Newsletter
Artificial Intelligence in perioperative medicine: hype, hope, or game-changer
A major shift is occurring in artificial intelligence in perioperative medicine. Previously, AI has been used to process large datasets, including patient records and research data, to identify associations and risk factors according to a narrow band of specific requirements. This worked well within the defined parameters of research and only if the data could be curated. (1) This approach, however, was difficult to integrate into clinical systems with their real-world and real-time demands and workflow.
Now developments in AI, especially machine learning (ML), are revealing features that promise to revolutionise preoperative risk prediction, intraoperative monitoring and automation, the detection of postoperative complications, and operational and workflow optimisation. (2)(3) ML can analyse vast datasets of high-dimensional data (data with multiple covariates), identify subtle patterns, and work with nuanced or specific clinical language from patient records. (4) However, there is still a gap between these promises and actual clinical application, and many concerns to eliminate before these models can become part of daily practice.
In the panel discussion, “Artificial Intelligence in perioperative medicine: hype, hope, or game-changer” the speakers will identify key areas and evaluate the evidence where we can implement Artificial Intelligence (AI) into anaesthetic practice. The session will also address the potential benefits and risks associated with integrating AI into clinical workflows.
Predicting potential risks in preoperative assessments with high reliability supports clinical decision-making to increase patient safety and reinforce improved outcomes. Now AI promises to predict personalised risks. Prof. Bettina Jungwirth, Chairman of the Department of Anesthesiology and Intensive Care Medicine at the University of Ulm, Germany, will discuss this topic in her presentation “AI for preoperative risk prediction”. Prof. Jungwirth has published extensively on machine learning in anaesthesia, particularly concerning its use for risk prediction, clinical decision support, and translating AI into perioperative workflows. In her publication “Enabling personalized perioperative risk prediction by using a machine-learning model based on preoperative data” (5) the authors designed an ML model to assess the patient’s individual risks of postoperative in-hospital mortality based solely on preoperative data. This ability to identify personalised risk for each patient, without intraoperative and postoperative data, is the evolution which takes this model beyond normal population-level predictions.
AI and ML also show promise in forecasting critical events during surgery. One specific area that is showing a great deal of promise is their role in predictive support for bleeding and transfusion risk. (6) These types of predictions can optimise workflow, estimate the duration of surgery and time in PACU, and prevent delays and cancellations. Prof. Jens Meier is the Chair Professor of Anaesthesiology and Intensive Care at the Johannes Kepler University in Linz, Austria, and has published on how ML can support clinical decision-making in anaesthesia and surgery. (7) His publications highlight clinical examples of predicting mortality after heart-valve surgery, transfusion needs, and massive blood loss in cardiac operations. They demonstrate the potential of AI to help anticipate blood loss, prepare resources, and intervene earlier during procedures. In his presentation “Applications of AI in the operating room”, he will discuss intraoperative AI applications such as predicting depth of anaesthesia, forecasting haemodynamic instability, and automated drug delivery in closed-loop systems. These developments are not without challenges. Any interaction between AI and humans in high-risk environments must facilitate focus on the work at hand and support for decision-making must be in balance with cognitive load. There are also regulatory and legal issues in letting non-human intelligence manage high-risk decision-making.
As we integrate AI systems into anaesthesia, it is reasonable to be cautious. We have seen that AI can hallucinate, providing false data with total confidence. Another issue is bias. Can AI show bias if it does not have human qualities like prejudice or discrimination? Dr. Sarah Saxena says, “Yes”. Dr Saxena, a professor of anaesthesiology at the University of Mons, Belgium, has published several articles on this topic. Her presentation “Is AI gender-biased?” will show that “AI models exhibited notable biases in gender, race/ethnicity, and age representation, failing to reflect the actual diversity within the anesthesiologist workforce.” (8) This evidence of bias warns against naivety when using AI. In one paper she co-authored, the authors note “medical students have started using generative artificial intelligence (GenAI), a subset of AI capable of generating text and synthesizing knowledge, to complement textbooks for clinical decision-making”. (9) Now, it is necessary to “urgently integrate critical AI literacy into its curricula. Students need training, not just on how to use GenAI tools, but on when not to rely on them.” (10)
The panel discussion “Artificial Intelligence in perioperative medicine: hype, hope, or game-changer” will take place at the Euroanaesthesia Congress 2026 on Monday, June 8 at 15:00–16:00 CEST in room PORT C.
References
- Graeßner M, Jungwirth B, Frank E, et al. Enabling personalized perioperative risk prediction by using a machine-learning model based on preoperative data. Sci Rep. 2023;13(1):7128. Published 2023 May 2. doi:10.1038/s41598-023-33981-8 https://pmc.ncbi.nlm.nih.gov/articles/PMC10153050/
- Bellini, V., Russo, M., Domenichetti, T. et al. Artificial Intelligence in Operating Room Management. J Med Syst 48, 19 (2024). https://doi.org/10.1007/s10916-024-02038-2 https://link.springer.com/article/10.1007/s10916-024-02038-2
- Habehh H, Gohel S. Machine Learning in Healthcare. Curr Genomics. 2021;22(4):291-300. doi:10.2174/1389202922666210705124359 https://pmc.ncbi.nlm.nih.gov/articles/PMC8822225/
- Graeßner M, Jungwirth B, Frank E, et al. Enabling personalized perioperative risk prediction by using a machine-learning model based on preoperative data. Sci Rep. 2023;13(1):7128. Published 2023 May 2. doi:10.1038/s41598-023-33981-8 https://pmc.ncbi.nlm.nih.gov/articles/PMC10153050/
- Meier, Jens M. MD; Tschoellitsch, Thomas MD. Artificial Intelligence and Machine Learning in Patient Blood Management: A Scoping Review. Anesthesia & Analgesia 135(3):p 524-531, September 2022. | DOI: 10.1213/ANE.0000000000006047 https://journals.lww.com/anesthesia-analgesia/abstract/2022/09000/artificial_intelligence_and_machine_learning_in.11.aspx
- Gisselbaek M, Minsart L, Köselerli E, Suppan M, Meco BC, Seidel L, Albert A, Barreto Chang OL, Saxena S and Berger-Estilita J (2024) Beyond the stereotypes: Artificial Intelligence image generation and diversity in anesthesiology. Front. Artif. Intell. 7:1462819. doi: 10.3389/frai.2024.1462819 https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2024.1462819/full
- Saxena S, Tsobgnie MN, Südy R, Gisselbaek M, Lechien JR, Carella M, Ingrassia PL, Dieckmann P, Berger-Estilita J. Generation Z versus generative artificial intelligence: a cross-sectional study assessing medical students’ confidence and over-reliance on artificial intelligence in perioperative clinical scenarios. Eur J Anaesthesiol. 2026 Mar 1;43(3):273-275. doi: 10.1097/ https://journals.lww.com/ejanaesthesiology/citation/2026/03000/generation_z_versus_generative_artificial.12.aspx
- Gisselbaek M, Köselerli E, Suppan M, Minsart L, Meco BC, Seidel L, Albert A, Barreto Chang OL, Berger-Estilita J, Saxena S. Gender bias in images of anaesthesiologists generated by artificial intelligence. Br J Anaesth. 2024 Sep;133(3):692-695. doi: 10.1016/j.bja.2024.05.027. Epub 2024 Jun 27. PMID: 38942642. https://www.bjanaesthesia.org/article/S0007-0912(24)00337-4/fulltext
