An early prediction of Acute Respiratory Distress Syndrome (ARDS) in patients with severe trauma based on clinical data can help nurse clinicians screen high-risk groups that would develop ARDS. To achieve this purpose, machine learning methods were adopted and tested. This retrospective cohort study was performed on the data of the severe trauma patients admitted to the ICU of the affiliated hospital of Zunyi Medical University from September 2021 to November 2022. The required data for construing the prediction models was collected from medical records of these patients. Univariate logistic regression was used first to achieve the purpose of reducing the data dimension. Then, twelve machine learning methods classified into four categories, which were neural network, logistic regression (LR), decision tree (DT) and support vector machine (SVM), were adopted in the early prediction of ARDS in patients with severe trauma. Internal cross-validation was conducted in 50 numerical experiments, and in each test, a training set consisted of 80% of the samples that were randomly selected, and the remaining 20% of the samples were in a validation data set. In the internal validation, 550 patients were involved. 250 cases developed ARDS within one week and 300 cases had no ARDS. Machine learning methods were also tested in external validation with 100 trauma patients who developed ARDS within one week and 101 controls. Based on the test results, the optimal machine learning model was investigated. Then, significant predictors associated with the development of ARDS were further examined with the help of SHAP (SHapley Additive exPlanations) analysis and causal inference. Tree models showed high discrimination in both internal and external validation. The model trained by the AdaBoost + DT (decision tree) algorithm had the most balanced results, and showed that AUC (area under the curve), accuracy, precision, specificity and sensitivity were 0.915, 0.833, 0.799, 0.823, 0.845, respectively, in the cross validation, and 0.851, 0.751, 0.734, 0.710, 0.793, respectively, in the external validation. The findings indicated that Glasgow Coma Scale (GCS), Injury Severity Score (ISS), Total protein (TP), and blood glucose (Glu) were the most important relevant factors for the ARDS prediction. The use of collected clinical data to predict the development of ARDS in patients with severe trauma has a certain value. Tree models have the best discrimination power in predicting ARDS after major trauma. Essential predictors at least contain GCS, ISS, TP, and Glu.

Jingdan Zhang, Xu Luo. Early Prediction of Acute Respiratory Distress Syndrome in Patients with Severe Trauma Based on Machine Learning. Academic Journal of Medicine & Health Sciences (2025), Vol. 6, Issue 8: 24-38. https://doi.org/10.25236/AJMHS.2025.060804.

[1] J. M. Jiménez Vizuete, J. M. Pérez Valdivieso, S. R. Navarro, et al., “Resuscitation damage control in the patient with severe trauma,” Revista Espanola de Anestesiologia y Reanimacion, vol. 59, no. 1, pp. 31-42, 2012. https://pubmed.ncbi.nlm.nih.gov/22429634/.

[2] X. Chun-Hong, J. Wan, H. Liu, et al., “Red blood cell distribution width is an independent risk factor in the prediction of acute respiratory distress syndrome after severe burns,” Burns, vol. 45, no. 5, pp. 1158-1163, 2019. https://doi.org/10.1016/j.burns.2019.01.006.

[3] G. Volpin, M. Cohen, M. Assas, et al.,“Ytokine levels (IL-4, IL-6, IL-8 and TGF beta) as potential biomarkers of systemic inflammatory response in trauma patients,” International Orthopaedics, vol. 38, no. 6, pp. 1303-1309, 2014. https://doi.org/10.1007/s00264-013-2261-2.

[4] A. Sauaia, F. A. Moore and E. E. Moore, “Postinjury inflammation and organ dysfunction,” Critical Care Clinics, vol. 33, no. 1, pp. 167-191, 2017. https://pubmed.ncbi.nlm.nih.gov/27894496/.

[5] L. D. Hudson, J. A. Milberg, D. Anardi, and R. J. Maunder, “Clinical risks for development of the acute respiratory distress syndrome,” American Journal of Respiratory and Critical Care Medicine, vol. 151, no. 2, pp. 293-301, 1995. https://doi.org/10.1164/ajrccm.151.2.7842182.

[6] X. Li, J. Yu, W. Huang, et al., “Risk factors analysis of acute respiratory distress syndrome in intensive care unit traumatic patients,” Chinese Critical Care Medicine, vol. 30, no. 10, pp. 978-982, 2018. http: //med. wanfangdata.com.cn/Paper/Detail?id=PeriodicalPaper_zgwzbjjyx201810015&dbid =WF_QK.

