CHART-ADAPT: Enabling Actionable Analytics at the Critical Care Unit Bedside

Moss, L. , Shaw, M., Piper, I., Kinsella, J. and Hawthorne, C. (2021) CHART-ADAPT: Enabling Actionable Analytics at the Critical Care Unit Bedside. In: 2021 34th International Symposium on Computer-Based Medical Systems (CBMS), 7-9 June 2021, ISBN 9781665441216 (doi: 10.1109/CBMS52027.2021.00032)

Full text not currently available from Enlighten.


The increasing amount of complex patient data stored and collected in critical care units, alongside future use of genomic and proteomic data at the patient's bedside, means that state-of-the-art data analysis techniques have the potential to revolutionise personalised care. However, technical, organisational and practical challenges remain; tools and technologies are required to enable the bedside analysis of this big data and translate the often computationally complex results into meaningful clinical impact at the point of care. This work describes a technological framework, CHART-ADAPT, implemented in a real-world critical care setting, to stream patient data to a high performance computing platform capable of the live deployment of sophisticated algorithms and models and presenting results at the patient bedside in real-time. Evaluation consisted of a prospective observational study carried out in a single-centre, adult Neurointensive Care Unit on 831 patients. Further, CHART-ADAPT was used to answer two important Neurointensive Care clinical research questions. Over the period of a year, live patient data was processed and results of models successfully reintegrated with existing monitoring equipment. Median timings of data flow between components ranged from 0.04 to 341 seconds and model computation was 40 seconds. No adverse outcomes were noted on either the hospital network, or the Neurointensive Care service. Implementation challenges were overcome and previously unavailable models and techniques can now be used at the patient bedside. CHART-ADAPT enables actionable analytics in clinically meaningful timescales and significantly improves clinicians' ability to develop, and gain access to, state-of-the-art models.

Item Type:Conference Proceedings
Glasgow Author(s) Enlighten ID:Piper, Dr Ian and Hawthorne, Dr Christopher and Kinsella, Professor John and Moss, Dr Laura and Shaw, Dr Martin
Authors: Moss, L., Shaw, M., Piper, I., Kinsella, J., and Hawthorne, C.
College/School:College of Medical Veterinary and Life Sciences > School of Medicine, Dentistry & Nursing

University Staff: Request a correction | Enlighten Editors: Update this record

Project CodeAward NoProject NamePrincipal InvestigatorFunder's NameFunder RefLead Dept
171524Connecting healthcare and research through data analysis provisioning technology (CHART-ADAPT)John KinsellaInnovate UK (INNOVATE)102113School Of Medicine