schliessen

Filtern

 

Bibliotheken

Can machine-learning improve cardiovascular risk prediction using routine clinical data?

Current approaches to predict cardiovascular risk fail to identify many people who would benefit from preventive treatment, while others receive unnecessary intervention. Machine-learning offers opportunity to improve accuracy by exploiting complex interactions between risk factors. We assessed whet... Full description

Journal Title: PloS one 2017, Vol.12(4), pp.e0174944
Main Author: Weng, Stephen F
Other Authors: Reps, Jenna , Kai, Joe , Garibaldi, Jonathan M , Qureshi, Nadeem
Format: Electronic Article Electronic Article
Language: English
Subjects:
ID: E-ISSN: 1932-6203 ; PMID: 28376093 Version:1 ; DOI: 10.1371/journal.pone.0174944
Link: http://pubmed.gov/28376093
Zum Text:
SendSend as email Add to Book BagAdd to Book Bag
Staff View
recordid: medline28376093
title: Can machine-learning improve cardiovascular risk prediction using routine clinical data?
format: Article
creator:
  • Weng, Stephen F
  • Reps, Jenna
  • Kai, Joe
  • Garibaldi, Jonathan M
  • Qureshi, Nadeem
subjects:
  • Machine Learning
  • Cardiovascular Diseases -- Etiology
ispartof: PloS one, 2017, Vol.12(4), pp.e0174944
description: Current approaches to predict cardiovascular risk fail to identify many people who would benefit from preventive treatment, while others receive unnecessary intervention. Machine-learning offers opportunity to improve accuracy by exploiting complex interactions between risk factors. We assessed whether machine-learning can improve cardiovascular risk prediction. Prospective cohort study using routine clinical data of 378,256 patients from UK family practices, free from cardiovascular disease at outset. Four machine-learning algorithms (random forest, logistic regression, gradient boosting machines, neural networks) were compared to an established algorithm (American College of Cardiology guidelines) to predict first cardiovascular event over 10-years. Predictive accuracy was assessed by area under the 'receiver operating curve' (AUC); and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) to predict 7.5% cardiovascular risk (threshold for initiating statins). 24,970 incident cardiovascular events (6.6%) occurred. Compared to the established risk prediction algorithm (AUC 0.728, 95% CI 0.723-0.735), machine-learning algorithms improved prediction: random forest +1.7% (AUC 0.745, 95% CI 0.739-0.750), logistic regression +3.2% (AUC 0.760, 95% CI 0.755-0.766), gradient boosting +3.3% (AUC 0.761, 95% CI 0.755-0.766), neural networks +3.6% (AUC 0.764, 95% CI 0.759-0.769). The highest achieving (neural networks) algorithm predicted 4,998/7,404 cases (sensitivity 67.5%, PPV 18.4%) and 53,458/75,585 non-cases (specificity 70.7%, NPV 95.7%), correctly predicting 355 (+7.6%) more patients who developed cardiovascular disease compared to the established algorithm. Machine-learning significantly improves accuracy of cardiovascular risk prediction, increasing the number of patients identified who could benefit from preventive treatment, while avoiding unnecessary treatment of others.
language: eng
source:
identifier: E-ISSN: 1932-6203 ; PMID: 28376093 Version:1 ; DOI: 10.1371/journal.pone.0174944
fulltext: fulltext
issn:
  • 19326203
  • 1932-6203
url: Link


