ICU sedation, delirium & recovery
A user evaluation of speech/phrase recognition software in critically ill patients: a DECIDE-AI feasibility study
Authors' conclusion
This feasibility evaluation of a novel speech/phrase recognition app using SRAVI demonstrated a good correlation between spoken phrases and app recognition.
Study context
- Population
- Critically ill patients with tracheostomies and speech impairments, tertiary hospital critical care unit, northwest England
- Participants
- 14 patients · 516 phrase identifiable recordings
- Exposure
- Deep neural networks (DNN) vs dynamic time warping (DTW) processing methods in the speech/phrase recognition app for voice impaired (SRAVI) speech/phrase recognition app
- Outcome
- Phrase recognition accuracy (top three ranked phrases and rank 1)
Methodological detailsDECIDE-AI, single-centre
- DECIDE-AI
- Single-centre
Main finding
As reported in the abstract
Recognition accuracy
| Group | Total recognition accuracy (ranks 1–3) | Rank 1 accuracy |
|---|---|---|
| DNN | 86% | 75% |
| DTW | 74% | 48% |
Total recognition accuracy (ranks 1–3)
DNN86%
DTW74%
Rank 1 accuracy
DNN75%
DTW48%
Observational data: associations, not effects.
Original publication
A user evaluation of speech/phrase recognition software in critically ill patients: a DECIDE-AI feasibility study. Crit Care. Published 10 July 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →