The debate about whether health insurance improves health has been ushered into prominence by the need of proponents and detractors of the ACA to make their case against the other side. It is an important discussion, regardless of why is has popped up (examples of good debate: one | two). Much of the best debate has been technical in nature, hinging on issues of research design, internal versus external validity, statistical power and the choice of measures. The latest round of skirmishes has ventured into the area of whether the benefits outweigh the costs, always a fair question in public policy as is seeking to fully identify the costs and benefits.
This morning I did something I may not have done since I graduated with my PhD 19 Mays ago-I read some of my dissertation “Alternative Measures of Medical Underservice: A Proposal and Assessment.”
Most of it is quite technical-it developed a more precise statistical means of allocating finite remedial resources such as funds for Community Health Centers and the National Health Service Corps using a system of structural equations, and tested the sensitivity and specificity of different measures using geo-coded data from the NHIS. I even proposed the Taylor Index of Medical Underservice (TIMU). And since you have never heard of TIMU, you get a sense of how that went.
However, leafing through my dissertation and reading through the dust, my attention stopped on section 1.4: Social Meaning of Medical Underservice, the beginning of a short section where I focused on why it was important to undertake policy to address medical underservice, essentially justifying the next 350 pages (~6 page pdf scanSocialMeaning). My reasoning was heavily influenced by Michael Walzer’s book Spheres of Justice: A Defense of Pluralism and Equality (1983; Basic Books). The essence of his argument is that the importance of a given good differs across cultures and time, as does the just distribution principle of a good, with this being determined based on the importance that members of a given society assign to a particular good. I think my thesis on why medical underservice deserved to be addressed is found in the middle of page 20 (scanSocialMeaning)
In this research, need is seen as the proper means of distributing health care services, not because it is absolutely clear that improved access will necessarily improve health, but because society as a whole is convinced that it will. Because of its importance in our culture, and the vast public expenditures to support the health infrastructure, access to the health care system should be provided to all members of this society on the basis of need “as a basic token of citizenship.” (Beauchamp, 1988, p. 4). [emphasis was in the original]
I spent very little time on the technical issue of whether you could prove increasing access to care would make people healthier, and I reviewed the uncertainties about this and noted that the evidence was mixed, though I think it tilts toward improved access being positive for health. (I even cited Ivan Illich’s book Medical Nemesis which can be briefly described as an invitation to run from modern medicine).
As I sit on my deck and think of the past 5 years, I am again reminded that one of the things that advocates of health reform learned to do was to make the case for reform in “business case” types of language (triple aim, reduction of cost shifting, transform toward sustainability, and the like). However, we seem to have buried or lost the ability or the willingness to make arguments like the one above, that the 19-year-ago-me would have started with.
cross posted in freeforall
Over those two decades, we've moved ever closer to a world where, if you can't buy it, or sell it, it doesn't exist; where, if you're not a buyer or a seller, you don't exist.
The idea of a basic token of citizenship has been replaced by an ante, which if you don't make, you don't even get dealt cards, never mind play and lose a hand.
Wow. Terrific comment
Medical insurance increases welfare by removing the risk of financial disaster from unpayable medical bills, or the health disaster of missing out on treatment because it's unaffordable. The proposition doesn't assume treatment is more than palliative. Formally, all you need to assume to prove the intuition is diminishing marginal utility of money or health for "rational" citizens, or loss aversion for "irrational" ones. This argument does work better for catastrophic than for routine care, for which you may need the case based on citizenship or fundamental rights, plus rather stronger assumptions about effectiveness.
Put it this way: denying a sick citizen the affordable medical care she needs is an affront to the values of the polis as a community of shared risks and opportunities. Denying it to a sick stranger is an affront to common humanity.
Thank you, James.
This thought clarifies one of the puzzles about the pre-PPACA health regime in the U.S. On the one hand, we had a failed attempt at a pair of (linked) free-market systems for health care provision (Market I) and health care finance/insurance (Market II). On the other hand, we had statutory provisions like EMTALA (Emergency Medical Treatment and Active Labor Act) that insisted that any hospital that received federal funds (e.g.., Medicare reimbursements, and so essentially all hospitals) that had an emergency department had to provide care to at least stabilize the patient.
This had the completely predictable effects on utilization patterns. Why such a completely irrational system would be allowed to come into existence baffled me. Now I get it.
Now, if we could only the GOTPers to get it, too. But as my other mother was fond of saying, "If wishes and buts were candy and nuts we'd all have a merry christmas."
As I sit on my deck and think of the past 5 years, I am again reminded that one of the things that advocates of health reform learned to do was to make the case for reform in “business case” types of language (triple aim, reduction of cost shifting, transform toward sustainability, and the like). However, we seem to have buried or lost the ability or the willingness to make arguments like the one above, that the 19-year-ago-me would have started with.
