Behaviour change techniques associated with adherence to prescribed exercise in patients with persistent musculoskeletal pain: systematic review

Key points

1. 01Five BCTs showed moderate evidence for improving adherence: social support (unspecified), goal setting (behaviour), instruction on how to perform the behaviour, demonstration of the behaviour, and behaviour practice/rehearsal.
2. 02Eight RCTs totalling 1,018 participants were included; five of eight trials reported significantly greater exercise adherence in the treatment group.
3. 03Interventions using seven or fewer BCTs distinct from the control group were more effective than those using more; the two low-risk-of-bias trials with the most BCTs (9 per treatment group) found no significant between-group difference.
4. 04Only three of eight trials reported using theory to guide intervention design, limiting conclusions about whether theoretical underpinning improves outcomes.
5. 05No trial reported an intervention fidelity assessment, so it is unknown whether the BCTs were delivered consistently as planned.

How it was conducted

Design

Systematic review of randomized controlled trials; narrative synthesis (meta-analysis not possible due to heterogeneity)

Databases searched

CINAHL, PsycINFO, Embase, MEDLINE, Cochrane CENTRAL, openDOAR, OpenGrey, Web of Science, and EThOS; inception to August 2017

Participants

Adults aged 18-65 with a clinician-diagnosed musculoskeletal pain condition persisting 3+ months (low back pain, fibromyalgia, whiplash-associated disorder); total 1,018 participants across 8 trials

Included trials

8 RCTs; 5 rated low risk of bias, 3 rated high risk of bias

BCT coding

Behaviour Change Taxonomy v1 (Michie et al. 2013) applied independently by two trained reviewers to both treatment and control groups

Primary outcome

Exercise adherence as defined by original authors (diaries, exercise logs, questionnaires, or sessions attended)

What they found

• 5 of 8 trials reported significant between-group differences in exercise adherence favouring the treatment group (Coppack 2012, Harkapaa 1990, Linton 1996, Reilly 1989, Vong 2011).
• Vong (2011) - low back pain: SMD 1.20 (95% CI 0.71, 1.63), p significant; treatment group completed 12.5 home exercise sessions/week versus 5.8 (SD 4.1) in control.
• Coppack (2012) - low back pain: SMD 1.34 (95% CI 0.56, 2.11), p significant; treatment group SIRAS score 13.5 versus 11.7 (SD 1.3) in control.
• Reilly (1989) - low back pain: SMD 4.66 (95% CI 3.42 to upper bound); treatment group attended 90.8 sessions versus 31.9 (SD 17.2) in control, p significant.
• Linton (1996): 52% of treatment group completed exercise versus 27% in control, p significant.
• Harkapaa (1990/1991): 51% of treatment group performed faultless exercise versus 37% in control at 1.5-year follow-up, p significant.
• Friedrich (1998): SMD -0.42 (95% CI -0.01 to upper); no significant between-group difference.
• Huyser (1997) - fibromyalgia: SMD 0.066 (95% CI -0.11 to upper); no significant between-group difference.
• Peterson (2015) - whiplash: 53% of treatment versus 60% of control completed 50% of prescribed home exercise sessions; no significant between-group difference.
• Total of 26 distinct BCTs (out of 93 possible) were identified across all 8 trials; treatment groups used 3-11 BCTs (median 7), control groups used 1-6 BCTs (median 3.5).
• BCT coding inter-rater agreement: 75.3% for treatment groups and 71.6% for control groups.

Limitations

• Only 8 trials met inclusion criteria and 3 were rated high risk of bias, making findings tentative and requiring cautious interpretation.
• Heterogeneous adherence outcome measures across trials prevented meta-analysis and limited direct comparisons.
• No trial reported an intervention fidelity assessment, so consistent delivery of BCTs cannot be confirmed.
• Most adherence data relied on self-report (diaries, questionnaires), introducing potential recall and social desirability bias.

Why it matters

For patients

If you have long-term musculoskeletal pain and struggle to keep up with prescribed exercises, adding structured support from a health professional - including goal setting, clear instruction, demonstration, and supervised practice - may meaningfully improve how consistently you exercise.

For clinicians

Physiotherapists and other health professionals should consider incorporating a small, targeted set of BCTs (particularly social support, goal setting, instruction, demonstration, and practice/rehearsal) into exercise prescriptions for PMSK patients, while avoiding overlap with techniques already present in usual care.

For readers

This review maps the active ingredients of effective exercise-adherence interventions, suggesting that a focused and distinct BCT package - rather than a larger undifferentiated set - may drive better adherence outcomes in people with persistent musculoskeletal pain.