A diagnostic prediction model for separating juvenile idiopathic arthritis and chronic musculoskeletal pain syndrome

Abstract

Objective: To develop and validate a diagnostic prediction model that can distinguish between juvenile idiopathic arthritis (JIA) and chronic musculoskeletal pain syndrome (CMPS) based on patient-reported outcomes. Study design: This retrospective cohort study evaluated whether the Juvenile Arthritis Multidimensional Assessment Report (JAMAR) performs well in distinguishing JIA from CMPS. We analyzed JAMARs completed by 287 patients at their first visit to the pediatric rheumatology department of Wilhelmina Children's Hospital in Utrecht, The Netherlands. Relevant JAMAR items for predicting a diagnosis of JIA were selected in a penalized multivariable model suitable for clinical application. The model was subsequently validated with new data from the same center. Results: A total of 196 JAMARs (97 JIA, 99 CMPS) were collected in the model development data, and 91 JAMARs (48 JIA, 43 CMPS) were collected in the validation data. Variables in the prediction model that were strongest associated with a diagnosis of JIA instead of CMPS were asymmetric pain/swelling in the shoulder (OR, 2.34), difficulty with self-care (OR, 2.41), skin rash (OR, 2.07), and asymmetric/pain swelling in the knee (OR, 2.29). Calibration and discrimination (area under the receiver operating characteristic curve, 0.83; 95% CI, 0.74-0.92) of the model in the validation data were good. Conclusions: Several items from the JAMAR questionnaire can potentially distinguish JIA from CMPS in patients with corresponding symptoms. We present an easy-to-use, adjusted, and validated model to separate these 2 diagnoses early at presentation based on patient-reported outcomes to facilitate proper referral and treatment.