RESUMEN

BACKGROUND: Pregnancy among women infected with HIV is classified as a high-risk pregnancy. While previous research has indicated an elevated likelihood of preterm birth, low birth weight, and early gestational age in infants born to mothers with HIV, the correlation between maternal HIV infection and different neonatal results remains unclear. OBJECTIVE: This study aims to investigate the impact of maternal HIV infection on after-birth neonatal outcomes using machine learning (ML) and statistical methods. METHODS: A case-control study is conducted through a three-stage process: Initially, the outcomes among newborns from HIV-positive mothers are identified through a combination of literature review and expert survey. Subsequently, data are paired at a 1:2 ratio based on gestational age with infants from HIV-positive mothers (n=48) compared to HIV-negative mothers (n=96) as the control group. Finally, various feature selection techniques are applied to identify outcomes that exhibit significant differences between the two groups. RESULTS: The statistical analysis showed that the rate of addiction among HIV-positive mothers is higher than that of the HIV-negative group. The need for mechanical ventilation and duration of ventilator-assisted breathing in infants born to HIV-positive mothers are significantly higher than in infants born to HIV-negative mothers. Moreover, based on feature selection methods, increasing the need for mechanical ventilation and reducing surfactant administration were two important outcomes. CONCLUSIONS: To investigate the impact of maternal HIV infection on neonatal outcomes, various statistical and machine learning-based feature selection techniques were implemented, and the results showed that the presented methods can be utilized to examine the potential impacts of different diseases contracted by the mother on the infant.