Attention conservation notice: Combines quibbles about what's in an academic paper on tooth-brushing with more quibbles about the right way to do causal inference.
Chris Blattman finds a new paper which claims not brushing your teeth is associated with higher risk of heart disease, and is unimpressed:
Toothbrushing is associated with cardiovascular disease, even after adjustment for age, sex, socioeconomic group, smoking, visits to dentist, BMI, family history of cardiovascular disease, hypertension, and diagnosis of diabetes.
...participants who brushed their teeth less often had a 70% increased risk of a cardiovascular disease event in fully adjusted models.
The idea is that inflamed gums lead to certain chemicals or clot risks.
In the past five days I've seen this study reported in five newspapers, half a dozen radio news shows, and several blogs. These researchers know how to use a PR firm.
Sounds convincing. What could be wrong there?
OH WAIT. MAYBE PEOPLE WHO BRUSH THEIR TEETH TWICE A DAY GENERALLY TAKE BETTER CARE OF THEMSELVES AND WATCH WHAT THEY EAT.
I'm consistently blown away by what passes for causal analysis in medical journals.
Now, I am generally of one mind with Blattman about the awfulness of causal inference in medicine — I must write up the "neutral model of epidemiology" sometime soon — but here, I think, he's being a bit unfair. (I have not read or listened to any of the press coverage, but I presume it's awful, because it always is.) If you read the actual paper, which seems to be open access, one of the covariates is actually a fairly fine-grained set of measures of physical activity, albeit self-reported. (I'm not sure why the didn't list it in the abstract.) It would be nice to have information about diet, and of course self-reports are always extra dubious for moralized behaviors like exercise. Still, it's not right to say, IN ALL CAPS, that the authors of the paper did nothing about this.
In fact, the real weakness of the paper is that they have a reasonably clear mechanism in mind, and enough information to test it, but didn't do so. As Blattman says, the idea is that not brushing your teeth causes tooth and gum disease, tooth and gum disease cause inflammatory responses, and inflammation causes heart disease. Because of this, the authors measured the levels of two chemical markers of inflammation, and found that they were positively predicted by not brushing, even adjusting for their other variables (including physical activity). So far so good. Following the logic of Pearl's front-door criterion, what they should have done next, but did not, was see whether conditioning on the levels of these chemical markers substantially reduced the dependence of heart disease on tooth brushing. (The dependence should be eliminated by conditioning on the complete set of chemicals mediating the inflammatory response.) This is what one would expect if that mechanism I mentioned actually works, but not if the association comes down to not brushing being a sign that one's an unhealthy slob.
The moral is: brush your teeth, for pity's sake, unless you want to end up like this poor soul.
Posted by crshalizi at June 01, 2010 13:25 | permanent link