Establishment and verification of a predictive model for feeding intolerance in premature infants
10.3760/cma.j.cn211501-20230920-00586
- VernacularTitle:早产儿喂养不耐受预测模型的构建与验证
- Author:
Zhengju CHEN
1
;
Jihong FANG
Author Information
1. 安徽医科大学护理学院，合肥 230032
- Keywords:
Infant, premature;
Feeding intolerance;
Prediction model;
Build;
Verify
- From:
Chinese Journal of Practical Nursing
2024;40(11):816-822
- CountryChina
- Language:Chinese
-
Abstract:
Objective:The prediction model of feeding intolerance in preterm infants was established and validated to provide guidance for clinical practice.Methods:This was a case-control study. A retrospective analysis was conducted on 210 premature infants with gestational age less than 34 weeks from September 2022 to May 2023. They were divided into training and validation sets in a 1∶1 ratio. The univariate and multivariate binary Logistic regression analysis were performed on training set samples, first identified the risk factors for feeding intolerance occurrence, and established a premature feeding intolerance risk prediction model based on these risk factors. Visualized the model using a column chart. The performance of the model was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and clinical decision curves in the training and validation sets, respectively. The ROC curve was used to evaluate the differentiation ability of the model, the calibration curve was used to evaluate the consistency of the model, and clinical decision-making was used to evaluate the net benefit status of patients when the model guides clinical interventions.Results:Among them, there were 84 cases in the feeding tolerance and 126 cases in the feeding intolerance. There were 53 males and 31 females with feeding tolerance aged (32.38 ± 1.37) weeks and 73 males and 53 females with feeding intolerance aged (30.01 ± 2.14) weeks. Through univariate Logistic regression analysis of 12 related variables, there were significant differences between the feeding tolerance premature infants and the feeding intolerance premature infants in 8 variables of premature birth weight, birth asphyxia, caffeine use, delayed defecation, gestational age, lactation time, non-invasive ventilation time, and invasive ventilation time ( OR values were 0.032-18.706, all P<0.05). Multiple Logistic regression ultimately screened out three variables, namely premature infant body mass, delayed defecation, and non-invasive ventilation time ( OR = 0.073, 4.926, 1.244, all P<0.05). The area under the ROC curve of the training and validation sets were 0.906 and 0.876, respectively. The calibration curves of the training and validation sets indicated that the model had high consistency. The Hosmer-Lemeshow goodness of fit test showed that χ2 = 7.92, P = 0.442. Conclusions:The prediction model established in this study has high discrimination, calibration, and clinical practical value, and can accurately predict the risk of feeding intolerance in premature infants, providing reference basis for timely nursing and clinical intervention.
