Development and validation of an explainable machine learning-based model for predicting mortality risk in polytrauma patients_Journal of Biomedical Engineering

Authors：

JIANG Yupeng ^1,2 , MAO Jinlong ³ , LI Lin ² , LIANG Chunyang ³ , KANG Hongjun ⁴ , JIA Zhilong ² , LIU Xiaoli ² ,  ZHANG Zhengbo ²

1. Medical School of Chinese PLA, Beijing 100853, P. R. China;
2. Medical Artificial Intelligence Research Center, Medical Innovation Research Department, Chinese PLA General Hospital, Beijing 100853, P. R. China;
3. Department of Neurosurgery, First Medical Center, Chinese PLA General Hospital, Beijing 100853, P. R. China;
4. Department of Critical Care Medicine, First Medical Center, Chinese PLA General Hospital, Beijing 100853, P. R. China;

Corresponding?author：

ZHANG Zhengbo, Email: zhengbozhang@126.com

Keywords：

Polytrauma; Prediction model; Machine learning; SHAP interpretability

DOI：

10.7507/1001-5515.202605025

Video：

Export PDF Favorites Scan Get Citation

Abstract Full text Figures/Tables Video References Cited by

Polytrauma is commonly defined as multisystem trauma involving at least two body regions with an Abbreviated Injury Scale (AIS) score ≥ 3, characterized by complex pathophysiological interactions and extreme clinical heterogeneity. This study retrospectively analyzed data from 171 274 polytrauma patients in the National Trauma Data Bank (NTDB) from 2018 to 2021. Key features were selected using the eXtreme Gradient Boosting (XGBoost) algorithm. Prediction models were then developed using XGBoost, logistic regression, random forest, naive Bayes, and support vector machine, respectively. Model performance was evaluated by receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). Subgroup standardized mortality ratio (SMR) validation and global and local interpretability analysis based on SHapley Additive exPlanations (SHAP) theory were also conducted. The results demonstrated that the XGBoost model incorporating Glasgow Coma Scale (GCS), Injury Severity Score (ISS), massive transfusion within 4 hours, respiratory support, unplanned intubation, prehospital care-limiting directives, unplanned intensive care unit (ICU) admission, sepsis, ventilator-associated pneumonia, and liver cirrhosis achieved optimal performance, with an area under the ROC curve (AUC) of 0.90. Its calibration curve closely aligned with the ideal diagonal, and decision curve analysis indicated the highest clinical net benefit across a wide range of threshold probabilities. Subgroup SMRs ranged from 1.07 to 1.18, all close to 1.0. In conclusion, the developed in-hospital mortality risk prediction model for polytrauma patients exhibits favorable discrimination and clinical net benefit, and can assist clinicians in identifying high-risk populations while providing evidence to support individualized treatment decisions.

Citation： JIANG Yupeng, MAO Jinlong, LI Lin, LIANG Chunyang, KANG Hongjun, JIA Zhilong, LIU Xiaoli, ZHANG Zhengbo. Development and validation of an explainable machine learning-based model for predicting mortality risk in polytrauma patients. Journal of Biomedical Engineering, 2026, 43(4): 769-776. doi: 10.7507/1001-5515.202605025 Copy

