Physician Aaron Carroll at The Incidental Economist has been doing a good job of demonstrating that the US pays too much for the health outcomes we actually attain. And he documents that the US spends more per capita on health and as a percent of national output than any other developed country here. It turns out that even though the US spends more, our population health indicators, such as life expectancy and infant mortality, are significantly lower than similar developed countries, such as the UK, France, Japan, and Germany. In fact, we rank lower than all but a handful of countries, such as the Czech Republic, Turkey, and Mexico. (See this Commonwealth Fund Report for a detailed recent update.)
Here's infant mortality rates, just in case you remain unconvinced that we spend more and get less than most comparable developed countries:
A related post by economist Uwe Reinhardt provides a very nice explanation of the components of health care costs and why "bending" any of them poses real problems in the current political and economic environment. Reinhardt also provides a very nice chart here showing that the overall objective in fine-tuning costs and health "output" is to produce optimal amounts of good health which in turn lead to higher levels of well-being.
My goal in this post is to highlight some features of the fine-tuning process that make it economically, ethically, and politically complicated. At the same time, I want to lay the ground work for my own conjecture that the US "system" has evolved in such a way that we are delivering too much (over-priced) care that yields very small increments in health to those who are insured and therefore able to pay for it. At the same time, we are delivering too little care to those who are uninsured and therefore unable to pay, but for whom that same amount of care or even less would deliver very large increments in health.
I'm going to start by doing something that always annoys non-economists; I'm going to assume a two-person world. In that world, one person (let's call her Person A) is healthier. By that I mean that she is born healthier and any defined amount of resources she uses to produce health result in a larger absolute gain in health for her than they would if I gave the same resources to the other person in this two-person world. By the same token, the other person (let's call him Person B) is born less healthy than A and any resources he uses result in a smaller absolute gain in health for him than they would if I gave the same amount of resources to Person A. In econ lingo, A's marginal product in producing health is higher than B's for any level of health-producing resources. I'm also going to assume that neither A nor B are over 65 and therefore out of the workforce and possibly in their twilight years (if the above graph on life expectancy is correct). By focusing on those under 65, I sidestep the economic and ethical dilemmas that surround the disproportionate share of health care costs that are expended in the last six months of life.
Imagine for a moment that A and B are your two children and you, their benevolent parent, have to decide how to allocate your income between the kids in a way that makes you happiest and them healthiest (and let's assume for the moment that there's no health insurance so you are stuck with allocating a fixed amount between your two kids who differ in both level of and ability to produce health.)
In what follows, I'm also going to assume that there is just one continuous health producing input. I don't think this is too much of a stretch if we imagine that what produces health is a mix of inputs such as nutrition, exercise, and some amount of health care. (Even if you object to this simplification, I don't think it is necessary for the final point I wish to make.)
Here is what each child's attainable level of health could be for different levels of resources up to 100% of your resources (and assuming that you could give 100% to each of them):
The vertical axis depicts the possible levels of health that could be achieved. Notice how at every level of input (expressed on the horizontal axis as a percent of total resources available), Person A is always at a higher level of health than Person B. Even if you devoted all your resources to your less healthy child, you could not make him as healthy as your more healthy child. The lines curve under the assumption that as health increases, the amount of health obtained from adding another unit of input increases but by less than the amount of health that was obtained from the preceding unit of input. In econ lingo, there is diminishing marginal product of the input as health improves.
But you can't give 100% of your resources to both A and B. You have to divide them somehow. How would a benevolent parent, someone who cares deeply about both A and B, make this decision?
Let's consider three plausible allocation "rules," all of which could be justified by some appeal to fairness.
1. Maximize total health produced from the scarce resources. Allocate resources between A and B so that total health (the sum of A's health plus B's health) is maximized. This is the type of rule one might choose if one had two factories that produced the same output, but with different technologies. For this reason, it fits well into neoclassical conceptualizations of household health production.
2. Equalize Inputs: allocate resources between A and B so that each receives an equal amount of the scarce input regardless of their health. In other words, a 50/50 split.
3. Equalize output: Allocate resources between A and B so that each has the exact same level of health.
Now the problem is that if Person A gets 100% of resources then Person B gets none and vice versa. They have to share and they need to figure out a rule that is fair to each individually and that makes their benevolent parent who cares deeply and equally about both as happy as she can be.
Here's the graph of the problem and the possible solutions for two people who start with the same level of health and differ only in their abilities to produce health (derived from very general and plausible health production functions):
When resources have to be shared, A experiences lower health as more of the resource goes to B and vice versa. The horizontal axis now depicts the amount of resources going to A. The amount going to B is just 100% minus the amount on the horizontal axis.
The yellow line indicates the health maximizing solution. It maximizes the total health of this two-person society. Notice that at this point, the person who is most efficient at producing health receives most of the health producing input. This will always be true if the objective is to maximize societal health.
The blue line is the equal inputs solution. Each person receives 50% of the available resources. At this point, we don't achieve equal health because the two people differ in their abilities to produce health, but we do achieve equality of inputs.
