EUVIMEDEUVIMED entdecken
KohortenstudieOpen Access

Machine Learning–Based Prediction of Culture-Confirmed Neonatal Sepsis in a Tertiary Neonatal Intensive Care Unit: Retrospective Cohort Study

Machine-Learning-basierte Vorhersage kulturbestätigter neonataler Sepsis in einer tertiären neonatologischen Intensivstation: retrospektive Kohortenstudie

Eman Badran iD, Oraib Al-Smadi iD, Loiy T Algazo iD, Alaa T Al Ghazo iD, Arwa Al Anber iD, Areej Sharaqa iD, Lena Abu-Argoub iD, Shatha Al Jaberi iD, Abdulrahman Alhanbali iD, Taimein Yacoub iD, Shahd Rihan iD, Hala Al-Jaberi iD

JMIR Medical Informatics · 14.09.2026 · 14 · e88732-e88732

DOI: 10.2196/88732 ↗

Auf einen Blick

Bei 3.274 Neugeborenen sagte XGBoost kulturbestätigte Sepsis im Testdatensatz am besten voraus. Die wichtigsten Prädiktoren waren C‑reaktives Protein, Thrombozytenzahl und Gestationsalter.

Fragestellung

Wie gut sagen ML-Modelle kulturbestätigte neonatale Sepsis bei Neugeborenen in einer tertiären NICU in Jordanien voraus?

Methode

Retrospektive Kohortenstudie mit 3 ML-Modellen auf strukturierten EHR-Daten von 3.274 Neugeborenen; Training/Test 80/20; SMOTE im Trainingsdatensatz.

Ergebnisse

XGBoost erreichte im unabhängigen Testset (n=655) eine Genauigkeit von 94 % (616/655; 95 % CI 92 % bis 96 %), Sensitivität 98 % (95/97; 95 % CI 96 % bis 99 %) und AUC 0,98 (95 % CI 0,97 bis 0,99). Neural Networks: AUC 0,81 (95 % CI 0,78 bis 0,84).

Einschränkungen

Externe Validierung und prospektive Studien sind vor der klinischen Anwendung erforderlich.

Schlussfolgerung der Autor:innen

XGBoost zeigte eine starke Vorhersageleistung und könnte eine Grundlage für künftige prospektive klinische Entscheidungshilfen sein.

Kernaussagen

  • XGBoost war das leistungsstärkste Modell im Testdatensatz.
  • Genauigkeit 94 %, Sensitivität 98 %, AUC 0,98.
  • Wichtige Merkmale: C‑reaktives Protein, Thrombozyten, Gestationsalter.

Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.

Kurz gefasst

Kohortenstudie in der Zeitschrift JMIR Medical Informatics, online erschienen am 14.09.2026. Verfasst von Eman Badran und 11 weiteren Autor:innen. Beteiligte Einrichtung: Department of Pediatrics, Neonatal Division, School of Medicine, University of Jordan, Queen Rania Street, Amman, 11942, Jordan, 962 795608288, 962 6…. Hauptthemen laut MEDLINE: Neonatal Sepsis, Machine Learning. Die Arbeit nennt 28 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.

Auf einen BlickAbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Neonatal sepsis remains a major cause of neonatal morbidity and mortality in low- and middle-income countries (LMICs). Early diagnosis is challenging because of nonspecific clinical manifestations and delays in laboratory confirmation. Machine learning (ML) approaches using structured electronic health record (EHR) data may improve early risk stratification in neonatal intensive care units (NICUs).

Objective

This study aimed to evaluate ML models for predicting culture-confirmed neonatal sepsis among neonates admitted to a tertiary NICU in Jordan, with the objective of addressing diagnostic gaps in resource-limited settings. Specifically, we aimed to identify key predictors through feature importance analysis, evaluate model performance with class-imbalanced data, and propose strategies to improve interpretability and generalizability in LMICs.

Methods

A retrospective cohort study was conducted using structured EHRs of 3274 neonates admitted to a tertiary NICU in Jordan between 2018 and 2024. Neonates who underwent blood culture testing were included. The dataset was divided into training (n=2619, 80%) and testing (n=655, 20%) subsets using stratified sampling. Three ML models—Extreme Gradient Boosting (XGBoost), decision trees, and neural networks—were trained using clinical, laboratory, and demographic variables. Class imbalance was addressed using the synthetic minority oversampling technique (SMOTE) applied to the training dataset. Model performance was evaluated using accuracy, sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC).

