In August of 2022, several prominent Australian physiotherapists published an article in the Journal of the American Medical Association (JAMA) titled “Effect of Graded Sensorimotor Retraining on Pain Intensity in Patients with Chronic Low Back Pain: a Randomized Clinical Trial”. The authors stated that “graded sensorimotor retraining is a novel intervention designed to alter how people think about their body in pain, how they process sensory information from their back, and how they move their back during activities.” This clinical trial1 evaluated the effects of “graded sensorimotor retraining” versus an attention control with sham procedures (i.e., sham laser therapy, sham short-wave diathermy, sham transcranial direct current stimulation) in patients with chronic low back pain.
We believe this 5-year trial, involving multiple prominent researchers from the Australian physiotherapy community (e.g., Drs. Lorimer Moseley, Chris Maher, Benedict Wand), deserves a critical review to put the findings within the context of real clinical practice. In short, we believe the conclusions made by the authors of this trial1 are not supported by the actual results.
1. MIMIMUM CLINICALLY IMPORTANT DIFFERENCE (MCID)
Regarding the 1-point mean between-group difference in pain reported in this trial, the authors concluded “the improvement in pain intensity was modest, but it was consistent with a clinically meaningful between-group difference.”1 We disagree with this claim. Notably, the authors of the trial1 choose to use an MCID of 1.0 point2 for the 11-point pain rating scale (NPRS, 0-10) in patients with chronic low back pain (LBP). However, the MCID for the NPRS has been shown to be 1.74 in patients with chronic pain conditions;3 thus, a change of 2 points or a 30% decrease in pain from baseline should be considered as the MCID in patients with chronic musculoskeletal pain.3 4
Moreover, in patients with LBP a change of 2 points is typically considered to be the smallest difference perceived by patients as beneficial (i.e., the MCID).5 Similarly, the MCID for the NPRS has been shown to be 2.5 points in patients with cervicogenic headache6 and 2.2 points in patients with shoulder pain receiving rehabilitation.7 Therefore, it seems the authors of the “sensorimotor retraining” trial1 selected the lowest possible MCID of 1.0 point in an attempt to give the appearance of clinical relevance to the very small between-group change in low back pain intensity that they found after a 5-year trial.
Interestingly, Bagg et al1 did not mention or discuss the minimum detectable change (MDC) for the NPRS in patients with LBP—i.e., the amount of change that must occur before it can be considered to have exceeded measurement error. Notably, the MDC for the NPRS (0-10 scale) in patients with LBP is considered to be 2 points.5 Therefore, when measuring change in pain intensity in patients with LBP, and according to Childs et al,5 “clinicians can be confident that a 2-point change on the NPRS represents clinically meaningful change that exceeds the bounds of measurement error.”
Thus, although statistical significance was shown, a between-group 1-point change in pain intensity (following 12 sessions of “graded sensorimotor retraining” over 12-18 weeks) is likely not worthwhile or perceived as beneficial by the majority of patients with chronic low back pain. Certainly the 26-week and 52-week between-group change scores reported for pain intensity (0.7 and 0.5, respectively) in this trial1 would not be considered meaningful by patients with chronic low back pain.
2. EFFECT SIZE
P values tell us “if” a treatment works; whereas effect sizes tell us “how well” an intervention works—i.e., the strength or magnitude of the treatment. Notably, the authors1 failed to report standardized measurements of effect size—i.e., Cohen’s d effect sizes or the Standardized Mean Difference (SMD).8 9 Instead, they chose to report unstandardized or raw change scores in the NPRS that, unlike the SMD statistic, do not take in to account the pooled standard deviation of both the experimental and control groups.8
On closer inspection, the between-group difference in the primary outcome (NPRS, 0-10)) at the primary endpoint (18 weeks) appears to be -0.7, not -1.0. Notably, the intervention group went from a mean baseline pain score of 5.6 to 3.1 at 18 weeks (i.e., a within-group change score of 2.5), whereas the control group went from a mean baseline pain score of 5.8 to 4.0 at 18 weeks (i.e., a within-group change score of 1.8). Therefore, using basic arithmetic, the unadjusted between-group change score for pain (NPRS, 0-10) at 18 weeks appears to be -0.7.
Notably, instead of using the more conservative “intention-to-treat” analysis that is typically employed in randomized controlled trials, the authors used a “compliers-only mean causal effect analysis” in an apparent attempt to adjust (i.e., inflate) the between-group change score of the primary outcome to -1.0, to enable them to then claim it met the MCID for this measure. In this regard, we believe the dataset for this study1 should be published online for other research teams to independently verify.
