TRÍ TUỆ NHÂN TẠO TRONG GÂY MÊ HỒI SỨC: TỔNG QUAN HỆ THỐNG VỀ ỨNG DỤNG, HẠN CHẾ VÀ KHẢ NĂNG TÍCH HỢP TẠI VIỆT NAM ĐẾN NĂM 2030

Abstract

Background: Artificial intelligence (AI) has emerged as a transformative tool in anesthesiology, with applications ranging from preoperative risk prediction to intraoperative monitoring and postoperative management. However, its clinical integration remains limited. Methods: We conducted a systematic review of eight published articles (2022–2025) and one bibliometric analysis that explored AI applications in anesthesiology. Studies included randomized controlled trials (RCTs), systematic reviews, and narrative reviews that reported on AI-based decision support, drug delivery, airway assessment, or perioperative risk prediction. Results: AI demonstrated significant promise in airway risk stratification (AUC > 0.9), real-time hypotension prediction, and closed-loop anesthesia systems. RCTs included in a meta-analysis (n = 568) showed no statistically significant difference in hypotension-related outcomes (MD = 0.22, 95% CI: –0.03 to 0.48, P = 0.215; I² = 93.8%), but AI-assisted systems reduced intraoperative hypotension duration (MD = 7.41%, 95% CI: 4.95–9.86, P < 0.001)​. Key limitations include algorithm transparency, data heterogeneity, ethical concerns, and limited interpretability. Conclusion: Although AI offers substantial benefits in anesthesia management, its integration in clinical practice remains in early stages. Vietnam should prioritize the development of localized data-driven models, promote AI literacy in anesthesiology training, and establish national guidelines for ethical AI use.