Development and Prospective Validation of a Deep Learning Algorithm for Predicting Need for Mechanical Ventilation.
artificial intelligence
artificial respiration
coronavirus
deep learning
lung
Journal
Chest
ISSN: 1931-3543
Titre abrégé: Chest
Pays: United States
ID NLM: 0231335
Informations de publication
Date de publication:
06 2021
06 2021
Historique:
received:
04
07
2020
revised:
19
11
2020
accepted:
04
12
2020
pubmed:
22
12
2020
medline:
22
6
2021
entrez:
21
12
2020
Statut:
ppublish
Résumé
Objective and early identification of hospitalized patients, and particularly those with novel coronavirus disease 2019 (COVID-19), who may require mechanical ventilation (MV) may aid in delivering timely treatment. Can a transparent deep learning (DL) model predict the need for MV in hospitalized patients and those with COVID-19 up to 24 h in advance? We trained and externally validated a transparent DL algorithm to predict the future need for MV in hospitalized patients, including those with COVID-19, using commonly available data in electronic health records. Additionally, commonly used clinical criteria (heart rate, oxygen saturation, respiratory rate, Fio We obtained data from more than 30,000 ICU patients (including more than 700 patients with COVID-19) from two academic medical centers. The performance of the model with a 24-h prediction horizon at the development and validation sites was comparable (AUC, 0.895 vs 0.882, respectively), providing significant improvement over traditional clinical criteria (P < .001). Prospective validation of the algorithm among patients with COVID-19 yielded AUCs in the range of 0.918 to 0.943. A transparent deep learning algorithm improves on traditional clinical criteria to predict the need for MV in hospitalized patients, including in those with COVID-19. Such an algorithm may help clinicians to optimize timing of tracheal intubation, to allocate resources and staff better, and to improve patient care.
Sections du résumé
BACKGROUND
Objective and early identification of hospitalized patients, and particularly those with novel coronavirus disease 2019 (COVID-19), who may require mechanical ventilation (MV) may aid in delivering timely treatment.
RESEARCH QUESTION
Can a transparent deep learning (DL) model predict the need for MV in hospitalized patients and those with COVID-19 up to 24 h in advance?
STUDY DESIGN AND METHODS
We trained and externally validated a transparent DL algorithm to predict the future need for MV in hospitalized patients, including those with COVID-19, using commonly available data in electronic health records. Additionally, commonly used clinical criteria (heart rate, oxygen saturation, respiratory rate, Fio
RESULTS
We obtained data from more than 30,000 ICU patients (including more than 700 patients with COVID-19) from two academic medical centers. The performance of the model with a 24-h prediction horizon at the development and validation sites was comparable (AUC, 0.895 vs 0.882, respectively), providing significant improvement over traditional clinical criteria (P < .001). Prospective validation of the algorithm among patients with COVID-19 yielded AUCs in the range of 0.918 to 0.943.
INTERPRETATION
A transparent deep learning algorithm improves on traditional clinical criteria to predict the need for MV in hospitalized patients, including in those with COVID-19. Such an algorithm may help clinicians to optimize timing of tracheal intubation, to allocate resources and staff better, and to improve patient care.
Identifiants
pubmed: 33345948
pii: S0012-3692(20)35454-4
doi: 10.1016/j.chest.2020.12.009
pmc: PMC8027289
pii:
doi:
Types de publication
Journal Article
Observational Study
Research Support, N.I.H., Extramural
Langues
eng
Sous-ensembles de citation
IM
Pagination
2264-2273Subventions
Organisme : NINDS NIH HHS
ID : R01 NS102190
Pays : United States
Organisme : NHLBI NIH HHS
ID : K24 HL132105
Pays : United States
Organisme : NHLBI NIH HHS
ID : R01 HL119201
Pays : United States
Organisme : NHLBI NIH HHS
ID : K23 HL154863
Pays : United States
Organisme : NIEHS NIH HHS
ID : K01 ES025445
Pays : United States
Organisme : NINDS NIH HHS
ID : R01 NS107291
Pays : United States
Organisme : NHLBI NIH HHS
ID : T32 HL134632
Pays : United States
Organisme : NHLBI NIH HHS
ID : R01 HL081823
Pays : United States
Organisme : NHLBI NIH HHS
ID : R01 HL085188
Pays : United States
Organisme : NIA NIH HHS
ID : RF1 AG064312
Pays : United States
Organisme : NIA NIH HHS
ID : R01 AG063925
Pays : United States
Organisme : NINDS NIH HHS
ID : R01 NS102574
Pays : United States
Organisme : NHLBI NIH HHS
ID : R01 HL154926
Pays : United States
Organisme : NHLBI NIH HHS
ID : R01 HL148436
Pays : United States
Commentaires et corrections
Type : UpdateOf
Informations de copyright
Copyright © 2020 American College of Chest Physicians. Published by Elsevier Inc. All rights reserved.
