Machine learning using the extreme gradient boosting (XGBoost) algorithm predicts 5-day delta of SOFA score at ICU admission in COVID-19 patients.

COVID-19 Clinical decision support system (CDSS) Extreme gradient boosting (XGBoost) Machine learning Multiple organ failure Organ dysfunction score

Journal

Journal of intensive medicine
ISSN: 2667-100X
Titre abrégé: J Intensive Med
Pays: China
ID NLM: 9918539389006676

Informations de publication

Date de publication:
Oct 2021
Historique:
received: 25 06 2021
revised: 20 08 2021
accepted: 06 09 2021
entrez: 14 2 2023
pubmed: 22 10 2021
medline: 22 10 2021
Statut: epublish

Résumé

Accurate risk stratification of critically ill patients with coronavirus disease 2019 (COVID-19) is essential for optimizing resource allocation, delivering targeted interventions, and maximizing patient survival probability. Machine learning (ML) techniques are attracting increased interest for the development of prediction models as they excel in the analysis of complex signals in data-rich environments such as critical care. We retrieved data on patients with COVID-19 admitted to an intensive care unit (ICU) between March and October 2020 from the RIsk Stratification in COVID-19 patients in the Intensive Care Unit (RISC-19-ICU) registry. We applied the Extreme Gradient Boosting (XGBoost) algorithm to the data to predict as a binary outcome the increase or decrease in patients' Sequential Organ Failure Assessment (SOFA) score on day 5 after ICU admission. The model was iteratively cross-validated in different subsets of the study cohort. The final study population consisted of 675 patients. The XGBoost model correctly predicted a decrease in SOFA score in 320/385 (83%) critically ill COVID-19 patients, and an increase in the score in 210/290 (72%) patients. The area under the mean receiver operating characteristic curve for XGBoost was significantly higher than that for the logistic regression model (0.86  The XGBoost model predicted the change in SOFA score in critically ill COVID-19 patients admitted to the ICU and can guide clinical decision support systems (CDSSs) aimed at optimizing available resources.

Sections du résumé

Background UNASSIGNED
Accurate risk stratification of critically ill patients with coronavirus disease 2019 (COVID-19) is essential for optimizing resource allocation, delivering targeted interventions, and maximizing patient survival probability. Machine learning (ML) techniques are attracting increased interest for the development of prediction models as they excel in the analysis of complex signals in data-rich environments such as critical care.
Methods UNASSIGNED
We retrieved data on patients with COVID-19 admitted to an intensive care unit (ICU) between March and October 2020 from the RIsk Stratification in COVID-19 patients in the Intensive Care Unit (RISC-19-ICU) registry. We applied the Extreme Gradient Boosting (XGBoost) algorithm to the data to predict as a binary outcome the increase or decrease in patients' Sequential Organ Failure Assessment (SOFA) score on day 5 after ICU admission. The model was iteratively cross-validated in different subsets of the study cohort.
Results UNASSIGNED
The final study population consisted of 675 patients. The XGBoost model correctly predicted a decrease in SOFA score in 320/385 (83%) critically ill COVID-19 patients, and an increase in the score in 210/290 (72%) patients. The area under the mean receiver operating characteristic curve for XGBoost was significantly higher than that for the logistic regression model (0.86 
Conclusions UNASSIGNED
The XGBoost model predicted the change in SOFA score in critically ill COVID-19 patients admitted to the ICU and can guide clinical decision support systems (CDSSs) aimed at optimizing available resources.

Identifiants

pubmed: 36785563
doi: 10.1016/j.jointm.2021.09.002
pii: S2667-100X(21)00032-3
pmc: PMC8531027
doi:

