What science can teach yoga, Part II
Below is Part II of our Q&A with Dr. Marshall Hagins of Long Island University, who is collaborating with Eddie Stern on a study of yoga’s effects on people with hypertension. The yoga being studied is Ashtanga-based.
If you missed Part I, which contains all the background, you can check it out here.
There is more genuine threat of bias in the scientific study of yoga than you might imagine—but it is pro-yoga, not anti-yoga! The vast majority of people investigating yoga from a scientific perspective are yogis who are also scientists. They love something about yoga and want to show that it works using the scientific lens they believe in and have been trained in. For most of them, the long term goal is encouraging individuals to practice yoga and increasing the number of politicians and policy makers who support yoga in our schools, communities and workplaces. Few people are driven to study something just in order to prove it is not true. And there are practical realities at work for scientists that promote the pro-yoga bias. For example, you are unlikely to obtain ongoing grant funding by showing in an early trial that yoga doesn’t work. No significant change with yoga? Too bad. Study something else. So yes, there is a bias—but if anything it is pro-yoga.
As for bias, good scientific study design is explicitly about eliminating bias to the degree possible. Minimizing bias is reinforced by the process of getting a study published or getting money for a grant. The scientist allows his/her work to be judged by other scientists as to the degree of bias in the methods. Less than 10% of National Institute of Health grant proposals get funded. The smartest scientists in the country review these applications eliminating studies that have methods allowing the potential for bias. We tried in our study, described above, to eliminate as much bias as possible. We explicitly measured and can therefore rule out or control bias in many of the variables that may impact on the outcome of interest (in this case blood pressure), like diet and level of physical activity. But there may be variables that impact on blood pressure that differ across the groups that we don’t know about. That is why subjects are randomized. Whatever those variables are, they are most likely to be even in both groups with randomization and consequently any effects they exert they exert equally on both groups—so no net effect for or against yoga. There are other variables, as you mention, like expectation of the participants, that might bias the outcomes. So…you measure that. In our study, after the first yoga and conventional exercise class we give the participants a survey which measures what they feel and believe about the intervention’s ability to help their blood pressure. Additionally, we control for the amount of time each group attends class and receives attention from the teachers. We control for the amount of home exercise practice each group does. We control for the level of energy used (calories burned) within each class. You cannot eliminate bias, but good science is all about minimizing it.
Q. I’m not sure you can do this briefly, but can you also give us a sense of where the science is on yoga’s impact on the human body (bodies is perhaps the better word choice)? Is there a general, growing consensus of the benefits — and have you seen that get stronger and do you expect it to continue?
We are in the early stages of the study of yoga and most current studies are not that well controlled and the findings may change as more rigorous studies proceed. This being said there is increasing evidence that yoga can decrease blood pressure, improve response to stressful events, improve psycho-social issues such as depression and anxiety and of course, such things as flexibility and balance. The number of research projects devoted to yoga is rapidly increasing. I have no doubt that this will continue over the next many years. Regardless of the fear some have in the yoga community about science mixing with yoga—the train has left the station. The best thing we can do is influence the studies that do occur to make sure they are as well designed as possible—being genuinely reflective of yoga practice and possibility.
Q. Finally, trying to put aside my question about bias, what are your expectations for the study? Do you have a statement or idea you hope to provide support for from your findings? Do you have a guess what you might find? And: What if the study suggests yoga doesn’t help with hypertension or somehow makes it worse?
Scientists are supposed to be objective but most will admit that when they start an experiment they have a preferred outcome. There is nothing wrong with having a desire as a scientist as long as your study methods are free of that bias. The preferred outcome for me in this study is that yoga will be shown to reduce blood pressure more than conventional exercise. In this study the conventional exercise is being performed at the same metabolic rate as the yoga practice so any benefits over and above the conventional exercise cannot be said to come simply from more intense physical activity in the yoga group. And yes, the study may not find that yoga is better. This is the risk of trying to systematically observe reality—sometimes your expectations do not match reality. If there was no risk there would be no question. However, you cannot separate your findings from your methods. That is, how you look determines what you find. So, for example, if my findings are negative I would then ask questions like: Did the participants take yoga long enough or frequently enough? Was it the “right kind” of yoga with the right kind of teachers? Was there enough personal attention/accommodation for participants? I really hope I don’t have to ask those questions.
Thanks again to Marshall and Eddie for taking time to answer the questions and connect us.
Posted by Steve