Explainable artificial intelligence model to predict acute critical illness from electronic health records

Abstract

Acute critical illness is often preceded by deterioration of routinely measured clinical parameters, e.g., blood pressure and heart rate. Early clinical prediction is typically based on manually calculated screening metrics that simply weigh these parameters, such as early warning scores (EWS). The predictive performance of EWSs yields a tradeoff between sensitivity and specificity that can lead to negative outcomes for the patient. Previous work on electronic health records (EHR) trained artificial intelligence (AI) systems offers promising results with high levels of predictive performance in relation to the early, real-time prediction of acute critical illness. However, without insight into the complex decisions by such system, clinical translation is hindered. Here, we present an explainable AI early warning score (xAI-EWS) system for early detection of acute critical illness. xAI-EWS potentiates clinical translation by accompanying a prediction with information on the EHR data explaining it.

Full text

Its licence does not allow us to show the full text. Read it at PubMed Central:

References 68

1. 1.

   Tomašev N, et al. A clinically applicable approach to continuous prediction of future acute kidney injury. Nature. 2019;572:116-119.

   Not in our library10.1038/s41586-019-1390-1

2. 2.

   Lauritsen SM, et al. Early detection of sepsis utilizing deep learning on electronic health record event sequences. Artif. Intell. Med. 2020;104:101820.

   Not in our library10.1016/j.artmed.2020.101820

3. 3.

   Barton C, et al. Evaluation of a machine learning algorithm for up to 48-hour advance prediction of sepsis using six vital signs. Comput. Biol. Med. 2019;109:79-84.

   Not in our library10.1016/j.compbiomed.2019.04.027

4. 4.

   Islam MM, et al. Prediction of sepsis patients using machine learning approach: a meta-analysis. Comput. Meth. Prog. Bio. 2019;170:1-9.

   Not in our library10.1016/j.cmpb.2018.12.027

5. 5.

   Moor, M., Horn, M., Rieck, B., Roqueiro, D. & Borgwardt, K. Temporal convolutional networks and dynamic time warping can drastically improve the early prediction of Sepsis. In Proceedings of the 4th Machine Learning for Healthcare Conference. PMLR 106 (2019).

   Not in our library

6. 6.

   Mao Q, et al. Multicentre validation of a sepsis prediction algorithm using only vital sign data in the emergency department, general ward and ICU. BMJ Open. 2018;8:e017833.

   Not in our library10.1136/bmjopen-2017-017833

7. 7.

   Shimabukuro DW, Barton CW, Feldman MD, Mataraso SJ, Das R. Effect of a machine learning-based severe sepsis prediction algorithm on patient survival and hospital length of stay: a randomised clinical trial. BMJ Open Respir. Res. 2017;4:e000234.

   Not in our library10.1136/bmjresp-2017-000234

8. 8.

   Kam HJ, Kim HY. Learning representations for the early detection of sepsis with deep neural networks. Comput. Biol. Med. 2017;89:248-255.

   Not in our library10.1016/j.compbiomed.2017.08.015

9. 9.

   Calvert JS, et al. A computational approach to early sepsis detection. Comput. Biol. Med. 2016;74:69-73.

   Not in our library10.1016/j.compbiomed.2016.05.003

10. 10.

    Futoma, J. et al. An improved multi-output gaussian process RNN with real-time validation for early Sepsis detection. In Proceedings of Machine Learning for Healthcare 2017. JMLR, 68 (2017).

    Not in our library

11. 11.

    Futoma, J., Hariharan, S. & Heller, K. Learning to detect sepsis with a multitask gaussian process RNN Classifier. In Proceedings of the 34th International Conference on Machine Learning. ICML vol. 70, 1174–1182 (2017).

    Not in our library

12. 12.

    Vellido, A., Ribas, V., Morales, C., Sanmartín, A. R. & Ruiz-Rodríguez, J. C. Machine learning for critical care: an overview and a Sepsis case study. In Bioinformatics and Biomedical Engineering 2017 (eds Rojas, I. & Ortuño, F.) (IWBBIO) vol. 10208, 15–30 (2017).

    Not in our library

13. 13.

    Shickel B, et al. DeepSOFA: a continuous acuity score for critically ill patients using clinically interpretable deep learning. Sci. Rep. 2019;9:1879.

    Not in our library10.1038/s41598-019-38491-0

14. 14.

    Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat. Med. 2019;25:44-56.

    Not in our library10.1038/s41591-018-0300-7

15. 15.

    Shickel B, Tighe PJ, Bihorac A, Rashidi P. Deep EHR: a survey of recent advances in deep learning techniques for electronic health record (EHR). Anal. IEEE J. Biomed. Health. 2017;22:1589-1604.

    Not in our library10.1109/jbhi.2017.2767063

16. 16.

    Rajkomar A, Dean J, Kohane I. Machine learning in medicine. N. Engl. J. Med. 2019;380:1347-1358.

    Not in our library10.1056/nejmra1814259

17. 17.

    Cabitza F, Rasoini R, Gensini GF. Unintended consequences of machine learning in medicine. JAMMA. 2017;318:517-518.

    Not in our library10.1001/jama.2017.7797

18. 18.

