Development of an AI Tool for Improved Self- and Caregiver-Diagnosis of Stroke Collecting Video-Clinical Data in Real-Life Settings

What this Trial Is About

Researchers are developing an artificial intelligence tool to improve early detection of stroke by enabling both self-diagnosis and diagnosis by others using video recordings. The study aims to collect video and clinical data in real-life settings to support the creation of AI algorithms that can help reduce disability after stroke by facilitating timely diagnosis. This research focuses on patients suspected of acute stroke or transient ischemic attack within 72 hours of symptom onset. Participants will use the AI-STROKE application to perform a complete neurological exam, which will be recorded on video by both healthcare workers and the patients themselves. The study involves initial video recordings on Day 0 and follow-up recordings at 3 months. The goal is to assess the feasibility and acceptability of patient self-recording, as well as the quality of videos taken by patients and hospital staff. During the study, participants will be assessed at the hospital and again within 4 months to evaluate stroke severity and disability. Researchers will collect data on the ease and acceptability of self-recording, along with video quality. Outcome measures include usable video recordings from patients and hospital workers on Day 0 and Month 3, patient-reported experiences, and clinical assessments of stroke severity and disability over time. The study is sponsored by Centre Hospitalier Universitaire de Nmes and does not involve masking or placebo control.

Clinical Development of a Tool for Optimized Self- and Hetero-diagnosis of Stroke Using Artificial Intelligence: Stage1- Collection of Video-clinical Data in a Pragmatic Situation.