[7] I.Bendib, A. Beldi-Ferchiou, F. Schlemmer, et al., “Alveolar compartmentalization of inflammatory and immune cell biomarkers in pneumonia-related ARDS,” Critical Care, vol.25, no.1, Article number: 23 (2021), 2021. https://doi.org/10.1186/s13054-020-03427-y.

[8] E. Y. Killien, B. Mills, R. S. Watson, et al., “Morbidity and mortality among critically injured children with acute respiratory distress syndrome,” Critical Care Medicine, vol. 47, no. 2, pp. e112-e119, 2019. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6336518/.

[9] J. M. Komisarow, F. Chen, M. S. Vavilala, et al., “Epidemiology and outcomes of acute respiratory distress syndrome following isolated severe traumatic brain injury,” Journal of Intensive Care Medicine, vol. 37, no. 19, Article number: 088506662097200, 2020. https: //doi. org/10. 1177%2F0885066620972001.

[10] C. M. Hendrickson, B. M. Howard, L. Z. Kornblith, et al., “The acute respiratory distress syndrome following isolated severe traumatic brain injury,” Journal of Trauma and Acute Care Surgery, vol. 80, no. 6, pp. 989–997, 2016. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5851280/.

[11] P. R. Miller, M. A. Croce, P. D. Kilgo, et al., “ Acute respiratory distress syndrome in blunt trauma: Identification of independent risk factors,” The American Surgeon, vol. 68, no. 10, pp. 845-850, 2002. https: //www. proquest. com/openview/b83cb8a3252b03812897e79be874c64d/1?pq-origsite =gscholar&cbl=49079.

[12] A. Daurat, I. Millet, J. P. Roustan, et al., “Thoracic trauma severity score on admission allows to determine the risk of delayed ARDS in trauma patients with pulmonary contusion,” Injury, vol. 41, no. 7, pp. 147-153, 2016. https://doi.org/10.1016/j.injury.2015.08.031.

[13] L. B. Ware, Z. Zhao, T. Koyama, et al., “Derivation and validation of a two-biomarker panel for diagnosis of ARDS in patients with severe traumatic injuries,” Trauma Surgery & Acute Care Open, vol. 2, no. 1, Article number: e000121, 2017. https://tsaco.bmj.com/content/2/1/e000121.abstract.

[14] P. Navarrete-Navarro, R. Rivera-Fernández, M. D. Rincón-Ferrari, et al., “Early markers of acute respiratory distress syndrome development in severe trauma patient,” Journal of Critical Care, vol. 21, no. 3, pp. 253-258, 2006. https://doi.org/10.1016/j.jcrc.2005.12.012.

[15] D. S. Heffernan, L. A. Dossett, M. A. Lightfoot, et al., “Gender and ARDS in critically injured adults: A prospective study,” Journal of Trauma, vol. 71, no. 4, pp. 878-885, 2011. https://www. ncbi. nlm. nih.gov/pmc/articles/PMC3201740/.

[16] H. Silverboard, I. Aisiku, G. S. Martin, et al., “The role of acute blood transfusion in the development of acute respiratory distress syndrome in patients with severe trauma,” The Journal of Trauma, vol. 59, no. 3, pp. 717-723, 2005. https://journals.lww.com/jtrauma/Abstract/2005/09000/ The_Role_of_Acute_Blood_Transfusion_in_the.24.aspx.

[17] C. J. Tignanelli, M. R. Hemmila, M. Rogers, et al., “Nationwide cohort study of independent risk factors for acute respiratory distress syndrome after trauma,” Trauma Surgery & Acute Care Open, vol. 4, no.1, 2019. http://dx.doi.org/10.1136/tsaco-2018-000249.

[18] P. Daher, P. G. Teixeira, T. B. Coopwood, et al., “Mild to moderate to severe: What drives the severity of ARDS in trauma patients?,” The American Surgeon, vol. 84, no. 6, pp. 808-812, 2018. https://doi.org/10.1177/000313481808400623.

[19] J. P. Reilly, S. Bellamy, M. Shashaty, et al., “Heterogeneous phenotypes of acute respiratory distress syndrome after major trauma,” Annals of the American Thoracic Society, vol. 11, no. 5, pp. 728-736, 2014. https://doi.org/10.1513/AnnalsATS.201308-280OC.

[20] Y. Qin, P. Wang, X. Xu, and Z. Li, “Analysis of risk factors of acute respiratory distress syndrome secondary to severe multiple trauma,” Chinese Critical Care Medicine, vol. 33, no. 3, pp. 299-304, 2021. http://med.wanfangdata.com.cn/Paper/Detail?id=PeriodicalPaper_zgwzbjjyx202103009.

