Extracting Interesting Rules from Gestation Course Data for Early Diagnosis of Neonatal Hypoxia

• Published in: Journal of medical systems Vol. 43; no. 1; p. 8
• Main Authors: Skarga-Bandurova, Inna, Biloborodova, Tetiana, Nesterov, Maksym
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.01.2019; Springer Nature B.V
• Subjects: Algorithms; Artificial Intelligence Application in Health Informatics; Brain damage; Childbirth & labor; Complications; Data Collection - methods; Data processing; Delivery, Obstetric; Diagnosis; Early Diagnosis; Encephalopathy; Female; Fetuses; Gestation; Gestational Age; Health Informatics; Health Sciences; Humans; Hypoxia; Hypoxia - diagnosis; Infant, Newborn; Ischemia; Medical diagnosis; Medical personnel; Medicine; Medicine & Public Health; Neonates; Newborn babies; Pregnancy; Pregnancy complications; Risk analysis; Risk factors; Statistics for Life Sciences; Systems-Level Quality Improvement; Measure of interestingness; Gestation course data; Mining association rules; Neonatal hypoxia
• ISSN: 0148-5598; 1573-689X; 1573-689X
• DOI: 10.1007/s10916-018-1125-0

The topic of neonatal hypoxia is of paramount importance to anyone who cares during pregnancy and childbirth. Modern medicine associates this pathology with severe problems in the prenatal period. Underlying diseases of the mother during pregnancy, her anamnesis of life are the leading causes of complications in the newborn. Nevertheless, patterns of fetal hypoxia and neonatal hypoxia, as well as mechanisms of hypoxic-ischemic encephalopathy in newborns, remains poorly known and require further research. This study is focused on finding risk factors related to the chronic fetal hypoxia and defining a group of signs for diagnosing neonatal hypoxia. The real data of 186 pregnant women at the gestation age from 12 to 38 weeks were analyzed. A methodology for discovering interesting associations in gestation course data is proposed. Technique for association rules mining and rules selection by the neonatal hypoxia under study is discussed. The rules suggest that a strong relationship exists between the specific sets of attributes and the diagnosis. As a result, we set up a profile of the pregnant woman with a high likelihood of hypoxia of the newborn that would be beneficial to medical professionals.