    Xiao C, Choi E, Sun J. Opportunities and challenges in developing deep learning models using electronic health records data: a systematic review. J. Am. Med. Inf. Assn. 2018;25:1419-1428.

    Not in our library10.1093/jamia/ocy068

19. 19.

    The Lancet Respiratory Medicine. Opening the black box of machine learning. Lancet Respir. Med.6, 837 10.1016/S2213-2600(18)30425-9 (2018).

    Not in our library10.1016/s2213-2600(18)30425-9

20. 20.
    Causability and explainability of artificial intelligence in medicine · 2019

    No notice
21. 21.

    Kalchbrenner, N. et al. Neural machine translation in linear time. Preprint at https://arxiv.org/abs/1610.10099 (2016).

    Not in our library

22. 22.

    Bai, S., Kolter, J. Z. & Koltun, V. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. Preprint at https://arxiv.org/abs/1803.01271 (2018).

    Not in our library

23. 23.

    Montavon G, Lapuschkin S, Binder A, Samek W, Müller K. Explaining nonlinear classification decisions with deep Taylor decomposition. Pattern Recogn. 2017;65:211-222.

    Not in our library10.1016/j.patcog.2016.11.008

24. 24.

    Montavon G, Samek W, Müller K-R. Methods for interpreting and understanding deep neural networks. Digit Signal Process. 2017;73:1-15.

    Not in our library10.1016/j.dsp.2017.10.011

25. 25.
    On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation · 2015

    No notice
26. 26.

    Samek, W., Wiegand, T. & Müller, K.-R. Explainable artificial intelligence: understanding, visualizing and interpreting deep learning models. The Impact of Artificial Intelligence (AI) on Communication Networks and Services1, 39–48 (2015)

    Not in our library

27. 27.

    Samek, W., Binder, A., Montavon, G., Lapuschkin, S. & Müller, K. Evaluating the visualization of what a deep neural network has learned. IEEE T. Neur. Net. Lear Syst.28, 2660–2673 (2017).

    Not in our library10.1109/tnnls.2016.2599820

28. 28.

    Lea, C., Vidal, R., Reiter, A. & Hager, G. D. Temporal convolutional networks: a unified approach to action segmentation. In Computer Vision – ECCV 2016 Workshops. 47–54 (Springer International Publishing, 2016).

    Not in our library

29. 29.

    Aksan, E. & Hilliges, O. STCN: Stochastic temporal convolutional networks. In International Conference on Learning Representations 2019, CoRR (2019).

    Not in our library

30. 30.

    Singer M, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315:801-810.

    Not in our library10.1001/jama.2016.0287

31. 31.

    Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin. Pract. 2012;120:c179-c184.

    Not in our library10.1159/000339789

32. 32.

    Zhang J, Kowsari K, Harrison JH, Lobo JM, Barnes LE. patient2Vec: a personalized interpretable deep representation of the longitudinal electronic health record. IEEE Access. 2018;6:65333-65346.

    Not in our library10.1109/access.2018.2875677

33. 33.
    Scalable and accurate deep learning with electronic health records · 2018

    No notice
34. 34.

    Lundberg SM, et al. From local explanations to global understanding with explainable AI for trees. Nat. Mach. Intell. 2020;2:56-67.

    Not in our library10.1038/s42256-019-0138-9

35. 35.

    Kaji DA, et al. An attention based deep learning model of clinical events in the intensive care unit. PLos ONE. 2019;14:e0211057.

    Not in our library10.1371/journal.pone.0211057

36. 36.

    Choi, E. et al. RETAIN: An interpretable predictive model for healthcare using reverse time attention mechanism. In 30th Annual Conference on Neural Information Processing Systems (NIPS 2016). Advances in Neural Information Processing Systems 3512–3520 (2016).

    Not in our library

37. 37.

    Holzinger A, Carrington A, Müller H. Measuring the quality of explanations: the system causability scale (SCS). Künstl Intell. 2020;34:193-198.

    Not in our library10.1007/s13218-020-00636-z

38. 38.

    Selbst AD, Powles J. Meaningful information and the right to explanation. Int Data Priv. Law. 2017;7:233-242.

    Not in our library10.1093/idpl/ipx022

39. 39.

    Seymour CW, et al. Assessment of clinical criteria for sepsis: for the third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315:762-774.

    Not in our library10.1001/jama.2016.0288

40. 40.

    Vincent J-L, et al. The SOFA (Sepsis-related organ failure assessment) score to describe organ dysfunction/failure. Intens. Care Med. 2018;22:707-710.

    Not in our library10.1007/bf01709751

Cited by papers we hold 5

• Explainable AI in Clinical Decision Support Systems: A Meta-Analysis of Methods, Applications, and Usability Challenges · 2025
  No notice
• Should AI models be explainable to clinicians? · 2024
  No notice
• Unbox the black-box for the medical explainable AI via multi-modal and multi-centre data fusion: A mini-review, two showcases and beyond · 2022
  No notice
• Machine Learning–Based Early Warning Systems for Clinical Deterioration: Systematic Scoping Review · 2021
  Not checked
• Accessing Artificial Intelligence for Clinical Decision-Making · 2021
  No notice

Reads

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.