Predicting Patient No-Shows: Situated Machine Learning with Imperfect Data

Christopher Gyldenkærne; Jakob Grue Simonsen; Gustav From; Morten Hertzum

doi:10.3233/SHTI240728

Predicting Patient No-Shows: Situated Machine Learning with Imperfect Data

Christopher Gyldenkærne
, Jakob Grue Simonsen
, Gustav From
, Morten Hertzum^*

^*Corresponding author

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

3 Citations (Scopus)

97 Downloads (Pure)

Abstract

Patients who do not show up for scheduled appointments are a considerable cost and concern in healthcare. In this study, we predict patient no-shows for outpatient surgery at the endoscopy ward of a hospital in Denmark. The predictions are made by training machine leaning (ML) models on available data, which have been recorded for purposes other than ML, and by doing situated work in the hospital setting to understand local data practices and fine-tune the models. The best performing model (XGBoost with oversampling) predicts no-shows at sensitivity = 0.97, specificity = 0.66, and accuracy = 0.95. Importantly, the situated work engaged local hospital staff in the design process and led to substantial quantitative improvements in the performance of the models. We consider the results promising but acknowledge that they are from a single ward. To transfer the no-show models to other wards and hospitals, the situated work must be redone.

Original language	English
Title of host publication	MIE2024: Proceedings of 34th Medical Informatics Europe Conference : Digital Health and Informatics Innovations for Sustainable Health Care Systems
Editors	John Mantas, Arie Hasman, George Demiris, Kaija Saranto, Michael Marschollek, Theodoros N. Arvanitis, Ivana Ognjanović, Arriel Benis, Parisis Gallos, Emmanouil Zoulias, Elisavet Andrikopoulou
Number of pages	5
Volume	316
Publisher	IOS Press
Publication date	22 Aug 2024
Pages	1598-1602
ISBN (Electronic)	978-1-64368-533-5
DOIs	https://doi.org/10.3233/SHTI240728
Publication status	Published - 22 Aug 2024
Event	34th Medical Informatics Europe Conference: Digital Health and Informatics Innovations for Sustainable Healthcare Systems - Eugenides Foundation, Athen, Greece Duration: 25 Aug 2024 → 29 Aug 2024 Conference number: 34 https://mie2024.org/

Conference

Conference	34th Medical Informatics Europe Conference
Number	34
Location	Eugenides Foundation
Country/Territory	Greece
City	Athen
Period	25/08/2024 → 29/08/2024
Other	The theme of the congress is “Digital Health and Informatics Innovations for Sustainable Health Care Systems” and we aim to advance international co-operation and dissemination of information in Medical Informatics in Europe and to promote research and development in medical informatics. The scientific programme will consist of keynote speeches by distinguished scholars, peer reviewed oral presentations, e-poster presentations, panel discussions, workshops, demonstrations, and tutorials showing the advancement of Biomedical and Health Informatics across the globe.
Internet address	https://mie2024.org/

Series	Studies in Health Technology and Informatics
Volume	316
ISSN	0926-9630

Keywords

Healthcare
Machine learning
Participatory design
Patient no-shows

Access to Document

10.3233/SHTI240728Licence: CC BY-NC

Predicting Patient No-Shows: Situated Machine Learning with Imperfect DataFinal published version, 168 KBLicence: CC BY-NC

Citation Styles

Gyldenkærne, C., Simonsen, J. G., From, G., & Hertzum, M. (2024). Predicting Patient No-Shows: Situated Machine Learning with Imperfect Data. In J. Mantas, A. Hasman, G. Demiris, K. Saranto, M. Marschollek, T. N. Arvanitis, I. Ognjanović, A. Benis, P. Gallos, E. Zoulias, & E. Andrikopoulou (Eds.), MIE2024: Proceedings of 34th Medical Informatics Europe Conference: Digital Health and Informatics Innovations for Sustainable Health Care Systems (Vol. 316, pp. 1598-1602). IOS Press. https://doi.org/10.3233/SHTI240728

