Advancing a machine learning-based decision support tool for pre-hospital assessment of dyspnoea by emergency medical service clinicians: a retrospective observational study.

Publisher: BioMed Central Country of Publication: England NLM ID: 100968543 Publication Model: Electronic Cited Medium: Internet ISSN: 1471-227X (Electronic) Linking ISSN: 1471227X NLM ISO Abbreviation: BMC Emerg Med Subsets: MEDLINE

Original Publication: London : BioMed Central, [2001-

Dyspnea*/diagnosis ; Emergency Medical Services*/methods ; Triage*/methods ; Machine Learning*; Humans ; Retrospective Studies ; Female ; Male ; Sweden ; Aged ; Middle Aged ; Aged, 80 and over ; Adult

Background: In Sweden with about 10 million inhabitants, there are about one million primary ambulance missions every year. Among them, around 10% are assessed by Emergency Medical Service (EMS) clinicians with the primary symptom of dyspnoea. The risk of death among these patients has been reported to be remarkably high, at 11,1% and 13,2%. The aim was to develop a Machine Learning (ML) model to provide support in assessing patients in pre-hospital settings and to compare them with established triage tools.
Methods: This was a retrospective observational study including 6,354 patients who called the Swedish emergency telephone number (112) between January and December 2017. Patients presenting with the main symptom of dyspnoea were included which were recruited from two EMS organisations in Göteborg and Södra Älvsborg. Serious Adverse Event (SAE) was used as outcome, defined as any of the following:1) death within 30 days after call for an ambulance, 2) a final diagnosis defined as time-sensitive, 3) admitted to intensive care unit, or 4) readmission within 72 h and admitted to hospital receiving a final time-sensitive diagnosis. Logistic regression, LASSO logistic regression and gradient boosting were compared to the Rapid Emergency Triage and Treatment System for Adults (RETTS-A) and National Early Warning Score2 (NEWS2) with respect to discrimination and calibration of predictions. Eighty percent (80%) of the data was used for model development and 20% for model validation.
Results: All ML models showed better performance than RETTS-A and NEWS2 with respect to all evaluated performance metrics. The gradient boosting algorithm had the overall best performance, with excellent calibration of the predictions, and consistently showed higher sensitivity to detect SAE than the other methods. The ROC AUC on test data increased from 0.73 (95% CI 0.70-0.76) with RETTS-A to 0.81 (95% CI 0.78-0.84) using gradient boosting.
Conclusions: Among 6,354 ambulance missions caused by patients suffering from dyspnoea, an ML method using gradient boosting demonstrated excellent performance for predicting SAE, with substantial improvement over the more established methods RETTS-A and NEWS2.
(© 2024. The Author(s).)

Declarations. Ethics approval and consent to participate: The study received approval from The Regional Ethical Review Board in Gothenburg, Sweden (Dnr 989–17) and adhered to the ethical standards of the Declaration of Helsinki [46]. Consent for data analysis was obtained from the chief executive officers of the two EMS organisations. In accordance with Swedish law, informed consent was not required for participation due to the disproportionate work effort involved. Also, some patients, couldn’t provide retrospective informed consent due to poor health or death, minimizing the risk of selection bias. Patient confidentiality was upheld by translating identification numbers into unique codes. Consent for publication: Not applicable. Competing interests: The authors declare no competing interests.

PLoS One. 2019 Feb 28;14(2):e0213145. (PMID: 30817792)
Scand J Trauma Resusc Emerg Med. 2020 Jan 29;28(1):7. (PMID: 31996233)
Eur Heart J Acute Cardiovasc Care. 2020 Mar;9(1_suppl):59-81. (PMID: 26315695)
iScience. 2023 Jul 17;26(8):107407. (PMID: 37609632)
Ann Intensive Care. 2018 Aug 13;8(1):82. (PMID: 30105416)
Clin Pract. 2023 Oct 23;13(5):1266-1285. (PMID: 37887090)
Comput Biol Med. 2022 Jan;140:105097. (PMID: 34864304)
JAMA. 2013 Nov 27;310(20):2191-4. (PMID: 24141714)
BMC Emerg Med. 2020 Nov 10;20(1):89. (PMID: 33172409)
J Gen Intern Med. 2020 Jan;35(1):220-227. (PMID: 31677104)
PLoS One. 2019 Dec 13;14(12):e0226518. (PMID: 31834920)
Am J Emerg Med. 2021 Jul;45:105-111. (PMID: 33684866)
Clin Med (Lond). 2022 Nov;22(6):509-513. (PMID: 36427875)
BMC Emerg Med. 2024 Jan 8;24(1):11. (PMID: 38191306)
Sci Rep. 2022 May 3;12(1):7180. (PMID: 35505048)
BMC Med Inform Decis Mak. 2023 Oct 9;23(1):206. (PMID: 37814288)
BMC Emerg Med. 2020 Aug 28;20(1):67. (PMID: 32859155)
Am J Cardiol. 2015 Jan 1;115(1):75-81. (PMID: 25456875)
Nat Mach Intell. 2020 Jan;2(1):56-67. (PMID: 32607472)
Scand J Trauma Resusc Emerg Med. 2018 Oct 29;26(1):91. (PMID: 30373652)
Open Access Emerg Med. 2021 Mar 25;13:107-116. (PMID: 33790664)
BMJ. 2015 Jan 07;350:g7594. (PMID: 25569120)
J Emerg Med. 2011 Jun;40(6):623-8. (PMID: 18930373)
Crit Care. 2019 Feb 22;23(1):64. (PMID: 30795786)
Resuscitation. 2022 Nov;180:99-107. (PMID: 36191809)
Front Cardiovasc Med. 2023 Mar 14;10:1087702. (PMID: 36998977)
Emerg Med Int. 2023 Jun 26;2023:1221704. (PMID: 37404873)
J Am Coll Emerg Physicians Open. 2020 Apr 01;1(3):163-172. (PMID: 33000031)
Medicine (Baltimore). 2021 May 14;100(19):e25813. (PMID: 34106618)
Resusc Plus. 2020 Dec 05;4:100046. (PMID: 34223321)
BMC Emerg Med. 2022 May 23;22(1):89. (PMID: 35606694)
Lakartidningen. 2021 Nov 30;118:. (PMID: 34861043)
Ann Emerg Med. 2014 Sep;64(3):292-8. (PMID: 24746846)
Int Urol Nephrol. 2023 Aug;55(8):2099-2109. (PMID: 36840802)
JMIR AI. 2023 Nov 6;2:e45257. (PMID: 38875543)
Ann Transl Med. 2023 Jan 31;11(2):95. (PMID: 36819553)
Eur Respir J. 2021 Sep 2;58(3):. (PMID: 33653806)
Scand J Trauma Resusc Emerg Med. 2019 Aug 29;27(1):84. (PMID: 31464622)
Clin Infect Dis. 2011 Feb 1;52(3):325-31. (PMID: 21217179)
Clin Med (Lond). 2019 Jan;19(1):94-95. (PMID: 30651271)
BMC Pulm Med. 2019 Dec 3;19(1):233. (PMID: 31795986)

Keywords: Ambulance; Decision support tool; Dyspnoea; Emergency medical services; Machine learning; Prehospital; Serious adverse event

Date Created: 20250105 Date Completed: 20250106 Latest Revision: 20250110

20260130

PMC11702062

10.1186/s12873-024-01166-9

39757181