The role of AI-driven communication in delirium prevention, detection, and care for critically ill ICU patients: A systematic review with inductive thematic synthesis

Abstract

BackgroundDelirium remains one of the most consequential complications among critically ill patients in ICUs, exerting profound effects on morbidity, mortality, and annual healthcare costs exceeding $81 billion. Communication barriers between sedated or mechanically ventilated patients, their families, and multidisciplinary teams frequently delay recognition and impair management of delirium. This systematic review examines how AI-driven communication technologies can address these barriers, enhance early detection, and promote more integrated, patient- and family-centered delirium care. MethodsA systematic review of literature published between 2015 and 2025 was conducted across five electronic databases: Scopus, PubMed, Web of Science, Embase, and IEEE Xplore. The search strategy employed keywords as “delirium,” “intensive care,” “artificial intelligence,” “AI-driven communication technologies”, “natural language processing”, “computer vision”, “multidisciplinary clinical collaboration”, and “family engagement”. Studies were eligible for inclusion if they focused on AI-enhanced communication in ICU delirium care. The included studies were analyzed using an inductive thematic synthesis approach. ResultsFrom 87 screened records, 16 studies demonstrated AI’s significant benefits across three clinical domains: 1) Prevention using AI-driven tools; 2) Early Detection via multimodal AI systems; and 3) Patient Care through Natural Language Processing (NLP)-powered support. An inductive thematic synthesis of these findings further delineated six core thematic domains: (1) inherent communication barriers; (2) AI as a multidirectional interface; (3) passive AI listening for early detection; (4) AI-enhanced family engagement; (5) AI-structured handovers for teamwork; and (6) ethical-regulatory-practical challenges. ConclusionAI-driven communication tools effectively bridge critical gaps in ICU delirium care, facilitating early detection, prevention, and patient-centered management. By enabling proactive interventions and fostering a collaborative care environment, these technologies demonstrate direct potential to reduce delirium duration, decrease antipsychotic use, improve long-term cognitive outcomes, and alleviate the substantial economic burden on healthcare systems. These findings validate AI’s role in transforming delirium care through enhanced multidirectional communication. Implications for Clinical PracticeICU nurses are pivotal in utilizing AI tools through interpreting NLP-generated alerts, calibrating computer vision outputs, and facilitating family engagement to translate AI insights into empathetic, tailored bedside interventions, thereby reinforcing human-AI collaboration.