[21] J. Wu, L. Sheng, S. Wang, et al., “The impact of clinical risk factors in the conversion from acute lung injury to acute respiratory distress syndrome in severe multiple trauma patients,” Journal of International Medical Research, vol. 36, no. 3, pp. 579-586, 2008. https: //doi. org/10. 1177%2F147323000803600325.

[22] T. R. Watkins, A. B. Nathens, C. R. Cooke, et al., “Acute respiratory distress syndrome after trauma: Development and validation of a predictive model,” Critical Care Medicine, vol. 40, no. 8, pp. 2295-2303, 2012. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3400931/.

[23] M. P. O’Leary, J. A. Keeley, A. Yule, et al., “Clinical predictors of early acute respiratory distress syndrome in trauma patients,” American Journal of Surgery, vol. 212, no. 6, pp. 1096-1100, 2016. https://doi.org/10.1016/j.amjsurg.2016.09.008.

[24] D. R. Amory, R. V. Burke, L. Joanna, et al., “Pediatric trauma-associated acute respiratory distress syndrome: Incidence, risk factors, and outcomes,” Journal of Pediatric Surgery, vol. 54, no. 7, pp. 1405-1410, 2019. https://doi.org/10.1016/j.jpedsurg.2018.07.005.

[25] P. Navarrete-Navarro, A. Rodriguez, N. Reynolds, et al., “Acute respiratory distress syndrome among trauma patients: trends in ICU mortality, risk factors, complications and resource utilization,” Intensive Care Medicine, vol. 27, no. 7, pp. 1133-1140, 2001. https://doi.org/10.1007/s001340100955.

[26] G. Kasotakis, B. Stanfield, K. Haines, et al., “Acute respiratory distress syndrome (ARDS) after trauma: Improving incidence, but increasing mortality,” Journal of Critical Care, vol. 64, pp. 213-218, 2021. https://doi.org/10.1016/j.jcrc.2021.05.003.

[27] X. Zhao, J. Wu, X. He, et al., “Risk factors of mortality in road traffic injury patients with acute respiratory distress syndrome,” Chinese Medical Journal, vol. 121, pp. 968-972, 2008. https://www. doc88.com/p-11747100309958.html.

[28] J. Wu, L. Sheng, Y. Ma, et al., “The analysis of risk factors of impacting mortality rate in severe multiple trauma patients with posttraumatic acute respiratory distress syndrome,” American Journal of Emergency Medicine, vol. 26, no. 4, pp. 419-424, 2008. https://doi.org/10.1016/j.ajem.2007.06.032.

[29] P. Navarrete-Navarro, M. Ruiz-Bailén, R. Rivera-Fernández, et al., “Acute respiratory distress syndrome in trauma patients: ICU mortality and prediction factors,” Intensive Care Medicine, vol. 26, no. 11, pp. 1624-1629, 2000. https://doi.org/10.1007/s001340000683.

[30] E. Apostolova, A. Uppal, J. E. Galarraga, et al., “Towards reliable ARDS clinical decision support: ARDS patient analytics with free-text and structured EMR data,” AMIA Annu Symp Proc, vol. 2019, pp. 228-237, 2020. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7153087/#.

[31] S. Le, E. Pellegrini, A. Green-Saxena, et al., “Supervised machine learning for the early prediction of acute respiratory distress syndrome (ARDS),” Journal of Critical Care, vol. 60, pp. 96-102, 2020. https://doi.org/10.1016/j.jcrc.2020.07.019.

[32] W. Xu, N. Sun, H. Gao, et al., “Risk factors analysis of COVID-19 patients with ARDS and prediction based on machine learning,” Scientific Reports, vol. 11, no. 1, Article number: 2933, 2021. https://doi.org/10.1038/s41598-021-82492-x.

[33] L. Singhal, Y. Garg, P. Yang, et al.,“eARDS: A multi-center validation of an interpretable machine learning algorithm of early onset Acute Respiratory Distress Syndrome (ARDS) among critically ill adults with COVID-19,” PLoS ONE, vol. 16, no. 9, 2021. https://doi.org/10.1371/journal.pone.0257056.

[34] Y. Fei, K. Gao and W. Li, “Prediction and evaluation of the severity of acute respiratory distress syndrome following severe acute pancreatitis using an artificial neural network algorithm model,” HPB (Oxford), vol. 21, no. 7, pp. 891-897, 2019. https://doi.org/10.1016/j.hpb.2018.11.009.