Types de publication

Journal Article

Langues

eng

Pagination

110-116

Investigateurs

Mario Alfaro-Farias (M)
Gerardo Vizmanos-Lamotte (G)
Thomas Tschoellitsch (T)
Jens Meier (J)
Hernán Aguirre-Bermeo (H)
Janina Apolo (J)
Alberto Martínez (A)
Geoffrey Jurkolow (G)
Gauthier Delahaye (G)
Emmanuel Novy (E)
Marie-Reine Losser (MR)
Tobias Wengenmayer (T)
Jonathan Rilinger (J)
Dawid L Staudacher (DL)
Sascha David (S)
Tobias Welte (T)
Klaus Stahl (K)
Agios Pavlos (A)
Theodoros Aslanidis (T)
Anita Korsos (A)
Barna Babik (B)
Reza Nikandish (R)
Emanuele Rezoagli (E)
Matteo Giacomini (M)
Alice Nova (A)
Alberto Fogagnolo (A)
Savino Spadaro (S)
Roberto Ceriani (R)
Martina Murrone (M)
Maddalena A Wu (MA)
Chiara Cogliati (C)
Riccardo Colombo (R)
Emanuele Catena (E)
Fabrizio Turrini (F)
Maria Sole Simonini (MS)
Silvia Fabbri (S)
Antonella Potalivo (A)
Francesca Facondini (F)
Gianfilippo Gangitano (G)
Tiziana Perin (T)
Maria Grazia Bocci (M)
Massimo Antonelli (M)
Diederik Gommers (D)
Raquel Rodríguez-García (R)
Jorge Gámez-Zapata (J)
Xiana Taboada-Fraga (X)
Pedro Castro (P)
Adrian Tellez (A)
Arantxa Lander-Azcona (A)
Jesús Escós-Orta (J)
Maria C Martín-Delgado (MC)
Angela Algaba-Calderon (A)
Diego Franch-Llasat (D)
Ferran Roche-Campo (F)
Herminia Lozano-Gómez (H)
Begoña Zalba-Etayo (B)
Marc P Michot (MP)
Alexander Klarer (A)
Rolf Ensner (R)
Peter Schott (P)
Severin Urech (S)
Nuria Zellweger (N)
Lukas Merki (L)
Adriana Lambert (A)
Marcus Laube (M)
Marie M Jeitziner (MM)
Beatrice Jenni-Moser (B)
Jan Wiegand (J)
Bernd Yuen (B)
Barbara Lienhardt-Nobbe (B)
Andrea Westphalen (A)
Petra Salomon (P)
Iris Drvaric (I)
Frank Hillgaertner (F)
Marianne Sieber (M)
Alexander Dullenkopf (A)
Lina Petersen (L)
Ivan Chau (I)
Hatem Ksouri (H)
Govind Oliver Sridharan (GO)
Sara Cereghetti (S)
Filippo Boroli (F)
Jerome Pugin (J)
Serge Grazioli (S)
Peter C Rimensberger (PC)
Christian Bürkle (C)
Julien Marrel (J)
Mirko Brenni (M)
Isabelle Fleisch (I)
Jerome Lavanchy (J)
Marie-Helene Perez (MH)
Anne-Sylvie Ramelet (AS)
Anja Baltussen Weber (AB)
Peter Gerecke (P)
Andreas Christ (A)
Samuele Ceruti (S)
Andrea Glotta (A)
Katharina Marquardt (K)
Karim Shaikh (K)
Tobias Hübner (T)
Thomas Neff (T)
Hermann Redecker (H)
Mallory Moret-Bochatay (M)
FriederikeMeyer Zu Bentrup (FZ)
Michael Studhalter (M)
Michael Stephan (M)
Jan Brem (J)
Nadine Gehring (N)
Daniela Selz (D)
Didier Naon (D)
Gian-Reto Kleger (GR)
Urs Pietsch (U)
Miodrag Filipovic (M)
Anette Ristic (A)
Michael Sepulcri (M)
Antje Heise (A)
Marilene Franchitti Laurent (M)
Jean-Christophe Laurent (JC)
Pedro D Wendel Garcia (PD)
Reto Schuepbach (R)
Dorothea Heuberger (D)
Philipp Bühler (P)
Silvio Brugger (S)
Patricia Fodor (P)
Pascal Locher (P)
Giovanni Camen (G)
Tomislav Gaspert (T)
Marija Jovic (M)
Christoph Haberthuer (C)
Roger F Lussman (RF)
Elif Colak (E)

Informations de copyright

© 2021 Chinese Medical Association. Published by Elsevier B.V.

