Artificial intelligence agent for delirium screening among patients in oncology and cardiac intensive care units: a proof-of-concept study

Агент искусственного интеллекта для скрининга делирия у пациентов в онкологических и кардиологических отделениях интенсивной терапии: исследование с доказательством концепции
Guolong Zhang, Fen Gu, Jian Chen, Yingchun Zeng, Hongxia Xie, Xinjun Jiang, Siew Tiang Lau
2026-01-27

autonomous AI agentdelirium knowledge graphdelirium screeningintensive care unitsretrieval-augmented generation
AIMS: This study aims to develop and initially test an autonomous artificial intelligence (AI) agent to support nurse-led delirium screening and guideline-concordant prevention and management. METHODS AND RESULTS: We constructed a delirium-specific knowledge graph from publicly available clinical guidelines and implemented an autonomous AI agent that integrates retrieval-augmented generation with validated delirium assessment tools to simulate decision support in nursing workflows. In this proof-of-concept evaluation, the agent was benchmarked on 20 clinical patient cases to assess (i) tool selection accuracy, (ii) fidelity of clinical conclusions, and (iii) adherence to guideline-based delirium care recommendations. Human experts review rated coherence, relevance, and clinical interpretability of the outputs. The 20 cases were drawn from post-operative care (n = 6), cardiac intensive care unit care (n = 5), and cancer wards (n = 9), including 6 males (30%) and 14 females (70%), with a mean age of 56.5 years (SD 17.9; range 21-82). Two experienced registered nurses independently validated the agent's outputs, benchmarked them against guideline-based recommendations. The agent achieved 100% accuracy in selecting appropriate tools and 90% overall accuracy in generating conclusions and care recommendations, compared with 45% for other large language model baselines. Recommendations were presented in a structured, actionable format and aligned with guideline-based delirium care. CONCLUSION: This proof-of-concept study suggests that an autonomous AI agent can deliver clinically interpretable, guideline-aligned delirium decision support and may help reduce missed or delayed recognition while standardizing nursing actions. Given the high delirium burden in cardiovascular pathways (e.g. cardiac intensive care units and post-operative care), prospective validation in cardiovascular settings is warranted to evaluate clinical impact, safety, and workflow integration.
1
Across 20 oncology, cardiac intensive care, and postoperative cases, the agent achieved 100% accuracy in selecting appropriate delirium assessment tools.
2
An autonomous AI agent was developed by combining a delirium-specific knowledge graph, retrieval-augmented generation, and validated delirium assessment tools.
3
Outputs were structured, actionable, clinically interpretable, and aligned with guideline-based delirium prevention and management recommendations.
4
Prospective validation is needed to assess clinical impact, safety, and workflow integration, particularly in cardiovascular care settings.
5
The agent generated clinically faithful conclusions and care recommendations with 90% overall accuracy, compared with 45% for other large language model baselines.

An autonomous artificial intelligence agent for delirium screening and care support in oncology and cardiac intensive care settings

The agent’s accuracy, clinical fidelity, interpretability, and adherence to guideline-based delirium prevention, management, and nursing workflow recommendations

Publication Details
Publication Date
2026-01-27
Journal
Publisher
ISSN
Cited by
4
Access Type
Author Information
Authors
Guolong Zhang
Fen Gu
Jian Chen
Yingchun Zeng
Hongxia Xie
Xinjun Jiang
Siew Tiang Lau
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%