Machine Learning Prediction Models for In-Hospital Mortality After Transcatheter Aortic Valve Replacement

Dagmar F. Hernandez-Suarez, Yeunjung Kim, P. Villablanca, Tanush Gupta, Jose Wiley, Brenda G. Nieves-Rodriguez, Jovaniel Rodriguez-Maldonado, Roberto Feliu Maldonado, Istoni da Luz Sant'Ana, Cristina Sanina, P. Cox-Alomar, Harish Ramakrishna, A. Lopez-Candales, William W. O'Neill, Duane S. Pinto, A. Latib, A. Roche-Lima

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Objectives: This study sought to develop and compare an array of machine learning methods to predict in-hospital mortality after transcatheter aortic valve replacement (TAVR) in the United States. Background: Existing risk prediction tools for in-hospital complications in patients undergoing TAVR have been designed using statistical modeling approaches and have certain limitations. Methods: Patient data were obtained from the National Inpatient Sample database from 2012 to 2015. The data were randomly divided into a development cohort (n = 7,615) and a validation cohort (n = 3,268). Logistic regression, artificial neural network, naive Bayes, and random forest machine learning algorithms were applied to obtain in-hospital mortality prediction models. Results: A total of 10,883 TAVRs were analyzed in our study. The overall in-hospital mortality was 3.6%. Overall, prediction models’ performance measured by area under the curve were good (>0.80). The best model was obtained by logistic regression (area under the curve: 0.92; 95% confidence interval: 0.89 to 0.95). Most obtained models plateaued after introducing 10 variables. Acute kidney injury was the main predictor of in-hospital mortality ranked with the highest mean importance in all the models. The National Inpatient Sample TAVR score showed the best discrimination among available TAVR prediction scores. Conclusions: Machine learning methods can generate robust models to predict in-hospital mortality for TAVR. The National Inpatient Sample TAVR score should be considered for prognosis and shared decision making in TAVR patients.

Original languageEnglish (US)
Pages (from-to)1328-1338
Number of pages11
JournalJACC: Cardiovascular Interventions
Volume12
Issue number14
DOIs
StatePublished - Jul 22 2019

Fingerprint

Hospital Mortality
Inpatients
Area Under Curve
Logistic Models
Transcatheter Aortic Valve Replacement
Machine Learning
Acute Kidney Injury
Decision Making
Databases
Confidence Intervals

Keywords

  • machine learning
  • mortality
  • transcatheter aortic valve replacement

ASJC Scopus subject areas

  • Cardiology and Cardiovascular Medicine

Cite this

Hernandez-Suarez, D. F., Kim, Y., Villablanca, P., Gupta, T., Wiley, J., Nieves-Rodriguez, B. G., ... Roche-Lima, A. (2019). Machine Learning Prediction Models for In-Hospital Mortality After Transcatheter Aortic Valve Replacement. JACC: Cardiovascular Interventions, 12(14), 1328-1338. https://doi.org/10.1016/j.jcin.2019.06.013

Machine Learning Prediction Models for In-Hospital Mortality After Transcatheter Aortic Valve Replacement. / Hernandez-Suarez, Dagmar F.; Kim, Yeunjung; Villablanca, P.; Gupta, Tanush; Wiley, Jose; Nieves-Rodriguez, Brenda G.; Rodriguez-Maldonado, Jovaniel; Feliu Maldonado, Roberto; da Luz Sant'Ana, Istoni; Sanina, Cristina; Cox-Alomar, P.; Ramakrishna, Harish; Lopez-Candales, A.; O'Neill, William W.; Pinto, Duane S.; Latib, A.; Roche-Lima, A.

In: JACC: Cardiovascular Interventions, Vol. 12, No. 14, 22.07.2019, p. 1328-1338.

