Construction of neonatal hypoglycemia risk prediction model and its effect verification
Objective To provide theoretical reference for early identification of neonatal hypoglycemia by constructing clinical risk factors for neonatal hypoglycemia and building a prediction model.Methods A total of 1330 newborns in a hospital in Nanjing from January 2020 to January 2022 were selected as the research subjects,and divided into a modeling group(n=931)and a validation group(n=399)according to 7:3.Logistic regression was used to build the model,and Hosmer-Lemeshow and ROC were used to evaluate the goodness of fit and prediction effect.Results The incidence of neonatal hypoglycemia in the modeling group and validation group were 12.57%(117/931)and 12.78%(44/399)respectively,there was no statistically significant difference between the two groups(P>0.05).Six factors,including maternal diabetes,maternal obesity,use of beta-blockers during pregnancy and lactation,prenatal use of corticosteroids,small for gestational age,and blood glucose testing methods were the independent influencing factors.The area under the ROC curve of the modeling group was 0.893(95%CI:0.855～0.931,P=0.000),and the H-L test result was P=0.155.The area under the curve of the validation group was 0.847(95%CI:0.828～0.9876,P=0.000),and the H-L test result was P=0.401.Conclusion:This model had good prediction effect and can effectively predict and classify neonatal hypoglycemia,providing a theoretical basis for clinical care of neonatal hypoglycemia.
万方数据
维普
