Objective Baseline characteristics of patients low back pain differ substantially between care settings, but it is largely unknown whether predictors are of equal importance across settings. The aim of this study was to investigate whether 8 known predictors relate differently to outcomes in chiropractic practice and in general practice and to which degree these factors may be helpful in selecting patients benefiting more from one setting or the other. Methods Patient characteristics were collected at baseline, and outcomes of pain intensity (numeric rating scale 0-10) and activity limitation (Roland-Morris Disability Questionnaire 0-100) after 2, 12, and 52 weeks. Differences in the prognostic strength between settings were investigated for each prognostic factor separately by estimating the interaction between setting and the prognostic factor using regression models. Between-setting differences in outcome in high-risk and low-risk subgroups, formed by single prognostic factors, were assessed in similar models adjusted for a propensity score to take baseline differences between settings into account. Results Prognostic factors were generally associated more strongly with outcomes in general practice compared with chiropractic practice. The difference was statistically significant for general health, duration of pain, and musculoskeletal comorbidity. After propensity score adjustment, differences in outcomes between settings were insignificant, but negative prognostic factors tended to be less influential in chiropractic practice except for leg pain and depression, which tended to have less negative impact in general practice. Conclusion Known prognostic factors related differently to outcomes in the 2 settings, suggesting that some subgroups of patients might benefit more from one setting than the other.< Réduire