• Cost effective v. cost saving

    Jeff Levin-Scherz nicely demonstrates that cost effective is not the same thing as cost saving, using the example of statins and their decrease in heart disease mortality risk.

    Statins are enormously effective drugs that, along with a decrease in cigarette smoking, have been responsible for a huge decrease in the incidence of cardiac death, especially in young men….Simvastatin has been a generic medication for a few years, and atorvastatin (Lipitor) is going generic this coming month. Have we finally reached a point where the use of statins is not merely cost-effective, but is actually cost-saving?

    In short, no.

    We don’t save money by using even generic statins for primary prevention of heart of vascular disease.   We only gain good health outcomes for a reasonable cost — which doesn’t seem to me to be a bad outcome at all.

    We need to learn to ask whether the use of all therapies are worth it, and we also need for the differences between the terms cost effective and cost saving to become more widely known in our culture, and not merely by technical experts. That is part of what it will mean to develop a sustainable health care system–one in which we can look at the benefits and costs rationally, and decide whether what we are doing is worth it.

    Update: I had forgotten about this TIE post from Austin ~ 6 weeks back that is relevant.

    • Worth it for whom? For the rare person who has his life saved, the answer is obviously yes. For everyone else – that’s a tougher question, but in general their welfare loss as a consequence of some other guy that they have no idea even exists dying prematurely is far less than what they’d suffer as a consequence of having a marginally smaller cell-phone budget each month.

      Consequently, whatever arbitrary set of variables deemed to constitute “welfare” and thus worthy of maximizing in whichever formal model serves as the basis for centralized health reform will be heavy on cost controls (masquerading as a concern for cost-effectiveness).

      This is already evident in the field’s heavy focus on metrics that say a lot about how well a population cares for itself, and little or nothing about how effectively doctors and hospitals treat specific diseases. Cutting all spending on psoriasis treatments won’t register as anything but a positive in the life expectancy-vs-spending stats.

    • @JayB
      Great points. We have to talk about it and decide. There are many values inherent in picking variables to operationalize as you note, but there are also values inherent in not taking the questions on.. There is a premise that we cannot sustain the current rate of health care growth.

    • It is frustrating because the alternative to admitting you are putting a price on human life is not admitting that you are putting a price on human life. There is simply no way to avoid putting a price on what the government pays for a human life. If you say that the government should pay $100,000 for a 1 in 10 chance of giving someone an extra two weeks of life during which they will be in pain 20% of the time and asleep the other 80% you are implicitly saying the government should not pay for other potentially life saving strategies (safer highways, more prevention, better law enforcement, stronger substance abuse programs, etc.)

      What we need is a national discussion of the moral dimension of how government-funded agencies define medical necessity. The moral question is what outcomes should we pay for. The scientific question is what outcomes would we get..for X treatment. Congress should sponsor that debate but doesn’t appear up to task. I am not sure who should do it, CMS, IOM ?

      • @George
        yes, need discussion and it needs to be informed by good data. As Austin was noting a few days back, good nonpartisan research is expensive and hard to come by

        • Good, nonpartisan research is critical for a long list of reasons, but we can begin to discuss the moral definition of medical necessity, e.g. should we pay for a treatment if it has a 1 in 10 chance of giving someone and extra 3 months to live, a 1 in 100 chance, a 70 %chance of shortening a painful but non fatal illness from 2 weeks to 1, etc.?

          Such a discussion could, in a small way, make funding policy research more acceptable by highlighting how little reliable data we sometimes have. Despite the barriers created by anecdotal evidence of the effectiveness of a treatment, it would probably be easier to convince the public of the need to study exactly how well a treatment works than to convince the public of the need to study health policy questions questions that Republican and Democrats believe they already know the answer to. I agree that it is a bit of a long shot, but increasing the visibility of acceptable research, might make more controversial research–where people think they already know the answers–more acceptable

    • FWIW I think that any time you funnel vast amounts of cash through a highly centralized distribution apparatus that’s rife with conflicts of interest and the potential for self dealing, two things are certain to happen.

      The first is that there will be an immense pressure to define the subset of variables that constitutes welfare with the values, preference, and interests of the people that have their hands on the levers that determine where the money goes.

      The second is that these public choice frictions will move the actual equilibria a long ways from the hypothetical optimum predicted by the model. Ergo even if were possible to specify an objectively superior set of welfare criteria that apply equally to everyone, the actual outcome would look nothing like the outcome predicted by the theory.

      These are intractable problems and the only way to avoid them is to abandon the assumption that you can objective determine what’s best for your neighbor according to his values and preferences and vice-versa. That has it’s own problems, but administering a simple transfer mechanism (think of foodstamps) is a moral and technical cakewalk than attempting something roughly akin to supervising the entire food economy and regulating people’s eating habits.

      • @JayB
        meta level, what do you suggest as a way forward in health policy/health care financing? Given the skew of health spending, most of it is on ‘your neighbor’ (probabilistically speaking) in a given year in any event. I realize the spending level is endogenous to the past insurance/financing choices, but it is unrealistic for individuals to decide via self insurance….I guess you are saying each person could get a fixed amount of $ and do as they see fit? This may be me putting words in your mouth.

    • FWIW I agree that if the public is footing the bill, “the public” (technocratic administrators who define cost and quality in terms of high level aggregates that have little or nothing to do with clinical efficacy, laboring under immense fiscal pressure to deny care) will have to decide when to pull the plug on granny. The default cost-to-QALY optimum under those constraints is clearly all palliative after the patient exceeds median life expectancy. I doubt that the folks crunching the numbers will ever come out and say that outright, but the incentives are such that this will be the operational reality, even though it’ll be cloaked behind a dense thicket of jargon-and-regression-heavy pretexts for discontinuing care.

      I’d personally much rather find myself in that position with a health-savings balance that’s accumulated for the entirety of my working life, that I’ve funded by making payments into my HSA instead of into 3-4X higher premiums for first-dollar insurance, making decisions about what’s worth paying for and what isn’t in the context of system where price transparency and competition are the rule rather than the exception, according to my own values and preferences. This hypothetical looks even better if I can transfer my balance to families or charities if I wish and vice versa. In that situation, with a hopeless diagnosis I’d gladly unplug myself and donate whatever balance I have left to someone else with better odds, my grandchildren, etc.

      You foot the bill, you make the call about when the costs of staying alive exceed the benefits. Couple it with income indexed transfers and you can mitigate but not eliminate income based disparities. Not perfect but less expensive, less authoritarian, and laden with far fewer landmines than the endpoint that we’re currently careening towards.