Gyldenkærne, Christopher ; Simonsen, Jakob Grue ; From, Gustav et al. / Predicting Patient No-Shows: Situated Machine Learning with Imperfect Data. MIE2024: Proceedings of 34th Medical Informatics Europe Conference: Digital Health and Informatics Innovations for Sustainable Health Care Systems. editor / John Mantas ; Arie Hasman ; George Demiris ; Kaija Saranto ; Michael Marschollek ; Theodoros N. Arvanitis ; Ivana Ognjanović ; Arriel Benis ; Parisis Gallos ; Emmanouil Zoulias ; Elisavet Andrikopoulou. Vol. 316 IOS Press, 2024. pp. 1598-1602 (Studies in Health Technology and Informatics, Vol. 316).

@inproceedings{54bd626f5630450592c0caea61a91f1f,

title = "Predicting Patient No-Shows: Situated Machine Learning with Imperfect Data",

abstract = "Patients who do not show up for scheduled appointments are a considerable cost and concern in healthcare. In this study, we predict patient no-shows for outpatient surgery at the endoscopy ward of a hospital in Denmark. The predictions are made by training machine leaning (ML) models on available data, which have been recorded for purposes other than ML, and by doing situated work in the hospital setting to understand local data practices and fine-tune the models. The best performing model (XGBoost with oversampling) predicts no-shows at sensitivity = 0.97, specificity = 0.66, and accuracy = 0.95. Importantly, the situated work engaged local hospital staff in the design process and led to substantial quantitative improvements in the performance of the models. We consider the results promising but acknowledge that they are from a single ward. To transfer the no-show models to other wards and hospitals, the situated work must be redone.",

keywords = "Healthcare, Machine Learning, Participatory Design, patient no-shows, Healthcare, Machine learning, Participatory design, Patient no-shows",

author = "Christopher Gyldenk{\ae}rne and Simonsen, \{Jakob Grue\} and Gustav From and Morten Hertzum",

year = "2024",

month = aug,

day = "22",

doi = "10.3233/SHTI240728",

language = "English",

volume = "316",

series = "Studies in Health Technology and Informatics",

publisher = "IOS Press",

pages = "1598--1602",

editor = "John Mantas and Arie Hasman and George Demiris and Kaija Saranto and Michael Marschollek and Arvanitis, \{Theodoros N.\} and Ivana Ognjanovi{\'c} and Arriel Benis and Parisis Gallos and Emmanouil Zoulias and Elisavet Andrikopoulou",

booktitle = "MIE2024: Proceedings of 34th Medical Informatics Europe Conference",

note = "34th Medical Informatics Europe Conference : Digital Health and Informatics Innovations for Sustainable Healthcare Systems, MIE2024 ; Conference date: 25-08-2024 Through 29-08-2024",

url = "https://mie2024.org/",

}

Gyldenkærne, C, Simonsen, JG, From, G & Hertzum, M 2024, Predicting Patient No-Shows: Situated Machine Learning with Imperfect Data. in J Mantas, A Hasman, G Demiris, K Saranto, M Marschollek, TN Arvanitis, I Ognjanović, A Benis, P Gallos, E Zoulias & E Andrikopoulou (eds), MIE2024: Proceedings of 34th Medical Informatics Europe Conference: Digital Health and Informatics Innovations for Sustainable Health Care Systems. vol. 316, IOS Press, Studies in Health Technology and Informatics, vol. 316, pp. 1598-1602, 34th Medical Informatics Europe Conference, Athen, Greece, 25/08/2024. https://doi.org/10.3233/SHTI240728

