Development and validation of an artificial neural network algorithm to predict mortality and admission to hospital for heart failure after myocardial infarction : a nationwide population-based study
(2022) In The Lancet Digital Health 4(1). p.37-45- Abstract
Background: Patients have an estimated mortality of 15–20% within the first year following myocardial infarction and one in four patients who survive myocardial infarction will develop heart failure, severely reducing quality of life and increasing the risk of long-term mortality. We aimed to establish the accuracy of an artificial neural network (ANN) algorithm in predicting 1-year mortality and admission to hospital for heart failure after myocardial infarction. Methods: In this nationwide population-based study, we used data for all patients admitted to hospital for myocardial infarction and discharged alive from a coronary care unit in Sweden (n=139 288) between Jan 1, 2008, and April 1, 2017, from the Swedish Web system for... (More)
Background: Patients have an estimated mortality of 15–20% within the first year following myocardial infarction and one in four patients who survive myocardial infarction will develop heart failure, severely reducing quality of life and increasing the risk of long-term mortality. We aimed to establish the accuracy of an artificial neural network (ANN) algorithm in predicting 1-year mortality and admission to hospital for heart failure after myocardial infarction. Methods: In this nationwide population-based study, we used data for all patients admitted to hospital for myocardial infarction and discharged alive from a coronary care unit in Sweden (n=139 288) between Jan 1, 2008, and April 1, 2017, from the Swedish Web system for Enhancement and Development of Evidence-based care in Heart disease Evaluated According to Recommended Therapies (SWEDEHEART) nationwide registry; these patients were randomly divided into training (80%) and testing (20%) datasets. We developed an ANN using 21 variables (including age, sex, medical history, previous medications, in-hospital characteristics, and discharge medications) associated with the outcomes of interest with a back-propagation algorithm in the training dataset and tested it in the testing dataset. The ANN algorithm was then validated in patients with incident myocardial infarction enrolled in the Western Denmark Heart Registry (external validation cohort) between Jan 1, 2008, and Dec 31, 2016. The predictive ability of the model was evaluated using area under the receiver operating characteristic curve (AUROC) and Youden's index was established as a means of identifying an empirical dichotomous cutoff, allowing further evaluation of model performance. Findings: 139 288 patients who were admitted to hospital for myocardial infarction in the SWEDEHEART registry were randomly divided into a training dataset of 111 558 (80%) patients and a testing dataset of 27 730 (20%) patients. 30 971 patients with myocardial infarction who were enrolled in the Western Denmark Heart Registry were included in the external validation cohort. A first event, either all-cause mortality or admission to hospital for heart failure 1 year after myocardial infarction, occurred in 32 308 (23·2%) patients in the testing and training cohorts only. For 1-year all-cause mortality, the ANN had an AUROC of 0·85 (95% CI 0·84–0·85) in the testing dataset and 0·84 (0·83–0·84) in the external validation cohort. The AUROC for admission to hospital for heart failure within 1 year was 0·82 (0·81–0·82) in the testing dataset and 0·78 (0·77–0·79) in the external validation dataset. With an empirical cutoff the ANN algorithm correctly classified 73·6% of patients with regard to all-cause mortality and 61·5% of patients with regard to admission to hospital for heart failure in the external validation cohort, ruling out adverse outcomes with 97·1–98·7% probability in the external validation cohort. Interpretation: Identifying patients at a high risk of developing heart failure or death after myocardial infarction could result in tailored therapies and monitoring by the allocation of resources to those at greatest risk. Funding: The Swedish Heart and Lung Foundation, Swedish Scientific Research Council, Swedish Foundation for Strategic Research, Knut and Alice Wallenberg Foundation, ALF Agreement on Medical Education and Research, Skane University Hospital, The Bundy Academy, the Märta Winkler Foundation, the Anna-Lisa and Sven-Eric Lundgren Foundation for Medical Research.
(Less)
- author
- Mohammad, Moman A.
LU
; Olesen, Kevin K.W.
; Koul, Sasha
LU
; Gale, Chris P.
; Rylance, Rebecca
LU
; Jernberg, Tomas
; Baron, Tomasz
; Spaak, Jonas
; James, Stefan
; Lindahl, Bertil
; Maeng, Michael
and Erlinge, David
LU
(Less) - organization
- publishing date
- 2022-01
- type
- Contribution to journal
- publication status
- published
- subject
- in
- The Lancet Digital Health
- volume
- 4
- issue
- 1
- pages
- 37 - 45
- publisher
- Elsevier
- external identifiers
-
- pmid:34952674
- scopus:85121486303
- ISSN
- 2589-7500
- DOI
- 10.1016/S2589-7500(21)00228-4
- language
- English
- LU publication?
