A personalized protocol involving self-managed cognitive functional therapy (CFT) was associated with sustained symptom improvement at 3 years in patients with chronic low back pain in a phase 3 follow-up study.

The three-armed RESTORE trial assessed adult patients in Australia with chronic disabling low back pain, showing effectiveness for CFT with or without motion sensor biofeedback compared with usual care for up to 1 year.

In this new extension analysis of RESTORE, both CFT treatments were still more effective than usual care at 3 years, with greater reductions in both pain intensity and activity limitation.

Interestingly, no differences in outcomes were found between the CFT-only group and the CFT plus biofeedback group.

“It shows that if we empower patients with knowledge and skills to manage back pain, including future flare-ups that are almost inevitable, then we can markedly reduce the massive personal and societal burden of back pain,” lead author Mark Hancock, PhD, Spinal Pain Research Center, Macquarie University, Sydney, Australia, told Medscape Medical News.

“We hoped to see these results, given the intervention design, but we were still somewhat surprised how well the effects lasted over time,” he added.

The findings were published online on August 5 in The Lancet Rheumatology.

Building on ‘Unusual’ Initial Findings

Hancock noted that most treatments for low back pain produce small, short-lasting benefits. 

“The CFT intervention is specifically designed to have long-term effects, as it gives people the skills, knowledge, and confidence to self-manage. The results we previously published showed the effects were sustained at 1 year, which is unusual, and we wanted to see if it remained effective at 3 years,” he said.

In the original RESTORE trial, the investigators recruited adult patients who had low back pain lasting more than 3 months and randomly assigned them to receive one of three treatments: CFT only (n = 164), CFT plus biofeedback (n = 163), or usual care (n = 165). 

The CFT groups received up to seven treatments over 12 weeks plus a booster session at 26 weeks, all of which were delivered by trained physiotherapists.

Investigators noted that CFT targets biopsychosocial barriers to recovery and has three components that are based on an individual patient’s goals. These components include “making sense of pain,” exposure with control, and lifestyle change.

Both CFT groups wore movement sensors during their sessions, but only the CFT plus biofeedback group and their physiotherapists had access to the data.

Usual care involved treatment that was chosen by the patient themselves and/or recommended by their healthcare provider.

The new 3-year follow-up included 312 of the original participants (mean age, 48 years; 60% women) and between 60% and 65% of each of the original treatment groups.

The extension’s primary outcome was self-reported pain-related physical activity limitation, as measured on the Roland-Morris Disability Questionnaire. The secondary outcome was pain intensity, as measured by the numeric pain rating scale.

Long-Term Pain Reduction

Results showed that CFT with biofeedback was more effective in reducing activity limitation at the 3-year mark vs usual care (mean difference, -4.1; P < .0001), as was CFT alone (mean difference, -3.5; P < .0001).

Additionally, CFT with or without biofeedback was more effective in reducing pain intensity vs usual care (mean differences, -1.5 and -1.0, respectively; P < .0001 for both).

For both measures, the differences between the two CFT groups were not significant.

Sensitivity analyses showed similar, although slightly smaller, effects.

In addition, 49% of the CFT plus biofeedback and 43% of the CFT-only groups maintained recovery level scores between the 1-year and 3-year follow-ups compared with 17% of the usual care group.

“These long-term effects are novel and provide the opportunity to markedly reduce the effect of chronic back pain if the intervention can be widely implemented,” the investigators wrote.

They noted that this would include an increase in clinician training and replication studies in other populations.

“This type of biopsychosocial intervention is widely recommended in guidelines, but the evidence [for it] is now becoming much stronger,” Hancock said.

“Many clinicians aim to deliver these interventions, but they often find it hard, especially if they were trained in a more pure biomedical way of thinking. There are now more resources to assist clinicians in upskilling in these approaches,” he added.

‘Cautious Optimism’

In an accompanying editorial, Dimitrios Lytras, PhD, Department of Physiotherapy, International Hellenic University, Thessaloniki, Greece, applauded how the study was conducted.

“Methodologically, RESTORE is exemplary: a pragmatic, multicenter trial embedded in routine care [and] supported by thorough therapist training,” Lytras wrote.

He added that the results “offer cautious optimism” and a shift from passive care models to ones that are more patient-centered.

The addition of biofeedback not leading to added benefit could be explained by the fact that CFT “already incorporates rich feedback mechanisms, making supplementary sensor input redundant,” he noted.

Lytras wrote that the intervention is also low risk and high value.

Durable Functional Gains

Commenting for Medscape Medical News, Sean Mackey, MD, PhD, Redlich professor and chief of the Division of Pain Medicine at Stanford Medicine, Palo Alto, California, said the study showed durable functional gains and modest pain changes.

The dose of seven visits over 12 weeks plus a 26-week booster was “lean,” but the effects held, which was “pleasantly surprising,” said Mackey, who was not involved with the research.

However, he did note a few caveats, including that only 63% of the randomized cohort contributed to the 3-year data. In addition, “those followed were less severe at baseline and did better at 1 year, though loss to follow-up was nondifferential by arm,” he said.

In reflecting on the study design, Mackey noted that an “attention-matched active comparator to blunt performance and expectancy bias” would have been helpful. He pointed out that the researchers themselves acknowledged that usual care was not contact-matched.

He added that it would also have been valuable to include data on 3-year adverse events, healthcare utilization, and costs.

If future research confirms its generalizability to other countries, Mackey said he would recommend clinicians “adopt the CFT package” of graded exposure with a control and lifestyle coaching, schedule 6-8 visits plus a 6-month booster, measure function as the primary outcome, and set expectations for about a one-point reduction in pain. He recommended against adding sensor biofeedback because “it didn’t help.”

Hancock reported being a member of the Australian Physiotherapy Association and holding research grants from the Australian National Health and Medical Research Council and Medical Research Future Fund, the Physiotherapy Research Foundation, the Australian Chiropractors Education and Research Foundation, and the Canadian Institutes of Health Research. The other investigators reported having a wide list of financial relationships, which are fully provided in the original article. Lytras and Mackey reported having no relevant financial relationships.