类风湿关节炎达标治疗人群临床特点及风险因素评估预测模型建立与验证OA北大核心CSTPCD

Objective To screen independent risk factors that affect standard treatment by analyzing the clinical characteristics of standard treatment populations in real-world studies,and to establish and validate an evaluation and prediction model for clinical standard treatment in Chinese rheumatoid arthritis(RA)patients.Methods A total of 3 310 RA patients admitted to 10 domestic hospitals from January 2015 to December 2021 were included.All received traditional antirheumatic drugs to improve their conditions.A total of 876 patients with complete clinical follow-up data were selected and enrolled in the cohort for 24 weeks.Univariate and multivariate Logistic regression analyses were used to screen out independent risk factors affecting the treatment outcome[Disease Activity Score-erythrocyte sedimentation rate(DAS28-ESR)＜2.6].Least absolute shrinkage and selection operator(LASSO)regression was used to construct the Normogram for predicting the risk of reaching the target outcome using the minimum absolute shrinkage and selection algorithm.Using Bootstrap internal verification concordance index(C-index),calibration graph and decision curve analysis(DCA),the accuracy and stability of the prediction model were verified and evaluated.Results(1)Univariate non-parametric test,Chi-square test,and multivariate Logistic regression analysis identified six independent risk factors for achieving RA treatment,including body mass index,self-assessment of rest pain,hemoglobin,exacerbation of cold exposure,limb numbness,short breath and fatigue.(2)R language constructed the Normogram of risk prediction.C-index=0.712 was calculated by internal sampling verification.The receiver operating characteristic curve was constructed by multiple factors,and the area under the curve=0.729 indicated that the model had good differentiation.(3)The Calibration curve fitted with the calibration prediction curve,which indicated that the calibration degree of the model was good.Hosmer-Lemeshow fit test of DCA curve showed that the model had good potential clinical application value.Conclusion This clinical prediction model was helpful for doctors to identify early risk factors affecting the standard treatment outcomes of RA patients in the baseline period,thus to intervene with timely and active treatment strategies and obtain better clinical treatment effects.