Clinical trials measure relative efficacy of a treatment compared to a control, placebo, or usual care.10 We believe that the authors of this trial1 should have used an active comparison group to determine whether or not the novel “sensorimotor retraining” works any better, or provides any different outcome than what physical therapists are already using for the treatment of low back pain (i.e., conventional physical therapy), rather than just comparing the experimental group to a sham or placebo treatment and finding an extremely small between-group difference (1 point!) in low back pain after 12-18 weeks of treatment).10-12
3. NUMBER-NEEDED-TO-TREAT (NNT)
Another reference standard and very commonly reported measure of treatment effect size in clinical trials is the NNT.8 13 The authors failed to report the NNT for the primary or any secondary outcomes at any time points. Therefore, given the data presented in the study, we calculated the NNT for the primary outcome measure for the reader.
At the 26-week follow-up, 24 of the 131 (18.3%) in the intervention group and 13 of 133 (9.8%) in the control group ‘met the criterion for recovery’, giving a NNT of 11.8. That is, on average, 12 patients would have to receive sensorimotor retraining instead of the control treatment for 1 additional patient to “recover” from their chronic LBP. At week 18, for the proportion of subjects achieving >30% decrease in pain intensity from baseline, the NNT is 5.4. Furthermore, at the 18-week follow-up, for the criterion of >50% decrease in pain intensity, the NNT is 8.2 –i.e., the clinician would need to provide sensorimotor retraining to 8 patients with chronic LBP (compared to receiving an attention control with sham procedures) to achieve 1 successful outcome.
In short, values of 5-12 for the NNT for an intervention strategy are just too high for real clinical practice. Most physical therapists are just not willing to see 5-12 patients with LBP to achieve 1 successful outcome. Furthermore, knowing those odds, most patients would not be willing to agree to 12 sessions of “sensorimotor retraining” over 12-18 weeks for their LBP.9 13
4. DISABILITY OUTCOMES
The between-group change scores for disability (RMDQ, 0-24) at 18 weeks, 26 weeks and 52 weeks were reported to be -2.6, -2.1, and -1.8 points, respectively;1 nevertheless, the MDC for the RMDQ and the MCID are considered to be 4 to 5 points14 and 3.5 to 5 points,15 16 respectively. Therefore, in this trial,1 and following 12 sessions of “graded sensorimotor retraining” over 12-18 weeks, the between-group difference in disability levels, did not exceed the bounds of measurement error (i.e., the MDC) or the smallest difference perceived by patients as beneficial (i.e., the MCID).
Twelve sessions of “graded sensorimotor retraining” over 12-18 weeks has a very small effect size on pain and disability levels that does not exceed the smallest difference perceived by patients as beneficial (i.e., the MCID) or the bounds of measurement error (i.e., the MDC) when compared with a sham treatment. The number-needed-to-treat (NNT) ranges between 5 and 12, which means, 5-12 patients with chronic low back pain would need to be treated with “sensorimotor retraining” to achieve 1 successful outcome.
Like prior calls from the Australian physiotherapy community to isolate or improve the feed-forward activation timing of the “transverse abdominus” and “deep neck flexors” for the treatment of low back pain and neck pain, respectively, this novel “sensorimotor retraining” approach, after a 5-year trial, doesn’t appear to give meaningful reductions in pain or disability in patients with chronic low back pain.
Notably, the commonality between the old hype over the transverse abdominus/deep neck flexors and the novel “sensorimotor retraining” is that these theoretical constructs and subsequent clinical trials were designed, in the main, by academics in Australia that haven’t been in real clinical practice for decades.