Research output: Contribution to journalArticle

Hernandez-Suarez, DF, Kim, Y, Villablanca, P, Gupta, T, Wiley, J, Nieves-Rodriguez, BG, Rodriguez-Maldonado, J, Feliu Maldonado, R, da Luz Sant'Ana, I, Sanina, C, Cox-Alomar, P, Ramakrishna, H, Lopez-Candales, A, O'Neill, WW, Pinto, DS, Latib, A & Roche-Lima, A 2019, 'Machine Learning Prediction Models for In-Hospital Mortality After Transcatheter Aortic Valve Replacement', JACC: Cardiovascular Interventions, vol. 12, no. 14, pp. 1328-1338. https://doi.org/10.1016/j.jcin.2019.06.013
Hernandez-Suarez DF, Kim Y, Villablanca P, Gupta T, Wiley J, Nieves-Rodriguez BG et al. Machine Learning Prediction Models for In-Hospital Mortality After Transcatheter Aortic Valve Replacement. JACC: Cardiovascular Interventions. 2019 Jul 22;12(14):1328-1338. https://doi.org/10.1016/j.jcin.2019.06.013
Hernandez-Suarez, Dagmar F. ; Kim, Yeunjung ; Villablanca, P. ; Gupta, Tanush ; Wiley, Jose ; Nieves-Rodriguez, Brenda G. ; Rodriguez-Maldonado, Jovaniel ; Feliu Maldonado, Roberto ; da Luz Sant'Ana, Istoni ; Sanina, Cristina ; Cox-Alomar, P. ; Ramakrishna, Harish ; Lopez-Candales, A. ; O'Neill, William W. ; Pinto, Duane S. ; Latib, A. ; Roche-Lima, A. / Machine Learning Prediction Models for In-Hospital Mortality After Transcatheter Aortic Valve Replacement. In: JACC: Cardiovascular Interventions. 2019 ; Vol. 12, No. 14. pp. 1328-1338.
@article{e7319b1e355847539581bd258bf54dd4,
title = "Machine Learning Prediction Models for In-Hospital Mortality After Transcatheter Aortic Valve Replacement",
abstract = "Objectives: This study sought to develop and compare an array of machine learning methods to predict in-hospital mortality after transcatheter aortic valve replacement (TAVR) in the United States. Background: Existing risk prediction tools for in-hospital complications in patients undergoing TAVR have been designed using statistical modeling approaches and have certain limitations. Methods: Patient data were obtained from the National Inpatient Sample database from 2012 to 2015. The data were randomly divided into a development cohort (n = 7,615) and a validation cohort (n = 3,268). Logistic regression, artificial neural network, naive Bayes, and random forest machine learning algorithms were applied to obtain in-hospital mortality prediction models. Results: A total of 10,883 TAVRs were analyzed in our study. The overall in-hospital mortality was 3.6{\%}. Overall, prediction models’ performance measured by area under the curve were good (>0.80). The best model was obtained by logistic regression (area under the curve: 0.92; 95{\%} confidence interval: 0.89 to 0.95). Most obtained models plateaued after introducing 10 variables. Acute kidney injury was the main predictor of in-hospital mortality ranked with the highest mean importance in all the models. The National Inpatient Sample TAVR score showed the best discrimination among available TAVR prediction scores. Conclusions: Machine learning methods can generate robust models to predict in-hospital mortality for TAVR. The National Inpatient Sample TAVR score should be considered for prognosis and shared decision making in TAVR patients.",
keywords = "machine learning, mortality, transcatheter aortic valve replacement",
author = "Hernandez-Suarez, {Dagmar F.} and Yeunjung Kim and P. Villablanca and Tanush Gupta and Jose Wiley and Nieves-Rodriguez, {Brenda G.} and Jovaniel Rodriguez-Maldonado and {Feliu Maldonado}, Roberto and {da Luz Sant'Ana}, Istoni and Cristina Sanina and P. Cox-Alomar and Harish Ramakrishna and A. Lopez-Candales and O'Neill, {William W.} and Pinto, {Duane S.} and A. Latib and A. Roche-Lima",
year = "2019",
month = "7",
day = "22",
doi = "10.1016/j.jcin.2019.06.013",
language = "English (US)",
volume = "12",
pages = "1328--1338",
journal = "JACC: Cardiovascular Interventions",
issn = "1936-8798",
publisher = "Elsevier Inc.",
number = "14",

}

TY - JOUR

T1 - Machine Learning Prediction Models for In-Hospital Mortality After Transcatheter Aortic Valve Replacement

AU - Hernandez-Suarez, Dagmar F.

