Cohort-Derived Machine Learning Models for Individual Prediction of Chronic Kidney Disease in People Living With Human Immunodeficiency Virus: A Prospective Multicenter Cohort Study.

Détails

Ressource 1Demande d'une copie Sous embargo indéterminé.
Accès restreint UNIL
Etat: Public
Version: de l'auteur⸱e
Licence: CC BY-NC-ND 4.0
ID Serval
serval:BIB_01426653908F
Type
Article: article d'un périodique ou d'un magazine.
Collection
Publications
Institution
Titre
Cohort-Derived Machine Learning Models for Individual Prediction of Chronic Kidney Disease in People Living With Human Immunodeficiency Virus: A Prospective Multicenter Cohort Study.
Périodique
The Journal of infectious diseases
Auteur⸱e⸱s
Roth J.A., Radevski G., Marzolini C., Rauch A., Günthard H.F., Kouyos R.D., Fux C.A., Scherrer A.U., Calmy A., Cavassini M., Kahlert C.R., Bernasconi E., Bogojeska J., Battegay M.
Collaborateur⸱rice⸱s
Swiss HIV Cohort Study (SHCS)
ISSN
1537-6613 (Electronic)
ISSN-L
0022-1899
Statut éditorial
Publié
Date de publication
13/10/2021
Peer-reviewed
Oui
Volume
224
Numéro
7
Pages
1198-1208
Langue
anglais
Notes
Publication types: Journal Article ; Multicenter Study ; Research Support, Non-U.S. Gov't
Publication Status: ppublish
Résumé
It is unclear whether data-driven machine learning models, which are trained on large epidemiological cohorts, may improve prediction of comorbidities in people living with human immunodeficiency virus (HIV).
In this proof-of-concept study, we included people living with HIV in the prospective Swiss HIV Cohort Study with a first estimated glomerular filtration rate (eGFR) >60 mL/minute/1.73 m2 after 1 January 2002. Our primary outcome was chronic kidney disease (CKD)-defined as confirmed decrease in eGFR ≤60 mL/minute/1.73 m2 over 3 months apart. We split the cohort data into a training set (80%), validation set (10%), and test set (10%), stratified for CKD status and follow-up length.
Of 12 761 eligible individuals (median baseline eGFR, 103 mL/minute/1.73 m2), 1192 (9%) developed a CKD after a median of 8 years. We used 64 static and 502 time-changing variables: Across prediction horizons and algorithms and in contrast to expert-based standard models, most machine learning models achieved state-of-the-art predictive performances with areas under the receiver operating characteristic curve and precision recall curve ranging from 0.926 to 0.996 and from 0.631 to 0.956, respectively.
In people living with HIV, we observed state-of-the-art performances in forecasting individual CKD onsets with different machine learning algorithms.
Mots-clé
Adult, Cohort Studies, Female, Glomerular Filtration Rate, HIV Infections/complications, HIV Infections/drug therapy, HIV Infections/epidemiology, Health Knowledge, Attitudes, Practice, Humans, Machine Learning, Male, Middle Aged, Predictive Value of Tests, Prospective Studies, Renal Insufficiency, Chronic/complications, Renal Insufficiency, Chronic/diagnosis, Renal Insufficiency, Chronic/epidemiology, Risk Factors, Switzerland/epidemiology, HIV, chronic kidney disease, digital epidemiology, machine learning, prediction
Pubmed
Web of science
Open Access
Oui
Création de la notice
15/06/2020 14:47
Dernière modification de la notice
23/11/2022 6:51
Données d'usage