1.	張連陽, 黃顯凱, 姚元章, 等. 多發傷病歷與診斷: 專家共識意見(2013版)[J]. 創傷外科雜志, 2014, 16(2): 192-193.
2.	李陽, 李輝, 陳駕君, 等. 多發傷病歷與診斷: 專家共識(2023版)[J]. 創傷外科雜志, 2023, 25(8): 561-568.
3.	Spahn D R, Bouillon B, Cerny V, et al. The European guideline on management of major bleeding and coagulopathy following trauma: fifth edition[J]. Crit Care, 2019, 23(1): 98.
4.	徐新, 尚峰, 曲鑫, 等. 創傷性腦損傷誘發的凝血功能障礙研究進展[J]. 中國現代神經疾病雜志, 2020, 20(8): 674-681.
5.	段紹斌, 馬磊, 董楊帆, 等. 創傷性休克復蘇后多器官功能障礙綜合征發生危險因素分析[J]. 中國全科醫學, 2010, 13(21): 2367-2369.
6.	王燕, 婁曼曼, 彭海琳, 等. 急診綠色通道創傷管理對急診多發傷患者的影響[J]. 國際護理學雜志, 2024, 43(5): 844-849.
7.	Topol E J. High-performance medicine: the convergence of human and artificial intelligence[J]. Nat Med, 2019, 25(1): 44-56.
8.	張鵬, 杜哲, 劉中砥, 等. 我國醫生對嚴重創傷評估能力的多中心調查[J]. 中華急診醫學雜志, 2021, 30(5): 533-536.
9.	苗振軍, 張登奎, 梁亞鵬, 等. 多發傷患者院內死亡獨立危險因素分析及預測模型的構建與驗證[J]. 中華創傷雜志, 2023, 39(7): 643-651.
10.	白麗爽, 王興義, 楊立山. 多發傷患者預后列線圖模型的構建和研究[J]. 中華急診醫學雜志, 2023, 32(4): 540-545.
11.	Iba T, Helms J, Neal M D, et al. Mechanisms and management of the coagulopathy of trauma and sepsis: trauma-induced coagulopathy, sepsis-induced coagulopathy, and disseminated intravascular coagulation[J]. J Thromb Haemost, 2023, 21(12): 3360-3370.
12.	袁強, 虞劍, 陳建, 等. 顱腦創傷急性期凝血功能障礙診治專家共識(2024版)[J]. 中華創傷雜志, 2024, 40(4): 310-322.
13.	王奕權, 田思佳, 牛升梅, 等. 基于機器學習構建創傷出血死亡風險預測模型研究[J]. 中華急診醫學雜志, 2025, 34(11): 1574-1578.
14.	孟祥福, 田友發, 張霄雁. 基于LightGBM模型的肺腺癌免疫相關基因篩選與患者生存率預測[J]. 生物醫學工程學雜志, 2024, 41(1): 70-79.
15.	Khera R, Haimovich J, Hurley N C, et al. Use of machine learning models to predict death after acute myocardial infarction[J]. JAMA Cardiol, 2021, 6(6): 633-641.
16.	Chen T, Guestrin C. XGBoost: a scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco: ACM, 2016: 785-794.
17.	Rajpurkar P, Chen E, Banerjee O, et al. AI in health and medicine[J]. Nat Med, 2022, 28: 31-38.
18.	齊戈堯, 徐進, 金志超. 基于循環神經網絡模型的創傷重癥患者臨床結局的動態預測[J]. 海軍軍醫大學學報, 2024, 45(10): 1241-1249.
19.	李肅義, 唐世杰, 李鳳, 等. 基于深度學習的生物醫學數據分析進展[J]. 生物醫學工程學雜志, 2020, 37(2): 349-357.
20.	Ciobanu-Caraus O, Aicher A, Kernbach J M, et al. A critical moment in machine learning in medicine: on reproducible and interpretable learning[J]. Acta Neurochir, 2024, 166: 14.
21.	Liu Z, Huang Y, Li L, et al. Development and clinical application of an automated machine learning-based delirium risk prediction model for emergency polytrauma patients[J]. Front Physiol, 2025, 16: 1629329.
22.	Xu Y, Huang C. Artificial intelligence for traumatic brain injury imaging: a translational review from algorithm development to clinical implementation[J]. Front Neurol, 2026, 17: 1798780.
23.	Reddy S. Explainability and artificial intelligence in medicine[J]. Lancet Digit Health, 2022, 4(4): e214-e215.
24.	Ghassemi M, Oakden-Rayner L, Beam A L. The false hope of current approaches to explainable artificial intelligence in health care[J]. Lancet Digit Health, 2021, 3(11): e745-e750.
