• What if Medicare’s drug benefit was more like the VA’s?

    In his budget deficit speech yesterday, Obama said, “We will cut spending on prescription drugs by using Medicare’s purchasing power to drive greater efficiency.” The accompanying White House fact sheet says, “The framework would limit excessive payments for prescription drugs by leveraging Medicare’s purchasing power – similar to what was called for by the bipartisan Fiscal Commission.”

    But the bipartisan Fiscal Commission report says no such thing. About the Medicare drug benefit (Part D), it says only,

    Drug companies are required to provide substantial rebates for prescription drugs purchased by Medicaid beneficiaries. We recommend extending these rebates to Medicaid beneficiaries who are also eligible for Medicare (individuals known as “dual eligibles”) and who receive prescription drug coverage through the Medicare Part D program.

    That’s got nothing to do with leveraging Medicare’s purchasing power. Thus, if Obama’s statement about “purchasing power” of Medicare means anything for the Part D program, it’s something neither he nor the Fiscal Commission has specified yet. But what? My sample is small, but some have speculated he wants to make Medicare’s drug program more like the VA’s. If you know anything about the VA’s drug benefit, you’ll know that this would be a huge change for beneficiaries and drug manufacturers, and therefore politically unlikely. Yet, it is worth asking, what would happen if Medicare did buy drugs like the VA does?

    Turns out, with colleagues Steve Pizer and Roger Feldman, I asked and answered that question in a paper titled “Should Medicare Adopt the Veterans Health Administration Formulary?” (funded by the Department of Veterans Affairs and the Robert Wood Johnson Foundation). It is soon to appear in the journal Health Economics. Journal editors have suggested I release a working paper version now (ungated) and discuss it in advance of publication on this blog. Below is a summary. The details and limitations are in the paper.


    Since January 2006 all Medicare beneficiaries have been eligible to obtain outpatient prescription drug coverage through private health plans. The Part D drug program made benefits available through HMOs and PPOs under the Medicare Advantage program (MA-PD plans). In addition, it includes new prescription drug plans (PDPs), which offer stand-alone drug coverage in multi-state regions.

    In the paper, we compute the savings to the Medicare program and the loss of value (formally, consumer surplus) to beneficiaries due to tightening Part D formularies to the level found in the Veterans Health Administration (VA). (A formulary is a list of drugs covered by a health plan.) We measure formulary generosity as the percentage of the 200 most popular drugs covered. The VA’s national formulary covers 59% of the top 200 drugs while Medicare PDPs cover between 68% and 93% of those drugs, averaging about 85% covered. So, if Medicare plans looked more like the VA, a lot fewer drugs would be covered.

    But, the tighter the formulary (the less drugs it covers) the more bargaining leverage a plan has with respect to drug manufacturers. Plans able to restrict drugs from their formularies have the clout to say “no” to high prices. This is one, but not the only, reason the VA can purchase drugs at prices 40% below those paid by Medicare Part D plans. If Medicare drug plans restricted their formularies to the level of generosity offered by the VA and obtained VA-like drug prices by doing so, we estimate that the program would save $510 per beneficiary per year or a total of $14 billion per year (2009 prices).

    However, in tightening formularies, beneficiaries would lose low-cost access to many drugs. That loss of choice is worth something, and it can be monetized using econometric techniques. Doing so, we estimate the loss of choice to be valued at $405 per beneficiary per year. Because the savings ($510 per beneficiary) exceeds the loss of value to beneficiaries ($405), they could, in principle, be made whole with $105 left over (= $510 – $405). This could be done by lowering premiums, for example.

    It’s worth asking, why are Part D formularies so generous? The reason is that a minimum of two drugs in each class must be included on formularies and six classes must include “all or substantially all” drugs on the market. Because of this, providing Medicare the authority to negotiate directly with manufacturers would not lead to price reductions on its own. To achieve savings, Medicare or its participating plans would also need the ability to exclude drugs from its formulary. This ability to tighten formularies would provide the leverage to bargain for lower prices.

    Medicare’s inability to negotiate program-wide prices and tighten plan formularies is in stark contrast to the VA, which negotiates directly with drug manufacturers and is not bound by the same formulary rules as Part D plans. That’s why the VA has been able to implement a national formulary more restrictive than those of Medicare plans and obtains lower drug prices. If Medicare plans could implement VA-like formularies and obtain commensurately lower prices, our paper shows that enough could be saved to compensate beneficiaries for the loss of choice, with savings to spare.

