• A very simple argument that we’re not on the “flat of the curve” in health care

    The curve in the chart below is the health care production possibility frontier. To keep this post brief, I will assume you know what it represents. If not, see this prior post and/or, perhaps better, that one.

    If the health system were on the “flat of the curve” we’d be at point B on the production possibility frontier. Some argue things are even worse, that we’re to the right of B on the curve. Actually, we’re not in any of those places. But if we were, the solution would be relatively simple: stop investing so many resources (factor inputs) in health care. Stop spending so much. Problem solved.

    Except, in fact, we know we cannot be at point B or to the right of point B. We know this because we have evidence that we’re not even on the production possibility frontier curve. We’re at some point below the frontier, like A or F. The evidence that this is so is that we have under-use of valuable care simultaneous with over-use of unnecessary care. That can’t happen on the production possibility frontier, by definition.

    The good news is that we’re not on the flat of the curve. The bad news is that solving the problems in health care requires something much harder than just spending less. We have to spend differently. To be on the flat of the curve we’d have to first be on the curve. And we’re not. That’s the problem.



    • Let me share you the information about cost of health insurance in US compared to Europe at the micro level.

      I am a graduate student in one of the universities in US where I have to health insurance cost of around $7000 per year. I am thankful to my university who pays 80% of that cost., so that the burden to me is just $1400 per year.

      Consider the situation in one of countries in Europe where my friend is enrolled in graduate program. She has to pay health insurance cost of only around Euro 500 per year.

      If you compare these two different scenarios, you can notice how costly is US health care system even for the students. Even after getting subsidy from university, the cost is still high in US compared to Europe.

      That being said, I totally agree with you that we need to spend differently but that difference should be reflected in cost.

    • Incredibly important point that I think gets lost in the discussions — namely that its not that we are spending too much or too little but that we are spending money in much less than an optimal way. The most frustrating part, however, is that trumpeting this argument leads one head-long into the counter that there’s a simple solution — completely deregulate the health care markets, get the government out of it, and the consumer will redirect things towards this optimal distribution.

      Thanks for reiterating this important point.

    • I would vote for “A” except I would move “A” far to the right (and still well below the dotted line).
      We aren’t even on the curve. We over pay for a lot of wasteful and actually damaging care. We also don’t provide good basic care (prenatal, etc.) to a significant number of people.
      So, some people are at “D” (or further back) and the insured folks are far to the right and low in the “quality” scale.

    • Given that under-use of valuable care simultaneous with over-use of unnecessary care will always occur (realistically, both cant be purged completely), the health care production possibility frontier will always be a theoretical discussion.

      Thus, so will flat of the curve medicine, at least in the context of the post. Having said that, I get it, and it conceptually conveys the appropriate visual of where our system is broke.

      • This is a good comment. I will follow up when I have figured something out.

      • Though we can’t be on the theoretically optimal production function, we could be on a lower, best practical one. If we were then we should not a lot of variation in the outcomes-spending relationship. We do see a lot of variation, so we know we’re not doing as well as we could. The signal that we are doing as well as we could is when we’re all on the same production function. And if that occurs, there wouldn’t be any discrepancy in the literature about whether more spending leads to more health or not. The studies would all point the same way.

    • -Does the definition of the “production function frontier” include all of the real outputs doctors and hospitals deliver that real people value? E.g reassurance that worrisome symptoms aren’t actually serious, mitigating the pain and discomfort caused by non-fatal conditions and diseases, convenience, etc, etc, etc, etc. These things aren’t easily abtracted into a numerical value and aggregated into a data set – so I suspect the answer in this case is “no.”

      Any theoretical optimum that fails to include these factors in its definition will always make it seem as though the real medical system is farther away from optimal production than it actually is when valuable outputs (things that real patients value enough to pay for) and/or that aren’t easily measured (or abstracted into numerical values and aggregated into a statistical data-set) don’t make it into the model that’s used to define the optimum.

