• The great Medicaid debates

    One of my passions that I put aside when kids arrived is Go. And, one of the great books on Go is titled “The Great Joseki Debates.” It’s a must read for any serious student of the game.

    We’re now having “the great Medicaid debates.” I wish it were all fun and games, like Go. But it isn’t. It’s serious, on at least two levels. The level that will get most attention pertains to policy. What should we do about Medicaid financing? How much flexibility should be afforded to states? How generous should the program be?

    In some sense there is no right answer to those questions. That brings me to the second level of the debate, the one I have spent and will spend more time on this blog writing about: what are the sound research methods for assessing the consequences of Medicaid policy (or any policy, for that matter)? There are right answers to this question. Some research methods are appropriate. Some are not.

    I have focused on this aspect of the debate because you can find the policy debate so many other places. What more could I say that isn’t said elsewhere, and why should my policy preference matter to you? What you won’t find elsewhere is a deeper understanding of the research and what studies are most credible and why. In fact, elsewhere you will mostly — but not entirely! — find misuse of research, cherry picking of studies and results within them. That is not what I do. I think you should care about sound research methods. That’s why I do what I do.

    Having said that, I recognize that I don’t always write about research in an accessible way. Some can follow. Some can’t. Yet I do want everyone to understand, so I’ll keep trying. I am encouraged and delighted that Jon Cohn understands. His new post nicely summarizes the key issues about studies of Medicaid outcomes.

    The basis for the claim [that Medicaid is worse than no insurance] is a handful of studies, chief among them a clinical study from researchers at the University of Virginia, in which people on Medicaid end up with worse outcomes than people with no insurance at all. As a general rule, the problem with studies like these is the underlying differences in the groups being studied: Simply put, the uninsured, overall, tend to be healthier than people on Medicaid. That’s going to skew the raw results, with the uninsured getting better medical outcomes. The results will suggest correlation, not causation.

    Good studies adjust for this fact and, to their credit, the University of Virginia researchers tried to do that. But, as Austin Frakt and Harold Pollack have written, the researchers controlled only for “observable” factors–age, disability, presence of certain medical conditions, that sort of thing–that were present in the medical records made available to them. Anybody who has studied the Medicaid population closely–and by that I include not just academics but also journalists, like myself, who have interviewed providers, patients, and social workers extensively over the years–will tell you that the differences in the populations go beyond these clinical markers.

    The University of Virginia researchers actually acknowledged as much in their paper, mentioning a whole list of factors they couldn’t address. Hospital staff, for example, are more likely to help the sickest patients navigate the enrollment process into Medicaid, which can be famously difficult. Medicaid patients may also end up in worse health because they have fewer family and community supports to keep them healthy or to get them help if something goes wrong. One key sign that the University of Virginia paper obscures the true distinctions among its patient populations is that Medicare patients also ended up with worse outcomes than people with no insurance, even after their adjustments. If there’s a theory for why seniors would be better off uninsured than on Medicare, I’m eager to hear it.

    Austin and Harold, bona fide experts both, offer more details in their exhaustive blog posts on the subject. In so doing, they draw on a large body of research published by some of the smartest and most respected scholars in the field.

    There is much more in Cohn’s piece and I encourage you to read it in full. Trust me, the great Medicaid debates are not over. If you want to keep up, read Cohn. I’ll handle the science, he and others should (and I hope will) stay on top of the policy developments.

    Share
    Comments closed
     
    • The American Enterprise Institute blog links to a WSJ article by Scott Gottlieb with the title “Medicaid Is Worse Than No Coverage at All” (http://online.wsj.com/article/SB10001424052748704758904576188280858303612.html)

      Once a meme like this takes hold, it can be impossible to dislodge (see also the CNBC’s bogus study purportedly showing that a third of all US wages now come in the form of welfare, debunked by Ryan Chittum here: http://www.cjr.org/the_audit/cnbc_misleads_on_welfare_state_1.php)

    • Hopefully the Oregon Medicaid Experiment will finally quash the debate. We’ll have to wait a bit, though.

      • @Aaron – I hope so. But there is attrition from the waiting list. So, if you believe that those who stay are (unobservably) prone to worse health, there’s still a problem.

    • I wish to raise an existential issue.

      A governor is desirous of a block grant system; wishes to cast MOE aside and loosen the HHS vice on its Mcaid regs, etc.

      Federal government studies the state’s population, agrees to a waiver [fantasy], and transfers a Medicare or commercial equivalent, risk adjusted sum to the state to approximate per capita cost for a set period of time.

      Federal government gives state #2, with FFS and managed mcaid infrastructure in place, actuarial equivalent dollars in order to upweight fees and recruit additional providers, Essentially,”tweak the old system.”

      After set period of time, on the net net, both sides perform about the same. Some groups, maybe disabled, do better on one side, those bedeviled with substance and psych issues do better on the other. Overall though, using costs and crude outcome measuress–not much difference. I have no evidence this will be so, but if I was a betting man…

      Would those that are waging a contentious battle against current framework then. ever concede that its not about health care, hospitals, providers, per se., but something bigger; the issues spoken of often on this blog and others but usually glossed over elsewhere–mainly the indisputable robustness of determinants of health and their impact,

      The disparities that most mention for the briefest of moments and usually in a cognitively dissonant way are where the greatest juice to squeeze ration resides.

      If someone proposing block grants (sorry to say, they have an image problem), could articulate with broad strokes some insight into this conundrum, many more folks would listen up.

      But again, I dont know if the Haley Barbour’s of the world are posturing, poorly counseled, or something else I am missing. Want to change status quo, you got to try harder and show me some substance. I am not seeing it, but I will wait.

      • @Brad F – If I’m following, I think I raised this very point near the end of one (or two) posts on Medicaid and outcomes. If one accepts observational controls as sufficient, what will be said when studies using those methods continue to show Medicaid with poor outcomes, even under any manner of reforms? I think that’s exactly what will happen and the IV studies and the Medicare results from the UVa surgical outcomes study suggests it will. When it is selection, not the program, driving the outcomes, you can’t change that by changing the program. That’s not to say the program can’t use reform, which is the ironic twist here. Never have I disagreed with that. Nor do I think it is necessary to disparage the IV studies to hold that position. Why then does one need to believe Medicaid kills people?

    • “Why then does one need to believe Medicaid kills people?”
      Wanting confirmation of strongly held political/ethical beliefs.

      Far too much of the political debate attempts to “prove” a point using science, when science is not really what’s at issue.

      Thank you for your work evaluating the adequacy of study designs and conclusions. I do think providing health care is a proper function of government, but no one benefits from misinformation.

    • According to my own monitoring, millions of people all over the world get the loans from various banks. Thence, there is good possibilities to receive a term loan in any country.