Can yoga be studied with “conventional” clinical trials?


As of September 2017, new Sweat Science columns are being published at Check out my bestselling new book on the science of endurance, ENDURE: Mind, Body, and the Curiously Elastic Limits of Human Performance, published in February 2018 with a foreword by Malcolm Gladwell.

- Alex Hutchinson (@sweatscience)


I’m generally a pretty skeptical guy. If someone tells me that X really works, but it can’t be verified by science because science is just one way of understanding things, I roll my eyes. But bear with me here. A couple of years ago, I interviewed a guy named Timothy McCall, the author of Yoga as Medicine, for an article about yoga research. He was a smart guy with some very interesting things to say, and I still get occasional e-mail updates from his website. The latest includes a link to this article, from the Spring 2011 issue of Yoga Therapy Today, by Nina Moliver, called “Yoga Research: Yes, No, or How?”

The article is a pretty wide-ranging look at the current debate (in yoga circles) about how yoga and science fit together. A couple of points caught my attention. First, McCall and Moliver argue, yoga is a very slow healing technique:

[E]ven six months is a drop in the bucket for a Yoga practice. By privileging short-term studies and standardized protocols, we are forever studying beginners, [and] we are systematically underestimating the healing potential of Yoga in our research…

The bigger argument is that randomized, controlled trials — of any length — to study yoga don’t work, for various complex, holistic reasons that don’t sound very convincing to me. The alternative is observational studies. And as it happens, Moliver completed an award-winning PhD thesis at Northcentral University last year that used an observational design — an online survey — to study yoga in 211 female yoga practitioners plus 182 controls. Observational studies have a lot of problems, in particular the inability to distinguish between cause and effect, as Moliver acknowledges:

For example, if a researcher didn’t randomly assign the participants, it is not possible to know if Yoga practitioners are happier because they practiced Yoga, or if people who were happier were naturally attracted to starting a Yoga practice.

But there are still ways of extracting useful data. For example, if you see a dose-response effect — the longer people have been yoga-ing, the happier they are — that’s pretty suggestive. And as for those confounding variables:

For the Yoga practitioner, these so-called confounders — a healthier diet, a simpler lifestyle, more time outdoors, more kindness and compassion, more loving relationships, more bike-riding, a better path to right livelihood — are not confusing. They are mutually enhancing and reinforcing.

In other words, who cares if you end up happier and healthier because you’ve aligned energy flows in your body or simply because you’ve spent more time being physically active and mindful — the result is what matters. And indeed, Moliver’s study did see evidence of a dose-response relationship in her subjects, some of whom had been practicing yoga for as a long as 50 years. (I’m hoping the study will be published, as the abstract isn’t very revealing about the “range of intercorrelated wellness measures” that demonstrated the dose-response effect.)

This all sounds very reasonable to me — and in fact, it’s very reminiscent of Paul Williams’ National Runners’ Health Study, which takes a similar observational approach (albeit with more than 100,000 subjects) to tease out dose-response relationships that would be nearly impossible to detect with conventional short-term intervention studies.

One caveat: this approach tells us what works, but it doesn’t tell us how it works. You can’t take an observational study that finds health benefits from yoga and conclude that this proves that we can indeed control the circulation of energy flow in our bodies. To make claims about cause and effect, you really do need proper randomized trials. Notably, Moliver’s study didn’t see any difference between different types of yoga: just doing it, and keeping at it for long periods of time, correlated to better levels of psychological and physical well-being.

10 Replies to “Can yoga be studied with “conventional” clinical trials?”

  1. I did yoga regularly when I was injured and I hated it because it wasn’t running. That said, I do think the key point is that yoga is more of a “culture” or “practice” than an exercise or an input that can be studied. And the benefits are very likely compounded (compounding?). Would you question the notion that the benefits of consistent, easy running (for example) are greater than the benefits of an individual run multiplied by the number of runs?

    I really like your blog and it is important for science to study things, but I think maybe in this case, it’s better to let it go. It would be like studying religion: the point isn’t how it works, the point is that it does work. I’m sure people have done that (there have been studies on prayer, I believe), but I don’t know how useful that is, because if you convert to religion because science proved it was true, well, then you are missing the point. I think the same could be said for yoga.

