RESUMO
Objective:To construct and validate a nomogram predictive model for readmission risk within 1 year of children with type 1 diabetes.Methods:A total of 395 children with type 1 diabetes who were hospitalized in four hospitals in Xi′an City from February 2019 to February 2021 were selected as the research subjects. The children were divided into training set ( n = 219) and verification set ( n = 175) in 5∶4 ratios. A nomogram prediction model for readmission risk within 1 year of children with type 1 diabetes was constructed based on the training set data, and internal validation was carried out. The external validation was carried out based on validation set data. Results:A total of 85 (21.5%) children were readmitted within 1 year. Mean glycohemoglobin A1c ≥ 7.5%, co-infection, complications of diabetes, and family history of diabetes were risk factors for readmission within 1 year of children with type 1 diabetes ( OR values were 4.010 - 5.510, P<0.05), and age of onset >7 years old was a protective factor ( OR = 0.070, P<0.05). The internal verification of nomogram model showed that the area under ROC curve was 0.778 (95% CI 0.703- 0.853), and the observed curve in calibration curve was basically consistent with the predicted curve. The external verification showed that the area under ROC curve was 0.748 (95% CI 0.642- 0.854), and the observed curve in calibration curve was basically consistent with the predicted curve. Conclusions:The nomogram predictive model for readmission risk within 1 year of children with type 1 diabetes is scientific and practical, and has certain clinical value in guiding targeted prevention and intervention of readmission of children with type 1 diabetes within one year.