Spot on Don. Many liberals (academic liberals particularly) have a habit of hiding behind Excel charts and saying "the data say that we must do X". But of course the data don't care what we do, that is a question of values. Conservatives in contrast tend to be comfortable making explicitly moral arguments. The public seems to respect the honesty of the conservative approach even if they don't agree with the value stances thereof, and I think they are right to do so - honesty is a good in the public square. The left would be more effective if it dropped its pose of amorality (which isn't fooling many people anyway) and argued the value premises more directly.
"I spent very little time on the technical issue of whether you could prove increasing access to care would make people healthier, and I reviewed the uncertainties about this and noted that the evidence was mixed, …"
Don, in my 70+ years I've also spent very little time on that technical question, which I never considered to be an "issue." It seemed so obvious to me that it never occupied my consciousness. But today, reading this post, I spent two minutes on it and came up with this simple, but incontrovertible, PROOF, which then took me thirty minutes to write up.
Consider a function, F(H,C) relating Health to Care. The statement of the question implies that the function has a measure. I.e., if we could collect ALL the data, and then we could do an experiment where we controlled an increase in the independent variable, we could measure the change (or non-change) in the dependent variable. So for the function F, we can write H=F(C), recognizing that this function may not behave as we'd like, but claiming that the function exists, which means it can be graphed from data, even if we can't resolve the algebraic expression of the relationship.
Now comes a little appeal to mathematical intuition, for those not versed in the math. Consider the possibility that we live at exactly the point of a singularity of the function. That is to say, consider that the function we have claimed to exist actually does exist over a wide range, but at precisely the point on the graph where we happen to live today, the function does NOT exist. My intuition rejects that possibility. There is nothing so unique about today that a relationship that holds generally over a large graph of possibilities does not hold specifically at my infinitessimal local neighborhood on the graph. So if there is some hitherto undiscovered relationship between Care and Health, my intuition rejects the possibility that the relationship does not hold true in our exact situation. This is a loosely stated non-mathematical argument that the function F is continuous in at least some neighborhood around Today.
By the same argument, I would propose that the derivative of the function F with respect to C is also continuous. It may be zero in Today's neighborhood, or it may be positive or negative, but it can't be infinite, which would imply a singularity in F.
OK, having established that we can consider a function F, let's do that. Let's test the hypothesis that changing C might NOT change H, i.e., that the function has a zero slope in the neighborhood of Today. But instead of trying to add C and find the change in H, lets SUBTRACT a little C. Let's consider the hypothesis that taking away a little C does not reduce H. But how do we take away a little C? Well, we can eliminate a med, or we can eliminate a doctor visit, or we can eliminate a little therapy, or … But one way or another, we need to eliminate a little something in the Care set.
Now, it is possible that any given reduction of one element in the Care set might have no effect. If I miss my physical therapy session, my sore back might hurt a week longer, but a decrement in my health is open to question. And clearly, another reduction might have a drastic effect. If I have a heart attack, and the ambulance doesn't show up, I'm probably a goner, which is a BIG negative for me.
So how does this relate to the question, or prove the answer to the question? Well, the question isn't about MY health, it's about OUR health, being a collective of 300 million individuals. So we can address the aggregate Health measure of us all, or we can individualize it by considering the measure to be an expected value measure. Either way, it's obvious-a negative change in C will inevitably include some component of Care that's essential, like ambulances and insulin injections. So it's clear, I think, that reducing C will reduce H.
And now we go back to the math. If F(H,C) and its first derivative are continuous in a neighborhood around Today, and if a small negative increment in C produces a small negative increment in H, then in that small neighborhood around Today, a small positive increment in C must produce a corresponding small positive increment in H.
It's not a health policy question. It's a math question.
I think there is too much analysis here, Ken.
The first problem is that any simple measures of H and C are composite measures. In reality, both are more probably vector-valued quantities. It's obvious to me that simple measures of C like dollars or fraction of GDP are wholly inadequate. We could have a reduction in expenditures by reducing elective surgery (say), and H would (probably) increase because nosocomial infections and other iatrogenic complications are zeroed. Similarly, we could increase C by encouraging known ineffective screening procedures (e.g., mammography or prostate specific antigen tests) in groups not at risk and H would decrease because the false positives result in still higher expenditure with little (population) benefit.
The second problem is that both H and C are statistical measures. We have to consider how changes in the (multivariate) distribution of C are reflected in changes in the (multivariate) distribution of H. Something like a Leontief matrix function is going to be necessary. It's possible that HRSA has some of the data to construct this, but I'm not even sure of that — the outcomes are the problem. Mortality can be found (sort of) from death certs, but the cause of death recorded there isn't entirely reliable. But mortality isn't a very good measure of H.
I think we are on stronger ground by making the ethical arguments. The mere existence of EMTALA and similar statutes provide strong argument that the polity believes that basic health care is a right or something like it. The question then becomes, "What is the right level of spending?" with a subsidiary question, "What is the right mix of services?" I think the macroeconomic data the G-whatever economies tell us we are above the optimal level of C.
The more interesting question to me is this: when will we be adding dental care to ACA mix? If you talk to anyone who has dealt with the uninsured and their issues at any length, you'll discover that dental care is as important as direct medical care.