Prediction using patient comparison vs. modeling: A case study for mortality prediction

M. Hoogendoorn, Ali El Hassouni, Kwongyen Mok, Marzyeh Ghassemi, Peter Szolovits

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

Abstract

Information in Electronic Medical Records (EMRs) can be used to generate accurate predictions for the occurrence of a variety of health states, which can contribute to more pro-active interventions. The very nature of EMRs does make the application of off-the-shelf machine learning techniques difficult. In this paper, we study two approaches to making predictions that have hardly been compared in the past: (1) extracting high-level (temporal) features from EMRs and building a predictive model, and (2) defining a patient similarity metric and predicting based on the outcome observed for similar patients. We analyze and compare both approaches on the MIMIC-II ICU dataset to predict patient mortality and find that the patient similarity approach does not scale well and results in a less accurate model (AUC of 0.68) compared to the modeling approach (0.84). We also show that mortality can be predicted within a median of 72 hours.

Original languageEnglish
Title of host publication2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016
PublisherInstitute of Electrical and Electronics Engineers, Inc.
Pages2464-2467
Number of pages4
Volume2016-October
ISBN (Electronic)9781457702204
DOIs
Publication statusPublished - 13 Oct 2016
Event38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016 - Orlando, United States
Duration: 16 Aug 201620 Aug 2016

Conference

Conference38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016
CountryUnited States
CityOrlando
Period16/08/1620/08/16

Fingerprint

Electronic medical equipment
Electronic Health Records
Mortality
Intensive care units
Learning systems
Health
Area Under Curve

Cite this

Hoogendoorn, M., El Hassouni, A., Mok, K., Ghassemi, M., & Szolovits, P. (2016). Prediction using patient comparison vs. modeling: A case study for mortality prediction. In 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016 (Vol. 2016-October, pp. 2464-2467). [7591229] Institute of Electrical and Electronics Engineers, Inc.. https://doi.org/10.1109/EMBC.2016.7591229
Hoogendoorn, M. ; El Hassouni, Ali ; Mok, Kwongyen ; Ghassemi, Marzyeh ; Szolovits, Peter. / Prediction using patient comparison vs. modeling : A case study for mortality prediction. 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016. Vol. 2016-October Institute of Electrical and Electronics Engineers, Inc., 2016. pp. 2464-2467
@inproceedings{234a7a833479436a86129517d18f12d5,
title = "Prediction using patient comparison vs. modeling: A case study for mortality prediction",
abstract = "Information in Electronic Medical Records (EMRs) can be used to generate accurate predictions for the occurrence of a variety of health states, which can contribute to more pro-active interventions. The very nature of EMRs does make the application of off-the-shelf machine learning techniques difficult. In this paper, we study two approaches to making predictions that have hardly been compared in the past: (1) extracting high-level (temporal) features from EMRs and building a predictive model, and (2) defining a patient similarity metric and predicting based on the outcome observed for similar patients. We analyze and compare both approaches on the MIMIC-II ICU dataset to predict patient mortality and find that the patient similarity approach does not scale well and results in a less accurate model (AUC of 0.68) compared to the modeling approach (0.84). We also show that mortality can be predicted within a median of 72 hours.",
author = "M. Hoogendoorn and {El Hassouni}, Ali and Kwongyen Mok and Marzyeh Ghassemi and Peter Szolovits",
year = "2016",
month = "10",
day = "13",
doi = "10.1109/EMBC.2016.7591229",
language = "English",
volume = "2016-October",
pages = "2464--2467",
booktitle = "2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016",
publisher = "Institute of Electrical and Electronics Engineers, Inc.",

}

Hoogendoorn, M, El Hassouni, A, Mok, K, Ghassemi, M & Szolovits, P 2016, Prediction using patient comparison vs. modeling: A case study for mortality prediction. in 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016. vol. 2016-October, 7591229, Institute of Electrical and Electronics Engineers, Inc., pp. 2464-2467, 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016, Orlando, United States, 16/08/16. https://doi.org/10.1109/EMBC.2016.7591229

Prediction using patient comparison vs. modeling : A case study for mortality prediction. / Hoogendoorn, M.; El Hassouni, Ali; Mok, Kwongyen; Ghassemi, Marzyeh; Szolovits, Peter.

2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016. Vol. 2016-October Institute of Electrical and Electronics Engineers, Inc., 2016. p. 2464-2467 7591229.

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Prediction using patient comparison vs. modeling

T2 - A case study for mortality prediction

AU - Hoogendoorn, M.

AU - El Hassouni, Ali

AU - Mok, Kwongyen

AU - Ghassemi, Marzyeh

AU - Szolovits, Peter

PY - 2016/10/13

Y1 - 2016/10/13

N2 - Information in Electronic Medical Records (EMRs) can be used to generate accurate predictions for the occurrence of a variety of health states, which can contribute to more pro-active interventions. The very nature of EMRs does make the application of off-the-shelf machine learning techniques difficult. In this paper, we study two approaches to making predictions that have hardly been compared in the past: (1) extracting high-level (temporal) features from EMRs and building a predictive model, and (2) defining a patient similarity metric and predicting based on the outcome observed for similar patients. We analyze and compare both approaches on the MIMIC-II ICU dataset to predict patient mortality and find that the patient similarity approach does not scale well and results in a less accurate model (AUC of 0.68) compared to the modeling approach (0.84). We also show that mortality can be predicted within a median of 72 hours.

AB - Information in Electronic Medical Records (EMRs) can be used to generate accurate predictions for the occurrence of a variety of health states, which can contribute to more pro-active interventions. The very nature of EMRs does make the application of off-the-shelf machine learning techniques difficult. In this paper, we study two approaches to making predictions that have hardly been compared in the past: (1) extracting high-level (temporal) features from EMRs and building a predictive model, and (2) defining a patient similarity metric and predicting based on the outcome observed for similar patients. We analyze and compare both approaches on the MIMIC-II ICU dataset to predict patient mortality and find that the patient similarity approach does not scale well and results in a less accurate model (AUC of 0.68) compared to the modeling approach (0.84). We also show that mortality can be predicted within a median of 72 hours.

UR - http://www.scopus.com/inward/record.url?scp=85009128708&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85009128708&partnerID=8YFLogxK

U2 - 10.1109/EMBC.2016.7591229

DO - 10.1109/EMBC.2016.7591229

M3 - Conference contribution

VL - 2016-October

SP - 2464

EP - 2467

BT - 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016

PB - Institute of Electrical and Electronics Engineers, Inc.

ER -

Hoogendoorn M, El Hassouni A, Mok K, Ghassemi M, Szolovits P. Prediction using patient comparison vs. modeling: A case study for mortality prediction. In 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2016. Vol. 2016-October. Institute of Electrical and Electronics Engineers, Inc. 2016. p. 2464-2467. 7591229 https://doi.org/10.1109/EMBC.2016.7591229