Examination of a subgroup of patients with chronic low back pain likely to benefit more from pilates-based exercises compared to an educational booklet

Objective: To investigate whether 2 previously published classification approaches, the updated treatment-based classification system and a Pilates subgroup defined by a preliminary clinical prediction rule, could identify patients with chronic low back pain who would benefit more from Pilates exercises compared to an educational booklet. Design: Secondary analysis of a randomized controlled trial. Methods: Two hundred twenty-two patients received advice and were randomly allocated to a group that received an educational booklet with no additional treatment (n = 74) or a group that received Pilates-based exercise treatment (n = 148) 2 or 3 times a week. At baseline, using a treatment-based classification system, patients were classified as having a good prognosis (positive movement control) or a poor prognosis. Similarly, using the Pilates clinical prediction rule, patients were classified as having a good prognosis (positive) or a poor prognosis (negative). The analysis was conducted using linear regression models to analyze the interaction between subgroup characteristics and treatment effect size, with changes in pain and disability from baseline to 6 weeks after randomization as dependent variables. Results: None of the interaction terms for pain and disability were statistically significant. The treatment effect of Pilates versus an educational booklet was similar in all subgroups. Conclusion: The treatment-based classification system and the Pilates clinical prediction rule did not differentiate subgroups of patients with chronic low back pain who were more or less likely to benefit more from Pilates compared to an educational booklet.