Rudely introduced »« Not again

Guest post: the real reason explanatory frameworks matter

Originally a comment by Chris Lawson on Folk remedies with lashings of meridian.

With due respect, I am becoming increasingly frustrated at seeing this argument trotted out repeatedly against acupuncture. I’m just about the strongest possible advocate for evidence-based medicine you can find and I think the vast majority of “alternative” medicine is bunkum, often dangerously so, but this particular argument has broken legs and ought to be taken out the back of the stables and put out of its misery.

What is true: there is a lot of published evidence favouring acupuncture, but most of it is very poor quality. The traditional Chinese theory behind acupuncture — that of Qi and meridians — is utterly wrong. We know this because (1) there are no known anatomical structures that follow the distribution of meridians, (2) even if some future scientist astonishes us by finding the appropriate structures, they certainly won’t be physically capable of moving air (which is what Qi means; modern quasi-scientific “spirit” or “energy” apologetics can be ignored) around the body, and (3) there are now several studies showing that sham acupuncture, that is random needling, is exactly as effective as traditional acupuncture, thus demonstrating that the traditional meridians have nothing to do with the medicinal effect of acupuncture. And yes, Chinese medical practitioners ought to be ashamed of themselves for sticking to a patently untrue theory. And yes, Asma’s pseudoscientific defence of Chinese medical theory deserves to be pilloried (and I’m pleased that Pigliucci and Boudry took it on). This much I agree with.

But…there is a big gaping hole at the center of this particular epistemic argument. Where Pigliucci and Boudry go wrong is in assuming that the explanatory framework is what makes a treatment evidence-based. It’s not. What makes medicine evidence-based is the strength of the evidence for a positive benefit. Counter-intuitively, I try to teach my medical students that they should be *distrustful* of theoretical explanations for why they should use a treatment. This kind of explanation is strongly favoured by pharmaceutical company marketing departments because they know full well that providing an explanation for why something works makes people more likely to believe they work. This line of thinking (not always due to pharmcos) has led to medical disasters in the past, including the use of flecainide routinely post-MI which probably caused upwards of 50,000 avoidable deaths per year in the US while it was in favour…purely because cardiologists thought they understood its mechanism of action well enough in prevent arrhythmias. In reality, when they tested it against the outcome that mattered, not arrhythmia frequency but deaths, flecainide was killing about 5% of the people it was given to post-infarct.

The flip-side is also important: there are many treatments in Western medicine that have poorly-understood mechanisms of action. Paracetamol (acetaminophen to Americans) is a case in point: a very effective analgesic sold by the millions of doses every day around the world, with a still only partially-unravelled mechanism. Even more surprising, we don’t really understand how general anaesthetics work.

Anyone who says “this is scientific because we understand how it works” is missing the point. Explanatory frameworks are important not because they prove what works (otherwise we’d still be treating ancient Greek humours), and even wrong theories can be useful (we can use Newtonian physics and ignore special relativity for most daily applications), but they’re important because they help us work out fruitful avenues for future research. There is an infinite number of possible experiments that can be performed. Explanatory frameworks help us direct our energies towards experiments that are likely to give us interesting answers. And in medicine, they help us weigh up risks and benefits when we don’t have as much evidence as we would like (which is often). That’s the real problem with meridians and Qi. It’s not just that the theory is wrong, it’s that the theory is so broken that it’s useless for designing future experiments or for deciding whether to use acupuncture for a given patient.

And follow-up comment:

Oh, and contra Pigliucci and Boudry, the mechanism of aspirin is not nearly as well understood as they present it. It’s certainly true that aspirin alters cyclo-oxygenase metabolism and that probably explains its anti-pyretic and analgesic action, but we still don’t understand why it can trigger the very nasty Reye’s syndrome, don’t know how important its effect on DNA transcription is, don’t know why some people develop allergic reactions to aspirin and others don’t, and so on. When I read their piece on how well we understand aspirin, it reminds me of A. A. Michelson’s 1894 opinion that science was so well known it was going to be about the “sixth place of decimals” from then on.

Comments

  1. leftwingfox says

    Good points.

    I see something similar happen a lot in politics, where pre-concieved notions of ideology, morality, or “common sense” are given weight over the evidence of the effectiveness of those policies (such as “Absinence-Only” sex-education)

  2. rq says

    This is great. I think it takes down every possible woo out there, quickly and clearly. I have friends who need to read this…

  3. quixote says

    Where [people] go wrong is in assuming that the explanatory framework is what makes a treatment evidence-based. It’s not. What makes medicine evidence-based is the strength of the evidence for a positive benefit.

    Yes, yes, yes, yes, yes. I’ve been trying to make this point here and everywhere, but I haven’t managed to do it this clearly. Thank you.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>