On Usage of Artificial Intelligence for Predicting Neonatal Diseases, Conditions, and Mortality: A Bibliometric Review

Abstract

Purpose: Care and attention during the neonatal period are crucial to preventing negative outcomes. The literature presents artificial intelligence models as promising tools to assist healthcare professionals in disease prediction and support clinical decision-making. Methods: This study conducts a bibliometric review of the use of artificial intelligence models in predicting neonatal diseases, conditions and mortality. The review analyzed publications from 2014 to 2024. A total of 629 studies were selected after applying selection criteria. Subsequently, analyses of collaboration networks, keyword co-occurrence, citations and cluster analysis were performed. Results: The results show that the United States, China and the United Kingdom lead scientific production and international collaborations. 12 neonatal diseases were identified, with emphasis on “retinopathy of prematurity”, “necrotizing enterocolitis” and “bronchopulmonary dysplasia”; 7 clinical conditions, including “prematurity”, “perinatal asphyxia” and “jaundice”; and 5 neonatal outcomes, mainly “sepsis”, “mortality” and “cerebral palsy.” Cluster analysis revealed that studies predominantly use clinical, laboratory, genetic and imaging data, with Logistic Regression, Random Forest and Convolutional. Conclusion: The study has growing interest in applying artificial intelligence to neonatal care. The models are increasingly used with clinical, laboratory, genetic and imaging data, enabling earlier and more accurate diagnoses. However, the study also underscores important ethical considerations, such as data quality, algorithmic transparency and equitable access to these technologies, particularly in underrepresented regions, with scientific production uneven and limited participation from low- and middle-income countries.

Ähnliche Arbeiten