AHRQ

Using AI, simulations to detect missed stroke diagnoses in emergency departments

January 8, 2026

Natural language processing of emergency department clinical records identified text markers that correlated positively with missed or delayed stroke diagnoses 30 days later, according to an AHRQ-funded study published in the Journal of Stroke and Cerebrovascular Diseases. Mention of “language,” “motor,” “sensory”, “cranial nerve”, and “coordination” were among the 11 markers. A predictive model created from these 11 markers performed well across two academic hospitals, suggesting potential for early identification of high-risk patients. Authors noted that validation in emergency department settings is warranted.

Sources:

(2026, January 4). Agency for Healthcare Research and Quality. Subject: Agency for Healthcare Research and Quality (AHRQ) Weekly Digest Bulletin [email]

Mayampurath A, et al. (2025, August). J Stroke Cerebrovasc Dis. Identification of neurological text markers associated with risk of stroke. https://pubmed.ncbi.nlm.nih.gov/40513788/