EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

A framework for assessing algorithmic discrimination risks in training data: a case of pediatric type 1 diabetes

Ioannis Bilionis, Ricardo C Berrios, Antonio de Arriba Muñoz, Luis Fernandez-Luque, Carlos Castillo

JAMIA Open · 07.07.2026 · 9 (4)

DOI: 10.1093/jamiaopen/ooag125 ↗

Kurz gefasst

Fachartikel in der Zeitschrift JAMIA Open, online erschienen am 14.07.2026. Verfasst von Ioannis Bilionis und 4 weiteren Autor:innen. Beteiligte Einrichtung: Adhera Health, Santa Cruz, CA, 95060, United States. Die Arbeit nennt 65 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: Spanish Ministry of Science and Innovation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objectives

To develop a generalizable framework for identifying algorithmic discrimination risks arising from subgroup imbalances in machine learning training data, with relevance to medical informatics applications where heterogeneous real-world data can bias model behavior. Materials and Methods We introduce a discrimination risk assessment framework for training datasets, a 4-step methodology integrating: (1) controlled representation sampling, (2) ensemble-based model training, (3) multilevel subgroup disparity quantification, and (4) mitigation-oriented interpretation. The framework systematically perturbs training subgroup composition while holding evaluation sets fixed to isolate representation effects. Validation was performed across 7 publicly available pediatric type 1 diabetes datasets using sociodemographic variables and continuous glucose monitoring data to assess robustness under heterogeneous data sources.

Results

Representation balance alone does not guarantee stable or equitable model outputs. Ensemble analyses revealed subgroup-dependent volatility, with some groups consistently contributing to model generalization, while others inducing instability despite increased representation. These findings demonstrate structural sensitivity to data composition that is not captured by standard performance-only evaluations.

Discussion

The framework exposes disparities and instability patterns that remain hidden under single-model or balanced-data evaluations. By quantifying representation-driven behaviors and subgroup-specific training value, it offers a diagnostic tool for understanding data-induced risks prior to model deployment.

Conclusion

This work provides a methodology for assessing discrimination risks in training datasets used in medical informatics. By integrating data composition effects, prediction stability, and subgroup-level disparity analysis, the framework supports more reliable and transparent development of machine learning systems across diverse clinical and nonclinical contexts. Clinical trial registration This research did not involve any new clinical trial. All analyses were performed on data from previously registered clinical trials, as listed below:

Schlagwörter

Stichwörter der Autor:innen

biasdiabetes mellitus type 1health equitymachine learningpediatrics

Autor:innen und Institutionen

  • Adhera Health, Santa Cruz, CA, 95060, United States.
  • Department of Engineering, Universitat Pompeu Fabra, Barcelona, 08018, Spain.
  • Department of Pediatric Endocrinology, University Hospital "Miguel Servet", Zaragoza, 50009, Spain.
  • ICREA, Catalonia, 08010, Spain.

Förderung

  • Spanish Ministry of Science and Innovation

Zitieren

APA

Bilionis, I., Berrios, R. C., de Arriba Muñoz, A., Fernandez-Luque, L., & Castillo, C. (2026). A framework for assessing algorithmic discrimination risks in training data: a case of pediatric type 1 diabetes. JAMIA Open, 9(4). https://doi.org/10.1093/jamiaopen/ooag125

Vancouver

Bilionis I, Berrios RC, de Arriba Muñoz A, Fernandez-Luque L, Castillo C. A framework for assessing algorithmic discrimination risks in training data: a case of pediatric type 1 diabetes. JAMIA Open. 2026;9(4). doi:10.1093/jamiaopen/ooag125

BibTeX
@article{Bilionis2026,
  author = {Bilionis, Ioannis and Berrios, Ricardo C and de Arriba Muñoz, Antonio and Fernandez-Luque, Luis and Castillo, Carlos},
  title = {A framework for assessing algorithmic discrimination risks in training data: a case of pediatric type 1 diabetes},
  journal = {JAMIA Open},
  year = {2026},
  volume = {9},
  number = {4},
  doi = {10.1093/jamiaopen/ooag125}
}
Kostenlos starten →