Apr 4 • Peter Malliaras

Adherence to exercise in musculoskeletal trials: current problems and how we can improve.

This short blog post will provide a clinical perspective on a recent scoping review from PhD student Matthew Kenny. I will attempt to summarise the two whys and two whats….

Why we did it:

Exercise is a recommended treatment for many pain and musculoskeletal conditions, including tendinopathy. It is really clear in other areas such as exercise for general health (e.g., heart health) that the more exercise you do, the better. Things don’t really work in this linear way for pain. Pain is complex. Maybe doing more exercise will make our pain and musculoskeletal patients stronger and more functional, but why should it improve their pain?

To get more certainty about whether we need to care about how much exercise we do to maximise certain outcomes, we need a really good way to measure exercise adherence. Exercise adherence is defined by the world health organisation as “the extent to which a person’s behaviour – taking medication, following a diet, and/or executing lifestyle changes, corresponds with agreed recommendations from a health care provider” (WHO, 2003). We anticipated that the problem currently is that many people doing reach either don’t bother measuring adherence, or they do but use crappy measures (like self-report, more on that later). 

We therefore did this scoping review to see how often researchers report adherence in trials of exercise for tendinopathy, and to report levels of adherence reported and also the type of adherence measures they used. 

What we did:

Instead of a systematic review we used a scoping review design which is used when you are scoping what is known in an area rather than answering specific questions (eg is treatment A better than treatment B). We included trials investigating exercise interventions in four common musculoskeletal conditions including low back pain, Achilles tendinopathy, shoulder pain and knee OA. Five databases were searched to identify trials. 

What we found:

A massive 321 trials were included, representing 25,893 participants (mean age 49.6 yrs and 61.8% were women). Here is a summary of the key findings:

  • 46.7% (150/321) said they assessed adherence (20% of the 150 did not report it though)
  • Registered trials were more likely to report adherence
  • Newer trials were not more likely to report adherence (see figure below)
  • Aside from direct measurement (supervision) the only other adherence measure was self-report. 
  • The median frequency of self-reported adherence was similar (80-86%) for all conditions, and slightly higher for supervised exercise.

Why is it important?

Key points to take away:

  • When it is reported, there is a reliance of self-report adherence which is prone to bias (such as recall bias, where people forget what they have actually done).
  • Adherence rates are really similar, on average, for the musculoskeletal conditions we looked at, but how much we can trust these estimates is not clear
  • Most importantly, we need better technology driven measures of home exercise adherence so we can truly understand whether how much people do influences outcomes for musculoskeletal conditions.