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

James Liley; Gergo Bohner; Samuel R. Emerson; Bilal A. Mateen; Katie Borland; David Carr; Scott Heald; Samuel D. Oduro; Jill Ireland; Keith Moffat; Rachel Porteous; Stephen Riddell; Simon Rogers; Ioanna Thoma; Nathan Cunningham; Chris Holmes; Katrina Payne; Sebastian J. Vollmer; Catalina A. Vallejos; Louis J. M. Aslett

doi:10.1038/s41746-024-01250-1

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

James Liley^*, Gergo Bohner, Samuel R. Emerson, Bilal A. Mateen, Katie Borland, David Carr, Scott Heald, Samuel D. Oduro, Jill Ireland, Keith Moffat, Rachel Porteous, Stephen Riddell, Simon Rogers, Ioanna Thoma, Nathan Cunningham, Chris Holmes, Katrina Payne, Sebastian J. Vollmer, Catalina A. Vallejos^*, Louis J. M. Aslett^*

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

Abstract

Emergency admissions (EA), where a patient requires urgent in-hospital care, are a major challenge for healthcare systems. The development of risk prediction models can partly alleviate this problem by supporting primary care interventions and public health planning. Here, we introduce SPARRAv4, a predictive score for EA risk that will be deployed nationwide in Scotland. SPARRAv4 was derived using supervised and unsupervised machine-learning methods applied to routinely collected electronic health records from approximately 4.8M Scottish residents (2013-18). We demonstrate improvements in discrimination and calibration with respect to previous scores deployed in Scotland, as well as stability over a 3-year timeframe. Our analysis also provides insights about the epidemiology of EA risk in Scotland, by studying predictive performance across different population sub-groups and reasons for admission, as well as by quantifying the effect of individual input features. Finally, we discuss broader challenges including reproducibility and how to safely update risk prediction models that are already deployed at population level.

Original language	English
Article number	277
Number of pages	13
Journal	npj Digital Medicine
Volume	7
DOIs	https://doi.org/10.1038/s41746-024-01250-1
Publication status	Published - 23 Oct 2024

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1038/s41746-024-01250-1Licence: CC BY

Liley-2024-Development-and-assessment-npjDigMed-7-277-CCBY
Copyright © The Author(s) 2024. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Final published version, 1.38 MBLicence: CC BY

Development and assessment of a machine learning tool for predicting emergency admission in Scotland (dataset)
Moffat, K. (Creator), GitHub, 2024
https://github.com/jamesliley/SPARRAv4
Dataset: Software

Cite this

Liley, J., Bohner, G., Emerson, S. R., Mateen, B. A., Borland, K., Carr, D., Heald, S., Oduro, S. D., Ireland, J., Moffat, K., Porteous, R., Riddell, S., Rogers, S., Thoma, I., Cunningham, N., Holmes, C., Payne, K., Vollmer, S. J., Vallejos, C. A., & Aslett, L. J. M. (2024). Development and assessment of a machine learning tool for predicting emergency admission in Scotland. npj Digital Medicine, 7, Article 277. https://doi.org/10.1038/s41746-024-01250-1

@article{044ea15a645c443dbbec4b6127df492e,

title = "Development and assessment of a machine learning tool for predicting emergency admission in Scotland",

abstract = "Emergency admissions (EA), where a patient requires urgent in-hospital care, are a major challenge for healthcare systems. The development of risk prediction models can partly alleviate this problem by supporting primary care interventions and public health planning. Here, we introduce SPARRAv4, a predictive score for EA risk that will be deployed nationwide in Scotland. SPARRAv4 was derived using supervised and unsupervised machine-learning methods applied to routinely collected electronic health records from approximately 4.8M Scottish residents (2013-18). We demonstrate improvements in discrimination and calibration with respect to previous scores deployed in Scotland, as well as stability over a 3-year timeframe. Our analysis also provides insights about the epidemiology of EA risk in Scotland, by studying predictive performance across different population sub-groups and reasons for admission, as well as by quantifying the effect of individual input features. Finally, we discuss broader challenges including reproducibility and how to safely update risk prediction models that are already deployed at population level.",

author = "James Liley and Gergo Bohner and Emerson, {Samuel R.} and Mateen, {Bilal A.} and Katie Borland and David Carr and Scott Heald and Oduro, {Samuel D.} and Jill Ireland and Keith Moffat and Rachel Porteous and Stephen Riddell and Simon Rogers and Ioanna Thoma and Nathan Cunningham and Chris Holmes and Katrina Payne and Vollmer, {Sebastian J.} and Vallejos, {Catalina A.} and Aslett, {Louis J. M.}",

