Predicting 120-day hospital readmission using medical administrative patient data

Marijn Van Wingerden; Jelle De Boer; Eric Postma

Predicting 120-day hospital readmission using medical administrative patient data

Marijn Van Wingerden, Jelle De Boer, Eric Postma

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Scientific › peer-review

Abstract

Hospitals and health-care insurers routinely use models to predict patient readmission, extrapolating from historical data. Subsequently, the predicted quantities can be used for contracting and pricing negotiations between these hospitals and healthcare insurers. The Dutch healthcare system uses unique standardized Care Trajectories (so-called DBCs) for administration and billing of care. Here, we compared supervised machine learning methods on predicting 120-day readmission as an operationally significant metric. We used administrative patient data from 21 common Care Trajectories, in combination with demographic information. A lightGBM model using undersampling to tackle class imbalance yielded an AUROC score of 0.86 and provided the highest recall score (73.8%).

Original language	English
Title of host publication	Proceedings of the Reference AI & ML Conference for Belgium, Netherlands & Luxemburg
Volume	2491
Publication status	Published - 1 Jan 2019
Event	28th Belgian Dutch Conference on Machine Learning - Brussels, Belgium Duration: 6 Nov 2019 → 8 Nov 2019 Conference number: 28 https://bnaic19.brussels/benelearn-call-for-papers/

Publication series

Name	CEUR Workshop Proceedings
Publisher	RWTH Aachen, Informatik 5
ISSN (Print)	1613-0073

Conference

Conference	28th Belgian Dutch Conference on Machine Learning
Abbreviated title	Benelearn
Country/Territory	Belgium
City	Brussels
Period	6/11/19 → 8/11/19
Internet address	https://bnaic19.brussels/benelearn-call-for-papers/

Keywords

Administrative patient data
Care trajectory
Health analytics
Hospital readmission

Cite this

@inproceedings{d5a91be288c94cdfa9c3907c8076538a,

title = "Predicting 120-day hospital readmission using medical administrative patient data",

abstract = "Hospitals and health-care insurers routinely use models to predict patient readmission, extrapolating from historical data. Subsequently, the predicted quantities can be used for contracting and pricing negotiations between these hospitals and healthcare insurers. The Dutch healthcare system uses unique standardized Care Trajectories (so-called DBCs) for administration and billing of care. Here, we compared supervised machine learning methods on predicting 120-day readmission as an operationally significant metric. We used administrative patient data from 21 common Care Trajectories, in combination with demographic information. A lightGBM model using undersampling to tackle class imbalance yielded an AUROC score of 0.86 and provided the highest recall score (73.8%).",

keywords = "Administrative patient data, Care trajectory, Health analytics, Hospital readmission",

author = "{Van Wingerden}, Marijn and {De Boer}, Jelle and Eric Postma",

note = "Additional event: 31st Benelux Conference on Artificial Intelligence and the 28th Belgian Dutch Conference on Machine Learning, BNAIC/BENELEARN 2019 6/11/19 → 8/11/19 Brussels, Belgium; 28th Belgian Dutch Conference on Machine Learning, Benelearn ; Conference date: 06-11-2019 Through 08-11-2019",

year = "2019",

month = jan,

day = "1",

language = "English",

volume = "2491",

series = "CEUR Workshop Proceedings",

publisher = "RWTH Aachen, Informatik 5",

booktitle = "Proceedings of the Reference AI & ML Conference for Belgium, Netherlands & Luxemburg",

url = "https://bnaic19.brussels/benelearn-call-for-papers/",

}

Predicting 120-day hospital readmission using medical administrative patient data. / Van Wingerden, Marijn; De Boer, Jelle; Postma, Eric.
Proceedings of the Reference AI & ML Conference for Belgium, Netherlands & Luxemburg. Vol. 2491 2019. (CEUR Workshop Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Predicting 120-day hospital readmission using medical administrative patient data

AU - Van Wingerden, Marijn

AU - De Boer, Jelle

AU - Postma, Eric

N1 - Conference code: 28

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Hospitals and health-care insurers routinely use models to predict patient readmission, extrapolating from historical data. Subsequently, the predicted quantities can be used for contracting and pricing negotiations between these hospitals and healthcare insurers. The Dutch healthcare system uses unique standardized Care Trajectories (so-called DBCs) for administration and billing of care. Here, we compared supervised machine learning methods on predicting 120-day readmission as an operationally significant metric. We used administrative patient data from 21 common Care Trajectories, in combination with demographic information. A lightGBM model using undersampling to tackle class imbalance yielded an AUROC score of 0.86 and provided the highest recall score (73.8%).

AB - Hospitals and health-care insurers routinely use models to predict patient readmission, extrapolating from historical data. Subsequently, the predicted quantities can be used for contracting and pricing negotiations between these hospitals and healthcare insurers. The Dutch healthcare system uses unique standardized Care Trajectories (so-called DBCs) for administration and billing of care. Here, we compared supervised machine learning methods on predicting 120-day readmission as an operationally significant metric. We used administrative patient data from 21 common Care Trajectories, in combination with demographic information. A lightGBM model using undersampling to tackle class imbalance yielded an AUROC score of 0.86 and provided the highest recall score (73.8%).

KW - Administrative patient data

KW - Care trajectory

KW - Health analytics

KW - Hospital readmission

UR - http://www.mendeley.com/research/predicting-120day-hospital-readmission-using-medical-administrative-patient-data

M3 - Conference contribution

VL - 2491

T3 - CEUR Workshop Proceedings

BT - Proceedings of the Reference AI & ML Conference for Belgium, Netherlands & Luxemburg

T2 - 28th Belgian Dutch Conference on Machine Learning

Y2 - 6 November 2019 through 8 November 2019

ER -

Predicting 120-day hospital readmission using medical administrative patient data

Abstract

Publication series

Conference

Keywords

Other files and links

Fingerprint

Cite this