Thursday, May 2, 2013

Read Stuff, You Should

Happy Birthday to Jo Ann Pflug, 73.

To the good stuff:

1. Aaron Carroll and Austin Frakt go in depth on yesterday's Oregon Medicaid study. I screwed this up in the Roundup at Greg's place last night, and I shouldn't have. The initial spin was that the study showed no health benefits (while also showing substantial financial benefits) from Medicaid, but that's an incorrect reading. It does show some benefits, but mainly the evidence is inconclusive -- and it's not clear, Carroll and Frakt explain, that the study was able to actually reach conclusions on health outcomes. And at any rate, what evidence there is points in the right direction (that is, in the direction of Medicaid helping people stay healthy), but is not statistically significant -- which is hardly the same thing as a claim of "no effect." Anyway, if you want to know more, click through; Carroll and Frakt walk everyone through it.

(I don't, and can't, read every study for which I link to a blog post or news article which summarizes it. But I need to do better than I did on this one).

2.  Dan Drezner, still a pessimist about political science and Congress.

3. And speaking of that first one: advice from Betsey Stevenson and Justin Wolfers, for non-experts, on how to know what to make of quantitative research.


  1. With all the Krugman stuff, the question of partisanship is very much in the air at the moment. Any chance of a post (or link to an old post) at some point on why being partisan isn't always bad?

  2. A republican health wonk claimed that Aaron Carroll made a foolish error in the last post of his to which you linked. It seems unlikely, but here's the quote:

    "This is only a change in purchasing venue for the plans. I don’t know whether the legislators interviewed for this story don’t understand the law, or are purposely making up facts about it for political gain. I also don’t know which of those two options is worse."

    "Oops, sorry Aaron. You are the one who is misinformed here. HHS has made it abundantly clear that employers may not use pre-tax dollars to pay premiums for their employees in the health insurance exchange. They may give employees after-tax dollars (say, by increasing their wages by the amount of the current health insurance benefit they must give up), but that would decrease its value to the employee by 30% to 50%, depending on the employee’s tax bracket."

  3. The study showed a very significant increase in diabetes diagnoses. This study did not show improvements in blood sugar levels, but we have many, many other studies that show current diabetes therapies significantly improve these parameters. So it is very reasonable to infer that the increase in diagnoses had a positive health impact, even though the trial was underpowered to prove this.


Note: Only a member of this blog may post a comment.

Who links to my website?