EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Development and Validation of Machine Learning Models to Predict Postoperative Delirium Using Clinical Features and Polysomnography Variables

Entwicklung und Validierung von Machine-Learning-Modellen zur Vorhersage eines postoperativen Delirs anhand klinischer Merkmale und Polysomnographie-Variablen

Woo-Seok Ha iD, Bo-Kyu Choi iD, Jungyeon Yeom, Seungwon Song, Soomi Cho iD, Min-Kyung Chu, Won-Joo Kim iD, Kyoung Heo, Kyung-Min Kim iD

Journal of Clinical Medicine · 16.09.2024 · 13 (18) · 5485

DOI: 10.3390/jcm13185485 ↗

Auf einen Blick

Bei 912 Erwachsenen sagte ein Machine-Learning-Modell ein postoperatives Delir mit klinischen und Schlafdaten besser voraus als mit PSG-Daten allein. Die höchste AUC lag bei 0,84.

Fragestellung

Lassen sich postoperative Delirien mit klinischen Merkmalen und Polysomnographie-Variablen per Machine Learning vorhersagen?

Methode

912 Erwachsene nach Operation in Allgemeinanästhesie an einem Tertiärkrankenhaus (2013–2024); PSG innerhalb von 5 Jahren vor der Operation. Aufteilung in Trainings- und Testset; Machine-Learning-Klassifikation.

Ergebnisse

912 Patient:innen; internes Trainingsset 700, externes Testset 212. AUC: 0,81 mit klinischen Variablen, 0,60 mit PSG allein, 0,84 mit beiden. Risikofaktoren u. a. Midazolam, OP-Dauer, Hypoalbuminämie.

Schlussfolgerung der Autor:innen

Das KI-Modell integriert klinische und Schlafvariablen und identifiziert ein postoperatives Delir zuverlässig; Schlaffaktoren tragen zur Identifikation bei.

Kernaussagen

  • AUC 0,84 mit klinischen und PSG-Daten
  • PSG allein: AUC 0,60
  • Wichtige Prädiktoren: OP-Dauer, Alter, Midazolam, O2-Nadir

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

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Clinical Medicine, online erschienen am 16.09.2024. Verfasst von Woo-Seok Ha und 8 weiteren Autor:innen. Beteiligte Einrichtung: Department of Neurology, Severance Hospital, Yonsei University College of Medicine, Seoul 03722, Republic of Korea. Die Arbeit wurde bisher 10-mal zitiert (Crossref) und nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Yonsei University College of Medicine.

Auf einen BlickAbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Delirium affects up to 50% of patients following high-risk surgeries and is associated with poor long-term prognosis. This study employed machine learning to predict delirium using polysomnography (PSG) and sleep-disorder questionnaire data, and aimed to identify key sleep-related factors for improved interventions and patient outcomes.

Methods

We studied 912 adults who underwent surgery under general anesthesia at a tertiary hospital (2013–2024) and had PSG within 5 years of surgery. Delirium was assessed via clinical diagnoses, antipsychotic prescriptions, and psychiatric consultations within 14 days postoperatively. Sleep-related data were collected using PSG and questionnaires. Machine learning predictions were performed to identify postoperative delirium, focusing on model accuracy and feature importance.

Results

This study divided the 912 patients into an internal training set (700) and an external test set (212). Univariate analysis identified significant delirium risk factors: midazolam use, prolonged surgery duration, and hypoalbuminemia. Sleep-related variables such as fewer rapid eye movement (REM) episodes and higher daytime sleepiness were also linked to delirium. An extreme gradient-boosting-based classification task achieved an AUC of 0.81 with clinical variables, 0.60 with PSG data alone, and 0.84 with both, demonstrating the added value of PSG data. Analysis of Shapley additive explanations values highlighted important predictors: surgery duration, age, midazolam use, PSG-derived oxygen saturation nadir, periodic limb movement index, and REM episodes, demonstrating the relationship between sleep patterns and the risk of delirium.

Conclusions

The artificial intelligence model integrates clinical and sleep variables and reliably identifies postoperative delirium, demonstrating that sleep-related factors contribute to its identification. Predicting patients at high risk of developing postoperative delirium and closely monitoring them could reduce the costs and complications associated with delirium.

Schlagwörter

Stichwörter der Autor:innen

machine learningpolysomnographypostoperative deliriumpredictive modelingsleep disorders

Autor:innen und Institutionen

  • Department of Neurology, Severance Hospital, Yonsei University College of Medicine, Seoul 03722, Republic of Korea.
  • Department of Biomedical Systems Informatics, Yonsei University College of Medicine, Seoul 03722, Republic of Korea.
  • Department of Neurology, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul 06273, Republic of Korea.

ORCID-Profile: Woo-Seok Ha, Bo-Kyu Choi, Soomi Cho, Won-Joo Kim, Kyung-Min Kim

Förderung

  • Yonsei University College of Medicine

Zitieren

APA

Ha, W. S., Choi, B. K., Yeom, J., Song, S., Cho, S., Chu, M. K., Kim, W. J., Heo, K., & Kim, K. M. (2024). Development and Validation of Machine Learning Models to Predict Postoperative Delirium Using Clinical Features and Polysomnography Variables. Journal of Clinical Medicine, 13(18), 5485. https://doi.org/10.3390/jcm13185485

Vancouver

Ha WS, Choi BK, Yeom J, Song S, Cho S, Chu MK, et al. Development and Validation of Machine Learning Models to Predict Postoperative Delirium Using Clinical Features and Polysomnography Variables. Journal of Clinical Medicine. 2024;13(18):5485. doi:10.3390/jcm13185485

BibTeX
@article{Ha2024,
  author = {Ha, Woo-Seok and Choi, Bo-Kyu and Yeom, Jungyeon and Song, Seungwon and Cho, Soomi and Chu, Min-Kyung and Kim, Won-Joo and Heo, Kyoung and Kim, Kyung-Min},
  title = {Development and Validation of Machine Learning Models to Predict Postoperative Delirium Using Clinical Features and Polysomnography Variables},
  journal = {Journal of Clinical Medicine},
  year = {2024},
  volume = {13},
  number = {18},
  pages = {5485},
  doi = {10.3390/jcm13185485}
}
Kostenlos starten →