The study compared doctors who maintained certification against those who didn't and found the former were better at ordering key screening tests.

So just last week I got this e-mail in my inbox:

Ugh.

So I went and paid my fee to the American Board of Internal Medicine, and realized I’m just about due to recertify my Internal Medicine Boards. Ok… so I paid for that:

Drip by drip.  After payment, I clicked on a link to schedule my exam and got this:

Ugh again.

That brief experience is just a taste of the frustration internal medicine physicians feel when dealing with maintenance of certification – an experience that, like renewing your license at the DMV, seems to be a practical necessity and also a complete waste of time.

Or is it?

This study, conducted by the American Board of Internal Medicine and led by ABIM-employed scientist Dr. Bradley Gray suggests that doctors who maintain their certification take better care of patients than doctors who don’t.

The authors identified 1,260 physicians who were initially certified by ABIM in 1991. This would make them due for recertification in 2001 and 2011. By 2011, 786 had recertified and 474 had not. After accounting statistically for a variety of patient and provider factors (including the docs’ score on the initial board exam), the authors the looked at a variety of patient-level quality metrics using Medicare data. You can see the basic results here.

The data tells a somewhat depressing story. Yes, the people who maintained certification did a bit better on things like ordering mammograms and whatnot, but overall it seems like patients are falling through the cracks– fewer than half of the diabetic patients in either group got all 3 recommended screenings. This data suggests we can be doing a lot better, but it doesn’t look like MOC is the way to greatness if you catch my drift.

But if you want to hang your hat on those small observed differences, the “E-Value” would like a word with you. An E-value is the size of association an unmeasured confounder would have to have with the exposure (in this case maintenance of certification) and an outcome (in this case a given quality metric) to get rid of the statistical significance.  If your E-value is really high, as a researcher you can say – sure – this wasn’t a randomized trial, but for my observation to not hold true there would have to be a crazy powerful, unmeasured third factor coloring the data.

In this case, the E-values ranged from 1.2 to 1.4.  That’s pretty low – just a puff of unmeasured confounding would make these results evaporate like the water in my basement doesn’t. One unmeasured confounder is brought up by Columbia’s Lee Goldman in his delightfully snarky editorial: the metrics “might be a marker of compulsiveness – the same type of compulsiveness required to complete the MOC process”.

But what really interests me most about this study is how the ABIM will handle it.  Will it be spun as a vindication for an onerous process that many physicians are fed up with, or will they (more appropriately in my opinion) use it as a battle cry to improve our care of patients across the board, regardless of board certification?