The red line is the equal health solution. At this point, the least healthy is as healthy as they can be given that A must have equal health. However, as you can see, we have imposed a large penalty on A to achieve this result.
Some of you who have read this far are already thinking of reasons why this can't possibly adequately describe the moral dilemma of how to allocate health producing resources. To you I say, "Perhaps not, but it's a good starting point for thinking about what is an ethically and economically complex problem because it explicitly shows the consequences of different resource allocations if you believe that we live in a world of individuals whose fates and well-being are not intertwined."
B will always prefer the equal outcomes solution, A will always prefer the health maximizing solution. If there are more A's than B's, we're likely to end up at the point closer to the yellow line than to the red line. A benevolent parent will be thinking things like: If I take too much from A, A will resent B and I need A to care for B after I'm gone. Since I need A to care for B, A needs to be as well off as possible, but the better off I can make B, the less B will impinge on A's resources. Also, I love B and I'm happier when B is better off. I have to find a middle ground. What is it?
Let's suppose that we do live in a world with more As and that our health care sector has evolved over time in a way that favors the more healthy. For example, healthier people are more likely to be employed full-time. What if we link health insurance to employment so that those who are healthier are also the ones most likely to be insured? Those who have higher incomes (and better health) will face lower prices for health care. What if the insured A's of the world face not only a lower price, but a system that pays providers for piecework so that the more care they deliver, the more money they make? But it is more lucrative to treat those with better insurance, which (since it is tied to employment) means that those they prefer to and do treat already have better health (the A's of the world). Moreover, the healthier people are, the less money providers will make, unless the healthy people face a reduced price.
Complicated and perverse incentives, yes?
I will suggest that in such a world, it is highly likely that the societal resource allocation could be to the right of the yellow line in the figure above. In other words, more will be consumed, more will be purchased by the healthiest individuals, but the (high) prices paid will yield very small increments in health for them. In addition, the health of the unhealthy will be worse than it would be if we could move to the health maximizing (yellow) point. And, last but certainly not least, total societal health in this world will be lower (since the maximum is at the yellow line and would require a redistribution of health-producing resources from the healthy to the less healthy to move left to it).
If this were the case, if our hypothetical (and heterogeneous in health) society is to the right of the health maximizing point, I would expect to see a country that spends more than other developed countries on health, but that has lower health outcomes. Just like the United States. I would expect to see economically and socially disadvantaged groups that have much worse health than they would have if they lived in a developed country that spent less on health, but allocated it more efficiently (if not more equitably).
Here's another chart from the OECD:
The curve in the above graph can be thought of as the amount of total health (measured by life expectancy) that a nation could expect to obtain from a given amount of expenditure on health. Points above the curve are highly efficient in that they are achieving higher levels of health than would be expected given their expenditures. As you can see, the US is high on expenditures and substantially below the curve, indicating a highly inefficient health production process. It's odd because it is also one of the few health "systems" with a large private sector presence. Remember that, at least in the US, the private sector is imagined always and everywhere to be more efficient than the public sector. I will leave it to you to ponder the meaning of this rather remarkable graph.
I glean several important points from this little back of the envelope two-person analysis:
1. The only technical solution is to equalize as much as possible each individual's marginal productivity with respect to health. If marginal products are equal, but health levels are different, then the 50/50 split is the same as health maximizing solution (the yellow and green lines would fall at the same point). If we could somehow equalize marginal product and level of health, then the health maximizing solution would be the same as the health equalizing solution and the 50/50 solution (the three lines would all fall at the same point).
2. If people differ in their abilities to produce health from a given amount of inputs, then eliminating health disparities will require us to improve the health production abilities of the disadvantaged. To achieve this, public and private resources will have to be devoted to improving prenatal, neonatal, and child health, particularly among the least advantaged. Public resources will also have to be devoted to preventing and managing chronic illness and improving education among the least advantaged since education has been strongly linked (yes, causally) to better health. Remember, if there are cumulative effects to the inter-generational transmission of bad and good health these will tend to widen the gap over time and make it that much more expensive for future generations to remedy.
3. A technical solution will require long-term strategic societal investment, which is not likely to occur in our time, and is not likely to eliminate health disparities entirely even if it did. This means that a discussion of where we want to be (between red and yellow) cannot be conducted meaningfully unless we first acknowledge that there is no simple short-term technical solution. As with the hypothetical benevolent parent, we are confronted with an ethical dilemma. Science and research can inform it, but in the final analysis, we will have to decide how we as a nation wish to live. Pretending that somehow a market could resolve it will only widen the disparities over time and make the narrowing of disparities even more costly in the future.
4. The national discussion about this must involve some recognition of the interdependencies of A and B. Obvious interdependencies that would lead us to something closer to a Rawlsian blisspoint are the benefits that B derives from A's increased health and productivity. If that translates into A creating innovations, health care technologies, and tax revenues that benefit B, then B should be willing to move right of the equal health point.