note = "Funding: L.J.M.A., and S.J.V. were partially supported by Health Data Research UK, an initiative funded by UK Research and Innovation, Department of Health and Social Care (England), the devolved administrations, and leading medical research charities; S.J.V., N.C., and G.B. were partially supported by the University of Warwick Impact Fund. S.R.E. is funded by the EPSRC doctoral training partnership (DTP) at Durham University, grant reference EP/R513039/1; L.J.M.A. was partially supported by a Health Programme Fellowship at The Alan Turing Institute; CAV was supported by a Chancellor{\textquoteright}s Fellowship provided by the University of Edinburgh. ",

year = "2024",

month = oct,

day = "23",

doi = "10.1038/s41746-024-01250-1",

language = "English",

volume = "7",

journal = "npj Digital Medicine",

issn = "2398-6352",

publisher = "Springer Nature",

}

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, vol. 7, 277. https://doi.org/10.1038/s41746-024-01250-1

TY - JOUR

T1 - Development and assessment of a machine learning tool for predicting emergency admission in Scotland

AU - Liley, James

AU - Bohner, Gergo

AU - Emerson, Samuel R.

AU - Mateen, Bilal A.

AU - Borland, Katie

AU - Carr, David

AU - Heald, Scott

AU - Oduro, Samuel D.

AU - Ireland, Jill

AU - Moffat, Keith

AU - Porteous, Rachel

AU - Riddell, Stephen

AU - Rogers, Simon

AU - Thoma, Ioanna

AU - Cunningham, Nathan

AU - Holmes, Chris

AU - Payne, Katrina

AU - Vollmer, Sebastian J.

AU - Vallejos, Catalina A.

AU - Aslett, Louis J. M.

N1 - Funding: L.J.M.A., and S.J.V. were partially supported by Health Data Research UK, an initiative funded by UK Research and Innovation, Department of Health and Social Care (England), the devolved administrations, and leading medical research charities; S.J.V., N.C., and G.B. were partially supported by the University of Warwick Impact Fund. S.R.E. is funded by the EPSRC doctoral training partnership (DTP) at Durham University, grant reference EP/R513039/1; L.J.M.A. was partially supported by a Health Programme Fellowship at The Alan Turing Institute; CAV was supported by a Chancellor’s Fellowship provided by the University of Edinburgh.

PY - 2024/10/23

Y1 - 2024/10/23

N2 - Emergency admissions (EA), where a patient requires urgent in-hospital care, are a major challenge for healthcare systems. The development of risk prediction models can partly alleviate this problem by supporting primary care interventions and public health planning. Here, we introduce SPARRAv4, a predictive score for EA risk that will be deployed nationwide in Scotland. SPARRAv4 was derived using supervised and unsupervised machine-learning methods applied to routinely collected electronic health records from approximately 4.8M Scottish residents (2013-18). We demonstrate improvements in discrimination and calibration with respect to previous scores deployed in Scotland, as well as stability over a 3-year timeframe. Our analysis also provides insights about the epidemiology of EA risk in Scotland, by studying predictive performance across different population sub-groups and reasons for admission, as well as by quantifying the effect of individual input features. Finally, we discuss broader challenges including reproducibility and how to safely update risk prediction models that are already deployed at population level.

AB - Emergency admissions (EA), where a patient requires urgent in-hospital care, are a major challenge for healthcare systems. The development of risk prediction models can partly alleviate this problem by supporting primary care interventions and public health planning. Here, we introduce SPARRAv4, a predictive score for EA risk that will be deployed nationwide in Scotland. SPARRAv4 was derived using supervised and unsupervised machine-learning methods applied to routinely collected electronic health records from approximately 4.8M Scottish residents (2013-18). We demonstrate improvements in discrimination and calibration with respect to previous scores deployed in Scotland, as well as stability over a 3-year timeframe. Our analysis also provides insights about the epidemiology of EA risk in Scotland, by studying predictive performance across different population sub-groups and reasons for admission, as well as by quantifying the effect of individual input features. Finally, we discuss broader challenges including reproducibility and how to safely update risk prediction models that are already deployed at population level.

U2 - 10.1038/s41746-024-01250-1

DO - 10.1038/s41746-024-01250-1

M3 - Article

SN - 2398-6352

VL - 7

JO - npj Digital Medicine

JF - npj Digital Medicine

M1 - 277

ER -

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

Abstract

UN SDGs

Access to Document

Fingerprint

Datasets

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

Cite this