    To repeat, the key findings are:

    • The VA pays 40% less than Medicare plans for prescription drugs.
    • Medicare plans cover about 85% of the most popular 200 drugs on average (ranging from a low of 68% to a high of 93%).
    • The VA’s national formulary includes 59% of the most popular 200 drugs.
    • If Medicare obtained the same drug prices as the VA, it would save $510 per beneficiary per year or a total of $14 billion per year (2009 prices).
    • If Medicare plans tightened formularies to the level of generosity available from the VA (59% of top 200 drugs covered), beneficiaries would lose $405 of value per year associated with the loss of choice of drugs. (The right way to interpret this is that the average beneficiary would be precisely indifferent between the loss of drug choice and $405 dollars in cash.)
    • Because the savings ($510 per beneficiary) exceeds the loss of value to beneficiaries ($405), they could, in principle, be made whole with $105 left over (= $510 – $405).
    • This is an interesting economic analysis but it ignores the therapeutic value of the formulary. There is a reason the VA excludes so many drugs from their formulary and it is because these drugs are ineffective or overpriced and can be replaced with cheaper drugs which have the same or better beneficial effect for the patient.
      Therefore, the statement that “there is a loss of value of $405 to patients” is only true in the sense that patients can’t buy overpriced and/or ineffective drugs. There is no loss of therapeutic value. If you take into account the fact that patients can be as healthy using the restricted formulary then there is no loss and no need to compensate them for the loss.

      • You bring an additional consideration, which is valid. I do not think we ignore it. Read the conclusion of the paper. However, the paper is economics. We are economists. You or someone equally qualified should write a paper on the point you have made.

      • Another value of the formulary is an increase in safety. A hot, new, expensive drug has no record of wide-spread use and may have side-effects yet to be discovered.

    • I am eligible for both Medicare and VA. The VA formulary sometimes seems needlessly burearocratic and grim but it is adequate. I am impressed with their all over results AND costs. I understand the VA provides medical care for about 20% less than the standard US amount. And I suspect the average person in VA is much older and likely in poorer health than the average US citizen.

      • Fascinating. I don’t think I’ve ever commented on the 20% number in blog public, so I’m about to do so — but cautiously. I’m curious to know why you in particular cite that VA is 20% cheaper and in what sense you understand it. Large numbers like these have floated around for at least 25 years, mostly incorrectly. It most definitely is not the case that for providing the same or equivalent set of health care services that the VA can do it for 20% cheaper. If it could, it would be a REAL argument to just shut down the rest of the US health care system and expand the VA. 20% is a HUGE number compared to real marginal differences in expenditure or cost per service.

        As an exercise left to the reader, who knows where these 20-30% cheaper numbers originate? There actually is an interesting “real” genesis to them, since repeated improperly by Congressmen and others.


    • I am sorry to say that I cannot find that link. It referred to costs about 5 years ago, and said that general population was $6300, and VA $5000. I will keep looking and post it when and if I find it. It was somewhere just yesterday. grrr.

    • I wonder if anyone else has had the experience of helping a parent with (undiagnosed) dementia work through the labyrinthine process of researching Medicare Part D and signing up for a plan. It is, to put it mildly, not fun. After you finally settle on a plan, you have endless worries that you may have picked the wrong one; that the next medical emergency may call for a drug that your chosen plan doesn’t cover. A VA-style single payer plan would not only save money, it would save seniors’ (and their families’) time and sanity. One more reason to run screaming from Ryan’s plan to privatize Medicare.

    • “However, the paper is economics. We are economists.”

      Usually I admire most of what Austin writes, but this comment is very disappointing.

      What is said, essentially, is that since the assessment of efficacy in measuring the value of drugs would require work outside the usual area, we will ignore it for now in favor of a more easily accessed number we can plug in, even if that makes the conclusions of the paper of very little value in the real world except to people trying to muddy the issues.

      This unfortunately plays into the hands of those who criticize economics as being too often concerned with Plato’s cave models and too often not concerned with real world facts.

      Writing about the “value” of drugs without considering their efficacy is like writing about the value of a car but ignoring the fact that it won’t run.

      The obvious easy solution would have been to locate a co-author with some experience in the issues of efficacy and comparative efficacy in drugs. I would have to believe that such people exist at the BU School of Public Health.

