Development and assessment of a machine learning tool for predicting emergency admission in Scotland

Aug 5, 2026

Upload by Aria

Emergency admissions (EA) requiring urgent in-hospital care pose major challenges for healthcare systems. The document presents SPARRAv4, a predictive risk score for EA in Scotland designed for nationwide deployment. SPARRAv4 is derived from supervised and unsupervised machine-learning on routinely collected electronic health records of about 4.8M residents (2013–2018). Results show improved discrimination and calibration versus prior Scottish scores, plus stability across three years.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages31

Reading Time1 h 18 min

AccessFree to download

Abstract

Introduction

Emergency admissions and healthcare challenge

Predictive risk models and potential interventions

SPARRA and motivation for local model development

Edinburgh Research Explorer Development and assessment of a machine learning tool for predicting emergency admission in Scotland Citation for published version: Liley, J, Bohner, G, Emerson, SR, Mateen, BA, Borland, K, Carr, D, Heald, S, Oduro, SD, Ireland, J, Moffat, K, Porteous, R, Riddell, S, Rogers, S, Thoma, I, Cunningham, N, Holmes, C, Payne, K, Vollmer, SJ, Vallejos, CA & Aslett, LJM 2024, 'Development and assessment of a machine learning tool for predicting emergency admission in Scotland', npj Digital Medicine. [https://doi.org/10.1038/s41746-024-01250-1](https://doi.org/10.1038/s41746-024-01250-1) Digital Object Identifier (DOI): 10.1038/s41746-024-01250-1 Link: Link to publication record in Edinburgh Research Explorer Document Version: Peer reviewed version Published In: npj Digital Medicine General rights Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights. Take down policy The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please [contact openaccess@ed.ac.uk](contact openaccess@ed.ac.uk) providing details, and we will remove access to the work immediately and investigate your claim. Download date: 25. Nov. 2025 1 Development and assessment of a machine learning tool 2 for predicting emergency admission in Scotland 3 James Liley 1,2,3,?,*, Gergo Bohner 1,4,?, Samuel R. Emerson3 , 4 Bilal A. Mateen 1,5 , Katie Borland6 , David Carr6 , Scott Heald6 , 5 Samuel D. Oduro6b , Jill Ireland6 , Keith Mo􀀋at6,7 , Rachel Porteous6 , Stephen 6 Riddell6b , Simon Rogers8 , Ioanna Thoma 1,2 , Nathan Cunningham 1,9 , Chris 7 Holmes 1,10 , Katrina Payne 1 , Sebastian J. Vollmer 1,4 , 8 Catalina A. Vallejos 1,2,*,y, and Louis J. M. Aslett 1,3,*,y 9 1 Alan Turing Institute, London, UK 10 2 MRC Human Genetics Unit, Institute of Genetics and Cancer, University of 11 Edinburgh, UK 12 3 Department of Mathematical Sciences, Durham University, UK 13 4 Mathematics Institute, University of Warwick, UK 14 5 Institute of Health Informatics, University College London, UK, and 15 Wellcome Trust, London, UK 16 6 Public Health Scotland (PHS) . (b): former employee 17 7 University of St Andrews, UK 18 8 NHS National Services Scotland, UK 19 9 Department of Statistics, University of Warwick, UK 20 10 Department of Statistics, University of Oxford, UK 21 ? Equal contribution 22 yEqual supervision 23 * Corresponding 24 25 JL: [james.liley@durham.ac.uk](james.liley@durham.ac.uk) 26 CAV: [catalina.vallejos@ed.ac.uk](catalina.vallejos@ed.ac.uk) 27 LJMA: [louis.aslett@durham.ac.uk](louis.aslett@durham.ac.uk) 28 September 2, 2024 29 Abstract 30 Emergency admissions (EA), where a patient requires urgent in-hospital care, are 31 a major challenge for healthcare systems. The development of risk prediction models 32 can partly alleviate this problem by supporting primary care interventions and public 33 health planning. Here, we introduce SPARRAv4, a predictive score for EA risk that 34 will be deployed nationwide in Scotland. SPARRAv4 was derived using supervised and 35 unsupervised machine-learning methods applied to routinely collected electronic health 36 records from approximately 4.8M Scottish residents (2013-18) . We demonstrate im- 37 provements in discrimination and calibration with respect to previous scores deployed 38 in Scotland, as well as stability over a 3-year timeframe. Our analysis also provides 39 insights about the epidemiology of EA risk in Scotland, by studying predictive perfor- 40 mance across di􀀋erent population sub-groups and reasons for admissio

FAQs

What problem does the document address?

It addresses emergency admissions in which patients need urgent in-hospital care, a major challenge for healthcare systems.

What is SPARRAv4 and what data was used to build it?

SPARRAv4 is a predictive score for EA risk in Scotland, derived using machine-learning methods on routinely collected electronic health records from about 4.8M residents (2013–2018).

How does SPARRAv4 perform compared with earlier tools?

The analysis demonstrates improvements in discrimination and calibration relative to previous deployed Scotland scores, and stability over a three-year timeframe.

Aug 5, 2026

Upload by Aria