Effect Modifiers of Graded Sensorimotor Retraining for Chronic Low Back Pain: A Secondary Analysis of the RESOLVE Randomized Trial.

Importance: Outcomes for people with low back pain (LBP) may be improved through targeting treatments to subgroups with particular experiences, characteristics, or symptoms and clinical factors. Objective: To investigate potential treatment effect modifiers of graded sensorimotor retraining affecting pain intensity and disability level in the short and long terms for people with chronic LBP. Design, Setting, and Participants: This post hoc exploratory secondary analysis of the RESOLVE randomized clinical trial in Sydney, Australia, was conducted from November 5, 2024, to May 27, 2025. Trial participants were recruited from primary care settings and randomly allocated (1:1) to receive either graded sensorimotor retraining (treatment group) or attention control and sham procedures (sham control group). Eligible participants were adults aged 18 to 70 years who reported chronic nonspecific LBP (lasting ≥12 weeks), with or without leg pain, that was rated at least a 3 out of 10 in pain intensity. Statistical analyses were conducted from November 11 to December 6, 2024. Interventions: The treatment group completed 12 weekly clinical sessions of the graded sensorimotor retraining package, which consisted of pain science education, premovement training, and graded movement and loading. The sham control group completed 12 weekly sessions, without advice or education, of sham electrotherapy to the back and sham noninvasive brain stimulation. Main Outcomes and Outcome Measures: Primary outcomes were pain intensity (measured with the 11-point Numerical Rating Scale ranging from 0 [no pain] to 10 [worst imaginable pain]) and disability level (measured with the 24-item Roland-Morris Disability Questionnaire with scores ranging from 0-24 [higher scores indicating greater levels of disability]) assessed at 18 weeks and 52 weeks after randomization. A formal moderation analysis was performed using a test for statistical interaction. Eight baseline variables-psychoactive medication use, pain intensity, disability level, beliefs about back pain consequences, kinesiophobia, pain catastrophizing, pain self-efficacy, and back perception-were investigated for their potential treatment effect modification. Results: The study included 276 participants (mean [SD] age, 46 [14.3] years; 138 females [50.0%]), 138 of whom were randomized to the treatment group and 138 of whom were randomized to the sham control group. Pain self-efficacy, pain catastrophizing, pain intensity, and psychoactive medication use showed no evidence of modifying the effect of the intervention. Impaired back perception was identified as a potential treatment effect modifier of pain intensity (β-coefficient = 0.18 [95% CI, 0.05-0.32]; P = .007) at the 52-week follow-up time point. Hypothesis-generating evidence (P < .20) indicated potential effect modification by kinesiophobia (on pain intensity at 18-week follow-up: β-coefficient = 0.06 [95% CI, -0.02 to 0.14], P = .15; 52-week follow-up: β-coefficient = 0.07 [95% CI, -0.02 to 0.16], P = .12), baseline disability level (on disability level at 18-week follow-up: β-coefficient = -0.15 [95% CI, -0.38 to 0.07], P = .17), beliefs about back pain consequences (on disability level at 52-week follow-up: β-coefficient = 0.16 [95% CI, -0.05 to 0.37], P = .14) and back perception (on pain intensity at 18-week follow-up: β-coefficient = 0.10 [95% CI, -0.02 to 0.22], P = .09; on disability level at 18-week follow-up: β-coefficient = 0.22 [95% CI, -0.04 to 0.48], P = .10; 52-week follow-up: β-coefficient = 0.27 [95% CI, 0.00-0.55], P = .05). Conclusions and Relevance: This secondary analysis found that the benefits of graded sensorimotor retraining are likely to be similar for all people with chronic nonspecific LBP presenting for care. Future clinical trials are needed to further explore and assess the role of the potential treatment effect modifiers identified in this analysis. Trial Registration: ANZCTR Identifier: ACTRN12615000610538.