And The Future

March 14th, 2011

Medicare Advantage (MA) has been at the center of partisan debate on the tradeoff between preserving Medicare solvency and protecting consumer choice. CMS’s recent release of preliminary payment policies for Medicare health plans — that is, plans offered through Medicare Advantage (MA) — is likely to refuel the ongoing debate on MA payments that continues despite efforts to resolve it in the Patient Protection and Affordable Care Act of 2010 (ACA).

At issue is the equity of MA payment reductions included in the ACA. Democrats argue that the changes establish closer parity between MA and the traditional program, correcting overpayments to MA that add to program costs and make it harder for the traditional program to compete. Republicans argue that the changes will remove important choices for beneficiaries that increase overall value in the Medicare program.  Both sides cite independent analysis to support their positions, particularly the Medicare Payment Advisory Board’s (MedPAC’s) documentation of the overpayments and the Congressional Budget Office’s (CBO’s) estimated dramatic declines in MA enrollment with the legislation.

With colleagues, I have been following the history of the MA program for many years both in its current incarnation and in its earlier lives (Medicare+Choice, Medicare risk contracting).  In this post, I try to cut through the rhetoric by summarizing what we know about the Medicare Advantage program and how beneficiaries are likely to be affected, positively or negatively, by health reform.

Medicare Advantage Today And Its Antecedents

MA payment policy has emerged over time through the cumulative, and perhaps unintentional, effect of policies seeking different goals. These policies converted a program that was originally viewed as generating savings for the Medicare program into one that paid more when beneficiaries selected a private plan over traditional Medicare. The health reform legislation sought to reverse this trend, generating savings that in part would offset the cost of expanded coverage to the under 65 population nationwide.

As of February 2011, 12.2 million Medicare beneficiaries were enrolled in Medicare Advantage and similar plans authorized under other authority. This reflects 25 percent of all Medicare beneficiaries and 36 percent  beneficiaries who enroll in Part D, Medicare’s prescription drug benefit. The MA market historically has been a particularly important source of coverage for beneficiaries who do not have access to subsidized retiree benefits through employers or Medicaid, particularly those with lower to moderate income.

Program enrollment has grown substantially since the Medicare Modernization Act of 2003 modified policies guiding plan payments. Between year-end 2003 and year-end 2010, enrollment more than doubled, growing from 5.5 million to 11.8 million. Though the majority of enrollees were in HMOs, the growth was heavily driven by enrollment in newer types of MA plans, particularly private fee for service (PFFS) plans. MA enrollment continues to grow despite legislative changes that have diminished the attraction of PFFS. Enrollment is increasingly is concentrated in a small number of major firms that historically have dominated the market  including UnitedHealthcare, Humana, Kaiser Permanente and Blue Cross-Blue Shield affiliates). MA penetration continues to be uneven across the country for a variety of sometimes idiosyncratic factors.

Evolution Of Current MA Payment Policy

The Medicare risk contracting program (1985-1997) was the first permanent private plan program in Medicare, providing access to HMOs. Program payments were linked to historical Medicare costs for similar beneficiaries in the same county, with payments pegged to 95 percent of costs and the plans required to use any savings between that and benefit costs to enhance options for beneficiaries (lower cost sharing, reduced premiums, extra benefits).

In 1997, this program was absorbed into the Medicare+Choice program as a result of the Balance Budget Act of 1997 (BBA). Medicare+Choice expanded the type of private plans available to Medicare beneficiaries and also made a series of changes in payment. While the intent was to gradually reduce payment variation across counties, the main operational effect in most counties was to limit payment increases to 2 percent annually. The BBA also added “floor”, or minimum payment levels, in rural counties. (Two years later authority for a second, and higher, “urban floor” was added.). These changes broke the link between fee-for-service (FFS) and MA payment levels in counties, and initiated changes which led to some counties being paid substantially more than FFS. The BBA also authorized the phase-in of a strengthened program of risk adjustments in rate setting that better reflected differences in the health status of enrollees in the programs. These adjustments are important since research on Medicare HMOs showed Medicare overpaying private plans by failing to adequately adjust for the health status of patients enrolled in them.

Enrollment in Medicare’s private plans declined substantially after the Medicare+Choice program was adopted, changes perceived as heavily driven by the slow growth in MA payments over this period. The Medicare Modernization Act of 2003 sought to stabilize and expand the MA program in anticipation of the introduction of the new Part D benefit in 2006. Under the Act,  the minimum annual increase would be the greater of either 2 percent  or the increase in the national growth rate, a change which led to substantially higher rates of increase than in previous years. The Act also raised the minimum county payment to at least 100 percent of FFS which, combined with other changes, meant that MA payments in each county exceeded those paid by the traditional program for similar beneficiaries.

According to MedPAC, the cumulative effect of all these changes in 2010 was that per capita payments to plans averaged an estimated  109-113 percent depending on assumptions made about physician payment in the traditional Medicare program. Based on MEDPAC  estimates, plans cost for Medicare Part A and Part B were at least the same as in traditional Medicare, with HMOs costing a few percent less and other plan types substantially more.

The health reform law sought to reduce the disparity in payment between traditional Medicare and MA. The Act froze MA plan payments in 2011 at the 2010 rate. In the following years, payments will be brought down by an amount that will vary with the current generosity of payments across geographic areas. Counties are divided into quartiles, based on current payments relative to fee for service in those counties. The highest paid counties will have benchmarks at 95 percent of FFS and the lowest at 115 percent, with the others in between. Plans that perform well on quality scores can offset some of the reduction with additional bonus payments for quality.

Likely Positive and Negative Effects of These Changes

Payment reductions in MA should enhance the fiscal solvency of the Medicare program because Medicare will no longer pay more when a beneficiary chooses Medicare Advantage over staying in the traditional program. The change also is likely to have a direct effect on the Part B premiums Medicare beneficiaries pay by slowing down the rate of growth in the program (According to MedPAC, the average Medicare beneficiary paid $3.34 more in Medicare Part B premiums each month in 2009 because of MA overpayments, with Medicare paying on average $1.08 for each $1 of extra benefits MA enrollees received.)

Beneficiaries currently enrolled in MA plans will be the ones to experience the main adverse effects of the payment changes. With payments rising less rapidly than in the past, plans are likely to eliminate or reduce some of the extra benefits that overpayments helped finance and premiums may rise.  However MA plans will still have to cover required Medicare benefits, including new preventive benefits authorized by the ACA. The change also creates more equity geographically since payment reductions will disproportionately be made in counties currently  most out of line with spending in the traditional program.

Proponents of the change believe this is equitable and creates a more level playing field between traditional Medicare and Medicare Advantage. Opponents disagree, arguing that payment reductions will make it less attractive for private plans to participate in Medicare, reducing beneficiary choice with effects targeted most heavily on the low to moderate income population attracted by MA.

Separating  Fact from Rhetoric

Proponents of higher MA payments argue that the differential is justified because private plans are more efficient, so it is worth paying more to encourage plan choices that may ultimately lead to a stronger Medicare program. However there is limited evidence to support this position.

  • Existing research provides no evidence that either traditional Medicare or Medicare Advantage perform consistently perform better than the other. From a cost perspective, HMOs in MA may be slightly more efficient on average, but other types of plans appear to cost more than traditional Medicare. Analysts might argue there is a wash between what some see as the greater flexibility in MA to innovate in delivery and traditional Medicare’s ability to use its market power to control provider payment rates. Obviously, there is considerable variation across the industry, with some plans undoubtedly better and some worse, at least on selected metrics.
  • The fear that MA payment reductions will dramatically reduce available private plan options for Medicare beneficiaries also appears, at least to date, overstated. While such a reduction may appear likely given historical experience under Medicare+Choice, MA enrollment to date shows continued growth in MA enrollment. Despite the freeze in payments, beneficiaries also still have many plan options in 2011 and though the number is less than in recent years it remains  far in excess of those generally available in the commercial market. Though benefit cuts through MA may target those less able to bear them, it probably is more equitable to enhance the overall Medicare benefit package or add to current programs targeting extra benefits for low income people than to tie receipt of such benefits to enrollment in MA and geographically uneven overpayments.

Projecting Future Change

While firms’ interest in MA participation obviously is affected by payment rates, the Medicare market may be more attractive now than it was previously. With an aging population, more and more beneficiaries ultimately will be in Medicare; at the same time privately sponsored insurance is declining. Further, there is the potential for growth in private plan products geared toward the dually eligible population enrolled both in Medicare and Medicaid.  If the alternative for beneficiaries is Medigap, firms also may have more opportunity to raise premiums without detracting from their market share. MA enrollment also is heavily concentrated in a small number of firms that are likely to find it hard to walk away from the Medicare market.

MA enrollment could hold steady or even increase. While CBO estimates assumed MA enrollment will decline dramatically as payment increases are limited, it is possible that CBO’s judgment that future trends will mirror historical experience under Medicare+Choice will not prove correct. For example, in their 2011 bids firms projected a net increase, rather than decrease,  in MA enrollment that at least some industry analysts project could continue in 2012. MA is arguably a more attractive product now to beneficiaries than it was previously as a result of changes in the economy. MA’s premiums arguably make the product more attractive now than previously compared with the higher priced Medigap benefit, especially for those with limited health care needs. MA enrollment also greatly simplifies choice for beneficiaries who enroll in an MA-PD plan and avoid the need to juggle Medigap supplements and, since 2006, Part D. MA’s PPO options also may be attractive to those previously discouraged by the closed HMO networks.

The Bottom Line

In conclusion, MA payment reductions included in the ACA have an empirical base and will lead to a stronger and more equitable Medicare program.  There also currently is little evidence to support concerns that the MA program is imploding, though that cannot be ruled out. While those opposing the MA cuts raise some legitimate points, their arguments seem insufficient to justify the additional resources it would take to offset the cuts at a time when federal budget expenditures of all kinds are under scrutiny.

Email This Post Email This Post Print This Post Print This Post

 to the #1 source of health policy research.

7 Trackbacks for “Medicare Advantage: Facts, Fallacies, And The Future”

  1. Health Policy Schizophrenia | The Health Care Blog
    August 22nd, 2011 at 9:15 am
  2. Health Policy Schizophrenia | John Goodman's Health Policy Blog |
    August 17th, 2011 at 8:12 am
  3. Medicare Advantage Redux | John Goodman's Health Policy Blog |
    March 24th, 2011 at 8:16 am
  4. Health Wonk Review Review: Medicaid, HRAs, & P4P - Avik Roy - The Apothecary - Forbes
    March 18th, 2011 at 10:31 am
  5. Health Wonk Review: Spring Training Edition | Pizaazz
    March 17th, 2011 at 7:12 am
  6. Medicare Advantage: Facts, Fallacies, And The Future – Health Affairs Blog « JHPPL News and Notes
    March 16th, 2011 at 6:06 am
  7. Reform Rodeo : HEALTH REFORM WATCH
    March 14th, 2011 at 11:56 pm

3 Responses to “Medicare Advantage: Facts, Fallacies, And The Future”

  1. Marsha Gold Says:

    Thank you, Jeff, for pointing out emerging work that may expand our future knowledge. In considering this work, I want to make three brief points.

    First, research on Medicare Advantage has been woefully limited in recent years despite the increasingly large role played by this program. Thus, new research on Medicare Advantage plans that AHIP is sponsoring is valuable and will be worth monitoring as it proceeds through the peer review process. Such review is critical given the known challenges in conducting high quality research on Medicare managed care where data are limited, plans unique, and systematic differences exist between populations in MA and fee-for-service that risk adjustment may not address. Independent peer review is particularly valuable when the authors or sponsors of the research, whoever they are, have a vital interest in the outcomes of that research.

    Second, these studies generally are not applicable to the most rapidly growing and least understood kinds of MA plans. They largely reflect the experience in HMOs, which are generally the tightest form of managed care. In contrast, much of the current growth in MA has been in PFFS plans and more recently PPOs. Very little is known about the effectiveness of these plans, but MedPAC analysis of their bids for covering traditional Medicare benefits show that they have higher costs than HMOs.

    Third, the new study results, whether confirmed or not, will not change the policy problem facing Congress, which is to reduce the rate of growth in Medicare spending. Even if care were marginally better in MA than in FFS, the rate Medicare was paying to MA before the recent change was not sustainable or justified. The promise of MA is that they should be able to reduce costs by improving care and reducing the incidence of preventable hospitalizations, and they have the incentive to do so, given the fixed monthly fee they receive from Medicare. If the papers cited are indeed credible evidence that MA accomplishes this objective of reducing unnecessary resources, a substantial share of these savings need to flow back to the federal government. For that to happen under a capitated arrangement, the rate must be at, and ultimately below, the rate that is equivalent to what FFS pays on average for beneficiaries with comparable health status.

  2. jeff lemieux Says:

    The author notes that “[e]xisting research provides no evidence that either traditional Medicare or Medicare Advantage perform consistently perform better than the other.”

    This is not an uncommon refrain.

    However, it may no longer be accurate.

    To be sure, the new evidence is preliminary. The studies and working papers are sponsored by the health insurance industry and are still in various stages of peer review. Most have focused on hospital readmission rates; the studies using data acquired directly from Medicare Advantage (MA) plans also included rates of hospitalization in general, office and outpatient visits, Emergency Room (ER) use, and AHRQ-defined potentially avoidable admissions.

    Study 1 examined diabetes and heart disease patients in eight mostly smaller or regional health plans, and compared with FFS patients in the same counties. The study population was full-year enrollees aged 65 through 89, and the risk adjustment used was the CMS Age, Sex, and Hierarchical Condition Category (HCC), or “risk-score” method.

    Working Paper: A Preliminary Comparison of Utilization Measures Among Diabetes and Heart Disease Patients in Eight Regional Medicare Advantage Plans and Medicare Fee-for-Service in the Same Service Areas (revised 09.01.2009) [link:

    Study 2 examined patients age 65-89 admitted to California and Nevada hospitals using AHRQ’s HCUP hospital discharge datasets, and a modified risk-score method of adjustment.
    Reductions in Hospital Days, Re-Admissions, and Potentially Avoidable Admissions Among Medicare Advantage Enrollees in California and Nevada, 2006 (revised 10.23.2009) [link:

    Study 3 added two large, multi-region, for-profit MA companies to the study 1 population.
    Working Paper: Comparisons of Utilization in Two Large Multi-State Medicare Advantage HMOs and Medicare Fee-for-Service in the Same Service Areas (12.11.2009) [link:

    Study 4 added seven states to the analysis of AHRQ data, using additional data from ARHQ and also data acquired directly from two large states (Texas and Pennsylvania), and featured an extended discussion of the issues involved with using the hospital discharge datasets for comparisons and an additional method of risk adjustment based on DRG distributions.

    Working Paper: Using State Hospital Discharge Data to Compare Readmission Rates in Medicare Advantage and Medicare’s Traditional Fee-for-Service Program (05.14.2010) [link:

    Study 5 used AHRQ’s new “revisit indicators” to precisely track risk-adjusted readmission rates by day length from six states.

    Using AHRQ’s ‘Revisit’ Data to Estimate 30-Day Readmission Rates in Medicare Advantage and the Traditional Fee-for-Service Program (10.12.2010) [link:

    Study 6 used data from 11 (mostly) large MA plans in the MedAssurant MORE2 database, and compared 30-day readmission rates using the “Jencks method” (as described in an article by Steve Jencks and colleagues in the New England Journal of Medicine, April 2009) and an alternative method developed by Gerry Anderson of Johns Hopkins and his colleagues that allowed contemporaneous comparisons and risk adjustment by DRG distribution. (This study is not yet published, but is available in draft form from the authors.)

    In addition, in 2010, AHIP published a qualitative study reporting on programs used to reduce readmission rates, unnecessary ER visits, and preventable hospitalizations.

    Innovations in Reducing Preventable Hospital Admissions, Readmissions, and Emergency Room Use (06.07.2010) [link:

    Some of the interventions described in the report are as fundamental as: having a nurse available to answer questions 24 hours a day; informing clinicians that patients have been hospitalized; calling the patients or their families shortly after discharge to check for problems and help schedule follow-up care; stationing nurses at hospitals to assist with coordination of discharge and follow-up care. Others are based on proved transitional care programs, such as the Naylor nurse practitioner model or the Coleman health coaching system./1 One plan sent a pharmacist to the home within 3 days of hospital discharge to ensure that patients’ medications were reconciled.