25.	Gunning D, Stefik M, Choi J, et al. XAI-explainable artificial intelligence[J]. Sci Robot, 2019, 4(37): eaay7120.
26.	Turky M, Samaranayake L, Osathanon T, et al. Explainable artificial intelligence in endodontics: concepts, clinical applications, and future directions[J]. J Dent, 2026, 171: 106733.
27.	劉冰, 賈國洪, 卜曉沛, 等. 反向休克指數與格拉斯哥昏迷評分乘積指數預測創傷患者死亡價值的Meta分析[J]. 中華創傷雜志, 2025, 41(11): 1094-1102.
28.	Han J, Yoon S Y, Seok J, et al. Predicting 30-day mortality in severely injured elderly patients with trauma in Korea using machine learning algorithms: a retrospective study[J]. J Trauma Inj, 2024, 37: 201-208.
29.	蔣宇鵬, 毛錦龍, 李林, 等. 創傷性腦損傷患者院內機械通氣危險因素預測模型構建與效能評估[J]. 中華創傷雜志, 2026, 42(5): 553-564.
30.	Niessen P D M, Krijnen P, Leijdesdorff H A, et al. Interpreting traumatic brain injury severity: analysis of the correlation between Glasgow coma scale and abbreviated injury scale[J]. Eur J Trauma Emerg Surg, 2025, 51(1): 239.
31.	Lin V S, Sun E, Yau S, et al. Definitions of massive transfusion in adults with critical bleeding: a systematic review[J]. Crit Care, 2023, 27(1): 265.
32.	Sigle M, Boos M, Weiss T, et al. AI-enabled forecasting of prehospital transfusion needs in patients with trauma: a multinational, registry-based, retrospective, machine learning development and validation study[J]. Lancet Digit Health, 2026, 8(1): 100945.
33.	王保榮, 鄒圣強, 周田田. 綜合 ICU 多發傷機械通氣患者 VAP 危險因素分析[J]. 中華現代護理雜志, 2017, 23(28): 3570-3575.
34.	Pape H C, Lefering R, Butcher N, et al. The definition of polytrauma revisited: an international consensus process and proposal of the new ‘Berlin definition’[J]. J Trauma Acute Care Surg, 2014, 77(5): 780-786.
35.	Liu D, Huang S Y, Sun J H, et al. Sepsis-induced immunosuppression: mechanisms, diagnosis and current treatment options[J]. Mil Med Res, 2022, 9(1): 56.
36.	劉雙慶, 祝筱梅, 姚詠明. 嚴重燒創傷膿毒癥免疫功能障礙的診療策略[J]. 中華創傷雜志, 2025, 41(5): 433-439.
37.	Dickmann P, Claus R A, Bauer M. Sepsis after trauma-evolving paradigms in stress biology and host response failure[J]. Burns Trauma, 2025, 13: tkaf014.
38.	Serrano E, Liu P, Nwabuo A I, et al. The effect of cirrhosis on trauma outcomes: a systematic review and meta-analysis[J]. J Trauma Acute Care Surg, 2020, 88(4): 536-545.
39.	李偉, 李書文, 沈鋒, 等. 術后計劃與非計劃轉入重癥監護病房患者的預后對比分析[J]. 中華危重病急救醫學, 2023, 35(7): 746-751.
40.	陳千益, 胡曉曼, 曹磊, 等. 早期機械通氣對重型顱腦損傷患者炎性因子表達及預后的影響[J]. 中華危重病急救醫學, 2021, 33(6): 727-730.
41.	梁英健, 馬曉春. 如何正確理解呼吸機相關性肺炎的診斷[J]. 中華醫學雜志, 2014, 94(5): 329-330.
42.	Zhang Y. The “survival of the fittest” bias: how excluding prehospital deaths distorts the impact of transport time in rural trauma[J]. Scand J Trauma Resusc Emerg Med, 2026, 34(1): 38.
43.	郭振霞, 王世垚, 劉亞濤, 等. 創傷登記系統研究進展[J]. 中華創傷雜志, 2022, 38(4): 374-379.
44.	殷菲, 劉云, 沈勰. 嚴重多發傷患者預后的影響因素研究及列線圖模型的建立和優化[J]. 中國全科醫學, 2022, 25(20): 2498-2506.
45.	周佳圻, 胡雨峰, 康洋波, 等. 老年嚴重創傷患者早期創傷性凝血病的危險因素分析[J]. 中華創傷雜志, 2022, 38(1): 61-66.

