EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Pooling cohorts for deep learning analysis: a potential source of bias for electrocardiogram analysis

Marie Mørch-Pedersen, Malene Nørregaard, Claus Graff, Oliver B. Vad, Pia R. Lundegaard, Jesper Hastrup Svendsen, Torben Hansen, Allan Linneberg, Peter G. Jørgensen, Magnus T. Jensen, Peter Rossing, Christina Ellervik und 6 weitere

Biomedical Signal Processing and Control · 2026-08 · 122 · 110424

DOI: 10.1016/j.bspc.2026.110424 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Biomedical Signal Processing and Control, 2026-08. Verfasst von Marie Mørch-Pedersen und 17 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Danish Cardiovascular Academy, Boston Children's Hospital, Danish Data Science Academy.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

• When pooling cohorts, bias can underlie seemingly good results. • Bias requires a difficult-to-predict outcome unevenly distributed between cohorts. • Strength of bias can be directly obtained to estimate unbiased results. Deep learning models can isolate device characteristics in medical images, enabling the models to identify the origin of a medical image, which creates a bias. It is unknown whether such bias can arise with raw medical signals such as electrocardiograms (ECGs), so we aimed to quantify to what extent deep learning models can identify the study cohort and site from which an ECG originates. We used 70,075 12-lead ECGs from four Danish cohorts and the UK Biobank. We trained a convolutional neural network (CNN) model using 5-fold cross-validation to identify the origin of each ECG. We also tested the effect of easy vs. difficult to predict outcomes (sex vs. diabetes) and equal vs. skewed outcome distributions. We reported accuracy and F 1 score, which is less sensitive to unequal sample sizes. The CNN model was able to distinguish ECGs from five different cohorts with an accuracy of 93.4% and an F 1 score of 77.1%. We found no bias with easily detected outcomes or equal distributions. We were unable to ascribe the CNN performance to any known factors, including ECG device, software, or population characteristics, including sex, age, and comorbidities. Deep learning models can identify the origin of an ECG. Combining studies can introduce confounding to the model if the rate of the outcome varies between studies and the outcome is difficult to identify. In deep learning models with potential cohort confounding, we recommend training a baseline model to separate these cohorts to assess the strength of confounding.

Förderung

  • Danish Cardiovascular Academy
  • Boston Children's Hospital
  • Danish Data Science Academy
  • Innovative Health Initiative
  • Shipowner Per Henriksen R and Wife Fund
  • Villum Fonden
  • Hjerteforeningen
  • Novo Nordisk Foundation

Zitieren

APA

Mørch-Pedersen, M., Nørregaard, M., Graff, C., Vad, O. B., Lundegaard, P. R., Svendsen, J. H., Hansen, T., Linneberg, A., Jørgensen, P. G., Jensen, M. T., Rossing, P., Ellervik, C., Linz, D., Thambawita, V., Maleckar, M. M., Riegler, M. A., Kanters, J. K., & Isaksen, J. L. (2026). Pooling cohorts for deep learning analysis: a potential source of bias for electrocardiogram analysis. Biomedical Signal Processing and Control, 122, 110424. https://doi.org/10.1016/j.bspc.2026.110424

Vancouver

Mørch-Pedersen M, Nørregaard M, Graff C, Vad OB, Lundegaard PR, Svendsen JH, et al. Pooling cohorts for deep learning analysis: a potential source of bias for electrocardiogram analysis. Biomedical Signal Processing and Control. 2026;122:110424. doi:10.1016/j.bspc.2026.110424

BibTeX
@article{MrchPedersen2026,
  author = {Mørch-Pedersen, Marie and Nørregaard, Malene and Graff, Claus and Vad, Oliver B. and Lundegaard, Pia R. and Svendsen, Jesper Hastrup and Hansen, Torben and Linneberg, Allan and Jørgensen, Peter G. and Jensen, Magnus T. and Rossing, Peter and Ellervik, Christina and Linz, Dominik and Thambawita, Vajira and Maleckar, Mary M. and Riegler, Michael A. and Kanters, Jørgen K. and Isaksen, Jonas L.},
  title = {Pooling cohorts for deep learning analysis: a potential source of bias for electrocardiogram analysis},
  journal = {Biomedical Signal Processing and Control},
  year = {2026},
  volume = {122},
  pages = {110424},
  doi = {10.1016/j.bspc.2026.110424}
}
Kostenlos starten →