APPLICATION OF ARTIFICIAL INTELLIGENCE IN INTENSIVE CARE: CLINICAL POTENTIAL AND CHALLENGES

Jagoda Maternia; Inga Jakubczyk; Kacper Szkodziński; Aleksandra  Łoś; Karolina Majowicz-Czaszyńska; Wiktoria Pempuś; Nikola Król; Barbara Tomaszek; Aleksandra  Blok; Dominik Wiater; Gabriela Płodzień

doi:10.31435/ijitss.3(51).2026.6013

Authors

Jagoda Maternia Saint Queen Jadwiga Clinical Regional Hospital in Rzeszów, Rzeszów, Poland https://orcid.org/0009-0000-9614-118X
Inga Jakubczyk Saint Queen Jadwiga Clinical Regional Hospital in Rzeszów, Rzeszów, Poland https://orcid.org/0000-0002-1825-0097
Kacper Szkodziński University Clinical Hospital F. Chopin in Rzeszów, Rzeszów, Poland https://orcid.org/0009-0009-1161-129X
Aleksandra Łoś Saint Queen Jadwiga Clinical Regional Hospital in Rzeszów, Rzeszów, Poland https://orcid.org/0009-0000-5094-1157
Karolina Majowicz-Czaszyńska Saint Queen Jadwiga Clinical Regional Hospital in Rzeszów, Rzeszów, Poland https://orcid.org/0009-0002-3240-8564
Wiktoria Pempuś Saint Queen Jadwiga Clinical Regional Hospital in Rzeszów, Rzeszów, Poland https://orcid.org/0000-0002-5410-0766
Nikola Król Saint Queen Jadwiga Clinical Regional Hospital in Rzeszów, Rzeszów, Poland https://orcid.org/0000-0002-5524-6476
Barbara Tomaszek The Independent Public Healthcare Institution in Przeworsk, Przeworsk, Poland https://orcid.org/0009-0001-8477-0007
Aleksandra Blok The Independent Public Healthcare Institution in Przeworsk, Przeworsk, Poland https://orcid.org/0009-0001-0813-0366
Dominik Wiater University Clinical Hospital F. Chopin in Rzeszów, Rzeszów, Poland https://orcid.org/0009-0005-3761-5118
Gabriela Płodzień Saint Queen Jadwiga Clinical Regional Hospital in Rzeszów, Rzeszów, Poland https://orcid.org/0009-0008-2193-9312

DOI:

https://doi.org/10.31435/ijitss.3(51).2026.6013

Keywords:

Artificial Intelligence, Intensive Care, Sepsis, Hemodynamic Instability, Machine Learning, Bioethics

Abstract

Background: Contemporary intensive care units (ICUs) generate vast amounts of real-time data, leading to staff cognitive overload and the phenomenon of alarm fatigue. Traditional prognostic scoring systems (APACHE II, SOFA) are inherently static and fail to account for dynamic clinical trends. Artificial intelligence (AI) algorithms, including machine learning (ML) and deep learning (DL), offer novel capabilities for continuous monitoring and early prediction of life-threatening conditions.

Objective: To analyze the clinical potential and implementation challenges associated with the deployment of AI algorithms in intensive care settings.

Methods: A comprehensive literature review was conducted across the PubMed, Embase, and IEEE Xplore databases for the years 2018–2026, focusing on studies utilizing multicenter databases (such as MIMIC-IV and eICU).

Results: AI algorithms demonstrate high discriminative performance (AUROC 0.79–0.96) in the early detection of sepsis, prediction of hemodynamic instability with a 5-to-15-minute lead time, and dynamic mortality risk estimation. In the context of mechanical ventilation, deep learning models (specifically convolutional neural networks, CNNs) enable automated detection of patient-ventilator asynchrony with an accuracy exceeding 90%, as well as precise forecasting of extubation success.

Conclusions: Artificial intelligence holds immense potential for optimizing ICU patient care. However, primary implementation barriers persist, including the lack of model interpretability (the "black box" problem), domain shift, and ethico-psychological dilemmas, such as clinician moral distress and the risk of care dehumanization. AI systems should function strictly as clinical decision support tools under physician supervision.

References

Goldberger, A. L., Evans, W., Harrison, W., et al. (2021). MIMIC-IV: A comprehensive ecosystem of clinical data from the intensive care unit. Scientific Data, 8, 142. https://doi.org/10.1038/s41597-021-00922-2

Sujan, M., Scott, P., Habli, I., et al. (2023). Digital technologies and alarm fatigue in intensive care: Systemic risks and solutions for patient safety. Anaesthesiology Intensive Therapy, 55(3), 214–221. https://doi.org/10.5114/ait.2023.129402

Vincent, J. L., Moreno, R., & Takala, J. (2021). The SOFA score at 25: Historical context, current utility, and future directions in clinical trials. Critical Care, 25(1), 210. https://doi.org/10.1186/s13054-021-03632-4

Schinkel, M., Paranjape, K., Nannan Panday, R. S., et al. (2022). Clinical artificial intelligence in the intensive care unit: A roadmap for implementation. Critical Care, 26(1), 145. https://doi.org/10.1186/s13054-022-04011-x

van de Sande, D., van Genderen, M. E., Huiskens, J., et al. (2021). Moving from predictive to prescriptive analytics in the clinic: Is clinical artificial intelligence ready for prime time? Intensive Care Medicine, 47(2), 123–124. https://doi.org/10.1007/s00134-020-06263-z

