SAMueL-2
Research type
Research Study
Full title
Stroke Audit Machine Learning
IRAS ID
322303
Contact name
Michael Allen
Contact email
Sponsor organisation
Senior Research Governance Officer, University of Exeter
Duration of Study in the UK
0 years, 9 months, 31 days
Research summary
Stroke is the commonest cause of severe adult disability (Xu XM et al. 2018) and outcomes can be significantly improved for selected patients though the use of time-critical treatments such as thrombolysis, which dissolves blood clots (NICE 2019). For thrombolysis to be useful, it needs to be given as soon after the stroke as possible, but it is not appropriate for all patients and can be risky. The Sentinel Stroke National Audit Programme (SSNAP) shows that the use of thrombolysis varies hugely (ranging from 5% and 25% between hospitals), even for patients with similar treatment pathways and with similar characteristics. In some hospitals it is rarely used, but in others it is given to a quarter of stroke patients (RCP 2013). One of the main reasons for uneven use of thrombolysis is clinician decision-making.
The overarching aim of the SAMueL-2 project is to build machine learning (ML) tools with the aim of supporting the optimal implementation of thrombolysis. ML is a type of artificial intelligence (AI) that learns patterns from historical data to predict new outcomes. SAMueL-2 will investigate how a ML-based approach that further enhance the analytical capabilities of SSNAP, can inform clinical practice, and be designed and adapted to suit physicians, in order to support their decision-making. The qualitative component will explore how physicians at a particular hospital use thrombolysis, in order to identify how use of ML applied to the national stroke audit will be of greatest value in reducing variation in thrombolysis use. We will learn from current initiatives in the NHS which are aiming to reduce variation in the treatment of thrombolysis, as well as investigate physicians’ and other stroke staffs’ experiences and understandings of the use of ML in healthcare. This will help us anticipate any barriers to implementation and change in practice.
REC name
East of England - Essex Research Ethics Committee
REC reference
23/EE/0124
Date of REC Opinion
25 Jul 2023
REC opinion
Further Information Favourable Opinion