The myth of ‘yoga as medicine'

NEW YORK — Fifteen years ago, a handful of poorly constructed, clearly biased studies purported to show that prayer was a legitimate medical tool. Americans fell for it, and we still haven’t learned our lesson. It’s hard to resist something we want to believe, especially when it comes in a science-shaped box.

Today, people want to believe that yoga will solve their problems. More than 200 studies were published about the health benefits of yoga last year. They appeared both in specialty journals like Ayu, a quarterly about Ayurvedic health research, and mainstream scientific publications, like the Journal of Clinical Oncology.

Yoga is supposed to cure everything from low back pain to short attention span to several forms of mental illness. Yoga is the new prayer: the risk-free, cost-free solution to all of your medical problems.

The evidence is shaky, and the methodology questionable, but we just can’t get enough.

We should, as a nation, revisit the “prayer as medicine” craze annually to remind us of the credulous depths we have plumbed. In 1988, the Southern Medical Journal published a study that asked “born again” Christians to pray for only half of participating patients in the coronary care unit. Those patients achieved better outcomes, aided by questionable statistical analysis, than the prayer-deprived. The author concluded that “prayer to the Judeo-Christian God has a beneficial therapeutic effect.” Similar studies followed. A 1999 paper in the highly regarded Archives of Internal Medicine claimed that prayer “may be an effective adjunct to standard medical care.”

The media was hyping prayer studies intensely by the late 1990s, but the article in Archives was the high water mark. The journal article caught the attention of skeptical scientists, who decried the absurdity of its methodology. The families and friends of patients in the nonprayer groups probably prayed for them, too, making it impossible to separate the control group from the test group. Does it matter how much time a patient is prayed for or to which God his prayers are directed? How do we know people were really praying? (A cheeky Dutch physician even claimed that he was telepathically influencing the results from across the Atlantic, thereby nullifying the research.) The statistical analysis, which should have been easy enough, was also deeply flawed. In the Archives study, the prayer group only won out in two of 35 outcomes measures, which is about what you’d expect from random chance. The authors, however, chose to focus on a small number of obscure cardiac outcomes to support their conclusions.

In 2006, a well-constructed study finally proved that praying to God confers no medical benefit. The Cochrane Collaboration published a thorough review reaching the same conclusion in 2009 and mercifully counseled, “We are not convinced that further trials of this intervention should be undertaken and would prefer to see any resources available for such a trial used to investigate other questions in health care.”

God’s medical career was over, but he left a void in the public discussion of medicine, and yoga has filled it. Studies come out on a near weekly basis trumpeting the benefits of yoga for any problem. Yoga for diabetes. Yoga for high blood pressure. Yoga for heart disease. Yoga for cancer. Yoga for slow reactions. Yoga for bad grades. The quasi-miraculous healing powers of yoga are, I concede, more credible than the truly miraculous healing power of a divine being. At least there is a nexus between health and yoga — the human body — which is something you can’t say for therapeutic prayer.

The yoga studies, however, contain myriad methodological problems, some of which are similar to those that plagued prayer research. First, what is yoga? That’s not a zen koan, but an honest question. In a real, practical sense, medical researchers have to agree on the elements essential to yoga practice before they can test it as a therapy. Is deep breathing or stretching the source of therapeutic benefit? Or maybe it’s simple exercise, which wouldn’t exactly be news. In addition, yoga, like prayer, can’t be dosed in milligrams. How much yoga do you need to do, and for how long, to achieve a benefit? There’s also significant individual variation at play. Some people breathe more deeply, hold poses for longer, and meditate “better” (I assume) than others. That’s going to muddy the statistics.

Control and blinding are also problematic. When you test a pill for heart disease, you give some people the pill and others a placebo (or an existing medication). Unless they’re extremely motivated and expert in chromatography, the patients can’t tell which group they’re in. It’s not easy to convince someone that they’ve been doing yoga for six weeks when they haven’t, so the placebo effect is always a problem. And in a surprising number of yoga studies, the researchers aren’t blinded, either, raising the risk of a second form of bias.

Take a look at systematic review articles and meta-analyses — studies that aggregate other studies — and you’ll see where the miraculous yoga cure really stands. The majority of such compilations both criticize the methodology of yoga research and find that yoga has little or no effect on serious illness.

A systematic review on the treatment of asthma with yoga was published in 2011. The author found that the methodology of the underlying studies was “mostly poor,” due to problems with blinding and randomization. High dropout rates also biased the results. In the only study included in the review that offered a credible placebo control — a nonyogic stretching regimen — yoga offered no benefit. The author of the review article concluded, “The belief that yoga alleviates asthma is not supported by sound evidence.”

A review of yoga for the treatment of schizophrenia, published in 2013, noted that none of the underlying studies blinded participants, and only three of the five studies blinded the researchers. Dropout rates were either high or unreported. The authors concluded, “No recommendation can be made regarding yoga as a routine intervention for schizophrenia patients.”

A 2013 review paper on yoga for hypertension complained that the “methodological quality of the included trials was evaluated as generally low,” and therefore “a definite conclusion about the efficacy and safety of yoga on hypertension cannot be drawn.”

To be fair, the folks who review existing studies occasionally do conclude that yoga may have modest benefits for sufferers of some afflictions, but they almost always include a laundry list of gripes about methodology and offer the weakest possible recommendation.

Why haven’t you already heard about all of these anti-yoga studies? They have no constituency, and therefore don’t interest the media much.

When a journal article showing that yoga improves quality of life in breast cancer patients came out earlier this month, hundreds of stories trumpeted the results in the mainstream media. Yet it’s difficult to find any mention of the review articles discussed above that question the efficacy of yogic practice as a health care tool. Few people wanted to read a skeptical take on therapeutic prayer in the 1990s, and there aren’t many people today who will click on stories about how yoga won’t solve their health problems. The negative studies never make it beyond medical journals.

Doctors eventually realized — most of them, at least — that prayer didn’t fit well into a clinical trial. Yoga doesn’t, either. That doesn’t mean you shouldn’t do yoga. By all means, do yoga, pray and eat lemons, if those things bring you contentment. Do yoga especially if it’s your preferred form of exercise — exercise is a health intervention supported by thousands of clinical trials. But recognize the “yoga as medicine” craze for what it is: an indicator of the zeitgeist, not a scientific discovery.

Brian Palmer is Slate’s chief explainer. He also writes “How and Why” and “Ecologic” for the Washington Post. Email him at explainerbriangmail.com.