• I’m screaming, because Austin can’t

    Austin is out of town, I believe, so I’m taking it on myself to ARGH for him.

    The NEJM has published a piece entitled, “What Business Are We In? The Emergence of Health as the Business of Health Care,” that’s getting a fair amount of press. It makes the argument that the health care industry should be more focused on the outcome people want – “health” – and less on the mechanism by which the industry currently tries to achieve that outcome.

    However, this paragraph caught my eye (emphasis mine):

    None of this evidence suggests that health care is not an important determinant of health or that it’s not among the most easily modifiable determinants. After all, we have established systems to support the writing of prescriptions and the performance of surgery or imaging but have found no easy way to cure poverty or relieve racial residential segregation. But the evidence does suggest that health care as conventionally delivered explains only a small amount — perhaps 10% — of premature deaths as compared with other factors, including social context, environmental influences, and personal behavior.4 If health care is only a small part of what determines health, perhaps organizations in the business of delivering health need to expand their offerings

    NO! No, no no, no, no!

    This factoid is commonly cited. But readers of the blog will know that Austin has thoroughly debunked it. He started here, when he followed up on the McGinnis study cited in the NEJM piece. It turned out that the “10%” number came from expert opinions. So Austin wrote:

    Still, it seems to me there is no good reason to accept the 10% figure at the top of the left-hand side of the infographic. I’d like to know more how it got there. I’d like to hear the best argument as to why it’s correct.

    Because Austin cares about, you know, facts, he followed up here. He found more citations that went nowhere, and “unpublished research”. He said:

    It seems like folks want to say that 10% of our health is due to the health system (or access to it) without actually being able to point to anything substantial to back it up. Maybe my readers with super Google skills can track this down. I’m not saying there’s nothing to this 10% figure. I’m saying we shouldn’t have much confidence in it unless and until we can see the research that supports it. Where is it, and why didn’t the Bipartisan Policy Center track this down and cite a source that has some actual work behind it?

    Excellent questions. Still, he wasn’t done, and wrote a piece at the AcademyHealth blog. You should go check it out. It was incredibly exhaustive, far more so than anything else I’ve seen. His conclusions:

    There are other papers one could read on these topics, but this is all I had time for and/or could get in full text. From these, it looks like reasonable figures are 40% of reduction in cardiovascular mortality are due to medical care, while 66% or more of reductions in infant mortality are due to medical care. Therefore, I think Cutler’s ballpark of 50% of longevity due to medical care is reasonable.

    When people want to discount the importance of access or minimize the problem of being uninsured, they say that it doesn’t matter anyway. They cite the 10% number to dismiss the actual value of getting health care. But it’s not true. Health care has much more value than that.

    It’s not all that matters. Even at 50%, that means there’s lots of other stuff that affects mortality as well. But here at TIE we like to argue from evidence. The “fact” that only 10% of longevity gains are gue to health care is a zombie idea. It needs to die.

    @aaronecarroll

    Share
    Comments closed
     
    • All well and good, but this doesn’t address the flip side of access to care – harmful effects of unneeded treatments and medical errors, or healthcare’s contribution to mortality..

      If you assume the newly-insured will experience comparable rates of medical errors and mortality from medical interventions as everyone else, then roughly as many newly-insured Americans will die from medical errors as are estimated to die from lack of health insurance.

      Increasing access to a system as highly fragmented and dangerous as America’s healthcare system is a double-edged sword – and the risks of it cutting both ways should not be underestimated.

    • While the 10% number may be unsubstantiated, it goes hand in hand with what the CDC has been saying for years: 75% of healthcare costs are preventable by simple lifestyle choices. Proper diet, exercise, not smoking or abusing substances, and other basic precepts that almost every person understands, but few follow. That’s what drives good health and keeps people out of the physician’s office, pharmacy, etc. I suspect Asch and Volpp are looking at that much broader picture.

    • You know it is not hard and fast what is healthcare. If you include vaccinations I think that the percent that is due to healthcare is very high. I would say that most expensive healthcare spending has little effect.

    • I am a new comer to the field, so forgive my uninformed question.

      Dr. Caroll- can you explain how such figures are calculated when there is no way to (definitively) tell if someone needed to die from a certain condition? That is, how do these studies incorporate the counterfactual in the analysis? Thanks for any clarification you can provide.

    • I think a good deal of this is ships passing in the night. Just what does each commentator define as “health care”? Starting from (say) 1800, it must certainly be the case that medical care accounts for a very small percentage of longevity gains compared to clean water, sewage systems, and insect control. Then add in vaccines and medical care per se (anyone can administer a vaccine) looms even smaller. On the other hand, all these life saving solutions were in place 50 years ago. What accounts for longevity gains since? Here I think Austin is entirely correct: improved emergency care, newborn care, etc. are surely at least half, and I think far more than half, of recent longevity gains. Of course, improved auto safety (seat belts and much more) are a major factor as well, particularly if one measures results in life years rather than lives saved. As to issues such as “lifestyle” choices, aka obesity, the matter is even more muddled, since statins (for example) do a great deal to increase longevity for hefty folks like me. What is the counterfactual? Then we have to deal with smoking cessation (a huge factor) and organic foods (zero effect). It would certainly be fascinating to see someone sort all this out in terms of counterfactuals and empirical data. Meanwhile, the 10% figure is nothing more than an arbitrary and tendentious placeholder.

    • Here:
      Supporters say the effort is long overdue in an age where preventable disease is the single largest cause of death. Indeed, unhealthy behaviors, like smoking and poor diet, account for 40 percent of premature deaths in the United States, while poor health care and limited access to the health care system accounted for a tenth of such deaths, according to an analysis of federal data and mortality studies by J. Michael McGinnis, a senior scholar at the Institute of Medicine.

      http://www.nytimes.com/2012/09/11/health/door-to-door-in-oklahoma-city-preaching-healthy-living.html?partner=rssnyt&emc=rss

      And Here:
      In South Carolina we are instead asking, “How do we most improve the health of our citizens?” and it leads us down a different path. First, when we focus on health and well-being, rather than health services and health insurance, we look to the social determinants of health. This well-documented model suggests that health services contribute 10-20 percent to overall health and well-being of an individual and community, while health behaviors and personal choices, income and employment, education, genetics, social supports, race, and place are much larger contributors

      http://www.nytimes.com/2012/09/11/health/door-to-door-in-oklahoma-city-preaching-healthy-living.html?partner=rssnyt&emc=rss