Clinical prediction model for interdisciplinary biopsychosocial rehabilitation in osteoarthritis patients

Abstract

BACKGROUND Osteoarthritis (OA) is a heterogenous condition, in which different subgroups are present. Individualized interdisciplinary multimodal pain treatments (IMPT) based on the biopsychosocial model have resulted in positive improvement of pain, health and disability in OA patients. Moreover, predictive factors for treatment success of IMPT in different musculoskeletal pain populations have been examined, but a clinical prediction model which informs whether an OA patient is expected to benefit or not from IMPT is currently lacking. AIM The aim was to develop and internally validate a clinical prediction model to inform patient-tailored care based on identified predictors for positive or negative outcomes of IMPT in patients with OA. DESIGN Longitudinal prospective cohort study. SETTING Center for Integral Rehabilitation at six locations in the Netherlands. POPULATION Chronic OA patients. METHODS Data in this study were collected during January 2019 until January 2022. Participants underwent a 10-week IMPT program based on the biopsychosocial model. Treatment success was defined by a minimal decrease from baseline of 9 points on the Pain Disability Index (PDI). Candidate predictors were selected by experts in IMPT and literature review. Backward logistic regression analysis was performed to develop the clinical predication model and bootstrap validation was performed for internal validation. RESULTS Overall, 599 OA patients were included, of which 324 experienced treatment success. Thirty-four variables were identified as possible predictors for good IMPT outcome. Age, gender, number of pain locations, PDI baseline score, maximal pain severity, use of pain medication and alcohol, work ability, brief illness perceptions questionnaire subscales timeline, consequences, identity and treatment control, pain catastrophizing scale and self-efficacy questionnaire score were found as predictors for treatment success. The internally validated model has an acceptable discriminative power of 0.71. CONCLUSIONS This study reports a specific clinical prediction model for good outcome of IMPT in patients with OA. The internally validated model has an acceptable discriminative power of 0.71. CLINICAL REHABILITATION IMPACT After external validation, this model could be used to develop a clinically useful decision tool.