From Data to Decisions: Harnessing Multi-Agent Systems for Safer, Smarter, and More Personalized Perioperative Care

Background/Objectives: Artificial intelligence (AI) is increasingly applied across the perioperative continuum, with potential benefits in efficiency, personalization, and patient safety. Unfortunately, most such tools are developed in isolation, limiting their clinical utility. Multi-Agent Systems for Healthcare (MASH), in which autonomous AI agents coordinate tasks across multiple domains, may provide the necessary framework for integrated perioperative care. This critical review synthesizes current AI applications in anesthesiology and considers their integration within a MASH architecture. This is the first review to advance MASH as a conceptual and practical framework for anesthesiology, uniquely contributing to the AI discourse by proposing its potential to unify isolated innovations into adaptive and collaborative systems. Methods: A critical review was conducted using PubMed and Google Search to identify peer-reviewed studies published between 2015 and 2025. The search strategy combined controlled vocabulary and free-text terms for AI, anesthesiology, perioperative care, critical care, and pain management. Results were filtered for randomized controlled trials and clinical trials. Data were extracted and organized by perioperative phase. Results: The 16 studies (6 from database search, 10 from prior work) included in this review demonstrated AI applications across the perioperative timeline. Preoperatively, predictive models such as POTTER improved surgical risk stratification. Intraoperative trials evaluated systems like SmartPilot and Navigator, enhancing anesthetic dosing and physiologic stability. In critical care, algorithms including NAVOY Sepsis and VentAI supported early detection of sepsis and optimized ventilatory management. In pain medicine, AI assisted with opioid risk assessment and individualized pain-control regimens. While these trials demonstrated clinical utility, most applications remain domain-specific and unconnected from one another. Conclusions: AI has broad potential to improve perioperative care, but its impact depends on coordinated deployment. MASH offers a unifying framework to integrate diverse agents into adaptive networks, enabling more personalized anesthetic care that is safer and more efficient.