What is it about?
Delirium is a common but routinely underdiagnosed neuropsychiatric disorder in hospitalized patients, characterized by sudden changes in attention and behavior. We have shown that the automated analysis of short video recordings can detect delirium by tracking patients' eye movements, facial expressions, and body postures. We enrolled 50 hospitalized patients and conducted daily delirium assessments to validate this approach. Our results are an important proof-of-concept for objective delirium monitoring that does not add burden to clinical staff.
Featured Image
Photo by Adhy Savala on Unsplash
Why is it important?
Identifying an objective and potentially continuous way to monitor delirium is important because delirium has a fluctuating course, which current diagnostic methods that rely on intermittent human assessments can miss. This is the first study to evaluate video-based delirium detection in a large, clinically diverse cohort of patients outside of the ICU. This work is timely given recent advancements in deep learning.
Perspectives
As a co-author, I find the prospect of scaling this approach to continuous monitoring of delirium exciting. Validating the machine learning models across larger and more diverse patient populations will be a key next step. I am also excited by the possibility of using this approach to better understand delirium pathophysiology by quantifying behavioral changes.
Maanasa Mendu
Yale University
Read the Original
This page is a summary of: Video-based detection of Delirium in hospitalized adults, PLOS Digital Health, May 2026, PLOS,
DOI: 10.1371/journal.pdig.0001462.
You can read the full text:
Contributors
The following have contributed to this page
