On Classifying Sepsis Heterogeneity in the ICU: Insight Using Machine Learning

Overview

Journal J Am Med Inform Assoc

Publisher Oxford University Press

Specialty Medical Informatics

Date 2020 Jan 18

PMID 31951005

Citations 28

Authors

Zina M Ibrahim

Honghan Wu

Ahmed Hamoud

Lukas Stappen

Richard J B Dobson

Andrea Agarossi

Affiliations

Soon will be listed here.

Abstract

Objectives: Current machine learning models aiming to predict sepsis from electronic health records (EHR) do not account 20 for the heterogeneity of the condition despite its emerging importance in prognosis and treatment. This work demonstrates the added value of stratifying the types of organ dysfunction observed in patients who develop sepsis in the intensive care unit (ICU) in improving the ability to recognize patients at risk of sepsis from their EHR data.

Materials And Methods: Using an ICU dataset of 13 728 records, we identify clinically significant sepsis subpopulations with distinct organ dysfunction patterns. We perform classification experiments with random forest, gradient boost trees, and support vector machines, using the identified subpopulations to distinguish patients who develop sepsis in the ICU from those who do not.

Results: The classification results show that features selected using sepsis subpopulations as background knowledge yield a superior performance in distinguishing septic from non-septic patients regardless of the classification model used. The improved performance is especially pronounced in specificity, which is a current bottleneck in sepsis prediction machine learning models.

Conclusion: Our findings can steer machine learning efforts toward more personalized models for complex conditions including sepsis.

Citing Articles

Sepsis subphenotypes: bridging the gaps in sepsis treatment strategies.

Zhang X, Zhang W, Zhang H, Liao X Front Immunol. 2025; 16:1546474.

PMID: 40013154 PMC: 11862915. DOI: 10.3389/fimmu.2025.1546474.

Application of causal forests to randomised controlled trial data to identify heterogeneous treatment effects: a case study.

Van Vogt E, Gordon A, Diaz-Ordaz K, Cro S BMC Med Res Methodol. 2025; 25(1):50.

PMID: 39987431 PMC: 11846376. DOI: 10.1186/s12874-025-02489-2.

Developing a rapid screening tool for high-risk ICU patients of sepsis: integrating electronic medical records with machine learning methods for mortality prediction in hospitalized patients-model establishment, internal and external validation, and....

Shi S, Zhang L, Zhang S, Shi J, Hong D, Wu S J Transl Med. 2025; 23(1):97.

PMID: 39838426 PMC: 11753157. DOI: 10.1186/s12967-025-06102-4.

A Multivariate Phenotypical Approach of Sepsis and Septic Shock-A Comprehensive Narrative Literature Review.

Tita A, Isac S, Isac T, Martac C, Teodorescu G, Jipa L Medicina (Kaunas). 2024; 60(11).

PMID: 39596925 PMC: 11596881. DOI: 10.3390/medicina60111740.

Early detection of sepsis using machine learning algorithms: a systematic review and network meta-analysis.

Yadgarov M, Landoni G, Berikashvili L, Polyakov P, Kadantseva K, Smirnova A Front Med (Lausanne). 2024; 11:1491358.

PMID: 39478824 PMC: 11523135. DOI: 10.3389/fmed.2024.1491358.

References

Lotsch J, Ultsch A . A machine-learned knowledge discovery method for associating complex phenotypes with complex genotypes. Application to pain. J Biomed Inform. 2013; 46(5):921-8. DOI: 10.1016/j.jbi.2013.07.010. View

Knox D, Lanspa M, Kuttler K, Brewer S, Brown S . Phenotypic clusters within sepsis-associated multiple organ dysfunction syndrome. Intensive Care Med. 2015; 41(5):814-22. PMC: 4607311. DOI: 10.1007/s00134-015-3764-7. View

Iwashyna T, Odden A, Rohde J, Bonham C, Kuhn L, Malani P . Identifying patients with severe sepsis using administrative claims: patient-level validation of the angus implementation of the international consensus conference definition of severe sepsis. Med Care. 2012; 52(6):e39-43. PMC: 3568444. DOI: 10.1097/MLR.0b013e318268ac86. View

Gaieski D, Edwards J, Kallan M, Carr B . Benchmarking the incidence and mortality of severe sepsis in the United States. Crit Care Med. 2013; 41(5):1167-74. DOI: 10.1097/CCM.0b013e31827c09f8. View