1. 張連陽, 黃顯凱, 姚元章, 等. 多發傷病歷與診斷: 專家共識意見(2013版)[J]. 創傷外科雜志, 2014, 16(2): 192-193.
2. 李陽, 李輝, 陳駕君, 等. 多發傷病歷與診斷: 專家共識(2023版)[J]. 創傷外科雜志, 2023, 25(8): 561-568.
3. Spahn D R, Bouillon B, Cerny V, et al. The European guideline on management of major bleeding and coagulopathy following trauma: fifth edition[J]. Crit Care, 2019, 23(1): 98.
4. 徐新, 尚峰, 曲鑫, 等. 創傷性腦損傷誘發的凝血功能障礙研究進展[J]. 中國現代神經疾病雜志, 2020, 20(8): 674-681.
5. 段紹斌, 馬磊, 董楊帆, 等. 創傷性休克復蘇后多器官功能障礙綜合征發生危險因素分析[J]. 中國全科醫學, 2010, 13(21): 2367-2369.
6. 王燕, 婁曼曼, 彭海琳, 等. 急診綠色通道創傷管理對急診多發傷患者的影響[J]. 國際護理學雜志, 2024, 43(5): 844-849.
7. Topol E J. High-performance medicine: the convergence of human and artificial intelligence[J]. Nat Med, 2019, 25(1): 44-56.
8. 張鵬, 杜哲, 劉中砥, 等. 我國醫生對嚴重創傷評估能力的多中心調查[J]. 中華急診醫學雜志, 2021, 30(5): 533-536.
9. 苗振軍, 張登奎, 梁亞鵬, 等. 多發傷患者院內死亡獨立危險因素分析及預測模型的構建與驗證[J]. 中華創傷雜志, 2023, 39(7): 643-651.
10. 白麗爽, 王興義, 楊立山. 多發傷患者預后列線圖模型的構建和研究[J]. 中華急診醫學雜志, 2023, 32(4): 540-545.
11. Iba T, Helms J, Neal M D, et al. Mechanisms and management of the coagulopathy of trauma and sepsis: trauma-induced coagulopathy, sepsis-induced coagulopathy, and disseminated intravascular coagulation[J]. J Thromb Haemost, 2023, 21(12): 3360-3370.
12. 袁強, 虞劍, 陳建, 等. 顱腦創傷急性期凝血功能障礙診治專家共識(2024版)[J]. 中華創傷雜志, 2024, 40(4): 310-322.
13. 王奕權, 田思佳, 牛升梅, 等. 基于機器學習構建創傷出血死亡風險預測模型研究[J]. 中華急診醫學雜志, 2025, 34(11): 1574-1578.
14. 孟祥福, 田友發, 張霄雁. 基于LightGBM模型的肺腺癌免疫相關基因篩選與患者生存率預測[J]. 生物醫學工程學雜志, 2024, 41(1): 70-79.
15. Khera R, Haimovich J, Hurley N C, et al. Use of machine learning models to predict death after acute myocardial infarction[J]. JAMA Cardiol, 2021, 6(6): 633-641.
16. Chen T, Guestrin C. XGBoost: a scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco: ACM, 2016: 785-794.
17. Rajpurkar P, Chen E, Banerjee O, et al. AI in health and medicine[J]. Nat Med, 2022, 28: 31-38.
18. 齊戈堯, 徐進, 金志超. 基于循環神經網絡模型的創傷重癥患者臨床結局的動態預測[J]. 海軍軍醫大學學報, 2024, 45(10): 1241-1249.
19. 李肅義, 唐世杰, 李鳳, 等. 基于深度學習的生物醫學數據分析進展[J]. 生物醫學工程學雜志, 2020, 37(2): 349-357.
20. Ciobanu-Caraus O, Aicher A, Kernbach J M, et al. A critical moment in machine learning in medicine: on reproducible and interpretable learning[J]. Acta Neurochir, 2024, 166: 14.
21. Liu Z, Huang Y, Li L, et al. Development and clinical application of an automated machine learning-based delirium risk prediction model for emergency polytrauma patients[J]. Front Physiol, 2025, 16: 1629329.
22. Xu Y, Huang C. Artificial intelligence for traumatic brain injury imaging: a translational review from algorithm development to clinical implementation[J]. Front Neurol, 2026, 17: 1798780.
23. Reddy S. Explainability and artificial intelligence in medicine[J]. Lancet Digit Health, 2022, 4(4): e214-e215.
24. Ghassemi M, Oakden-Rayner L, Beam A L. The false hope of current approaches to explainable artificial intelligence in health care[J]. Lancet Digit Health, 2021, 3(11): e745-e750.
25. Gunning D, Stefik M, Choi J, et al. XAI-explainable artificial intelligence[J]. Sci Robot, 2019, 4(37): eaay7120.
