Prediction of newborn assessment according to maternal nutritional status: a model based on decision trees.

Ramos-Purihuaman K.C.; Santos-Rosales Y.R. · Revista de Obstetricia y Ginecologia de Venezuela (2024) · DOI: 10.51288/00840205

Resumen

Objective: To determine whether it is possible to predict the assessment of the newborn according to maternal nutritional status through a decision tree model. Methods: Cross-sectional analytical study. A total of 326 medical records of pregnant women from a Peruvian public hospital were reviewed, in 2021. The newborn was assessed using the APGAR score, gestational age at birth, birth weight, weight and height for gestational age. Maternal nutritional status included pregestational body mass index and gestational weight gain. The prediction was made using a supervised machine learning model called a “decision tree.” Results: The APGAR score at one minute and height for gestational age were not possible to predict by maternal nutritional status. The probability of having full-term gestational age at birth is 97.2% when gestational weight gain is > 5.4 kg (p = 0.007). The highest probabilities of adequate birth weight were with gestational weight gain between 4.5 kg (p < 0.001) and 17 kg (p < 0.001) and with pregestational body mass index ≤ 36.523 kg/m2 (p = 0.004). Finally, the highest probability of adequate weight for gestational age is when gestational weight gain is < 11.8 Kg (p < 0.001) and with a pregestational body mass index ≤ 36.523 Kg/m2 (p = 0.005). Conclusions: It is possible to predict the assessment of the newborn based on the mother’s nutritional status using machine learning.

Instituciones: Universidad Nacional Mayor de San Marcos