Predicting Patient No-Shows: Situated Machine Learning with Imperfect Data. / Gyldenkærne, Christopher; Simonsen, Jakob Grue; From, Gustav et al.
MIE2024: Proceedings of 34th Medical Informatics Europe Conference: Digital Health and Informatics Innovations for Sustainable Health Care Systems. ed. / John Mantas; Arie Hasman; George Demiris; Kaija Saranto; Michael Marschollek; Theodoros N. Arvanitis; Ivana Ognjanović; Arriel Benis; Parisis Gallos; Emmanouil Zoulias; Elisavet Andrikopoulou. Vol. 316 IOS Press, 2024. p. 1598-1602 (Studies in Health Technology and Informatics, Vol. 316).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - Predicting Patient No-Shows: Situated Machine Learning with Imperfect Data

AU - Gyldenkærne, Christopher

AU - Simonsen, Jakob Grue

AU - From, Gustav

AU - Hertzum, Morten

N1 - Conference code: 34

PY - 2024/8/22

Y1 - 2024/8/22

N2 - Patients who do not show up for scheduled appointments are a considerable cost and concern in healthcare. In this study, we predict patient no-shows for outpatient surgery at the endoscopy ward of a hospital in Denmark. The predictions are made by training machine leaning (ML) models on available data, which have been recorded for purposes other than ML, and by doing situated work in the hospital setting to understand local data practices and fine-tune the models. The best performing model (XGBoost with oversampling) predicts no-shows at sensitivity = 0.97, specificity = 0.66, and accuracy = 0.95. Importantly, the situated work engaged local hospital staff in the design process and led to substantial quantitative improvements in the performance of the models. We consider the results promising but acknowledge that they are from a single ward. To transfer the no-show models to other wards and hospitals, the situated work must be redone.

AB - Patients who do not show up for scheduled appointments are a considerable cost and concern in healthcare. In this study, we predict patient no-shows for outpatient surgery at the endoscopy ward of a hospital in Denmark. The predictions are made by training machine leaning (ML) models on available data, which have been recorded for purposes other than ML, and by doing situated work in the hospital setting to understand local data practices and fine-tune the models. The best performing model (XGBoost with oversampling) predicts no-shows at sensitivity = 0.97, specificity = 0.66, and accuracy = 0.95. Importantly, the situated work engaged local hospital staff in the design process and led to substantial quantitative improvements in the performance of the models. We consider the results promising but acknowledge that they are from a single ward. To transfer the no-show models to other wards and hospitals, the situated work must be redone.

KW - Healthcare

KW - Machine Learning

KW - Participatory Design

KW - patient no-shows

KW - Healthcare

KW - Machine learning

KW - Participatory design

KW - Patient no-shows

U2 - 10.3233/SHTI240728

DO - 10.3233/SHTI240728

M3 - Article in proceedings

VL - 316

T3 - Studies in Health Technology and Informatics

SP - 1598

EP - 1602

BT - MIE2024: Proceedings of 34th Medical Informatics Europe Conference

A2 - Mantas, John

A2 - Hasman, Arie

A2 - Demiris, George

A2 - Saranto, Kaija

A2 - Marschollek, Michael

A2 - Arvanitis, Theodoros N.

A2 - Ognjanović, Ivana

A2 - Benis, Arriel

A2 - Gallos, Parisis

A2 - Zoulias, Emmanouil

A2 - Andrikopoulou, Elisavet

PB - IOS Press

T2 - 34th Medical Informatics Europe Conference

Y2 - 25 August 2024 through 29 August 2024

ER -

Gyldenkærne C, Simonsen JG, From G, Hertzum M. Predicting Patient No-Shows: Situated Machine Learning with Imperfect Data. In Mantas J, Hasman A, Demiris G, Saranto K, Marschollek M, Arvanitis TN, Ognjanović I, Benis A, Gallos P, Zoulias E, Andrikopoulou E, editors, MIE2024: Proceedings of 34th Medical Informatics Europe Conference: Digital Health and Informatics Innovations for Sustainable Health Care Systems. Vol. 316. IOS Press. 2024. p. 1598-1602. (Studies in Health Technology and Informatics, Vol. 316). doi: 10.3233/SHTI240728