Références
Am J Respir Crit Care Med. 2020 May 15;201(10):1299-1300
pubmed: 32228035
JAMA. 2016 Dec 13;316(22):2402-2410
pubmed: 27898976
Crit Care. 2018 Oct 30;22(1):286
pubmed: 30373653
Intensive Care Med. 2020 Jun;46(6):1099-1102
pubmed: 32291463
Intensive Care Med. 2020 Oct;46(10):1924-1926
pubmed: 32671470
Lancet Respir Med. 2020 Apr;8(4):e19
pubmed: 32105633
Crit Care. 2020 Apr 28;24(1):176
pubmed: 32345343
Lancet. 2020 Apr 25;395(10233):1321-1322
pubmed: 32277876
Nat Med. 2019 Jan;25(1):24-29
pubmed: 30617335
NPJ Digit Med. 2018 May 8;1:18
pubmed: 31304302
Nature. 2019 Aug;572(7767):116-119
pubmed: 31367026
Intensive Care Med. 2020 May;46(5):837-840
pubmed: 32123994
N Engl J Med. 2020 Apr 30;382(18):1687-1695
pubmed: 32286748
Ann Intern Med. 2009 Jan 20;150(2):132-8
pubmed: 19153413
Artif Intell Med. 2021 Mar;113:102036
pubmed: 33685592
JAMA. 2020 Mar 17;323(11):1061-1069
pubmed: 32031570
N Engl J Med. 2020 May 21;382(21):2049-2055
pubmed: 32202722
N Engl J Med. 2020 Apr 30;382(18):1708-1720
pubmed: 32109013
Zhonghua Jie He He Hu Xi Za Zhi. 2020 Apr 12;43(4):288-296
pubmed: 32294813
Br J Surg. 2015 Feb;102(3):148-58
pubmed: 25627261
Lancet. 2020 Mar 28;395(10229):1054-1062
pubmed: 32171076
Curr Cardiol Rep. 2014 Jan;16(1):441
pubmed: 24338557
Clin Chest Med. 2008 Jun;29(2):323-8, vii
pubmed: 18440440
J Thorac Dis. 2013 Aug;5 Suppl 2:S122-6
pubmed: 23977432
Am J Respir Crit Care Med. 2002 Apr 1;165(7):867-903
pubmed: 11934711
Science. 2006 Jul 28;313(5786):504-7
pubmed: 16873662
J Am Med Inform Assoc. 2018 Oct 1;25(10):1273
pubmed: 30312446
Am J Respir Crit Care Med. 2019 Jun 1;199(11):1368-1376
pubmed: 30576221
N Engl J Med. 2020 May 21;382(21):1973-1975
pubmed: 32202721
Anesthesiology. 2020 Jun;132(6):1317-1332
pubmed: 32195705
Ann Acad Med Singap. 2020 Jan;49(3):108-118
pubmed: 32200400
Lancet. 2020 Feb 15;395(10223):497-506
pubmed: 31986264
Crit Care Med. 2018 Apr;46(4):547-553
pubmed: 29286945
Chest. 2014 Oct;146(4 Suppl):e145S-55S
pubmed: 25144262
Crit Care Med. 2020 Feb;48(2):210-217
pubmed: 31939789
Lancet. 2020 Jun 6;395(10239):1763-1770
pubmed: 32442528
Lancet Respir Med. 2020 Jun;8(6):e53
pubmed: 32444270
Biometrics. 1988 Sep;44(3):837-45
pubmed: 3203132
Resuscitation. 2017 Dec;121:76-80
pubmed: 29032298