icu_heart_accelerometer_study1
Research type
Research Study
Full title
ICU-Heart Advancing Artifact Detection Informatics
IRAS ID
302750
Contact name
Annemarie Docherty
Contact email
Sponsor organisation
University of Edinburgh
Duration of Study in the UK
0 years, 3 months, 1 days
Research summary
Current best practice for cleaning ICU high resolution monitoring data of artifacts still rely upon hiring experienced ICU research nursing staff to manually annotate data streams, highlighting periods of “invalid data”. Such costly and time-consuming methods impede the ability to share high resolution data and conduct rapid multicenter intensive care research. Application of informatics based approaches including the use of machine learning methods have in recent years shown promise, yet there remains a lack of uptake of these methods in general intensive care most likely due to a paucity of studies systematically comparing methods both against each other and with “gold-standard” research nurse based validation.
This proposal aims to fill an unmet need to develop a state of the art infrastructure for comparing ICU artifact detection methods and to generate pilot “labelled” data suitable for fostering collaborations between academia, clinical practice and industry based approaches to addressing this key bottle-neck. Specifically, advancing the feasibility of use of higher resolution monitoring data in collaborative fast-track ICU research.
The first step is to collect some pilot data for fostering initial collaborations across academia and industry.
REC name
Scotland A: Adults with Incapacity only
REC reference
21/SS/0079
Date of REC Opinion
16 Dec 2021
REC opinion
Favourable Opinion