VBID in Medicare: Will CMS Look Before It Leaps?

In recent testimony before the U.S. House Ways and Means Subcommittee on Health, A. Mark Fendrick, Director of the Center for Value-Based Insurance Design (VBID) at the University of Michigan, advocated for what he described as “the essential role of clinical nuance” in Medicare, specifically “reducing financial barriers to evidence-based services and high-performing providers” while “imposing disincentives to discourage use of low-value care."[1] Fendrick’s proposal included two major components as “an interim step” to be applied initially only within Medicare Advantage Plans (MAPs): first, allowing MAPs to use higher copayments to deter use of “low-performing” providers, rather than simply excluding those providers from their networks as MAPs are currently permitted to do; and second, allowing MAPs to vary copayments for medical services or drugs based on their clinical value. The latter idea is relatively old, as current health policy proposals go—it was first described in 2001 as a “benefit-based copay for prescription drugs,” in which medication copayments were to be based on clinical value rather than drug cost. The theory is that if higher-value treatments (e.g., beta blockers following a heart attack) are available for a lower copayment, patients will be more likely to use them.

Fendrick set forth his proposal for what he described as “the common-sense step of allowing co-payments to vary based on whether an intervention is high-value or low-value” as part of a broader vision for the U.S. health care system as a whole. If we were to “[reallocate] health care dollars to services for which there is clear evidence for improving clinical outcomes,” Fendrick argued, “we could simultaneously enhance quality and reduce the amount we spend.” Quite a promise—but assuming that this assessment is accurate, will lowering copayments for “high-value” services and increasing copayments for “low-value” services help us get there?

In an evidence-based health care system, the answer should depend on the findings of high-quality research on the degree to which copayment changes actually influence the behavior of insured beneficiaries. I’ll discuss this research—and what it teaches us about how to evaluate and use study findings in making policy decisions—in more detail in coming posts. But for now, consider that the “growing body of evidence” for VBID mentioned in Fendrick’s testimony includes primarily observational (nonexperimental) studies, which earn generally low quality ratings in evidence-grading systems.[3] Stronger, experimental evidence about VBID is limited to a single randomized controlled trial (RCT) conducted in a commercially insured, non-Medicare population of heart attack survivors.[4] Some good-quality observational data on cost-sharing in Medicare samples are available.[5] However, the refutation of even the highest-quality observational findings by stronger RCT evidence has become so commonplace as to be the rule rather than the exception in health care—as is well known to those familiar with ACCORD, ADVANCE, and the WHI, to cite just a few examples.[6]

So, in considering VBID at the present time, the Centers for Medicare & Medicaid Services (CMS) is faced with limited evidence coupled with enthusiastic advocacy. It may seem like an unusual combination, but CMS has been here before, about a decade ago. At that time, advocates of providing population disease management (PDM) programs to Medicare beneficiaries with chronic diseases pointed to numerous, mostly observational, studies documenting positive health outcomes and return-on-investment (ROI) estimates of up to $11-$14 in medical cost reductions.[7] The PDM proposal made good common sense because it was widely recognized that beneficiaries with chronic disease were “a large and costly subgroup of the Medicare population” that struggled to “navigate a system … structured and financed to manage acute, rather than long-term, health problems.”[7] PDM, which provides user-friendly health education and a readily available source of advice outside of physician office hours, seemed ideally suited to address the problem. According to Senate testimony provided in 2002 by Dan Crippen, then Director of the Congressional Budget Office, “the long-range fiscal challenges facing the Medicare program” made the prospect of adding a PDM benefit to Medicare “tantalizing” because “proponents claim that such a benefit would improve the quality of care that beneficiaries receive and at the same time reduce federal costs.”[8]

Sound familiar?

But rather than adopt the widely touted PDM benefit without rigorously testing it first, CMS initiated the Medicare Health Support (MHS) experiment in 2005 to randomize about 240,000 beneficiaries with diabetes and/or congestive heart failure to receive PDM provided by nurse-based call centers or usual care. The MHS results were both disappointing and inconsistent with the weaker research evidence that had been enthusiastically highlighted by PDM benefit supporters. Far from producing an enormous ROI because of better medical outcomes, PDM had no significant effect on hospitalizations, emergency room use, or total health care cost.[9]

What can we take away from the MHS example? As Casazza et al. observed in a recent report on “myths, presumptions, and facts about obesity,” it has long been recognized in medical science that circumstances sometimes necessitate policy adoption absent experimental evidence—specifically, when conducting an RCT would be infeasible, unethical, or unnecessary because “observed associations are not plausibly due to confounding.”[10] However, Casazza et al. pointed out that this recognition has sometimes been used as an excuse for accepting weak evidence “by those who are eager to garner support for a proposal in the absence of strong data from randomized studies.” In other words, when RCTs are feasible and there is a reasonable possibility that observational evidence has been contaminated by confounding effects (usually baseline differences among study groups that cannot be statistically controlled in analysis), methodologically weaker evidence should be viewed as unacceptable for the purposes of public policy making. The MHS experiment reflected this view.