    Two follow-up reports are in the works: 1) a study discussing methods of assessing risk-adjusted, hospital-specific readmission rates and tracking them over time, and 2) a study that attempts to determine whether transitional care programs or “network steering” to more effective hospitals explains the difference in readmission rates among MA and FFS patients – this study examines risk-adjusted MA and FFS readmission rates at the same hospitals.

    The comparison methods developed for these studies are now being expanded beyond readmission measures. Follow-on studies are also underway to examine rates of adverse events (such as C. Difficile or MRSA). Future studies are planned to study mortality rates associated with these events directly, for example, mortality associated with extended ICU stays.

    The industry studies have been presented in a series of seminars, beginning prior to the publication of the first working paper [link:, to public discussions at industry conferences, such as the 2010 National Readmissions Summit [link:
    We have also given public seminars at the Harvard medical school department of health care policy, and at a faculty workshop on health care policy sponsored by the Dartmouth Institute for Health Policy and Clinical Practice and the Nelson A. Rockefeller Center. The data have been presented at private seminars for the staffs of MedPAC, CBO, and CMS.

    Thus far, the new data have proven consistent across many states and datasets, different measures, and different methods of risk adjustment. A recent, unrelated paper by Niall Brennan and Mark Shepard of the Brookings Institution used an innovative approach to compare HEDIS scores for various quality process measures across MA and FFS Medicare — the results also showed gains in MA.

    Of course, one could always argue that risk adjustment could be improved, and this has been mentioned often in review, especially by clinicians. While risk adjustment will never be perfect, we believe that the DRG-distribution method, in particular, is a very important methodological advance, especially when measuring readmission rates per discharge or directly following a discharge (as opposed to across a population of enrollees, some of whom might not have been hospitalized). And no plausible method of risk adjustment could explain the magnitude of some of the cohort differences we are measuring.

    We are, of course, well aware of the profoundly disturbing implications of these data — this alone is reason for caution. For example, even with risk adjustment, we are finding that readmission rates per admission consistently fall between 13 and 20 percent lower in MA than in FFS. The data imply that patient treatment and outcomes are far from exogenous. The Dartmouth Atlas has shown unexplained differences in treatment patterns across regions, which may not always be beneficial to patients or related to local preferences or health care needs. These new comparisons between MA and FFS imply that many patients in Medicare’s FFS program may be getting lower quality care relative to their neighbors in the same states and regions.

    However, this may be changing. Preliminary data from Medicare’s Care Transitions Project indicate that patient interventions used by MA plans to reduce readmission rates can also work in FFS settings. These efforts, which are sponsored by several Quality Improvement Organizations (QIOs) across the country, indicate that improving communication among health care providers and using proven transitional care interventions can play a key role in reducing readmission rates. For example, in one part of western Pennsylvania, FFS readmission rates were reduced from 18 percent to 14 percent using staff of the local Area Agencies on Aging trained in the Coleman health coaching model./2 Preliminary results from similar interventions provided to FFS Medicare patients in Colorado are also encouraging./3

    Jeff Lemieux

    /1 Boutwell A, Hwu S. Effective Interventions to Reduce Rehospitalizations: A Survey of the Published Evidence. Cambridge, MA: Institute for Healthcare Improvement. 2009. Available from:

    /2 Personal communications with Dr. David Wenner, November 30, 2010. Dr. Wenner’s preliminary results from the Care Transitions Project in Pennsylvania were first presented at the conference: Optimizing Home Health in Care Transitions, 2010 Summit, October 26, 2010, Philadelphia PA. Slides from that conference are available from the authors or from Dr. Wenner. For more information about the Care Transitions Project, please see

    /3 Booth M, “Medicare spreads savings from Denver program,” Denver Post, 2010 Dec 13. Available from: Postings of preliminary results and statements from the Care Transitions Quality Improvement Organization Support Center (QIOSC) are available from: A slide presentation of preliminary results was posted on December 16, 2010 at:

  3. TRL Says:

    One point you left out: the opportunity for health plans to make up for the cuts in Medicare Advantage by hitting quality targets and qualifying for quality bonuses.

Leave a Reply

Comment moderation is in use. Please do not submit your comment twice -- it will appear shortly.

Authors: Click here to submit a post.