Commentary on "Cognitive Functional Therapy in Patients with Non-Specific Chronic Low Back Pain".

Abstract

In this issue, you will find a paper by Vibe Fersum et al., entitled “Cognitive Functional Therapy in Patients with Non Specific Chronic Low Back Pain: A Randomized Controlled Trial 3-year Follow Up” (Vibe Fersum et al., 2019). This interesting and well done paper reports on 3 year outcomes from a previously reported randomized trial. The primary findings are lower self-reported disability for those randomized to receive cognitive functional therapy instead of manual therapy and exercise. This is a strong contribution to the literature as there is growing interest in psychologically informed approaches to managing musculoskeletal pain. In particular, the authors are to be commended for reporting 3-year data, as follow up beyond 12 or 24 months is rare in conservative treatment trials for low back pain. This commentary will introduce some issues for readers to consider when interpreting the results of this study. The primary issue when reading this paper is to consider the potential impact that loss to follow up has on the long-term effects of cognitive functional therapy. This issue is mentioned by the authors as a primary limitation in the paper, and I will expand on it briefly. Only 52% (63/121) of those from the original trial provided follow up data at 3 years. The authors have taken appropriate care statistically in accounting for this loss to follow up via sensitivity analyses using mixed models and multiple imputation. However, the fact remains that this was a relatively small trial to begin with and the 3-year attrition rate has substantially eroded the original sample size. The absolute number of subjects providing follow up data is small enough (n< 50 per treatment group) that the precision of reported treatment effects is limited. Further drawing attention to this issue is the evidence of differential loss to follow up that suggests those providing 3-year data had lower psychosocial distress. This is a clinical relevant baseline difference because psychological distress is often an important mediator of treatment outcome. The authors do account for this by adjusting analyses for baseline scores. While that is an acceptable statistical solution, it doesn’t entirely account for the impact selective follow-up could have on the 3-year results. Another issue to consider when reading this paper is the timing of the treatment effects. At 3-years group differences still exist for disability and most of this gain appears to have much earlier (i.e. at 3 month follow up). This is evidence supporting that after an initial period of receiving a treatment a favourable trajectory is established. Statistically the treatment effect was sustained at 3 years but the size of the effect is cut in half from when the benefit was largest (i.e. 3 month follow up). Even with a smaller absolute effect at 3 years, the effect sizes reported remain large when compared to others in this field. This deserves mention both in the “good news” category (i.e. even at 3 years these effect sizes are larger than what you see in the typical trial) but also in the “need more studies” category (i.e. can these effect sizes be replicated when the treatment is delivered by others in different settings). Several of the next steps one could take from this clinical trial are well aligned with recent research priorities identified for psychological informed practice. For example, cognitive functional therapy was associated with improved self-reported disability but identifying specific factors involved in the process of achieving better outcomes were not part of these analyses. Future studies could consider measures of pain intensity, self-efficacy, fear of pain and catastrophizing and identify which are key mediators of favourable outcomes when receiving cognitive functional therapy. Another seemingly informative study would be to investigate the impact of “booster sessions” for cognitive functional therapy that could be delivered as part of group sessions or tele-health. Use of such post-treatment booster sessions could help determine if the size of treatment effects initially obtained able to be maintained at later follow up times. I appreciate the chance to provide additional insight into this article. The authors are to be congratulated for presenting work that provides valuable information on longer term effects of cognitive functional therapy and also helps to plan future clinical research. None declared.