• TIE-U, Fall 2012: Mortensen’s Introduction to Health Systems

    This semester I’ll be blogging along with Karoline Mortensen’s Introduction to Health Systems (UMD’s  HLSA 601). The syllabus is here (PDF). This first post will serve as an index to all subsequent ones. Below the following index are a few comments about one of the first week’s readings.

    Post Index for UMD’s HSLA 601 (Mortensen)

    Comments on How Public Health Policy is Created, by Alfred Sommer (American Journal of Epidemiology, 2001)

    The paper by Sommer is both short and ungated. I recommend you read it. Though there’s plenty to agree with, I want to push back on this sentence in the third paragraph:

    In an ideal world, health policy would be formulated in a rational, linear process, moving from data collection, to interpretation, to scientific consensus.

    This is a popular sentiment among researchers. I may have expressed the same some time ago. But I now think it’s wrong.

    In a TIE guest post, Keith Humphreys commented on his paper Scientific evidence alone is not sufficient basis for health policy, coauthored by Peter Piot (BMJ, 2012). Keith wrote,

    What we do with scientific evidence is always a political and moral judgment. We don’t provide health care to the sick because the evidence forces us to. We provide health care to the sick because such activity is in keeping with our values. Likewise, we might choose to morally oppose certain policies (e.g., capital punishment) even if there is solid scientific evidence of benefit (e.g., if it is ever shown conclusively that capital punishment reduces crime). It would be dishonest to hide behind the evidence and say, for example, that science made us put a helpless human being to death; that moral judgment falls on all of our heads.

    Since his BMJ paper is ungated and but his post brief post is not.

    Despite where it starts, Sommer’s paper seems to wind up agreeing with Humphreys and Piot. The final sentence is,

    Health policy involves far more issues than epidemiologists and other scientists know or care about.

    Agreed. So it is not clear in what sense a rational, data-driven world guided by scientific consensus would be ideal. More than that, it is not so clear how such a world could exist. Decisions are informed by data but not made by them.

    Postscript: The other paper assigned this week is a summary of the IoM’s Crossing the Quality Chasm (2001). TIE has at least one  post about that report.


    • The link to the Humphreys paper is gated and I didn’t easily find an ungated version on the Web.

    • I appreciate the distinction between values and scientific evidence. In the ideal world perhaps we choose _what_ to do based on our values while evidence can help us figure out _how_ to do it. Of course as can be seen with health care (and as you point out with crime) the how has a value component as well. For many people an individual mandate has become an unacceptable approach based on values (and political positions) rather than any judgement of whether or not it will work to cover the uninsured.

      I find it interesting that most of the opposition to the mandate is on grounds like “infringement of liberty” and “un-American” rather than “that’s a dumb idea because it won’t work”. Perhaps that proves your point about values.

    • “Decisions are informed by data but not made by them.” is an unsupported assertion. Surely how decisions are made depend, in part, on whether the thinking is “fast” or “slow.”

      I find in my own decision making processes — if I clearly witness them — that sometimes data also influences my values, and may change them. I believe both values and evidence-based rational processes are susceptible to change. Though surely susceptible to a different degree.

      Two determinants of change are the strength of the evidence and how it is presented. A personal example of the former is a brief 90%-done essay which is meant to weigh on and persuade regarding values and policy. The abstract:

      This essay reveals the population effects associated with our market society and individualism. To accomplish this, I use a measure of conservatism as a proxy for the two ideologies. This proxy is then compared with outcomes of well-being and health. In every case higher concentrations of conservatism strongly correlate with worse average outcomes. The results are put in context and suggestions for the future are made.

      The current draft can be found here:

      Furthermore, unless evidence is presented so it is easy to accept it will likely be ignored. This from Hilary Mantel’s “Bring Up the Bodies”,

      Truth can break the gates down, truth can howl in the street; unless Truth is pleasing, personable, and easy to like, she is condemned to stay whimpering at the back door.

      These, strength and presentation, are also essential ingredients for the efficacy of evidence. However, in the history they are usually missing.