[35] S. Röhrich, J. Hofmanninger, L. Negrin, et al.,“Radiomics score predicts acute respiratory distress syndrome based on the initial CT scan after trauma,” European Radiology, vol. 31, no. 8, pp. 5443-5453, 2021. https://doi.org/10.1007/s00330-021-07995-7.

[36] A. Esteva, A. Robicquet, B. Ramsundar, V. Kuleshov, et al., “A guide to deep learning in healthcare,” Nature Medicine, vol. 25, pp. 24-29, 2019. https://doi.org/10.1038/s41591-018-0316-z.

[37] G. Bonaccorso, Machine learning algorithms, Packt Publishing Ltd, Birmingham, UK, 2017. https://xueshu.studiodahu.com/books?hl=zh-CN&lr=&id=_-ZDDwAAQBAJ&oi=fnd&pg=PP1&dq= Machine+learning+algorithms&ots=epkBC4IG3B&sig=UOfHrF8m1pTrfK92tRPwBeZMzQ4.

[38] X. Dong, Z. Yu, W. Cao, et al., “A survey on ensemble learning,” Frontiers of Computer Science, vol. 14, pp. 241-258, 2020. https://doi.org/10.1007/s11704-019-8208-z.

[39] S. M. Lundberg and S. I. Lee, “A unified approach to interpreting model prediction,” in Proceedings of the 31st Conference on Neural Information Processing System, pp. 4768–4777, ACM, Long Bench, USA, December 2017. https://dl.acm.org/doi/10.5555/3295222.3295230.

[40] Z. Zhongheng, J. Peng, F. Menglin, et al., “Causal inference with marginal structural modeling for longitudinal data in laparoscopic surgery: A technical note,” Laparoscopic, Endoscopic and Robotic Surgery, vol. 5, no. 4, pp. 146-152, 2022. https://doi.org/10.1016/j.lers.2022.10.002.

[41] P. Judea, G. Madelyn and P. J. Nicholas, Causal Inference in Statistics - A Primer, Wiley press, Hoboken, USA, 2016. https://deepblue.lib.umich.edu/bitstream/handle/2027. 42/154956/insr12369.pdf? sequence=1.

[42] V. M. Ranieri, G. D. Rubenfeld, B. T. Thompson, et al., “Acute respiratory distress syndrome: the Berlin Definition,” JAMA, vol. 307, no. 23, pp. 2526-2533, 2012. https://jamanetwork.com/journals/ jama/article-abstract/1160659.

[43] Microsoft, DoWhy, available at: https://github.com/py-why/dowhy, accessed at 3rd January, 2023.

[44] P. F. Michael and A. P. Michael, “Wilcoxon–Mann–Whitney or t-test? On assumptions for hypothesis tests and multiple interpretations of decision rules,” Statistics Surveys, vol. 4, pp. 1-39, 2010.https://pubmed.ncbi.nlm.nih.gov/20414472/.

[45] F. Y. Hsieh, D. A. Bloch and M. D. Larsen, “A simple method of sample size calculation for linear and logistic regression, ” Statistics in Medicine, vol. 17, no. 14, pp. 1623-1634, 1998. https://doi. org/10.1002/(SICI)1097-0258(19980730)17:14<1623::AID-SIM871>3.0.CO;2-S.

[46] S. J. Raudys and A. K. Jain, “Small sample size effects in statistical pattern recognition: recommendations for practitioners,” IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 13, no. 3, pp. 252-264, 1991. https://ieeexplore.ieee.org/document/75512/citations#citations.

[47] J. S. Leboeuf, F. Leblanc and M. Marchand, “Decision trees as partitioning machines to characterize their generalization properties,” 10.48550/arXiv.2010.07374, 2020. https: //arxiv. org/ abs/2010.07374.

[48] Z. Zhang, Y. Song, H. H. Cui, et al., “Topological analysis and Gaussian decision tree: Effective representation and classification of biosignals of small sample size,” IEEE Transactions on Biomedical Engineering, vol. 64, no. 9, pp. 2288 - 2299, 2017. https://ieeexplore. ieee.org/abstract/ document/7765109.

[49] C. M. Hendrickson, J. Abbott and H. Zhuo, “Higher mini-BAL total protein concentration in early ARDS predicts faster resolution of lung injury measured by more ventilator-free days,” American Journal of Physiology-Lung Cellular and Molecular Physiology, vol. 312, no. 5, pp. L579-L585, 2017. https://doi.org/10.1152/ajplung.00381.2016.

[50] L. Luo, Clinical progress of acute respiratory distress syndrome, People's Medical Publishing House, Beijing, China, 2017. https://www.pmphmall.com/gdsdetail/604843-279471.