Construction of Intraoperative Blood Glucose Abnormality Risk Prediction Model and Nursing Strategies for Nondiabetic Elderly Patients.

Abstract

Purpose: To construct a risk prediction model for intraoperative blood glucose abnormalities in nondiabetic elderly patients and explore nursing strategies.

Design: A retrospective study design was employed.

Methods: A total of 200 nondiabetic elderly surgical patients admitted to our hospital between March 2022 and December 2024 were retrospectively selected for the study. The modeling group comprised 160 patients, with 40 allocated to the validation group. In the modeling group, patients were further divided into the normal blood glucose group (n = 110) and the abnormal blood glucose group (n = 50) based on intraoperative blood glucose status. We analyzed predictors of blood glucose abnormalities and developed a predictive model.

Findings: In the modeling group, there were statistically significant differences in age, surgical duration, and preoperative blood glucose between the two groups, multivariate logistic regression identified age, surgical duration, and preoperative glucose as independent predictors of intraoperative glycemic abnormalities in nondiabetic elderly patients, and the model was successfully established. The predicted and actual risks of postoperative recurrence showed good agreement. The receiver operating characteristic analysis revealed an area under the curve of 0.891 for the validation set, and the standard error was 0.031 (95% CI: 0.8633 to 0.9416). Sensitivity and specificity were 72.9% and 77.1%, respectively. Decision curve analysis demonstrated significant net benefits, indicating good clinical utility.

Conclusions: A risk prediction model for intraoperative blood glucose abnormalities in nondiabetic elderly patients has been successfully established and validated.