• You just can’t see this often enough

    This is from the NICHM, in their recent brief, “The Concentration of Health Care Spending“:

    That’s the cumulative distribution of personal health care spending from 2009. There are so many ways to talk about this. One thing to note is that the top 5% of spenders (some of the sickest among us) account for about half of all health care spending. More significantly, the bottom half of spenders (ie the healthier half) account for less than 3% of all health care spending.

    When we talk about incentivizing people to forego care, we’re talking mostly about healthy people. When we talk about consumer directed health care, we’re talking mostly about healthy people. We don’t want sick people to avoid care. We want to stop healthy people from consuming it. The problem is that healthy people consume so little care to begin with. If we could incentivize the healthier half of people to forego all their personal health care spending, we’d spend $36 billion less out of a total $1.259 trillion in personal health care spending. That would be a drop in the bucket. And no one – no one at all – thinks we can get people to stop all their health care spending.


    • Hi Aaron,
      I noted that, too. See this story, which is the lead story in this morning’s Fiscal Times:

    • With so much spend in the tail, I could also see this supporting a government-sponsored basic health coverage for all, coupled with mandatory catastrophic coverage. The trick would be getting catastrophic coverage for those already super-sick …

    • We don’t want sick people to avoid care.

      But we do not want them to get care that will not help them significantly.

      My wife one year had a total of $60,000 in medical spending (we only spent $10,000 out of pocket insurance spent the rest). Another year she had an expense of $30,000 all out of it out of pocket, but I am still not sure that she will life time be above average in healthcare spending, and if she is not she will have been a looser for having had insurance. Also in the $30,000 year would have been higher had her doctor not let her leave the hospital earlier because we were paying out of pocket.

    • Maybe someone will come up with a plan to deal with the five percenters? One can only dream. Thanks for pointing this out.

      • Love actuary has come up with the beginnings of such a plan.
        You might be interested to know I and 3 of my partners are working with Milliman, an actuarial firm, on having 3 separate pools.
        !. 0-$25,000
        2. $25,000-$50,000.
        3. $50,000 and over.
        By accumulating reserves in the insurer which can never be cashed in and can only be used for medical claims, healthy people can reduce their premiums by 60% by virtue of a paid-up $25,000 policy in 3 years..
        They can reduce premiums by 80% by having a $50,000 paid-up in 5 years.
        Both these objectives can be accomplished for very “affordable ” premiums.
        And, they continue to pay catastrophic premiums so that large claims can continue to be paid.
        We will have Milliman’s paper and numbers in 3 weeks to show interested life insurers.
        The goal is to have a break-even health insurance product, so the life insurer can sell more of its life insurance products and annuities. We will form a 501(c)(4) health insurer as a subsidiary of the life insurer.
        Don Levit

    • Aaron, surely you must know that the patients in the top 1% this year are not the same as the 1% next year or the year after that. if they were the same people year after year, then our health care spending problems would boil down to the treatment of a small group of people. That, however, is not the case. See: http://healthblog.ncpa.org/persistence/

    • I’d be interested in seeing research about interventions targeting the top spending cohort or subgroups within this cohort. Better coordination of care is the main theme I have seen for reducing costs with this group. The issue is that many of these patients seem to be high utilizers because they are very sick and thus most of these costs are unavoidable.

    • “When we talk about consumer directed health care, we’re talking mostly about healthy people. We don’t want sick people to avoid care. We want to stop healthy people from consuming it”

      I want sick people to better manage their care and the expenses associated with it. We need better healthcare delivery methods/models, better care coordination, better medication management (taking the drugs that are Rx’d and paid for). If all this improvement starts with the top 5% of spenders (in any given year) then the benefits (cost savings) will be that much more dramatic.
      We cannot continue to insulate the top 5% of spenders from the costs and consequences of their healthcare consumption. That doesn’t mean we can’t provide generous subsides for their care, but it does mean that we structure those subsides in such a way as to make these spenders functional (and very influential) market participants.

    • This is one of the most important charts to keep in mind when discussing health care. It is also one of the reasons I wish more docs would participate in health care policy. Most spending comes for those big ticket problems like cancer, heart disease and trauma. People are going to exceed any reasonable deductible. People are not going to worry about costs. When you tell a patient they need their valve replaced, they simply do not think about costs.

      Yes, there may be savings to be gained from the left side of that chart, but that is not enough to bend the curve like we need to do.


    • It seems that no one has commented here on the obvious fact that consumption of health care rises with age. My recent $100,000 shoulder joint replacement happened after years of increasing pain. The diagnosis was “advanced degenerative arthritis.” That condition must be relatively rare among young people.To the extent that high usage of medical resources by the top 5% is a consequence of aging, that expense may be more or less fixed, unless we resort to putting old people on ice floes and pushing them out to sea.

    • This is why adverse selection is such a problem when Medicare and Medicaid agencies purchase health care for beneficiaries through HMOs. If the most expensive beneficiaries never enroll, or leave their plans and go back to Fee-for-Service, the programs will significantly overpay for services. This is especially true for Medicaid agencies that don’t risk adjust, but use age-sex and aid category averages to calculate capitation rates.

      Medicare does risk adjust, however, according to Henry Aaron’s recent testimony to Congress the risk adjustment system is increasingly ineffective in preventing overpayment. He said, “Recent research has shown that the Medicare risk adjustment algorithm actually increased program costs by as much as $30 billion or 8 percent in 2006. ” If Aaron is right, this does not bode well for the State-organized exchanges which will also attempt to use risk adjustment to prevent adverse selection.