5. The interdependencies work both ways. If inequalities of income and health require higher taxes to cover B's health care costs, then the As (and their grandchildren) would benefit from B and B's grandchildren's having better health. There may also be other losses associated with large or growing health disparities. If they foster mistrust and societal discord or inefficiency in health production and national output, or if they force the A's of the world to employ guards and live behind walls, thereby curtailing their liberty and safety in significant ways, then A should be willing to move closer to the red or green line.
6. If we have a value system that requires us to provide equal opportunities to pursue life, liberty and happiness, then commitment to those values would almost certainly move us left of the yellow line and right of the red line, but would probably not inform us on where exactly we should be.
It should be noted that as Reinhardt has recently pointed out, our over-arching goal is to improve well-being, not to maximize health. All of us as individuals make trade-offs and even sacrifice some health in our efforts to achieve something that approximates what we think will make us happy or that will deliver a good life for us and those we love. Indeed, some of us trade off large amounts of health for the short-term "happiness" of tobacco or excessive alcohol consumption, but I have neglected self-chosen bad health here (it clearly muddies the ethical waters). The fact that all of us are willing to sacrifice some health to obtain other things that we believe will make us happy and that some of us make bad choices are other reasons why in an interdependent society, the solution almost certainly is not at either the red line or the yellow line. However, all else equal, to the extent that we derive happiness from the happiness and health of others, we may be willing to sacrifice at least minor gains in health, especially if we are near our individual peak.
I'm pretty sure we spend more per capita than any other developed nation and get less for it because we are at a very inefficient point to the right of the yellow line in Figure 2 (at least among those of us under 65). Add to this the inefficiencies created by the undertreatment of (uninsured) unhealthy people under 65 who age into Medicare with a lifetime of poorly managed often multiple chronic diseases, which then drive up costs during the last six months of life when the marginal benefit of care is often of necessity way below its cost and you have a highly inefficient system. Exactly what we saw in the last OECD chart above.
The fact that we pay too much for all care including care for which the marginal benefit for the mostly healthy is (much, much) less than the marginal cost only exacerbates the problem. We could at a minimum move leftward back to the societal health maximizing point (yellow line) with very little loss among the already healthy insured and with large gains among those who are not healthy or insured (and who are not also elderly and near-terminal), but it would mean restructuring incentives in the private sector and it would mean curtailing use somehow among the healthy and insured. At least making the healthy insured and their providers more prudent. The latter could to some extent be fixed by restructuring incentives in the private sector. Reinhardt has already described how unlikely that is to happen unless US voters become resistant to fear mongering by health care providers and insurance companies and more adept at managing both their own and the public's health. He has also begun to explain pay-for-performance (P4P) incentives and why they may not offer an easy solution.
We have a "system" that evolved to deliver rescue care and curative care, not prevention. And it has insulated most healthy people under 65 (and nearly everyone over 65) from price. Right now, the money continues to be in expensive rescue and curative care. One good thing about the Patient Protection and Affordable Care Act of 2010 is that Medicare payment mechanisms have been altered to make it more profitable for providers to prevent illness and promote wellness. This is a Good Thing for us all, both in terms of health and in terms of future health costs. Others are that elimination of pre-existing condition exclusions from health insurance and increased funding for federally qualified community health centers will assure that the less healthy get the care they would need to align our health indicators more closely with those of other comparable developed countries. Measures to improve quality should over time also improve efficiency.
Ironically, isolation from market forces has also tended to dampen our own estimation of the extent to which we as private individuals could control health care costs if only we would recognize, for example, the importance of prevention and of the externalities of infectious disease (I say this while sneezing and sniffling from a cold). Why worry about prevention when a $30 dollar copay covers a $480 dollar antibiotic Rx? A colleague who consults for physician group practices tells me that last year's heightened disease control measures to reduce the spread of H1N1 flu resulted in major declines in income among his clients. Think of it this way: every time you sneeze or grab a doorknob with the hand you sneezed into, a physician or a hospital makes more money. At the same time, productivity losses rise and health status falls. Think about it. Now go wash your hands for at least 30 seconds, sneeze into your elbow, and stay home when you're sick.
Lest you think this is an argument for higher deductibles and copays, it is not. It's nice to tell stories in which a higher out of pocket price will solve everything, but it won't. Especially when prices are so badly distorted and information so asymmetric. Moreover, among the elderly and the disadvantaged, such policy changes are likely to lead to delayed care, higher morbidity and mortality, higher health costs, and widening health disparities. It would put us to the right of the yellow line, maybe further right than we already are if I'm right about this.
We would do much better, all of us, with a true, integrated system that is designed to align private incentives to make money with public objectives to have a healthier, happier populace. Such policies would also be likely to reduce the high cost of end of life care (which I have for the most part ignored in this explication) since a healthier population will have fewer high-cost complications and comorbidities that drive up costs at the end of life. It's not unlike finance. Aligning private incentives with public objectives for improved health and well-being and lower health-related costs and expenditures would make us all better off. Since many existing private incentives run counter to this, it almost certainly will require a bigger, not a smaller, role for government. This is why I continue to believe that PPACA is an important first step in the right direction and that we, the lucky, healthy and insured, are unlikely to suffer much from the changes that result. It would be a real mistake to repeal this groundbreaking law.