Massage Today
Massage Today dotted line
dotted line

dotted line
Share |
  Forward PDF Version  
Massage Today
March, 2012, Vol. 12, Issue 03

Trying to Get Something From Nothing

By Keith Eric Grant, PhD, NCTMB

"Sometimes nothing can be a real cool hand" – Paul Newman in the movie, "Cool Hand Luke."

You are going to be hearing more and more about evidence-based massage therapy (EBMT). Partly, this reflects a current trend in health care to re-evaluate treatment and to determine what has a sound basis for use and what doesn't. One example of this is the Institute of Medicine report, "Evidence-Based Medicine and the Changing Nature of Healthcare."

A second factor is the existence of the Massage Therapy Foundation (MTF), which has goals of making those in the profession aware of research, promoting research literacy and integrating research with practice. I would count the forthcoming book on such integration by Dryden and Moyer as among the efforts facilitated by the MTF.

A third factor is massage therapy now being regulated by the majority of states, combined with the legal presumption that such regulation is done for the protection of the public. If we consider that training is necessary for safe practice, the presumption of public protection can only be fulfilled when state-mandated training is based upon a solid foundation of objectively validated knowledge.

A final factor is the modern technology embodied in an interactive web. Communication without regard to physical proximity is facilitating extensive discussions among those both with a stake in massage therapy and with backgrounds in research and statistics. Alice Sanvito discusses evidence-based massage on her web site and provides a number of links there to additional resources. I like the definition she gives for EBMT.

Evidenced based massage therapy is massage therapy founded on ideas and principles supported by evidence. Many of the claims made and practices used by massage therapists are founded on tradition rather than evidence. Since there is not yet a large body of knowledge documenting the physiology of and effects of massage therapy, if we were only able to make statements strictly on the basis of scientific studies, we would be severely limited indeed. Some people prefer the term "evidence informed practice" as more accurate. An evidence informed practice takes into consideration scientific evidence, clinical experience and careful observation.

The concept of being evidence-based, however, necessitates having methods to collect such evidence. In this, we also need to be careful to distinguish between whether or not an intervention can be shown to work (beyond random chance) and the model that we believe is the mechanism underlying the intervention. It is fully possible, as with massage relieving muscle soreness and the lactic acid myth, for an intervention to be effective while the supposed mechanism is incorrect. The randomized controlled trial (RCT) is taken as the "gold standard" of clinical proof, but how does that work? We need three things: a study population, a methodology for the study and the ability to analyze the results for effectiveness.

For example, our study population might be those diagnosed with high blood pressure, over the age of 40, not having other medical complications and not knowing Morse code. The goal of our study might be to determine whether or not listening to relaxing messages keyed in Morse code by a live practitioner were effective in reducing blood pressure. A complicating factor for the study is that people respond to the presence of other people. As put by Ravensara Travillian recently, "As psychosocial beings, we respond psychologically and socially in ways that can be described as healing body and mind due to presence and caring attention from others." Thus, our study will need a means of differentiating the effects of the Morse code from those effects simply from the practitioner's presence and the setup of the trial itself.

After gaining a sufficient number of suitable participants, we would fulfill the "randomized" concept of the trial by randomly dividing them into three groups: control, sham Morse-code and Morse code. The intervention might be three-times per week for 12 weeks. Controls would come in, have their blood pressure (BP) measured, wait 30 minutes, and have their BP measured again. Those in the Morse-code group would, in the 30 minutes wait, listen to Morse-code keyed by one of several live practitioners. Those in the "sham-group" would listen to 30 minutes of keying, similar to Morse-code, but keyed by "practitioners" unfamiliar with Morse-code. With this protocol, we can look for short-term effects between sessions, beginning and session end, as well as for longer term effects over the length of the study. By taking follow-up measurements after the end of the 12 weeks, we can also look for persistence of any changes.

Now we get to the point of getting something from nothing. We assume the null hypothesis, that any differences between the groups is simply from random chance, and calculate the probability (p) that this assumption is true. We conclude that there is a statistically significant difference between groups only when the probability of our observations being due entirely to chance is less than 5% (p<0.05). We have three separate probabilities to check: whether the sham group is statistically different from the control group (psychosocial effect), whether the Morse code group is different from the control group (psychosocial plus Morse code), and whether the difference between the sham and code groups is significant (Morse code effect). In a recent paper, Bakker and Wicherts underscore the importance of doing the third test explicitly, even when the differences between the sham and control groups is not significant.

If there is no difference between the three groups, the study would conclude that, in the clinical trial as designed, neither psychosocial factors nor messages in Morse-code were effective. If the sham group differed statistically from the control, we would conclude that there was a psychosocial effect. Because the psychosocial effect would also be present in the code group, only if the code group was statistically different from the sham group could we conclude that there was an effect from Morse-code itself. Note that this code effect could be of either sign, adding to a psychosocial effect or negating it.

There you have the outline of a randomized control/clinical trial. Assuming initially that we get no difference, we end up with information. Sometimes "nothing can be a real cool hand."


  1. Bakker M and Wicherts JM. "The (mis)reporting of statistical results in psychology journals." Behav Res Meth, 2011;43( 3): 666-678,
  2. Dryden, Trish, and Christopher Moyer. "Massage Therapy – Integrating Research and Practice." Human Kinetics. Available May 2012. ISBN-13: 9780736085656,
  3. Institute of Medicine. "Evidence-Based Medicine and the Changing Nature of Healthcare." October 2008,
  4. Sanvito, Alice. "What Is Evidence Based Massage Therapy?" Accessed 15 January 2012,
  5. Travillian, Ravensara. "Let's read together: A Randomised Controlled Single-Blind Trial of the Efficacy of Reiki at Benefitting Mood and Well-Being—Introduction." Accessed 17 January 2012,

Click here for previous articles by Keith Eric Grant, PhD, NCTMB.


Join the conversation
Comments are encouraged, but you must follow our User Agreement
Keep it civil and stay on topic. No profanity, vulgar, racist or hateful comments or personal attacks. Anyone who chooses to exercise poor judgement will be blocked. By posting your comment, you agree to allow MPA Media the right to republish your name and comment in additional MPA Media publications without any notification or payment.
comments powered by Disqus
dotted line