    • Here is the part of Aaron’s testimony related to the distribution of costs throughout the population:

      “Failure of Risk Adjustment. A necessary element of successful competition under a voucher is effective risk adjustment. It is well known that health expenses are highly concentrated and are much higher, on the average, for Medicare enrollees than for the general population. Insurers who ‘get stuck’ with a lot of very sick people can lose a lot of money or even grow broke. Shareholders do not hire administrators to lose money or go broke. Accordingly, insurance administrators have a duty to the people who hired them to try to enroll healthier than average people. Of perhaps greater importance, they need to retain healthier-than-average enrollees. Because of these incentives, all competent health analysts have long recognized that, if premiums are uniform or vary less than expected cost, the key to a successful health insurance market is risk adjustment. Risk adjustment consists of financial transfers among insurers to offset the variations in expected health costs related to the characteristics of enrollees. Insurers that enroll people with comparatively low expected health care use would pay money to insurers that enroll people with high expected use.

      In the 1990s, risk adjustment was inadequate. It was not then ‘good enough to discourage competition based on risk selection. But it was getting better. I assumed, perhaps too facilely, that it soon would get ‘good enough.’ Well, to date it hasn’t. Recent research has shown that the Medicare risk adjustment algorithm actually increased program costs by as much as $30 billion or 8 percent in 2006.[9] The problem is that risk selection increased along lines that were not included or could not be included in the risk adjustment formula. Plans have available many ways to attract customers expected to have low costs (“the X health plan is offering a free golf weekend”). They can also use the quality and availability of services to discourage high cost enrollees from remaining (“we are sorry, but our oncologist is booked solid for the next six weeks; no, he is the only one on staff”). They can also take steps to encourage low-cost enrollees to stay (“all current enrollees who remain in our plan will receive a free gym club membership”). The challenge of defeating such behaviors is never easy. But in an atmosphere hostile to aggressive regulation, it is impossible, particularly when the stakes are as high as they are with the Medicare population, whose costly patients are very costly indeed. ”

    • Although I have never managed a hospital, my impression of the highest-spending 5% is this:

      – some of them are just in the hospital a long time, due to spinal injuries or psychosis or a coma or many other traditional illnesses;

      – and some are in the hospital a short time, but receive newer treatment such as transplants, high-intensity cancer treatment, etc.

      One way to lower costs for the first group would be to transfer some of them to VA hospitals. I am completely serious. This assumes that the VA hospitals have open beds, which they do where I live in MN.
      These places have staff that is already paid for.

      As for the second group, there does need to be a realistic national fee schedule and a limit to balance billing. If the govt said that a cancer drug like Avastin could only be charged at $1000 a month, I assume that the manufacturer would go on making it. Yes, there will be a slowdown in drug research if that happens. I would contend that America can take a break in this area, in order to keep 95% of health care still affordable.

      Some of what happens in the second group, the high-tech cases, is just price gouging by hospitals. (rarely by surgeons) We do need politicians who will stand up to hospitals and their pricing practices.

    • I would be very interested in seeing this same graph with all individuals and costs associated with the last 3-6-12 months of life removed from the picture. I wonder how this would affect the extreme steepness on the right side. It would seem to give us a sense of how a policy shift toward end-of-life palliative care would affect our overall cost situation, because the problem with the curve as depicted is that it’s always going to have this basic shape, almost by definition. It’s just a question of how extreme the slope gets on the far right side.

    • The same kind of curve probably exists in any enterprise that saves lives and handles stark emergencies.

      1% of homeowners account for 99% of fire department spending each year. I once read that about 5% of homes account for over 80% of police calls also.

      So the question is not, how to alter the curve, but rather how to just get the spending down.

      There is no force of nature that requires American health insurers to pay
      $4000 a day for intensive care. What prevents them from doing so are two fears:

      a. hospitals need the money, quite a few would go broke without
      their intensive care revenue;


      b. a small number of people will die if we cut back emergency services.
      Even a four hour wait can prove fatal in some situations.

      In orher words, a serious cost reform of end-of-life care might cost 500,000 nursing jobs and 25,000 extra deaths.Can the American political system handle that?

    • Bob:
      Those are good questions.
      As a society and as individuals, we have to come to grips with our mortality.
      Doing so will lead to more courageous, bold action on our part.
      To die well is to also live well.

      If the insured or his family wishes some financial compensation for the denial of life-prolonging care, why not have the insurer provide a death benefit equal to a percentage of the potential medical benefits?
      Any one can die with all that life-saving paraphernalia.
      The courageous person dies without all that periphery stuff, being grateful for the life he was provided, of which death is merely the flip side of life.
      Don Levit

    • “We don’t want sick people to avoid care.”

      Of course we do. How much of the care received by the 5% does any good? How much is totally futile? Terri Schiavo anyone? Take a look at the stats for DES (Drug Eluting Stents). Massively expensive but not notably more effective than regular stents. How about the latest generation of anti-cancer drugs? Two months of end-of-life misery for $200K? Doctor friends tell me that EPO is massively over prescribed and useless in most (by not all) cases. MRI utilization has rise by 100s of percent in recent years with no justification.

      For fun take a look at the actual C-section rate versus the medically indicated C-section rate.

      Privately doctors have endless lists of treatments they know are marginal or useless. They employ them anyway. Why? Patients expect doctors to “do something” and there’s always a third party (Medicare, Medicaid, insurance, etc.) to pay the bills.

      • Frank,
        Excellent points. Doctors have a huge financial incentive to do procedures with little value. Patients, expecting the doctors to “do something” as you note, don’t care since they don’t pay.

        In a utopian world with infinite resources, none of this would matter. Since much of the massive spending for useless end of life care is ultimately paid for with deficit financing, we should all care very much since we are mortgaging the future of our children and grandchildren.

        Ultimately, this chart becomes the poster child for a discussion on the rationing of health care.