  2. “I really like your blog and it is important for science to study things, but I think maybe in this case, it’s better to let it go. It would be like studying religion: the point isn’t how it works, the point is that it does work.”

    I don’t agree, John. How do you “know” yoga works if you don’t study it, any more than people 10,000 years ago “knew” that doing a rain dance made it rain? The point of this post is that yes, yoga’s holism makes it difficult to study HOW it works using conventional randomized trials, but (as the article I linked to suggests) it’s still possible to study WHETHER it works using large-scale observational trials that accept yoga’s holism. That sounds pretty reasonable to me — so I’m not sure why it would be “better to let it go.”

  3. It seems to me that the main problem is not something inherent in yoga, but that the wholistic benefits it claims are difficult to quantify.

    It doesn’t claim to make you faster, jump higher or anything like that, which are pretty easy to measure, but stuff like making you more balanced, and unity of mind, body and spirit. You’d have a hard time measuring that for running, rugby and formula 1 racing too. Making you ‘hapier’ or ‘healthier’ in the long run is probably only a very rough translation of that, but the most concrete claim it makes.

    It seems to be perfectly feasable however to measure whether yoga makes certaing joints more flexible. Not the stated goal of yoga, but probably one of the ‘colateral’ benefits it can claim.

  4. I don’t really think you could get WHETHER it works from a study because its holistic nature means that it’s pretty near impossible to tell if it is yoga working or something else. I didn’t mean to be dismissive of scientific study generally, just that in this case, it seems like whatever grant money is being used in these studies might be more wisely spent on something that could eventually result in some kind of application. I didn’t mean that you shouldn’t write about it. You write about lots of “bad” science (like the recent cardio/heart disease study you debunked). I guess I just expected more skepticism from you on this topic.

  5. @RH
    Yes, there are plenty of yoga studies that measure easily quantifiable parameters like stress hormones. They do find effects, but as you rightly point out, these are not the “main point” of yoga. Finding a dose-response effect in an observational trial, as the Moliver study did, strikes me as an intriguing way of studying yoga in a more holistic way.

    @John Lofranco
    I’m honestly baffled by your comments here — I really don’t understand what you’re arguing. That we simply shouldn’t study yoga at all? This blog post was about a novel approach to studying yoga in a more holistic fashion (seeking dose-response relationships in an observational study rather than traditional randomized trials). I think it’s very good science. What is it that you want me to be skeptical about? Yoga itself? The results of the Moliver study? If so, what is it in the Moliver study you object to?

  6. I do agree with you that the design of the study is very well chosen for the type of question it tries to tackle, and, as far as general feelings of well-being are concerned, studying the dose relationship is a very smart way to get most out of the data.

    That said, I do think John Lofranco has a point when he says that studying certain variables like ‘feelings of trancedence’ is a bit like studying if catholics experience more transsubstation. They are particular to the practice to begin with.

  7. A bit of topic, but the beauty you find in Moliver’s the study of the dose effect, reminds me of an article I once read about the birth of evidence based medicine early in the nineteenth century. It is not exactly the same, but has the same clever indirect approach.

    Here, the question was (well, turned out to be), whether bloodletting worked against pneumonia. In those days I suppose that patients would object to being sent away without a proper bloodletting, and anyway, it would be unethical to do so, so a control group was out of the question. The question was settled by comparing the mortality and duration of the disease in those who were bled in the first days after the onset of the first symptoms with those that were bled several days later.

  8. Very, very cool paper, RH — thanks for posting the link. Interesting story about the history, and also some insightful reflections about the nature of “evidence-based” medicine.

    And John, I didn’t mean to jump all over you. Trying to scientifically study yoga is a fraught topic, both from the yoga perspective and the science perspective, so I understand that many people have reservations about it.

  9. No worries. I agree, my disagreement was poorly worded. What I probably should have written was, very simply: “Studying yoga? I’m not convinced that’s do able.” But it’s an interesting try, for sure.

Comments are closed.