AU - Kim, Yeunjung

AU - Villablanca, P.

AU - Gupta, Tanush

AU - Wiley, Jose

AU - Nieves-Rodriguez, Brenda G.

AU - Rodriguez-Maldonado, Jovaniel

AU - Feliu Maldonado, Roberto

AU - da Luz Sant'Ana, Istoni

AU - Sanina, Cristina

AU - Cox-Alomar, P.

AU - Ramakrishna, Harish

AU - Lopez-Candales, A.

AU - O'Neill, William W.

AU - Pinto, Duane S.

AU - Latib, A.

AU - Roche-Lima, A.

PY - 2019/7/22

Y1 - 2019/7/22

N2 - Objectives: This study sought to develop and compare an array of machine learning methods to predict in-hospital mortality after transcatheter aortic valve replacement (TAVR) in the United States. Background: Existing risk prediction tools for in-hospital complications in patients undergoing TAVR have been designed using statistical modeling approaches and have certain limitations. Methods: Patient data were obtained from the National Inpatient Sample database from 2012 to 2015. The data were randomly divided into a development cohort (n = 7,615) and a validation cohort (n = 3,268). Logistic regression, artificial neural network, naive Bayes, and random forest machine learning algorithms were applied to obtain in-hospital mortality prediction models. Results: A total of 10,883 TAVRs were analyzed in our study. The overall in-hospital mortality was 3.6%. Overall, prediction models’ performance measured by area under the curve were good (>0.80). The best model was obtained by logistic regression (area under the curve: 0.92; 95% confidence interval: 0.89 to 0.95). Most obtained models plateaued after introducing 10 variables. Acute kidney injury was the main predictor of in-hospital mortality ranked with the highest mean importance in all the models. The National Inpatient Sample TAVR score showed the best discrimination among available TAVR prediction scores. Conclusions: Machine learning methods can generate robust models to predict in-hospital mortality for TAVR. The National Inpatient Sample TAVR score should be considered for prognosis and shared decision making in TAVR patients.

AB - Objectives: This study sought to develop and compare an array of machine learning methods to predict in-hospital mortality after transcatheter aortic valve replacement (TAVR) in the United States. Background: Existing risk prediction tools for in-hospital complications in patients undergoing TAVR have been designed using statistical modeling approaches and have certain limitations. Methods: Patient data were obtained from the National Inpatient Sample database from 2012 to 2015. The data were randomly divided into a development cohort (n = 7,615) and a validation cohort (n = 3,268). Logistic regression, artificial neural network, naive Bayes, and random forest machine learning algorithms were applied to obtain in-hospital mortality prediction models. Results: A total of 10,883 TAVRs were analyzed in our study. The overall in-hospital mortality was 3.6%. Overall, prediction models’ performance measured by area under the curve were good (>0.80). The best model was obtained by logistic regression (area under the curve: 0.92; 95% confidence interval: 0.89 to 0.95). Most obtained models plateaued after introducing 10 variables. Acute kidney injury was the main predictor of in-hospital mortality ranked with the highest mean importance in all the models. The National Inpatient Sample TAVR score showed the best discrimination among available TAVR prediction scores. Conclusions: Machine learning methods can generate robust models to predict in-hospital mortality for TAVR. The National Inpatient Sample TAVR score should be considered for prognosis and shared decision making in TAVR patients.

KW - machine learning

KW - mortality

KW - transcatheter aortic valve replacement

UR - http://www.scopus.com/inward/record.url?scp=85068443858&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068443858&partnerID=8YFLogxK

U2 - 10.1016/j.jcin.2019.06.013

DO - 10.1016/j.jcin.2019.06.013

M3 - Article

C2 - 31320027

AN - SCOPUS:85068443858

VL - 12

SP - 1328

EP - 1338

JO - JACC: Cardiovascular Interventions

JF - JACC: Cardiovascular Interventions

SN - 1936-8798

IS - 14

ER -