JMIR Publications

ABSTRACT

Background:

Ventilator-associated pneumonia (VAP) is one of the serious complications of mechanical ventilation therapy, which seriously affects the treatment and prognosis of patients. Artificial intelligence (AI) has been increasingly used to predict VAP due to its excellent data mining capabilities.

Objective:

This article reviews the prediction models for VAP based on AI, providing reference for early identification of high-risk groups in future clinical practice.

Methods:

A scoping review was conducted in line with the guidelines of the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) extension for scoping reviews. WANFANG Database, CBM, The Cochrane Library, Web of Science, PubMed, MEDLINE, and EMBASE were explored to identify relevant articles. Study selection and data extraction were independently conducted by 2 reviewers. Data extracted from the included studies were synthesized narratively.

Results:

Of the 137 publications, 11 studies were included in the scoping review. The included studies reported the use of AI in the prediction of VAP. Of the 11 studies, all were predictive studies of VAP occurrence and did not include studies of VAP prognosis. All studies were modelled using text data, and none involved imaging data. Public databases are used as the primary data choice for model building (6/11, 55%), with the remaining studies having sample sizes <1000. Machine learning is the main algorithm for studying VAP prediction models. Deep learning and large language models have not yet been used to construct VAP prediction models. Random forest is the most commonly used algorithm (5/11, 45%). All studies were internal validations and none addressed how the model was actually used.

Conclusions:

This review presents an overview of studies based on AI used to predict and diagnose VAP. AI models has better predictive performance than traditional methods, and is expected to provide an indispensable tool for risk prediction of VAP in the future. However, the current research is in the stage of model construction and validation, and how to implement and provide guidance for clinical prediction of VAP still needs further research.