IDENTIFYING TREATMENT EFFECT MODIFIERS IN THE START BACK TRIAL: A SECONDARY ANALYSIS

Purpose and Background: To identify treatment effect modifiers within the STarT Back Trial which demonstrated prognostic stratified care was effective in comparison to standard care for patients with low back pain. Methods: Secondary analysis of the STarT Back Trial using 688 patients with available 4-month follow-up data. Disability (baseline and 4 months) was assessed using the Roland Morris Disability Questionnaire (RMDQ) using continuous and dichotomized (>7) outcome scores. Potential treatment effect modifiers were evaluated with group x predictor interaction terms using linear and logistic regression models. Modifiers included: age, gender, education, socio-economic status (SES), employment status, work satisfaction, episode duration, general health (SF-12), number of pain medications, and treatment expectations. Results: No significant treatment effect modifiers were identified using linear regression, however logistic regression analysis identified SES as a treatment modifier (interaction: OR = 0.65; 95% confidence interval = 0.42 to 0.98). Comparing those receiving stratified care versus best current care, the proportion of patients with poor 4 month outcome was similar for lower SES patients (35.4% vs. 37.2%), yet different for high SES patients (19.3% vs. 38.9%). No other treatment effect modifiers were identified; however exploratory analyses indicated trends for worse outcomes with increased age, patients taking ≥3 medications, those with lower education, and those with poor treatment expectations as potential treatment effect modifiers. Conclusion: SES was identified as a treatment effect modifier in the STarT Back Trial with patients with lower SES not benefiting from stratified care. The explanations for this finding need further exploration.