The multivariate cox regression model for complete enteral nutrition after primary anastomosis in neonates with intestinal atresia

OBJECTIVE: Enteral feeding after intestinal atresia has always been a concern for clinicians. But the present studies mainly focused on single factors. This research aimed to comprehensively analyze the multiple factors on complete enteral nutrition after primary anastomosis, and establish the convenient prediction model. METHODS: We retrospectively collected reliable information in neonates with intestinal atresia form January 2010 to June 2022. The cox regression analysis was performed to select independent risk factors and develop nomogram. Subsequently, ROC curve, calibration curve and decision curve were drawn to thoroughly evaluate the accuracy and applicability of the model. RESULTS: The predictors finally included in the model were gestational age, meconium peritonitis, distance from the anastomosis to the ileocecal region, diameter ratio of proximal to distal bowels, and time of initial feeding. The nomogram of predicting the probability of week 2, week 3 and week 4 was drawn and their area under the curve were 0.765, 0.785 and 0.747, respectively. Similarly, calibration and decision curve indicated that the prediction model had a great prediction performance. CONCLUSION: The clinical value of predictive models can be recognized. The hope is that the predictive model can help pediatricians reduce hospital costs and parental anxiety.