Mar 01 2011

That treatment is not based on science? Don’t Worry, says KevinMD.

A guest blog post on the popular health site, KevinMD, by Dr. Peter J. Weiss, promises in the title an explanation for Why alternative care seems to work. The post does not deliver on the promise, and instead offers a naive and confused justification for using unscientific treatments.

He begins:

People sort of want to try alternative care.  They’re working up to it, but then they read more about it; they read about the theory and they say, “No I don’t believe that theory.  That can’t possibly work.  There’s nothing to it. I’m not going to do it.”  But, the problem is that the theory of why something works or the explanation is not necessary related to the effect of how things actually work.

I have written many blog posts and articles, and I know you need to start off with some hook, and sometimes they are a bit contrived. But you should try to avoid just making stuff up – using unverified assertions. Weiss begins with a premise – that people want to try so-called “alternative care” (which he makes no attempt to define) but they cannot get past that pesky lack of any possible mechanism. I would like to see the survey that even suggests this is the case.

I don’t pretend to have the answers either, and when I offer my opinion or personal experience I try to make that clear. Public attitude toward so-called CAM is not a minor point – since it is a major premise of his article. I am not aware of any evidence, and it is not my personal experience, that lack of a scientific mechanism is a major hurdle to public acceptance of specific CAM modalities. Most people I talk to just want to know if something works.

While Weiss acts as if he is speaking to the skeptical public, he really isn’t. His article is targeted at the defenders of science-based medicine (whether he knows it or not).

Science-based medicine essentially holds that the best way to evaluate specific treatments in medicine is to consider both clinical and basic science evidence. Clinical evidence is important, but is very problematic – and so we like to back it up with at least a minimal level of scientific plausibility. It is that notion that Weiss challenges, in a manner that I find particularly naive.

He starts off with a point that is true, but fails to put it into the proper context. He argues that because someone or some group offers an erroneous explanation for how something works, that does not necessarily mean that it does not work. Therefore, chiropractors may argue that manipulation works by correcting non-existent spinal subluxations and restoring the flow of non-existent life energy, but that does not mean that manipulation does not work.

From the point of view of logic, this is true. An unsound argument does not render the conclusion false. I might argue that the sky must be blue because turtles are really slow – a non sequitur. But the sky is still blue.

What Weiss misses, however, is that when a putative mechanism turns out to be false it does call into question the underlying claim. This is especially true when the treatment was developed to address the putative mechanism. In other words, it is not as if chiropractic manipulation was observed to work first, and then chiropractors later tried to figure out how it might work. Rather, D. D. Palmer developed his notions of chiropractic, based on his pre-existing beliefs in life-energy medicine – and developed chiropractic techniques based upon his claims of the underlying mechanism.

Again, this does not mean absolutely that chiropractic does not work. But it does mean it would be an incredible coincidence if it did – at least for the kinds of things often claimed for chiropractic. Chiropractic is a tricky profession to discuss, because it covers so much ground. The above applies best to straight chiropractic – using spinal manipulation to treat asthma, for example. This is a very different kind of claim than that spinal manipulation causes a temporary decrease in back pain by relaxing muscles – a fairly non-specific effect of mechanical manipulation of muscles and not specific to chiropractic. In other words, it is highly unlikely that chiropractic would end up having very specific medical effects, even though the beliefs that led to the treatment turned out to be entirely fictitious. But it would not be surprising if there turned out to be some non-specific effects, like temporary muscle relaxation.

We can also look at this point another way – if a claimed mechanism of action turns out to be wrong, then what reason are we left with for believing that a treatment works? We can then rely upon careful rigorous clinical studies – and Weiss glosses over the fact that for many of the treatments he discusses there is a lack of this as well.

Weiss also discusses acupuncture, writing:

The same is true, I think, of acupuncture, which is another common modality — increasingly common in the western world — that was really developed in ancient China in a prescientific era.  The Chinese developed these theories of energy flow (or Ch’i) in the body, along channels related to meridians; and all of that define the practice of acupuncture.

It does seem that the beliefs came first, and acupuncture evolved as a technique to address those beliefs about health and illness. These beliefs were not just related to energy flow, but also blood and a mystical astrological connection to the heavens.

In fact blood letting is a perfect analogy – letting and purging were developed as techniques to address the philosophy of humoral-based medicine – the notion that health and illness relate to the balance of the four humors – blood, phlegm, green bile, and black bile. Wouldn’t it be amazing if the treatments developed to address humoral balance happened to work through some other completely unrelated mechanism? It’s not logically impossible -I’m just not going to hold my breath.

Weiss continues on his theme, writing:

So this is kind of what I’m talking about.  Don’t get so hung up on the explanation that you don’t believe in, that you’re unwilling to try a practice that might actually help you.  Just keep an open mind.  You don’t have to know everything about how things work; you just have to know that they work. Just like, do I really understand electricity or do I just know that if I turn the light switch, the light comes on?