      -Imagine a patient choosing between two cancer treatments. One has greater odds of increasing longevity but will require enduring many bouts of excruciating pain. The other will minimize suffering and maximize function but is much less likely to prolong life effectively. They both cost the same. Isn’t defining which of the two is closer to the optimum production frontier completely dependent on the patients own values and preferences?

      Here’s to hoping that the people constructing these models and evaluating their significance understand their limitations as well as their utility.

      • Independent of their merits, I fail to see how your concerns are relevant to the rather simple point I was making. Are we on the flat of the curve? I think clearly not.

        • -See below. There’s no such curve that can be objectively defined in a manner that’s independent of the subjective value judgments that inform its construction. The distance between where we are and the PPF depends on how its defined.

          It is possible to get better at diagnosing and treating diseases, it’s possible to disseminate that information, it’s possible to calculate the costs of doing so, and it’s possible to better align the incentives of patients, payers, and providers. All without any of the participants knowing what a PPF is, much less explicitly trying to move medicine towards a particular PPF defined by any external authority.

          The aggregate effect of doing so is much more likely to bring society closer to the closest real-world approximation of the PPF than contemplating where the real world stands relative to an arbitrarily defined PPF – so here’s to hoping that’s ultimately what we’ll emphasize in any reform efforts.

          • Do you guys (JayB, Mark Spohr) understand that this is a conceptual argument, that the production function shown in the post is just a sketch to show the idea, not something derived from some specific model or data? Of course one can never reach the production possibility frontier. It’s an idealization. Yet people debate whether we’re on the flat of the curve or not. Which curve do they have in mind? My point is that we’re not on any single curve. Different health systems have different production functions. None of them are anywhere near the ideal. All systems can be more efficient, some much more. Hence, we’re not on the flat of “the” curve. We’re not on “the” curve.

      • It looks like the production function model is stochastic and can include as many terms as necessary. I don’t know the details of the model in this graph but I would not assume that it doesn’t include measures for quality, patient satisfaction, etc.
        Your choice is a real one but in real life the costs are usually far apart. The choice is for a painful, debilitating, expensive course of treatment which “may” prolong life by some small amount or an inexpensive palliative care model which offers much greater comfort and a somewhat shorter life. In our current medical model, the doctors, hospitals, pharma, and even insurance companies (which work on cost+) have a great incentive to push the first option but patients may be better off with the second.

    • -FWIW, does anyone reading this actually believe that there’s a single, unique production possibility frontier that can actually be defined in a manner that’s free from arbitrary value judgments about which outputs ought to be measured and how?

      Is there any objective way to determine whether treating, say, atrial fibrillation with Pradaxa (higher unit cost, lower stroke risk, requires less monitoring, no blood thinning, no dietary restrictions) is closer to the production possibility frontier than treating the same symptoms with generic warfarin?

      I can see how one would determine an optimum solution from the payer’s side, and from the patient’s side, but it’s hard to see any single definition converging on a single solution unless the payer and the patient are the same person.

    • -It’s clear that it’s purely conceptual.

      It just seems like an overly abstract and formalized way of restating an uncontroversial and obvious point that could be written in plain language -e.g. “We should spend more on what works and less on what doesn’t.”

      No one disagrees with that. The point I was making above is that there’s no unique, unambiguous definition of “what works” with regards to all of the outputs that medicine delivers or their relative importance to real patients.

      It’s clear how generating better data on clinical efficacy, developing better mechanisms to distribute that data to clinicians, etc can help spend more on what works and what doesn’t. It’s not clear how the PFF concept will help anyone do those things better than “spend more on what works, less on what doesn’t.”

      • Well, I’m talking to those who use the “flat of the curve” language. Hard to speak to that without drawing a curve. I fail to see how it helps not to draw a curve and discuss how we’re not on it. So, you’ve lost me. I think you want this post to be something else. Maybe that’s a different post.