Capan M, Hoover S, Miller K, Pal C, Glasgow J, Jackson E . Data-driven approach to Early Warning Score-based alert management. BMJ Open Qual. 2018; 7(3):e000088. PMC: 6109824. DOI: 10.1136/bmjoq-2017-000088. View

Rothman M, Levy M, Dellinger R, Jones S, Fogerty R, Voelker K . Sepsis as 2 problems: Identifying sepsis at admission and predicting onset in the hospital using an electronic medical record-based acuity score. J Crit Care. 2016; 38:237-244. DOI: 10.1016/j.jcrc.2016.11.037. View

Askim A, Moser F, Gustad L, Stene H, Gundersen M, Asvold B . 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. Scand J Trauma Resusc Emerg Med. 2017; 25(1):56. PMC: 5466747. DOI: 10.1186/s13049-017-0399-4. View

Johnson A, Pollard T, Shen L, Lehman L, Feng M, Ghassemi M . MIMIC-III, a freely accessible critical care database. Sci Data. 2016; 3:160035. PMC: 4878278. DOI: 10.1038/sdata.2016.35. View

Islam M, Nasrin T, Walther B, Wu C, Yang H, Li Y . Prediction of sepsis patients using machine learning approach: A meta-analysis. Comput Methods Programs Biomed. 2019; 170:1-9. DOI: 10.1016/j.cmpb.2018.12.027. View

10.

Olenick E, Zimbro K, DLima G, Ver Schneider P, Jones D . Predicting Sepsis Risk Using the "Sniffer" Algorithm in the Electronic Medical Record. J Nurs Care Qual. 2016; 32(1):25-31. DOI: 10.1097/NCQ.0000000000000198. View

11.

McCoy A, Das R . Reducing patient mortality, length of stay and readmissions through machine learning-based sepsis prediction in the emergency department, intensive care unit and hospital floor units. BMJ Open Qual. 2018; 6(2):e000158. PMC: 5699136. DOI: 10.1136/bmjoq-2017-000158. View

12.

Singer M, Deutschman C, Seymour C, Shankar-Hari M, Annane D, Bauer M . The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016; 315(8):801-10. PMC: 4968574. DOI: 10.1001/jama.2016.0287. View

13.

Boudier A, Curjuric I, Basagana X, Hazgui H, Anto J, Bousquet J . Ten-year follow-up of cluster-based asthma phenotypes in adults. A pooled analysis of three cohorts. Am J Respir Crit Care Med. 2013; 188(5):550-60. DOI: 10.1164/rccm.201301-0156OC. View

14.

Vincent J . The Clinical Challenge of Sepsis Identification and Monitoring. PLoS Med. 2016; 13(5):e1002022. PMC: 4871479. DOI: 10.1371/journal.pmed.1002022. View

15.

Vanfleteren L, Spruit M, Groenen M, Gaffron S, van Empel V, Bruijnzeel P . Clusters of comorbidities based on validated objective measurements and systemic inflammation in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2013; 187(7):728-35. DOI: 10.1164/rccm.201209-1665OC. View

16.

Nikkila J, Toronen P, Kaski S, Venna J, Castren E, Wong G . Analysis and visualization of gene expression data using self-organizing maps. Neural Netw. 2002; 15(8-9):953-66. DOI: 10.1016/s0893-6080(02)00070-9. View

17.

Ferrer R, Martin-Loeches I, Phillips G, Osborn T, Townsend S, Dellinger R . Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guideline-based performance improvement program. Crit Care Med. 2014; 42(8):1749-55. DOI: 10.1097/CCM.0000000000000330. View

18.

Seymour C, Gesten F, Prescott H, Friedrich M, Iwashyna T, Phillips G . Time to Treatment and Mortality during Mandated Emergency Care for Sepsis. N Engl J Med. 2017; 376(23):2235-2244. PMC: 5538258. DOI: 10.1056/NEJMoa1703058. View

19.

Strobl C, Boulesteix A, Kneib T, Augustin T, Zeileis A . Conditional variable importance for random forests. BMC Bioinformatics. 2008; 9:307. PMC: 2491635. DOI: 10.1186/1471-2105-9-307. View

20.

Seymour C, Kennedy J, Wang S, Chang C, Elliott C, Xu Z . Derivation, Validation, and Potential Treatment Implications of Novel Clinical Phenotypes for Sepsis. JAMA. 2019; 321(20):2003-2017. PMC: 6537818. DOI: 10.1001/jama.2019.5791. View