Ah – the “open mind” gambit. Keeping an open mind does not mean being gullible, ignoring evidence and plausibility. It also means being open to the possibility that something does not work, despite popular belief that it does.

When he writes, “you just have to know that they work,” he is begging the question. There is no compelling evidence that acupuncture works for anything. There is no compelling evidence that chiropractic works for any disease, and there is only weak evidence that it is helpful for back strain (and no more effective than cheaper modalities, like physical therapy or massage).

But he next makes his worst argument, falsely equating the lack of a known specific mechanism with the lack of any possible mechanism.

A good example, I think from western medicine is this drug Lunesta.  You’re probably familiar with it.  It’s the little butterfly that flies around on the commercial while the lady is trying to get a good night sleep and the commercial talks about how Lunesta can help you sleep.  And at the very end of the commercial, the very very end, there’s this little statement, something like, “The mechanism of action of Lunesta is unknown” or “How Lunesta actually works is unknown.”

Weiss should have gone beyond the commercial and indulged in a PubMed search or some other research. Lunesta is not a great example for his point – but it is a great example for criticism of his point. Lunesta (eszopiclon) is a benzodiazepine receptor agonist – it binds to and activates these receptors. Further, it is known that activating these receptors causes sedation. There is also ongoing research looking into all the downstream effects on brain function and sleep regulation (for example).

It is therefore not true that we have no idea how Lunesta works – we have a pretty good idea. We don’t have complete knowledge of its mechanisms, just as we don’t have complete knowledge of brain function and sleep regulation. There are always deeper levels of scientific understanding to plumb – knowledge is not black or white. But Weiss presents this false dichotomy, as if we either know how something works or we don’t.

Worse than this false premise is his major conceptual fallacy – equating not knowing precisely how something works with not knowing how something could possibly work. Even if we don’t know how a medication works, we know that pharmacologically active substances get into the body and then can act through a number of possible mechanisms to influence physiology.

This is not an analogous situation to that of  things like homeopathy. Homeopathic preparations generally have no active ingredient – there is nothing there to influence physiology. Chiropractic has a similar lack of plausibility when it comes to treating diseases, like asthma. There is plausibility when it comes to musculoskeletal problems. Acupuncture is similar in that it completely lacks plausibility for treating diseases (so-called medical acupuncture), but there is weak plausibility for non-specific symptomatic effects, like pain and nausea.

What Weiss is essentially saying is that we should forget any evaluation of plausibility, and focus entirely on the clinical evidence (and this is being generous, because he doesn’t specifically say this – you could interpret his post as advocating for reliance on anecdotal evidence, but since he does not say this explicitly either I will give him the benefit of the doubt).

This approach – ignoring plausibility – fails for two main reasons. The first is that clinical evidence, while essential to support a specific treatment claim, is a very problematic and slippery form of evidence with many false positives. Clinical evidence can become rigorous, repeatable, and therefore highly reliable. But we often do not have such iron-clad clinical evidence. And therefore we need to consider prior plausibility from our basic science understanding of biology in order to put the imperfect clinical evidence into context.

The second major flaw is the failure to realize that plausibility does not simply range from neutral (lack of information) to positive (a known mechanism). There is also implausibility – scientific information that suggests a treatment should not work or cannot work. It is logically impossible to prove a negative – but plausibility is not about proof, it’s about estimating probability from what is known.

There is not only no known mechanism by which manipulating the spine can improve lung function – there is no plausible mechanism by everything we know about biological function. The same goes for curing cancer by sticking tiny needles in the skin or having any biological effect from taking sugar pills treated with magic water.

Having said that, we also acknowledge that scientific knowledge is tentative and subject to revision. So if people want to research treatments that seem implausible, it may not be unreasonable. But we have to set different thresholds for different actions. Doing some preliminary exploratory research (as long as subjects are not being put at risk or being denied standard medical care) can have a very low threshold. Doing expensive and elaborate clinical trials with taxpayer money needs to have a higher threshold for plausibility and evidence. Offering a treatment to patients needs to also have a higher threshold, and mandating that insurance companies pay for such treatments an even higher threshold still.

For patient care, the risk vs benefit analysis is key- what is the probability for benefit vs the probability for harm (we might also include expense, effort, emotional investment, and opportunity on the “harm” side of the scale). For highly implausible treatments the potential for benefit may be negligible, so that no expense or risk of harm is justified.


Weiss’s post on KevinMD is very disappointing, and unfortunately indicates that the filter on that blog for guest posts does not appear to be adequate. I hope it does not indicate a shift in philosophy away from science-based medicine, which would be worse.

Weiss is essentially saying to his readers – don’t think too much about alternative therapies. Don’t get caught up in all that science. Just keep an open (i.e. gullible) mind and give it a try. This is CAM apologetics, not meaningful analysis.

21 responses so far