Ghassemi, M., Oakden-Rayner, L., & Beam, A. L. (2021). The false hope of current explainable AI in medicine: Unlocking the clinical black box. The Lancet Digital Health, 3(11), e745–e750. https://doi.org/10.1016/S2589-7500(21)00208-9

Page, M. J., McKenzie, J. E., Bossuyt, P. M., et al. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71

Johnson, A. E. W., Bulgarelli, L., Pollard, T. J., et al. (2022). Deep learning with electronic health records in the intensive care unit: A systematic review. IEEE Journal of Biomedical and Health Informatics, 26(4), 1890–1904. https://doi.org/10.1109/JBHI.2021.3124556

Amirsavadkouhi, A., & Mirtajani, S. B. (2024). The role of artificial intelligence in the prediction, diagnosis, and management of sepsis in the intensive care unit (ICU). Medical Research Archives, 12(10). https://doi.org/10.18103/mra.v12i10.5847

Schinkel, M., Paranjape, K., Nannan Panday, R. S., Skyttberg, N., & Nanayakkara, P. W. B. (2022). Clinical applications of artificial intelligence in sepsis: A narrative review. Computers in Biology and Medicine, 147, 105761. https://doi.org/10.1016/j.compbiomed.2022.105761

Zhou, P., Duan, J., Li, J., et al. (2024). Improving sepsis mortality prediction with machine learning: A comparative study of advanced classifiers and performance metrics. Advances in Clinical and Experimental Medicine, 33(11), 1245–1255. https://doi.org/10.17219/acem/189432

Greenstein, Y. Y., Maheshwari, K., Khanna, A. K., et al. (2023). Proactive management of hemodynamic instability using machine learning-derived early warning systems: A multicenter randomized trial. JAMA Network Open, 6(8), e2328105. https://doi.org/10.1001/jamanetworkopen.2023.28105

Vistisen, S. T., Hatib, F., & Scheeren, T. W. L. (2022). Machine learning algorithm based on arterial waveforms predicts hypotension in critical care settings. Anesthesia & Analgesia, 134(3), 512–521. https://doi.org/10.1213/ANE.0000000000005820

Maheshwari, K., Shimada, T., Yang, D., et al. (2023). Hypotension Prediction Index for prevention of hypotension during noncardiac surgery: A registry-based cohort study. Anesthesiology, 139(3), 285–295. https://doi.org/10.1097/ALN.0000000000004634

Chang, L., Zhou, S., Geng, J., et al. (2023). Application of artificial intelligence in mechanical ventilation: A systematic review and meta-analysis. Frontiers in Medicine, 10, 1225564. https://doi.org/10.3389/fmed.2023.1225564

Rehm, G. B., Han, J., Kuhn, B. T., et al. (2022). Explicable machine learning for the prediction of patient-ventilator asynchrony in the intensive care unit. Computer Methods and Programs in Biomedicine, 218, 106720. https://doi.org/10.1016/j.cmpb.2022.106720

Otaguro, T., Tanaka, H., Igarashi, Y., et al. (2024). Machine learning models for predicting reintubation in intensive care unit patients after extubation: A retrospective cohort study. Journal of Intensive Care, 12(1), 14. https://doi.org/10.1186/s40560-024-00723-x

Vincent, J. L., Baron, D. M., & Reinhart, K. (2022). Evaluation of ICU severity scoring systems: From APACHE to modern dynamic benchmarking. Intensive Care Medicine, 48(4), 412–421. https://doi.org/10.1007/s00134-022-06615-1

Sheikhalishahi, S., & Osmani, V. (2021). Real-time mortality tracking and benchmarking of machine learning models on the eICU database. Scientific Reports, 11(1), 12450. https://doi.org/10.1038/s41598-021-91820-y

Thorsen-Meyer, H. C., Nielsen, A. B., et al. (2022). Dynamic mortality prediction in the ICU using attention-based transformers. The Lancet Digital Health, 4(3), e182–e191. https://doi.org/10.1016/S2589-7500(22)00012-4

Kundu, S. (2021). AI in medicine must be explainable. Nature Medicine, 27(8), 1328. https://doi.org/10.1038/s41591-021-01461-z

Nestler, S., Goldmann, A. C., Richter, T., et al. (2024). Generalizability of machine learning models in critical care medicine: A systematic review. The Lancet Digital Health, 6(2), e112–e123. https://doi.org/10.1016/S2589-7500(23)00244-5

Challen, R., Denny, J., Pitt, M., et al. (2022). Bias, algorithmic safety, and the self-fulfilling prophecy in medical artificial intelligence. BMJ Quality & Safety, 31(4), 289–296. https://doi.org/10.1136/bmjqs-2021-013450

Borenstein, J., & Howard, A. Y. (2021). Emerging challenges in AI and the need for moral resilience in healthcare. The American Journal of Bioethics, 21(7), 53–55. https://doi.org/10.1080/15265161.2021.1925790

Montgomery, K., & Charon, R. (2022). The automation of clinical thought: Clinical judgment and artificial intelligence. Oxford University Press.

Morley, J., Machado, C. C., Burr, C., et al. (2023). The ethics of AI in critical care: A scoping review. Intensive Care Medicine, 49(9), 1043–1055. https://doi.org/10.1007/s00134-023-07134-3