Skip to main content
SpringerLink
Log in

Early warning model for death of sepsis via length insensitive temporal convolutional network

  • Original Article
  • Published:
Medical & Biological Engineering & Computing Aims and scope Submit manuscript

Abstract

Sepsis is a life-threatening systemic syndrome characterized by various biological, biochemical, and physiological abnormalities. Due to its high mortality, identifying sepsis patients with high risk of in-hospital death early and accurately will help doctors make optimal clinical decisions and reduce the mortality of sepsis patients. In this paper, we propose a length insensitive TCN-based model to predict sepsis patient’s death risk in the future k hours, which is the first work for sepsis death risk early warning model only based on vital signs time series to our best knowledge. Furthermore, we design residual connections between temporal residual blocks to improve the prediction performance and stability especially on short input sequences. We validate and evaluate our model on two freely-available datasets, i.e., MIMIC-IV and eICU, from which 16,520 and 29,620 patients are selected respectively. The experiment results show that our model outperforms LSTM and other machine learning methods, as it has the highest sensitivity and Youden index in almost all cases. Meanwhile, the Youden index of the TCN-based model only slightly decreases by 0.0233 and 0.0307 when the time range of the input sequence changes from 24 to 4 h for k equal to 6 and 12, respectively.

Graphical abstract

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7

Similar content being viewed by others

Data availability

The databases used in our work, MIMIC-IV and eICU, are freely-available and can be accessed at https://physionet.org/content/mimiciv/0.4 and https://physionet.org/content/eicu-crd/2.0 respectively.

References

  1. Singer M, Deutschman CS, Seymour CW et al (2016) The third international consensus definitions for sepsis and septic shock (Sepsis-3) [J]. Jama 315(8):801–810

    Article  CAS  Google Scholar 

  2. Perman SM, Goyal M, Gaieski DF (2012) Initial emergency department diagnosis and management of adult patients with severe sepsis and septic shock[J]. Scand J Trauma Resuscitation Emerg Med 20(1):1–11

    Article  Google Scholar 

  3. Gauer R (2013) Early recognition and management of sepsis in adults: the first six hours[J]. Am Fam Physician 88(1):44–53

    PubMed  Google Scholar 

  4. Fleischmann C, Scherag A, Adhikari NKJ et al (2016) Assessment of global incidence and mortality of hospital-treated sepsis. Current estimates and limitations[J]. Am J Respir Crit Care Med 193(3):259–272

    Article  CAS  Google Scholar 

  5. Rudd KE, Johnson SC, Agesa KM et al (2020) Global, regional, and national sepsis incidence and mortality, 1990–2017: analysis for the Global Burden of Disease Study[J]. Lancet 395(10219):200–211

    Article  Google Scholar 

  6. Esper AM, Martin GS (2009) Extending international sepsis epidemiology: the impact of organ dysfunction[J]. Crit Care 13(1):1–3

    Article  Google Scholar 

  7. Blanco J, Muriel-Bombín A, Sagredo V et al (2008) Incidence, organ dysfunction and mortality in severe sepsis: a Spanish multicentre study[J]. Crit Care 12(6):1–14

    Article  Google Scholar 

  8. Harrison DA, Welch CA, Eddleston JM (2006) The epidemiology of severe sepsis in England, Wales and Northern Ireland, 1996 to 2004: secondary analysis of a high quality clinical database, the ICNARC Case Mix Programme Database[J]. Crit Care 10(2):1–10

    Article  Google Scholar 

  9. Danai P, Martin GS (2005) Epidemiology of sepsis: recent advances[J]. Curr Infect Dis Rep 7(5):329–334

    Article  Google Scholar 

  10. Kaukonen KM, Bailey M, Suzuki S et al (2014) Mortality related to severe sepsis and septic shock among critically ill patients in Australia and New Zealand, 2000-2012[J]. JAMA 311(13):1308–1316

    Article  CAS  Google Scholar 

  11. Rangel-Frausto MS, Pittet D, Costigan M et al (1995) The natural history of the systemic inflammatory response syndrome (SIRS): a prospective study[J]. JAMA 273(2):117–123

    Article  CAS  Google Scholar 

  12. Goulden R, Hoyle MC, Monis J et al (2018) qSOFA, SIRS and NEWS for predicting inhospital mortality and ICU admission in emergency admissions treated as sepsis[J]. Emerg Med J 35(6):345–349

    Article  Google Scholar 

  13. Khwannimit B, Bhurayanontachai R, Vattanavanit V (2019) Comparison of the accuracy of three early warning scores with SOFA score for predicting mortality in adult sepsis and septic shock patients admitted to intensive care unit[J]. Heart Lung 48(3):240–244

    Article  Google Scholar 

  14. Brink A, Alsma J, Verdonschot RJCG et al (2019) Predicting mortality in patients with suspected sepsis at the Emergency Department; a retrospective cohort study comparing qSOFA, SIRS and National Early Warning Score[J]. PLoS One 14(1):e0211133