Déclaration de conflit d'intérêts

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Références

JAMA. 2016 Feb 23;315(8):801-10
pubmed: 26903338
IEEE J Biomed Health Inform. 2020 Jan;24(1):235-246
pubmed: 30762572
JAMA. 2021 Jun 1;325(21):2149-2150
pubmed: 33818587
Crit Care. 2019 Aug 22;23(1):284
pubmed: 31439010
Nat Med. 2021 Feb;27(2):186-187
pubmed: 33526932
Artif Intell Med. 2020 May;105:101847
pubmed: 32505428
Crit Care. 2021 May 25;25(1):175
pubmed: 34034782
Crit Care. 2019 Apr 8;23(1):112
pubmed: 30961662
Ann Transl Med. 2020 Apr;8(7):443
pubmed: 32395487
Anesthesiol Clin. 2021 Jun;39(2):265-284
pubmed: 34024430
Intensive Care Med. 1996 Jul;22(7):707-10
pubmed: 8844239
EClinicalMedicine. 2020 Aug;25:100449
pubmed: 32838231
Nat Med. 2020 Mar;26(3):364-373
pubmed: 32152583
Crit Care Med. 2021 Apr 1;49(4):661-670
pubmed: 33405410
Front Endocrinol (Lausanne). 2021 Jun 17;12:649525
pubmed: 34220706
Crit Care Med. 2021 Mar 1;49(3):e279-e290
pubmed: 33470778
Nature. 2021 Dec;600(7889):472-477
pubmed: 34237774
Diabetologia. 2020 Aug;63(8):1500-1515
pubmed: 32472191
IEEE Trans Biomed Eng. 2021 Jan;68(1):148-160
pubmed: 32406821

Auteurs

Jonathan Montomoli (J)

Department of Anaesthesia and Intensive Care, Infermi Hospital, AUSL della Romagna, Rimini 47923, Italy.

Luca Romeo (L)

Department of Information Engineering, Università Politecnica delle Marche, Ancona 60131, Italy.

Sara Moccia (S)

Department of Information Engineering, Università Politecnica delle Marche, Ancona 60131, Italy.
The BioRobotics Institute and Department of Excellence in Robotics and AI, Scuola Superiore Sant'Anna, Pisa 56127, Italy.

Michele Bernardini (M)

Department of Information Engineering, Università Politecnica delle Marche, Ancona 60131, Italy.

Lucia Migliorelli (L)

Department of Information Engineering, Università Politecnica delle Marche, Ancona 60131, Italy.

Daniele Berardini (D)

Department of Information Engineering, Università Politecnica delle Marche, Ancona 60131, Italy.

Abele Donati (A)

Anesthesia and Intensive Care Unit, Azienda Ospedaliero Universitaria Ospedali Riuniti, Ancona 60126, Italy.
Department of Biomedical Sciences and Public Health, Università Politecnica delle Marche, Ancona 60126, Italy.

Andrea Carsetti (A)

Anesthesia and Intensive Care Unit, Azienda Ospedaliero Universitaria Ospedali Riuniti, Ancona 60126, Italy.
Department of Biomedical Sciences and Public Health, Università Politecnica delle Marche, Ancona 60126, Italy.

Maria Grazia Bocci (MG)

Department of Anaesthesia and Intensive Care, Fondazione Policlinico Universitario A. Gemelli IRCCS, Università Cattolica del Sacro Cuore, Rome 00168, Italy.

Pedro David Wendel Garcia (PD)

Institute of Intensive Care Medicine, University Hospital of Zurich, Zurich 8091, Switzerland.

Thierry Fumeaux (T)

Swiss Society of Intensive Care Medicine, Basel 4001, Switzerland.

Philippe Guerci (P)

Department of Anesthesiology and Critical Care Medicine, University Hospital of Nancy, Nancy 54511, France.

Reto Andreas Schüpbach (RA)

Institute of Intensive Care Medicine, University Hospital of Zurich, Zurich 8091, Switzerland.

Can Ince (C)

Department of Intensive Care Erasmus MC, University Medical Center Rotterdam, Rotterdam, 3015 GD, Netherlands.

Emanuele Frontoni (E)

Department of Information Engineering, Università Politecnica delle Marche, Ancona 60131, Italy.

Matthias Peter Hilty (MP)

Institute of Intensive Care Medicine, University Hospital of Zurich, Zurich 8091, Switzerland.

Classifications MeSH