Were there just two groups in the RESTORE trial: CFT and usual care?

In the original RESTORE study, participants were randomly allocated to one of three groups: usual care, CFT only or CFT plus biofeedback.

However, for the purposes of this analysis, the CFT and CFT plus biofeedback groups were combined because we did not find any clinically meaningful differences between the two CFT groups in treatment effectiveness.

What were the five baseline moderators that you analysed?

In order to limit the chance of spurious findings, we prospectively selected five baseline moderators to investigate: activity limitation,
cognitive flexibility, pain intensity, self-efficacy and catastrophising.

Did baseline activity limitation identify patients who responded better to CFT?

We found that baseline level of activity limitation was associated with different effects of CFT treatment in both the short and long term.

Patients who had higher baseline activity limitation had greater effects from CFT at both 13 and 52 weeks compared with those with lower levels of baseline activity limitation.

For each additional Roland-Morris Disability Questionnaire point at baseline, the treatment effect increased by 0.18 points (95% CI 0.01 to 0.34) at 13 weeks and by 0.23 points (95% CI 0.04 to 0.42) at 52 weeks.

For example, a person with a baseline Roland-Morris Disability Questionnaire score of 18 would on average be 6.1 points (95% CI 4.8 to 7.4) better at 13 weeks if they received CFT compared to usual care.

However, a person with a baseline score of seven would on average only be 3.6 points (95% CI 2.6 to 4.6) better at 13 weeks if they received CFT compared to usual care.

Dr Mark Hancock

What about baseline cognitive flexibility?

Patients who had higher baseline cognitive flexibility may have greater effects of CFT at 13 weeks compared to those with lower levels of baseline cognitive flexibility.

However, the CIs include no moderating effect so caution is needed.

For each additional point on the Cognitive Flexibility Scale at baseline, the treatment effect increased by 0.10 points (95% CI –0.04 to 0.25) at 13 weeks.

For example, a person with a baseline score of 69 would on average be 6.1 points (95% CI 4.5 to 7.7) better at 13 weeks if they received CFT compared to usual care.

However, a person with a baseline score of 49 would on average only be 4.1 points (95% CI 2.2 to 5.8) better at 13 weeks if they received CFT compared to usual care. Cognitive flexibility did not appear to have a moderating effect at 52 weeks.

What about the other potential moderators you assessed?

Pain intensity, pain self-efficacy and catastrophising did not appear to have a moderating influence in the short or long term.

Where do you think research in this area needs to go next?

CFT is most effective in patients who have the most disability, experience the greatest health burden and are often the most challenging for clinicians.

This strongly suggests that CFT should be considered as a treatment for these patients, especially before progressing to more expensive and risky interventions.

The results of moderation analyses are always relative to the control group.

Therefore, our moderator results may be different if CFT was compared to a different control.

Future studies investigating moderators of CFT when compared to other interventions will be important in guiding clinicians.

We are just starting a new large trial comparing CFT to standardised physiotherapy care and will investigate similar moderators in that study.

>>Dr Mark Hancock APAM is a professor of physiotherapy at Macquarie University. Mark has over 20 years of clinical experience as a musculoskeletal physiotherapist. His research focuses on the diagnosis, prevention and treatment of low back pain.