@attributes
ID964558484
RANK0.07
NO1
SEARCH_ENGINEprimo_central_multiple_fe
SEARCH_ENGINE_TYPEPrimo Central Search Engine
LOCALfalse
PrimoNMBib
record
control
sourcerecordid28376093
sourceidmedline
recordidTN_medline28376093
sourceformatXML
sourcesystemPC
pqid1884098417
galeid488519999
display
typearticle
titleCan machine-learning improve cardiovascular risk prediction using routine clinical data?
creatorWeng, Stephen F ; Reps, Jenna ; Kai, Joe ; Garibaldi, Jonathan M ; Qureshi, Nadeem
ispartofPloS one, 2017, Vol.12(4), pp.e0174944
identifier
subjectMachine Learning ; Cardiovascular Diseases -- Etiology
descriptionCurrent approaches to predict cardiovascular risk fail to identify many people who would benefit from preventive treatment, while others receive unnecessary intervention. Machine-learning offers opportunity to improve accuracy by exploiting complex interactions between risk factors. We assessed whether machine-learning can improve cardiovascular risk prediction. Prospective cohort study using routine clinical data of 378,256 patients from UK family practices, free from cardiovascular disease at outset. Four machine-learning algorithms (random forest, logistic regression, gradient boosting machines, neural networks) were compared to an established algorithm (American College of Cardiology guidelines) to predict first cardiovascular event over 10-years. Predictive accuracy was assessed by area under the 'receiver operating curve' (AUC); and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) to predict 7.5% cardiovascular risk (threshold for initiating statins). 24,970 incident cardiovascular events (6.6%) occurred. Compared to the established risk prediction algorithm (AUC 0.728, 95% CI 0.723-0.735), machine-learning algorithms improved prediction: random forest +1.7% (AUC 0.745, 95% CI 0.739-0.750), logistic regression +3.2% (AUC 0.760, 95% CI 0.755-0.766), gradient boosting +3.3% (AUC 0.761, 95% CI 0.755-0.766), neural networks +3.6% (AUC 0.764, 95% CI 0.759-0.769). The highest achieving (neural networks) algorithm predicted 4,998/7,404 cases (sensitivity 67.5%, PPV 18.4%) and 53,458/75,585 non-cases (specificity 70.7%, NPV 95.7%), correctly predicting 355 (+7.6%) more patients who developed cardiovascular disease compared to the established algorithm. Machine-learning significantly improves accuracy of cardiovascular risk prediction, increasing the number of patients identified who could benefit from preventive treatment, while avoiding unnecessary treatment of others.
languageeng
source
version8
lds50peer_reviewed
links
openurl$$Topenurl_article
backlink$$Uhttp://pubmed.gov/28376093$$EView_this_record_in_MEDLINE/PubMed
openurlfulltext$$Topenurlfull_article
addlink$$Uhttp://exlibris-pub.s3.amazonaws.com/aboutMedline.html$$EView_the_MEDLINE/PubMed_Copyright_Statement
search
creatorcontrib
0Weng, Stephen F
1Reps, Jenna
2Kai, Joe
3Garibaldi, Jonathan M
4Qureshi, Nadeem
titleCan machine-learning improve cardiovascular risk prediction using routine clinical data?
description
0Current approaches to predict cardiovascular risk fail to identify many people who would benefit from preventive treatment, while others receive unnecessary intervention. Machine-learning offers opportunity to improve accuracy by exploiting complex interactions between risk factors. We assessed whether machine-learning can improve cardiovascular risk prediction.
1Prospective cohort study using routine clinical data of 378,256 patients from UK family practices, free from cardiovascular disease at outset. Four machine-learning algorithms (random forest, logistic regression, gradient boosting machines, neural networks) were compared to an established algorithm (American College of Cardiology guidelines) to predict first cardiovascular event over 10-years. Predictive accuracy was assessed by area under the 'receiver operating curve' (AUC); and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) to predict 7.5% cardiovascular risk (threshold for initiating statins).
224,970 incident cardiovascular events (6.6%) occurred. Compared to the established risk prediction algorithm (AUC 0.728, 95% CI 0.723-0.735), machine-learning algorithms improved prediction: random forest +1.7% (AUC 0.745, 95% CI 0.739-0.750), logistic regression +3.2% (AUC 0.760, 95% CI 0.755-0.766), gradient boosting +3.3% (AUC 0.761, 95% CI 0.755-0.766), neural networks +3.6% (AUC 0.764, 95% CI 0.759-0.769). The highest achieving (neural networks) algorithm predicted 4,998/7,404 cases (sensitivity 67.5%, PPV 18.4%) and 53,458/75,585 non-cases (specificity 70.7%, NPV 95.7%), correctly predicting 355 (+7.6%) more patients who developed cardiovascular disease compared to the established algorithm.
3Machine-learning significantly improves accuracy of cardiovascular risk prediction, increasing the number of patients identified who could benefit from preventive treatment, while avoiding unnecessary treatment of others.
subject
0Machine Learning
1Cardiovascular Diseases -- Etiology
general
028376093
1English
2MEDLINE/PubMed (U.S. National Library of Medicine)
310.1371/journal.pone.0174944
4MEDLINE/PubMed (NLM)
sourceidmedline
recordidmedline28376093
issn
019326203
11932-6203
rsrctypearticle
creationdate2017
addtitlePloS one
searchscope
0medline
1nlm_medline
2MEDLINE
scope
0medline
1nlm_medline
2MEDLINE
lsr412017
citationpf e0174944 vol 12 issue 4
startdate20170101
enddate20171231
lsr30VSR-Enriched:[pqid, galeid]
sort
titleCan machine-learning improve cardiovascular risk prediction using routine clinical data?
authorWeng, Stephen F ; Reps, Jenna ; Kai, Joe ; Garibaldi, Jonathan M ; Qureshi, Nadeem
creationdate20170000
lso0120170000
facets
frbrgroupid4706836107974718352
frbrtype5
newrecords20190701
languageeng
creationdate2017
topic
0Machine Learning
1Cardiovascular Diseases–Etiology
collectionMEDLINE/PubMed (NLM)
prefilterarticles
rsrctypearticles
creatorcontrib
0Weng, Stephen F
1Reps, Jenna
2Kai, Joe
3Garibaldi, Jonathan M
4Qureshi, Nadeem
jtitlePlos One
toplevelpeer_reviewed
delivery
delcategoryRemote Search Resource
fulltextfulltext
addata
aulast
0Weng
1Reps
2Kai
3Garibaldi
4Qureshi
aufirst
0Stephen F
1Jenna
2Joe
3Jonathan M
4Nadeem
au
0Weng, Stephen F
1Reps, Jenna
2Kai, Joe
3Garibaldi, Jonathan M
4Qureshi, Nadeem
atitleCan machine-learning improve cardiovascular risk prediction using routine clinical data?
jtitlePloS one
date2017
risdate2017
volume12
issue4
spagee0174944
pagese0174944
issn1932-6203
eissn1932-6203
formatjournal
genrearticle
ristypeJOUR
abstractCurrent approaches to predict cardiovascular risk fail to identify many people who would benefit from preventive treatment, while others receive unnecessary intervention. Machine-learning offers opportunity to improve accuracy by exploiting complex interactions between risk factors. We assessed whether machine-learning can improve cardiovascular risk prediction.
doi10.1371/journal.pone.0174944
pmid28376093