Results

Among 3274 neonates included in the study, the XGBoost model demonstrated the best predictive performance on the independent test set (n=655, 20%), achieving an accuracy of 94% (616/655 correct predictions, 95% CI 92% to 96%), sensitivity of 98% (95/97 sepsis cases correctly identified, 95% CI 96% to 99%), and an AUC of 0.98 (95% CI 0.97 to 0.99). Decision trees provided interpretable classification rules with moderate performance, whereas neural networks showed lower discriminative ability, with an AUC of 0.81 (95% CI 0.78 to 0.84). Important predictive features included C-reactive protein, platelet count, and gestational age.

Conclusions

XGBoost demonstrated strong predictive performance in this retrospective cohort, supporting its potential as a foundation for future prospective clinical decision support tools. External validation and prospective studies are required before clinical implementation.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Neonatal SepsisMachine LearningHumansRetrospective StudiesIntensive Care Units, NeonatalJordanInfant, NewbornPredictive Learning ModelsFemaleBoosting Machine Learning AlgorithmsTertiary Care CentersClassification AlgorithmsROC CurvePrediction AlgorithmsElectronic Health Records

Stichwörter der Autor:innen

EHRExtreme Gradient BoostingXGBoostclinical decision support toolsdecision treeselectronic health recordmachine learning modelsneonatal sepsisneural networks

Autor:innen und Institutionen

  • Department of Pediatrics, Neonatal Division, School of Medicine, University of Jordan, Queen Rania Street, Amman, 11942, Jordan, 962 795608288, 962 6 5300431.
  • Department of Neonatology, Princess Rahmah Teaching Hospital, Ministry of Health, Irbid, Jordan.
  • Department of Mechatronics Engineering, Faculty of Engineering, Hashemite University, Zarqa, Jordan.
  • Department of Pharmacology Public Health and Clinical Skills, Faculty of Medicine, Hashemite University, Zarqa, Jordan.
  • Department of Pediatrics, The University of Texas Medical Branch at Galveston, Galveston, TX, United States.

ORCID-Profile: Eman Badran, Oraib Al-Smadi, Loiy T Algazo, Alaa T Al Ghazo, Arwa Al Anber, Areej Sharaqa, Lena Abu-Argoub, Shatha Al Jaberi, Abdulrahman Alhanbali, Taimein Yacoub, Shahd Rihan, Hala Al-Jaberi

Zitieren

APA

Badran, E., Al-Smadi, O., Algazo, L. T., Al Ghazo, A. T., Al Anber, A., Sharaqa, A., Abu-Argoub, L., Al Jaberi, S., Alhanbali, A., Yacoub, T., Rihan, S., & Al-Jaberi, H. (2026). Machine Learning–Based Prediction of Culture-Confirmed Neonatal Sepsis in a Tertiary Neonatal Intensive Care Unit: Retrospective Cohort Study. JMIR Medical Informatics, 14, e88732-e88732. https://doi.org/10.2196/88732

Vancouver

Badran E, Al-Smadi O, Algazo LT, Al Ghazo AT, Al Anber A, Sharaqa A, et al. Machine Learning–Based Prediction of Culture-Confirmed Neonatal Sepsis in a Tertiary Neonatal Intensive Care Unit: Retrospective Cohort Study. JMIR Medical Informatics. 2026;14:e88732-e88732. doi:10.2196/88732

BibTeX
@article{Badran2026,
  author = {Badran, Eman and Al-Smadi, Oraib and Algazo, Loiy T and Al Ghazo, Alaa T and Al Anber, Arwa and Sharaqa, Areej and Abu-Argoub, Lena and Al Jaberi, Shatha and Alhanbali, Abdulrahman and Yacoub, Taimein and Rihan, Shahd and Al-Jaberi, Hala},
  title = {Machine Learning–Based Prediction of Culture-Confirmed Neonatal Sepsis in a Tertiary Neonatal Intensive Care Unit: Retrospective Cohort Study},
  journal = {JMIR Medical Informatics},
  year = {2026},
  volume = {14},
  pages = {e88732--e88732},
  doi = {10.2196/88732}
}
Kostenlos starten →