      • This is silly. A single paper cannot do everything. Read the whole paper, including the discussion. I am an economist. I do economics. There’s nothing wrong with that. I make no claim that it is the only, or even the most important, perspective.

        • I do not suggest that there is anything wrong with being an economist. I do suggest that there is something wrong with assigning value to drugs based not on their comparative effectiveness in attaining the desired result but instead on the price that their manufacturers assign to them.

          I would suggest that a high blood pressure drug that costs $1200 a year but does not work as well as a drug that costs $15 a year has less, not more, value than the cheaper drug. That happens to be one exact situation from the VA formulary.

          I understand the economic models you are using. It just seems to me that in the face of a growing crisis in health care, a model basing calculations of value in health care independent of comparative effectiveness obscures the choices that we must make to change the health care system to make it affordable to both individuals and society as well as even more effective. A calculation that obscures waste is not a step in the right direction.

          • ” I do suggest that there is something wrong with assigning value to drugs based not on their comparative effectiveness in attaining the desired result but instead on the price that their manufacturers assign to them.”

            That’s is not at all what is done in the paper. I believe you have misunderstood it. That said, I am fully supportive of comparative effectiveness and will be doing just that type of work in other areas. This paper doesn’t do that because that’s not what we were funded to do. As you seem not to have understood the paper I do not think you are in a position to determine the extent of its value to the debate.

          • (coming in a bit late but there is a link-back to this article from a recent one and I think I might help Pat through his discontent.)

            There is a disconnect here that is common to conversations between economists and non-economists (muggles?). As is often the case, a particular assumption, ceteris paribus, held by the economist is left out of the discussion. Likewise the purpose and value of this assumption appears under-appreciated by the muggle.

            Pat S says, “A calculation that obscures waste is not a step in the right direction.”

            Pat, the paper which is the substance of this article is not commentary or a prescription for better healthcare, it was commissioned to determine one thing; all else staying the same, what would be the effect of changing medicare’s RX system to that of the VA. The paper includes a “calculation that obscures waste” because the status quo includes a “calculation that obscures waste”.

            The status quo also includes a drug benefit which is very different from the VA. If the paper was intended to determine the savings or costs of covering RX based on effectiveness then it would not ALSO change the manner in which the drugs are sourced.

            Comparing a complex system to itself but with a single variable or subset of variables changed is necessary when studying the effect that variable or subset has on the system.

            Keeping research specific and narrow in focus is the hallmark of a scientist. Your criticism is misplaced.

    • It is the role of FDA to weigh in Efficacy vs. Toxicity for an approval. A drug company goes through a long process of this approval. What expertise does VA have in determining this issue? This is what FDA (also a federal agency states)

      “The main consumer watchdog in this system is the U.S. Food and Drug Administration’s Center for Drug Evaluation and Research (CDER). The center’s best-known job is to evaluate new drugs before they can be sold.” If so, why should VA or Medicare get into other federal agencies role of which drugs should be made available to public?

      After an approval, of course there is a trained doctor who has to provide the prescription. Of course we cannot go to a pharmacy and buy whatever we want. If FDA has determined the safety and efficacy and the doctor wants to prescribe, I do not think VA or Medicare has any expertise in this area (except to reduce their own costs).


    • The FDA determines the safety of a drug but has nothing to say about the relative effectiveness or cost of the drug. The VA makes its decisions based on effectiveness and cost. If there are two drugs which both work then they can negotiate to get the one with the best price. Patients get an effective drug and the VA saves money. The VA has real doctors and pharmacists who make these decisions about the best therapy. The FDA just qualifies drugs to make sure they are not too toxic.

    • The FDA’s protocol for new drugs is pretty compromised in favor of big pharmacy. IIRC, drugs are compared against a placebo, not the best performing generics as they ought to be. Also they do insufficient screeing for rarer or long term usage side affects.

      VA will Rx drugs not on the formulary, one just has to cover the cost without the assitance of VA. What is wrong with that? I would prefer to see more people covered with basic medical care, than just a few lucky one getting gold plated medical care. Not that I am opposed to allowing people to buy gold plate plans.

    • When I asked my original question on ask.com, I did not mean that the “number” of drugs for Medicare would “have to be” reduced, merely that they should not have to purchased on the open market. They should all be purchased on bid, which is the method used by the VA.