In launching the new Massage Therapy Journal Club, I’ve made the decision to learn some things as I go and to refine and improve things as the project progresses. A couple days after we completed the first episode, it occurred to me that it would be worthwhile to make a blog post that summarizes some of the key discussion points and which includes helpful links for folks who might want to learn more.
We talked about a lot of different things as we considered Field’s 1998 article entitled “Massage Therapy Effects” and I would like to try and summarize and expand on three of those here. They are (1) publication bias; (2) the possibility that reduction of cortisol is the mechanism by which massage therapy works; and (3) models and methods that massage therapy research could borrow and adapt from other research areas.
Publication bias. In her article Field mentioned the likelihood that the body of massage therapy research might make massage therapy appear to be more uniformly effective than it actually is in reality due to publication bias. This can happen in many areas of science and research. Briefly, it is much more likely that a study will be published if it reports upon a positive finding, and especially if that finding is strong, unexpected, or very positive. Studies with null findings or an absence of findings may never be written up or submitted for publication by the persons who conduct them, and when they are submitted for publication they are more likely to be rejected because they may be judged to be uninteresting and/or a poor use of a scientific journal’s limited space for presenting studies. This is an important concept to be aware of when evaluating clinical research. Fortunately, there are quite a few good ways to account for it in quantitative reviews of research. For more detail on what publication bias is, and also how we can avoid being misled by it, have a look at the Wikipedia entry on publication bias.
Cortisol. Cortisol is a stress hormone produced by the adrenal glands. In her 1998 review article Field argues that massage therapy may work by reducing the recipient’s cortisol level; notably, she continues to strongly assert that this is how massage therapy works in many of her subsequent publications. My own research counters this. When my colleagues and I have carefully analyzed the data examining the effect of massage on cortisol, we consistently find that massage therapy does not significantly reduce cortisol. (Here is a detailed scientific paper on this, and here is Part 1 and Part 2 of a YouTube presentation that gives a slightly less technical summary of that research.) Some of the participants noted, also, that even though cortisol is a “stress hormone” it was probably not accurate to conceptualize it as having anything like a 1:1 relationship to stress. They are exactly right to be thinking this way – the relationship between stress and cortisol is complicated and is only one part of a complicated system for responding to a range of stressors.
Notably, this exact subject was raised at the end of Robert Sapolsky’s keynote address at last year’s San Diego Pain Summit. After Sapolsky – one of the world’s experts on stress and health – finished his talk, my friend Ruth Werner asked him about the nature of the stress-cortisol relationship including how it might relate to the discrepant conclusions in massage therapy research. You can view the exchange in the first seven minutes of this video taken during the Q&A portion of his talk. (And if you’ve never seen it, consider watching the National Geographic documentary “Stress: Portrait of a Killer” which features Dr. Sapolsky’s work; here is a link to it on YouTube.
Models for massage therapy research. This is a complicated topic so I am only going to touch on it here. Briefly, our discussion touched on the fact that people interested in massage therapy research sometimes believe that this area of research presents unique problems, such as how can massage therapy ever be understood by science if science demands that it be studied under controlled and standardized conditions? And, if the success of massage depends on there being a good therapeutic relationship between the provider and the recipient, doesn’t that mean we can never really measure what it most important in treatment?
These are very important things to consider, and they can make it difficult to do good research. However, they are not insurmountable. More importantly, they are not problems that are unique to massage therapy research. There are other areas of clinical research and of intervention research that also confront these and related problems and which have developed clever and effective methods for addressing them. I have argued that the well-developed are of psychotherapy research can be an excellent model for massage therapy research. The fields of medicine, education, and occupational therapy are some other areas that have dealt with these and similar challenges and are ones that the field of massage therapy research can look to, and borrow from, to hasten its own progress. Massage therapy research does have some challenging elements due to the fact that it is based on interpersonal contact and may work best when it is not performed according to a standardized protocol; however, massage therapy researchers do not need to ‘reinvent the wheel’ to address these issues. They can and should look to these other areas of clinical and intervention research for partial or even complete solutions to these challenges.
P.S. I am conducting a promotion to help ensure the success of the Massage Therapy Journal Club. Have a look at this short video to learn how you can be entered to win a free copy of my book, Massage Therapy: Integrating Research and Practice, and please share the link if you can.