Embracing cohort heterogeneity in clinical machine learning development: a step toward generalizable models.
Journal
Scientific reports
ISSN: 2045-2322
Titre abrégé: Sci Rep
Pays: England
ID NLM: 101563288
Informations de publication
Date de publication:
24 05 2023
24 05 2023
Historique:
received:
10
10
2022
accepted:
20
05
2023
medline:
26
5
2023
pubmed:
25
5
2023
entrez:
24
5
2023
Statut:
epublish
Résumé
This study is a simple illustration of the benefit of averaging over cohorts, rather than developing a prediction model from a single cohort. We show that models trained on data from multiple cohorts can perform significantly better in new settings than models based on the same amount of training data but from just a single cohort. Although this concept seems simple and obvious, no current prediction model development guidelines recommend such an approach.
Identifiants
pubmed: 37225751
doi: 10.1038/s41598-023-35557-y
pii: 10.1038/s41598-023-35557-y
pmc: PMC10209202
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
8363Informations de copyright
© 2023. The Author(s).
Références
JAMA Intern Med. 2021 Aug 1;181(8):1040-1041
pubmed: 34152360
Stat Med. 2021 Jul 10;40(15):3533-3559
pubmed: 33948970
J Am Med Inform Assoc. 2019 Dec 1;26(12):1651-1654
pubmed: 31373357
Stat Med. 2019 Sep 30;38(22):4290-4309
pubmed: 31373722
JAMA Intern Med. 2021 Aug 1;181(8):1065-1070
pubmed: 34152373
Nat Med. 2019 Jan;25(1):44-56
pubmed: 30617339
EBioMedicine. 2022 Aug;82:104176
pubmed: 35853298