Development and assessment of a predictive model for early diagnosis of rheumatoid arthritis in southwest China: A new nomogram

Abstract

Rheumatoid arthritis (RA) is a disease complicated with inflammatory synovitis, which seriously affects the life quality of patients. Early diagnosis is important for prognosis of RA. Here, we aimed to develop and assess a model for early diagnosis of RA in southwest China. A nomogram including 44 patients with an early diagnosis of RA was developed. Variables were filtered by least absolute contraction selection operator and multiple logistic regression. The efficiency and clinical application range were evaluated. This nomogram showed that rheumatoid factor, erythrocyte sedimentation rate, RA33, facet joint and knee joint had high positive predictive value for RA. The area under curve was 0.920 [95% confidence interval (CI): 0.865–0.975]. In the validation model, area under curve was 0.942 (95% CI: 0.893–0.991). Calibration and decision curve suggested that this nomogram was helpful within the threshold probability range of 0.02 to 1.00. Using this nomogram will help clinicians in the early diagnosis of RA. Laboratory indicators such as rheumatoid factor, erythrocyte sedimentation rate, RA33, and clinical symptoms such as morning stiffness, facet joint and knee joint are very important, which deserves the attention of clinicians.