Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Jul 16, 2026

A Bedside, Single Burr Hole Approach to Multimodality Monitoring in Severe Brain Injury
06:18

A Bedside, Single Burr Hole Approach to Multimodality Monitoring in Severe Brain Injury

Published on: March 26, 2019

Time-Updated Prognostic Modeling in ICU Patients with Documented Coma or Unresponsiveness Using Routine Arterial

Pompiliu Mircea Bogdan1, Camer Salim2, Roxana Elena Bogdan-Goroftei3

  • 1Doctoral School of Biomedical Sciences, Faculty of Medicine and Pharmacy, Research Center in the Medical-Pharmaceutical Field, "Dunărea de Jos" University of Galați, 800008 Galați, Romania.

Journal of Clinical Medicine
|July 15, 2026
PubMed

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Epilepsy in Paediatric Palliative Care: Prevalence, Clinical Correlations and the Development of a Consensus-Based Seizure Management Protocol.

Medicina (Kaunas, Lithuania)·2026
Same author

Carotid Atherosclerosis in Chronic Kidney Disease: Pathophysiological Mechanisms, Prevention and Tailored Clinical Management.

Journal of clinical medicine·2026
Same author

Oral Health and Gut-Targeted Microbial Marker Changes Associated with Prolonged Hospitalization in Cardiac Patients: An Integrative Risk Analysis.

Life (Basel, Switzerland)·2026
Same author

Syphilis Co-Infection Among People Living with HIV in Romania: Epidemiological and Clinical Characteristics in a Single-Center Retrospective Study.

Pathogens (Basel, Switzerland)·2026
Same author

Vitamin D and Vitamin B<sub>12</sub> in Psychiatric Disorders: An Exploratory Systematic Review and Meta-Analysis of Nutrient-Specific Status and Supplementation Evidence.

Diseases (Basel, Switzerland)·2026
Same author

Admission Biomarkers as Predictors of Mortality in Comatose Patients in the Intensive Care Unit: A Retrospective Pilot Study.

Diagnostics (Basel, Switzerland)·2026
Summary

Explainable machine learning models using routine arterial blood gas (ABG) and SpO2 trajectories can predict outcomes in ICU patients with coma. While trajectory data shows promise, its added value over simpler models needs further validation.

Area of Science:

  • Critical Care Medicine
  • Machine Learning in Healthcare
  • Prognostic Modeling

Background:

  • Prognostication for ICU patients with coma is challenging due to limitations of static scoring systems.
  • Routine arterial blood gas (ABG) and SpO2 measurements offer dynamic physiological data.
  • Explainable machine learning can potentially leverage these data for improved predictions.

Purpose of the Study:

  • To develop and validate explainable machine learning models for ICU outcome prediction in comatose patients.
  • To assess the prognostic value of physiological trajectories derived from ABG/SpO2.
  • To compare staged models incorporating baseline and time-updated data.

Main Methods:

  • Retrospective study of 108 adult ICU patients with coma or unresponsiveness.
Keywords:
ICU mortalitySHAParterial blood gascomacross-validationdecision curve analysisexplainable artificial intelligenceintensive care unitmachine learningoxygenationphysiological trajectoriesprognostication

More Related Videos

Cerebral Ischemic Coma Model Induced by Modified Four-Vessel Occlusion
03:37

Cerebral Ischemic Coma Model Induced by Modified Four-Vessel Occlusion

Published on: July 5, 2024

Related Experiment Videos

Last Updated: Jul 16, 2026

A Bedside, Single Burr Hole Approach to Multimodality Monitoring in Severe Brain Injury
06:18

A Bedside, Single Burr Hole Approach to Multimodality Monitoring in Severe Brain Injury

Published on: March 26, 2019

Cerebral Ischemic Coma Model Induced by Modified Four-Vessel Occlusion
03:37

Cerebral Ischemic Coma Model Induced by Modified Four-Vessel Occlusion

Published on: July 5, 2024

  • Utilized demographics, biomarkers, and ABG/SpO2 trajectories at admission, 24h, and 72h.
  • Employed logistic regression, random forest, and XGBoost with cross-validation; interpreted using SHAP.

    • Main Results:

    • ICU mortality was 65.7%.
    • Models incorporating 72h trajectories (Model C_noRS) showed strong discrimination (AUC-ROC 0.895).
    • Oxygenation, acid-base status, and ΔPaO2 at 72h were key predictors; respiratory support intensity was also significant.

    Conclusions:

    • Explainable ML models using ABG/SpO2 trajectories are feasible for prognostication in comatose ICU patients.
    • Trajectory-enriched models showed potential but require further validation against simpler benchmarks.
    • Findings are exploratory and necessitate external validation before clinical deployment.