26. Turky M, Samaranayake L, Osathanon T, et al. Explainable artificial intelligence in endodontics: concepts, clinical applications, and future directions[J]. J Dent, 2026, 171: 106733.
27. 劉冰, 賈國洪, 卜曉沛, 等. 反向休克指數與格拉斯哥昏迷評分乘積指數預測創傷患者死亡價值的Meta分析[J]. 中華創傷雜志, 2025, 41(11): 1094-1102.
28. Han J, Yoon S Y, Seok J, et al. Predicting 30-day mortality in severely injured elderly patients with trauma in Korea using machine learning algorithms: a retrospective study[J]. J Trauma Inj, 2024, 37: 201-208.
29. 蔣宇鵬, 毛錦龍, 李林, 等. 創傷性腦損傷患者院內機械通氣危險因素預測模型構建與效能評估[J]. 中華創傷雜志, 2026, 42(5): 553-564.
30. Niessen P D M, Krijnen P, Leijdesdorff H A, et al. Interpreting traumatic brain injury severity: analysis of the correlation between Glasgow coma scale and abbreviated injury scale[J]. Eur J Trauma Emerg Surg, 2025, 51(1): 239.
31. Lin V S, Sun E, Yau S, et al. Definitions of massive transfusion in adults with critical bleeding: a systematic review[J]. Crit Care, 2023, 27(1): 265.
32. Sigle M, Boos M, Weiss T, et al. AI-enabled forecasting of prehospital transfusion needs in patients with trauma: a multinational, registry-based, retrospective, machine learning development and validation study[J]. Lancet Digit Health, 2026, 8(1): 100945.
33. 王保榮, 鄒圣強, 周田田. 綜合 ICU 多發傷機械通氣患者 VAP 危險因素分析[J]. 中華現代護理雜志, 2017, 23(28): 3570-3575.
34. Pape H C, Lefering R, Butcher N, et al. The definition of polytrauma revisited: an international consensus process and proposal of the new ‘Berlin definition’[J]. J Trauma Acute Care Surg, 2014, 77(5): 780-786.
35. Liu D, Huang S Y, Sun J H, et al. Sepsis-induced immunosuppression: mechanisms, diagnosis and current treatment options[J]. Mil Med Res, 2022, 9(1): 56.
36. 劉雙慶, 祝筱梅, 姚詠明. 嚴重燒創傷膿毒癥免疫功能障礙的診療策略[J]. 中華創傷雜志, 2025, 41(5): 433-439.
37. Dickmann P, Claus R A, Bauer M. Sepsis after trauma-evolving paradigms in stress biology and host response failure[J]. Burns Trauma, 2025, 13: tkaf014.
38. Serrano E, Liu P, Nwabuo A I, et al. The effect of cirrhosis on trauma outcomes: a systematic review and meta-analysis[J]. J Trauma Acute Care Surg, 2020, 88(4): 536-545.
39. 李偉, 李書文, 沈鋒, 等. 術后計劃與非計劃轉入重癥監護病房患者的預后對比分析[J]. 中華危重病急救醫學, 2023, 35(7): 746-751.
40. 陳千益, 胡曉曼, 曹磊, 等. 早期機械通氣對重型顱腦損傷患者炎性因子表達及預后的影響[J]. 中華危重病急救醫學, 2021, 33(6): 727-730.
41. 梁英健, 馬曉春. 如何正確理解呼吸機相關性肺炎的診斷[J]. 中華醫學雜志, 2014, 94(5): 329-330.
42. Zhang Y. The “survival of the fittest” bias: how excluding prehospital deaths distorts the impact of transport time in rural trauma[J]. Scand J Trauma Resusc Emerg Med, 2026, 34(1): 38.
43. 郭振霞, 王世垚, 劉亞濤, 等. 創傷登記系統研究進展[J]. 中華創傷雜志, 2022, 38(4): 374-379.
44. 殷菲, 劉云, 沈勰. 嚴重多發傷患者預后的影響因素研究及列線圖模型的建立和優化[J]. 中國全科醫學, 2022, 25(20): 2498-2506.
45. 周佳圻, 胡雨峰, 康洋波, 等. 老年嚴重創傷患者早期創傷性凝血病的危險因素分析[J]. 中華創傷雜志, 2022, 38(1): 61-66.

Journal of Biomedical Engineering

Development and validation of an explainable machine learning-based model for predicting mortality risk in polytrauma patients

Abstract Full text Figures/Tables Video References Cited by

Previous Article

Next Article

Format

Content