- yes
- additional info
- Publisher Copyright: © 2022 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license
- id
- be78c94b-12ef-48f4-b810-a8289423682e
- date added to LUP
- 2022-04-12 16:04:50
- date last changed
- 2024-06-18 03:15:28
@article{be78c94b-12ef-48f4-b810-a8289423682e,
abstract = {{<p>Background: Patients have an estimated mortality of 15–20% within the first year following myocardial infarction and one in four patients who survive myocardial infarction will develop heart failure, severely reducing quality of life and increasing the risk of long-term mortality. We aimed to establish the accuracy of an artificial neural network (ANN) algorithm in predicting 1-year mortality and admission to hospital for heart failure after myocardial infarction. Methods: In this nationwide population-based study, we used data for all patients admitted to hospital for myocardial infarction and discharged alive from a coronary care unit in Sweden (n=139 288) between Jan 1, 2008, and April 1, 2017, from the Swedish Web system for Enhancement and Development of Evidence-based care in Heart disease Evaluated According to Recommended Therapies (SWEDEHEART) nationwide registry; these patients were randomly divided into training (80%) and testing (20%) datasets. We developed an ANN using 21 variables (including age, sex, medical history, previous medications, in-hospital characteristics, and discharge medications) associated with the outcomes of interest with a back-propagation algorithm in the training dataset and tested it in the testing dataset. The ANN algorithm was then validated in patients with incident myocardial infarction enrolled in the Western Denmark Heart Registry (external validation cohort) between Jan 1, 2008, and Dec 31, 2016. The predictive ability of the model was evaluated using area under the receiver operating characteristic curve (AUROC) and Youden's index was established as a means of identifying an empirical dichotomous cutoff, allowing further evaluation of model performance. Findings: 139 288 patients who were admitted to hospital for myocardial infarction in the SWEDEHEART registry were randomly divided into a training dataset of 111 558 (80%) patients and a testing dataset of 27 730 (20%) patients. 30 971 patients with myocardial infarction who were enrolled in the Western Denmark Heart Registry were included in the external validation cohort. A first event, either all-cause mortality or admission to hospital for heart failure 1 year after myocardial infarction, occurred in 32 308 (23·2%) patients in the testing and training cohorts only. For 1-year all-cause mortality, the ANN had an AUROC of 0·85 (95% CI 0·84–0·85) in the testing dataset and 0·84 (0·83–0·84) in the external validation cohort. The AUROC for admission to hospital for heart failure within 1 year was 0·82 (0·81–0·82) in the testing dataset and 0·78 (0·77–0·79) in the external validation dataset. With an empirical cutoff the ANN algorithm correctly classified 73·6% of patients with regard to all-cause mortality and 61·5% of patients with regard to admission to hospital for heart failure in the external validation cohort, ruling out adverse outcomes with 97·1–98·7% probability in the external validation cohort. Interpretation: Identifying patients at a high risk of developing heart failure or death after myocardial infarction could result in tailored therapies and monitoring by the allocation of resources to those at greatest risk. Funding: The Swedish Heart and Lung Foundation, Swedish Scientific Research Council, Swedish Foundation for Strategic Research, Knut and Alice Wallenberg Foundation, ALF Agreement on Medical Education and Research, Skane University Hospital, The Bundy Academy, the Märta Winkler Foundation, the Anna-Lisa and Sven-Eric Lundgren Foundation for Medical Research.</p>}},
author = {{Mohammad, Moman A. and Olesen, Kevin K.W. and Koul, Sasha and Gale, Chris P. and Rylance, Rebecca and Jernberg, Tomas and Baron, Tomasz and Spaak, Jonas and James, Stefan and Lindahl, Bertil and Maeng, Michael and Erlinge, David}},
issn = {{2589-7500}},
language = {{eng}},
number = {{1}},
pages = {{37--45}},
publisher = {{Elsevier}},
series = {{The Lancet Digital Health}},
title = {{Development and validation of an artificial neural network algorithm to predict mortality and admission to hospital for heart failure after myocardial infarction : a nationwide population-based study}},
url = {{http://dx.doi.org/10.1016/S2589-7500(21)00228-4}},
doi = {{10.1016/S2589-7500(21)00228-4}},
volume = {{4}},
year = {{2022}},
}