    Article  CAS  Google Scholar 

  15. Hwang SY, Jo IJ, Lee SU et al (2018) Low accuracy of positive qSOFA criteria for predicting 28-day mortality in critically ill septic patients during the early period after emergency department presentation[J]. Ann Emerg Med 71(1):1–9 e2

    Article  Google Scholar 

  16. Fernando SM, Tran A, Taljaard M et al (2018) Prognostic accuracy of the quick sequential organ failure assessment for mortality in patients with suspected infection: a systematic review and meta-analysis[J]. Ann Intern Med 168(4):266–275

    Article  Google Scholar 

  17. Jaimes F, Leon A, Ascuntar J et al (2016) 1414: Prospective validation of qSOFA in emergency services: a useless bedside clinical score[J]. Crit Care Med 44(12):429

    Article  Google Scholar 

  18. Chen FC, Kung CT, Cheng HH et al (2019) Quick sepsis-related organ failure assessment predicts 72-h mortality in patients with suspected infection[J]. Eur J Emerg Med 26(5):323

    Article  Google Scholar 

  19. Askim Å, Moser F, Gustad LT et al (2017) Poor performance of quick-SOFA (qSOFA) score in predicting severe sepsis and mortality–a prospective study of patients admitted with infection to the emergency department[J]. Scand J Trauma Resuscitation Emerg Med 25(1):1–9

    Article  Google Scholar 

  20. Taylor RA, Pare JR, Venkatesh AK et al (2016) Prediction of in-hospital mortality in emergency department patients with sepsis: a local big data–driven, machine learning approach[J]. Acad Emerg Med 23(3):269–278

    Article  Google Scholar 

  21. Khojandi A, Tansakul V, Li X et al (2018) Prediction of sepsis and in-hospital mortality using electronic health records[J]. Methods Inf Med 57(04):185–193

    Article  Google Scholar 

  22. Rodríguez A, Mendoza D, Ascuntar J et al (2020) Supervised classification techniques for prediction of mortality in adult patients with sepsis[J]. Am J Emerg Med 45:392–397

    Article  Google Scholar 

  23. Perng JW, Kao IH, Kung CT et al (2019) Mortality prediction of septic patients in the emergency department based on machine learning[J]. J Clin Med 8(11):1906

    Article  Google Scholar 

  24. Lea C, Flynn M D, Vidal R, et al. (2017) Temporal convolutional networks for action segmentation and detection[C]. proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 156-165.

  25. Bai S, Kolter J Z, Koltun V. (2018) Convolutional Sequence Modeling Revisited. International Conference on Learning Representations (ICLR). Workshop.

  26. Johnson A et al (2020) "MIMIC-IV" (version 0.4). PhysioNet. https://doi.org/10.13026/a3wn-hq05

  27. Johnson AEW, Pollard TJ, Shen L et al (2016) MIMIC-III, a freely accessible critical care database[J]. Sci Data 3(1):1–9

    Article  Google Scholar 

  28. Johnson AEW, Aboab J, Raffa JD et al (2018) A comparative analysis of sepsis identification methods in an electronic database[J]. Crit Care Med 46(4):494

    Article  Google Scholar 

  29. Pollard TJ, Johnson AEW, Raffa JD et al (2018) The eICU Collaborative Research Database, a freely available multi-center database for critical care research[J]. Sci Data 5(1):1–13

    Article  Google Scholar 

  30. Hochreiter S, Schmidhuber J (1997) Long short-term memory[J]. Neural Comput 9(8):1735–1780

    Article  CAS  Google Scholar 

  31. Chen T, Guestrin C. (2016) Xgboost: A scalable tree boosting system[C]. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM, 785-794.

Download references

Funding

This work is supported partially by China NSFC under Grant 61672309.

Author information

Author notes

  1. Minghui Gong and Jingming Liu contributed equally to this work.

Authors and Affiliations

  1. School of Software, Tsinghua University, Beijing, China

    Minghui Gong, Chunping Li & Ruolin Wang

  2. Emergency Department in Beijing Tiantan Hospital, Capital Medical University, Beijing, China

    Jingming Liu, Wei Guo & Zheng Chen

Authors
  1. Minghui Gong
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Jingming Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Chunping Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Wei Guo
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Ruolin Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Zheng Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Jingming Liu and Wei Guo designed the experiments and provided the clinical expertise and context. Minghui Gong and Chunping Li pre-processed the data and implemented the warning model and experiment. Ruolin Wang contributed to analyses of the data. Zheng Chen provided the clinical expertise and discussion, and revised the paper further.

Corresponding authors

Correspondence to Chunping Li or Wei Guo.

Ethics declarations

Ethical approval and consent to participate

Before we gained access to the MIMIC-IV database, we had passed CITI's ethics test (Record ID: 33389955).

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Gong, M., Liu, J., Li, C. et al. Early warning model for death of sepsis via length insensitive temporal convolutional network. Med Biol Eng Comput 60, 875–885 (2022). https://doi.org/10.1007/s11517-022-02521-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11517-022-02521-3

Keywords

Search

Navigation