Establishment and verification of an osteoporosis risk model in patients with rheumatoid arthritis: a valuable new model.

To establish a model for osteoporosis risk in patients with rheumatoid arthritis and validate the model. A newly generated predictive model has been suggested to have good differentiation, calibration, and clinical validity and may be a useful clinical model for predicting osteoporosis in patients with rheumatoid arthritis. To establish a prediction model for osteoporosis risk in patients with rheumatoid arthritis and validate the model internally and externally. A total of 270 patients with rheumatoid arthritis who underwent bone mineral density measurement at our hospital from June 2019 to June 2020 were enrolled in the study. The patients were divided into two groups according to their entry time: a training set containing the first 2/3 of the patients (n = 180) and a validation set containing the remaining 1/3 of the patients (n = 90). Binary logistic regression analysis was used to establish the regression models, and the concordance index (C-index), calibration plot, and decision curve analysis were used to evaluate the prediction model. Five variables, including age (X1), course of disease (X2), the disease activity score using 28 joint counts (DAS28) (X4), anti-cyclic citrullinated peptide antibody (CCP) (X7), and 7-joint ultrasonic bone erosion (X14), were selected to enter the model. The prediction model is Logit Y = - 12.647 + 0.133X1 + 0.011X2 + 0.754X4 + 0.001X7 + 0.605X14. The model had good differentiation; the C-index in the internal verification was 0.947 (95% CI is 0.932-0.977) and the C-index in the external verification was 0.946 (95% CI is 0.940-0.994). The calibration plot of the model showed excellent consistency between the prediction probability and actual probability. When > 0.483 was taken as the cutoff value for the diagnosis of osteoporosis, the sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and Jordan index of the model were 90.24%, 87.76%, 7.37, 0.11, and 78.00%, respectively. A newly generated predictive model has been suggested to have good differentiation, calibration, and clinical validity and may be a useful clinical model for predicting osteoporosis in patients with rheumatoid arthritis.