On Survivors' Costs

Though I would expect nothing less, Frank Pasquale has a fascinating post over at Concurring Opinions.

Excerpt:

The conservative Manhattan Institute recently commissioned a study of a gap in life-expectancy gains over the past 20 years . . .

To make a long story short, the researcher found that found that "longevity increased the most in those states where access to newer drugs . . . in Medicaid and Medicare programs has increased the most."

Unfortunately, budgetary rules often make the federal government concentrate more on the costs of such interventions than their benefits. For example, the CBO counts "increased costs to the Medicare program for extending the life of its beneficiaries" as “survivors' costs.”

While I certainly believe that biologics are a moral as well as a social good (and, a fortiori, so is clinical research), the history of medicine and public health teaches quite well that the therapeutic revolution had very little to do with the tremendous shift in mortality from the mid 19th century to about 1940.  Mortality rates were already in dramatic decline well before 1940, and the first truly efficacious chemotherapeutics were the sulfa drugs of the early 1930s.  Penicillin was not developed until 1943, even though Alexander Fleming discovered it in the late 1920s.  Ergo, biologics simply could not be the primary cause of the major gains in health indices.

This is part of the controversial McKeown thesis, though it is the uncontroversial part, one generally accepted by historians, epidemiologists, and public health scholars alike.  In addition, the findings of the social epidemiologists -- the Whitehall study in Britain, conducted by Marmot -- and much of the work by Daniels, Kawachi, and Kennedy in the U.S. suggests that the most important determinant of public health is far and away socioeconomic status.  This correlation, which was observed as long ago as 1713 by Ramazzini, and more famously by Chadwick and Shattuck in the mid 19th century, is both extremely robust and persists, strikingly, even when variables like access to care are controlled for.

Thus, without having pored through the study itself, I'd be extremely dubious of the claim that the major causal factor in the increased longevity was access to newer biologics.  As Gostin et al point out, it is difficult to contest the claim that we could have a far greater effect on mortality AND morbidity in the aggregate if we devoted more than 5% of our health care dollars to public health and prevention medicine, as opposed to the 95% we devote to acute care (rescue medicine).

That said, the dynamic of "euthanasia by budget" Frank describes is, IMO, both real and deeply problematic, and is in part a consequence of rationing health care surreptitiously, rather than above board, as Oregon tried to do in the early 1990s.  The disability rights community has led the vanguard in watching for these kinds of value judgments masquerading as cost-benefit decisions, and has rightly, IMO, assailed the use of QALYs and DALYs as assuming the value of a disabled person's life is worth less than the value of an able-bodied person's life.

As to the Texas law Frank mentions, I am quite familiar with the inner workings of that law and the surrounding discourse, and there's no question that cost considerations are (and, IMO, ought to be) a relevant factor in end-of-life care.  That said, I'm dubious of claims that hospital ethics committees, at least the ones I'm familiar with, make the decisions they do in end-of-life care solely motivated by a concern to save money.  If not for moral reasons, the liability exposure involved in doing so would exceed even the enormous costs involved in sustained ICU care, not to mention the loss of goodwill that would follow.

None of the above should be taken as an argument that we should avoid seeking ways to expand access to biologics.  But if we really want to increase longevity, I think dollars would be much better aimed at reducing socioeconomic inequality, or, barring that, at least directing the flow of health care dollars to public health and prevention measures at greater than a 19:1 ratio (acute vs. public health preventive care).

Comments

[Patrick S. O'Donnell]

Lots of wonderful stuff here: a different perspective, important facts, references, etc. Thanks.

[Frank]

These are all very interesting points, and I agree on the broad theme. But I think that larger efforts toward equality are on the backburner for the types of reasons Westmoreland gives in his budget article, and I think the expansion of extant entitlements may be the only political option left.

(I admit that focus on this goal probably distracts fro m the more useful goals you mention. Larry Solum's "Constitutional Possibilities" is interesting on this score--should one seek a "comprehensive solution" that is very unlikely or tinker with a higher likelihood of success?)

I also wonder about the degree to which things like SES and health entitlements can be disentangled. Isn't the guarantee of healthcare to someone a way of raising their SES? Certainly a low-income worker is going to feel less anxiety, and more valued as a member of society, if he is assured that society won't just let him die of a condition that could be treated for tens of thousands of dollars.

[Daniel Goldberg]

Hey Frank,

Oh, I agree that as a pragmatic matter, we aren't going to be getting any closer to redistributing income any time soon.

But you do raise an important question, and it is a very serious problem, indeed. Namely, if the social epidemiologists are right and the single most important factor in health and disease is SES, "tinkering" with other measures virtually guarantees that progress -- if it occurs at all -- will be incremental.

OTOH, if we cannot pragmatically reduce SES, then should we simply flog the dead horse instead of trying to at least make incremental progress?

This, I submit, is the paradox of ethics and health policy, one I touch open in my latest paper, actually.

My crude response is that there many things we can do to ameliorate social conditions that do not involve redistribution of wealth. At least we can begin by changing the 19:1 ratio of health care spending. There is no excuse, IMO, for our current imbalanced approach to health.

You're also right to point out the problems in assessing SES, which the social epidemiologists admit is one of the limitations of their theory. While health entitlements and SES are not totally independent variables, I also tend to think that guaranteeing access to care would not, in aggregate, have a really significant effect on raising SES.

I guess what I mean on that point is that the Whitehall studies in the UK -- where access to care is guaranteed -- displayed just such a robust socioeconomic gradient of health. This is partly what Sreenivasan is pointing out in his essay on the social epidemiologists' point. Daniels uses this data as a justification for universal health care. Sreenivasan's argument is that by their own data, efforts would (in theory) be better directed to reducing income inequality, seeing as how universal health care doesn't seem to affect the gradient all that much.