James Dunning, PhD, DPT, MSc, FAAOMPT, Dip. Osteopractic
Director, AAMT Fellowship in Orthopaedic Manual Physical Therapy
Montgomery Osteopractic Physical Therapy & Acupuncture, Montgomery, AL
Megan Wong, PT, DPT, Cert. DN
Fyzical Therapy & Balance Center, Albuquerque, NM
Fellow-in-training, AAMT Fellowship in Orthopaedic Manual Physical Therapy
Paul Bliton, DPT, OCS, SCS, FAAOMPT, Dip. Osteopractic
Associate Director, AAMT Fellowship in Orthopaedic Manual Physical Therapy
William S. Middleton VA Hospital, Madison, WI
Patrick Gorby, DPT, FAAOMPT, Dip. Osteopractic
Assistant Director, AAMT Fellowship in Orthopaedic Manual Physical Therapy
Gorby Osteopractic Physical Therapy, Phoenix, AZ
Casey Charlebois, PhD(c), DPT, MSc, FAAOMPT, Dip. Osteopractic
Director of Clinical Research, AAMT Fellowship in Orthopaedic Manual Physical Therapy
1. Bagg MK, Wand BM, Cashin AG, et al. Effect of Graded Sensorimotor Retraining on Pain Intensity in Patients With Chronic Low Back Pain: A Randomized Clinical Trial. JAMA 2022;328(5):430-39. doi: 10.1001/jama.2022.9930
2. Busse JW, Bartlett SJ, Dougados M, et al. Optimal Strategies for Reporting Pain in Clinical Trials and Systematic Reviews: Recommendations from an OMERACT 12 Workshop. J Rheumatol 2015;42(10):1962-70. doi: 10.3899/jrheum.141440 [published Online First: 20150515]
3. Farrar JT, Young JP, Jr., LaMoreaux L, et al. Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain 2001;94(2):149-58. [published Online First: 2001/11/03]
4. Salaffi F, Stancati A, Silvestri CA, et al. Minimal clinically important changes in chronic musculoskeletal pain intensity measured on a numerical rating scale. Eur J Pain 2004;8(4):283-91. doi: 10.1016/j.ejpain.2003.09.004
5. Childs JD, Piva SR, Fritz JM. Responsiveness of the numeric pain rating scale in patients with low back pain. Spine (Phila Pa 1976) 2005;30(11):1331-4. doi: 10.1097/01.brs.0000164099.92112.29
6. Young IA, Dunning J, Butts R, et al. Psychometric properties of the Numeric Pain Rating Scale and Neck Disability Index in patients with cervicogenic headache. Cephalalgia 2019;39(1):44-51. doi: 10.1177/0333102418772584 [published Online First: 20180419]
7. Michener LA, Snyder AR, Leggin BG. Responsiveness of the numeric pain rating scale in patients with shoulder pain and the effect of surgical status. J Sport Rehabil 2011;20(1):115-28. doi: 10.1123/jsr.20.1.115 [published Online First: 2011/03/18]
8. Durlak JA. How to select, calculate, and interpret effect sizes. J Pediatr Psychol 2009;34(9):917-28. doi: 10.1093/jpepsy/jsp004 [published Online First: 2009/02/19]
9. Faraone SV. Interpreting estimates of treatment effects: implications for managed care. P T 2008;33(12):700-11. [published Online First: 2009/09/15]
10. Kamper SJ. Control Groups: Linking Evidence to Practice. J Orthop Sports Phys Ther 2018;48(11):905-06. doi: 10.2519/jospt.2018.0706 [published Online First: 2018/11/02]
11. Fritz JM, Cleland J. Effectiveness versus efficacy: more than a debate over language. J Orthop Sports Phys Ther2003;33(4):163-5. doi: 10.2519/jospt.2003.33.4.163 [published Online First: 2003/05/02]
12. MacPherson H, Vertosick E, Lewith G, et al. Influence of control group on effect size in trials of acupuncture for chronic pain: a secondary analysis of an individual patient data meta-analysis. PLoS One 2014;9(4):e93739. doi: 10.1371/journal.pone.0093739 [published Online First: 2014/04/08]
13. Kamper SJ. Interpreting Outcomes 2-Statistical Significance and Clinical Meaningfulness: Linking Evidence to Practice. J Orthop Sports Phys Ther 2019;49(7):559-60. doi: 10.2519/jospt.2019.0704 [published Online First: 2019/07/02]
14. Stratford PW, Binkley J, Solomon P, et al. Defining the minimum level of detectable change for the Roland-Morris questionnaire. Phys Ther 1996;76(4):359-65; discussion 66-8. doi: 10.1093/ptj/76.4.359
15. Jordan K, Dunn KM, Lewis M, et al. A minimal clinically important difference was derived for the Roland-Morris Disability Questionnaire for low back pain. J Clin Epidemiol 2006;59(1):45-52. doi: 10.1016/j.jclinepi.2005.03.018 [published Online First: 20051104]
16. Ostelo RW, de Vet HC. Clinically important outcomes in low back pain. Best Pract Res Clin Rheumatol2005;19(4):593-607. doi: 10.1016/j.berh.2005.03.003