Such should be the case with consideration of VBID in Medicare. An RCT is feasible, there is a possibility that observational evidence will be overturned by stronger experimental data, and the stakes are high. Consider that, at about the time that the decision to conduct the MHS trial was made, the most recent data available showed prevalence rates of 14% and 18% for CHF and diabetes, respectively, in the Medicare fee-for-service beneficiary population.[7] With monthly administrative fees of $74 to $159 paid to PDM vendors in the MHS[9], 85% participation[9], and a 2005 enrollee count of about 43.4 million[11], a “back-of-the-envelope” calculation assuming an overall prevalence of about 25% for both diseases (combined) suggests that CMS would have spent an additional $8-$18 billion annually had a PDM benefit been adopted as suggested by advocates—with no discernible health improvement for Medicare beneficiaries. For this reason, the decision to conduct the MHS provides us with a good example of evidence-based policy making in action and, perhaps even more importantly, of why health care research methods matter.

[1] Fendrick AM. The essential role of clinical nuance in Medicare’s benefit design. Testimony before U.S. House of Representatives Ways and Means Subcommittee on Health. February 26, 2013.

[2] Fendrick AM, Smith DG, Chernew ME, Shah SN. A benefit-based copay for prescription drugs: patient contribution based on total benefits, not drug acquisition cost. Am J Manag Care. 2001;7(9):861-67.
[3] McAlister FA, van Diepen S, Padwal RS, Johnson JA, Majumdar SR. How evidence-based are the recommendations in evidence-based guidelines? PLoS Med. 2007;4(8):1325-32; and Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) Working Group. Grading quality of evidence and strength of recommendations. BMJ. 2004;328:1490-94.
[4] Choudhry NK, Avorn J, Glynn RJ, et al. Full coverage for preventive medications after myocardial infarction. N Engl J Med. 2011;365:2088-97.
[5] For example: Tseng CW, Brook RH, Keller E, Steers WN, Mangione CM. Cost-lowering strategies used by Medicare beneficiaries who exceed drug benefit caps and have a gap in drug coverage. JAMA. 2004;292(8):952-60; Raebel MA, Delate T, Ellis JL, Bayliss EA. Effects of reaching the drug benefit threshold on medicare members’ healthcare utilization during the first year of Medicare part D. Med Care. 2008;46(10):1116-22; Trivedi AN, Moloo H, More V. Increased ambulatory care copayments and hospitalizations among the elderly. N Engl J Med. 2010;362:320-28.
[6] Rossouw JE, Anderson GL, Prentice RL, et al.; Writing Group for the Women’s Health Initiative Investigators. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women’s Health Initiative randomized controlled trial. JAMA. 2002;288(3):321-33; ADVANCE Collaborative Group, Patel A, MacMahon S, Chalmers J, et al. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med. 2008;358(24):2560-72; Action to Control Cardiovascular Risk in Diabetes Study Group, Gerstein HC, Miller ME, Byington RP, et al. Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med. 2008;358(24):2545-59.
[7] Cromwell J, McCall N, Burton J. Evaluation of Medicare Health Support chronic disease pilot program. Health Care Finance Rev. 2008;30(1):47-60; Foote S. Population-based disease management under fee-for-service Medicare. Health Aff (Millwood). 2003 Jul-Dec;Suppl Web Exclusives:W3-342-56.
[8] Crippen DL. Disease management in Medicare: data analysis and benefit design issues. CBO Testimony before Special Committee on Aging, United States Senate. September 19, 2002.
[9] McCall N, Cromwell J. Results of the Medicare Health Support disease-management pilot program. N Engl J Med. 2011;365(18):1704-12.
[10] Casazza K, Fontaine KR, Astrup A, et al. Myths, presumptions, and facts about obesity. N Engl J Med. 2013;368:446-54.
[11] Henry J. Kaiser Family Foundation. State Health Facts. Total number of Medicare beneficiaries, 2005.