• In which I agree with Megan McArdle

    I began my editorial on the already famous Massachusetts mortality study,

    How certain are we that health insurance reduces mortality rates?

    Weeks before the Patient Protection and Affordable Care Act (ACA) became law, a provocative column in The Atlantic posed this question. It drew on an unrepresentative subset of the literature that failed to find an insurance–mortality relationship.

    That column was by Megan McArdle. I was critical of her approach and conclusions, and in that I had some company. (See my unpublished letter, written with Aaron and other colleagues, which we posted here.)

    But, in the spirit with which McArdle wrote a reaction to the Massachusetts study, let me update my priors. For, she wrote,

    I still basically think that health insurance improves mortality rates, but that that improvement is unlikely to be huge if you can get results like Oregon. However, after yesterday’s report, I’ve revised the probability of “huge benefits” upward, and you should do the same. And beware of those who are only willing to revise their beliefs in one direction. [Emphasis added.]

    So wise, who could not concur?

    Ross Douthat also quotes McArdle in agreement, then sagely adds that we should stay calm and gather more data.

    [T]he reality is that on most questions — mortality, deficits, cost containment, and coverage — we are by definition years away from remotely dispositive assessments. The Massachusetts research is only significant because the authors had years of post-Romney data to examine; going forward, on just about every metric that matters for Obamacare, we’re looking at a similar timeframe for meaningful numbers and outcomes and results. So if either side, on any of these issues, is going to need to give substantial ground, the only intellectually-responsible way to do it will be … slowly.

    I fully agree. Only there’s one thing we should all understand. With respect to Massachusetts, there was the possibility of constructing a plausibly unbiased control group (counties in states with similar characteristics before the Massachusetts reform). With respect to Obamacare, there isn’t, apart from exploiting variations in Medicaid expansion, maybe (which is not all of Obamacare).

    That’s not to say we can’t do some good research on the effects of health insurance on health care going forward. Just don’t expect studies without limitations at least as significant as those to date. If you’re waiting for that one, dispositive assessment, be prepared to wait a long time. We’re not merely “years away” from it, it’ll probably never occur.


    Comments closed