Predictors of functional outcome in patients with chronic low back pain undergoing back school, individual physiotherapy or spinal manipulation.

Background: Recent studies on chronic low back pain (cLBP) rehabilitation suggest that predictors of treatment outcome may be differ according to the considered conservative treatment.
Aim: To identify predictors of response to back school (BS), individual physiotherapy (IP) or spinal manipulation (SM) for cLBP.
Population: outpatients with cLBP.
Setting: Outpatient rehabilitation department.
Design: Retrospective analysis from a randomized trial.
Methods: Two hundred and ten patients with cLBP were randomly assigned to either BS, IP or SM; the Roland Morris Disability Questionnaire (RM) was assessed before and after treatment: those who decreased their RM score <2.5 were considered non-responders. Baseline potential predictors of outcome included demographics, general and cLBP history, life satisfaction.
Results: Of the 205 patients who completed treatment (140/205 women, age 58+14 years), non-responders were 72 (34.2%). SM showed the highest functional improvement and the lowest non-response rate. In a multivariable logistic regression, lower baseline RM score (OR 0.82, 95% CI 0.76-0.89, P<0.001) and received treatment (OR 0.32, 95% CI 0.21-0.50, P<0.001) were independent predictors of non-response. Being in the lowest tertile of baseline RM score (<6) predicted non response to treatment for BS and IP, but not for SM (same risk for all tertiles).
Conclusions: In our patients with cLBP lower baseline pain-related disability predicted non-response to physiotherapy, but not to spinal manipulation.
Clinical Rehabilitation Impact: Our results suggest that, independent form other characteristics, patients with cLBP and low pain-related disability should first consider spinal manipulation as a conservative treatment.