EUVIMEDEUVIMED entdecken
BeobachtungsstudieOpen Access · CC BY 4.0

Prediction of Postprandial Blood Glucose Variability Using Machine Learning in Frequent Insulin Injection Therapy with a Simplified Carbohydrate Counting Model

Hiroyuki Tominaga, Masahide Hamaguchi iD, Youji Hamaguchi, Ren Yashiki, Aki Yamaguchi, Tadaharu Arai, Masahiro Yamazaki, Noriyuki Kitagawa iD, Yoshitaka Hashimoto iD, Hiroshi Okada iD, Michiaki Fukui iD

Nutrients · 07.12.2025 · 17 (24) · 3832

DOI: 10.3390/nu17243832 ↗

Kurz gefasst

Beobachtungsstudie in der Zeitschrift Nutrients, online erschienen am 07.12.2025. Verfasst von Hiroyuki Tominaga und 10 weiteren Autor:innen. Beteiligte Einrichtung: Department of Endocrinology and Metabolism, Graduate School of Medicine, Kyoto Prefectural University of Medicine, Kyoto 602-8566, Japan. Hauptthemen laut MEDLINE: Machine Learning, Postprandial Period, Blood Glucose, Insulin. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Kyoto Innovative Medical Technology Research and Development Program (FY2022), Japan IDDM Network 2022 Research Grant, Japan Science and Technology Agency (JST) 2023 Feasibility Verification.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background/Objectives: Postprandial glucose variability is a key challenge in diabetes management for patients receiving multiple daily insulin injections (MDI). This study evaluated transformer-based machine-learning models for predicting post-prandial glucose peaks and nadirs using pre-meal glucose, insulin dose, and nutritional input.

Methods

In this observational study, 58 adults with diabetes provided dietary records, insulin logs, and continuous glucose monitoring data. After preprocessing and participant-level splitting (64:16:20), model-ready datasets comprised 6155/1449/1805 (train/validation/test) meal events for the Full-Nutrition model and 6299/1484/1849 for the Carbohydrate and Available-Carbohydrate models. We evaluated three transformer-based models and assessed performance using MAE, R2, and the Clarke error grid.

Results

The Full Nutrition Model achieved MAEs of 32.2 mg/dL (peak) and 21.8 mg/dL (nadir) with R2 values of 0.58 for both. Carbohydrate-based models showed similar accuracy. Most predictions fell within Clarke error grid Zones A and B.

Conclusions

Transformer-based machine-learning models can accurately predict postprandial glucose variability in MDI-treated patients. Carbohydrate-only inputs performed comparably to full-nutrient data, supporting the feasibility of simplified dietary inputs in clinical applications.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Machine LearningPostprandial PeriodBlood GlucoseInsulinDietary CarbohydratesHypoglycemic AgentsHumansFemaleMaleMiddle AgedAdultAged

Stichwörter der Autor:innen

blood glucose variabilitycarbohydrate counting methoddiabetesfrequent insulin injection therapymachine learningtransformer model

Substanzen

Blood GlucoseInsulinDietary CarbohydratesHypoglycemic Agents

Autor:innen und Institutionen

  • Department of Endocrinology and Metabolism, Graduate School of Medicine, Kyoto Prefectural University of Medicine, Kyoto 602-8566, Japan.
  • S.I.P., Tokyo 141-0021, Japan.
  • Persol Avc Technology Co., Ltd., Osaka 569-1194, Japan.
  • Department of Metabolism and Endocrinology, Japanese Red Cross Society Kyoto Daini Hospital, Kyoto 602-8026, Japan.
  • Department of Diabetes, Kameoka Municipal Hospital, Kameoka 621-8585, Japan.
  • Department of Diabetes and Endocrinology, Matsushita Memorial Hospital, Osaka 570-8540, Japan.

ORCID-Profile: Masahide Hamaguchi, Noriyuki Kitagawa, Yoshitaka Hashimoto, Hiroshi Okada, Michiaki Fukui

Förderung

  • Kyoto Innovative Medical Technology Research and Development Program (FY2022)
  • Japan IDDM Network 2022 Research Grant
  • Japan Science and Technology Agency (JST) 2023 Feasibility Verification
  • Fiscal Year 2024 Research and Development Support Program for Growth-Oriented Small and Medium-Sized Enterprises
  • Tanuma Green House Foundation Research Grant (2024)

Zitieren

APA

Tominaga, H., Hamaguchi, M., Hamaguchi, Y., Yashiki, R., Yamaguchi, A., Arai, T., Yamazaki, M., Kitagawa, N., Hashimoto, Y., Okada, H., & Fukui, M. (2025). Prediction of Postprandial Blood Glucose Variability Using Machine Learning in Frequent Insulin Injection Therapy with a Simplified Carbohydrate Counting Model. Nutrients, 17(24), 3832. https://doi.org/10.3390/nu17243832

Vancouver

Tominaga H, Hamaguchi M, Hamaguchi Y, Yashiki R, Yamaguchi A, Arai T, et al. Prediction of Postprandial Blood Glucose Variability Using Machine Learning in Frequent Insulin Injection Therapy with a Simplified Carbohydrate Counting Model. Nutrients. 2025;17(24):3832. doi:10.3390/nu17243832

BibTeX
@article{Tominaga2025,
  author = {Tominaga, Hiroyuki and Hamaguchi, Masahide and Hamaguchi, Youji and Yashiki, Ren and Yamaguchi, Aki and Arai, Tadaharu and Yamazaki, Masahiro and Kitagawa, Noriyuki and Hashimoto, Yoshitaka and Okada, Hiroshi and Fukui, Michiaki},
  title = {Prediction of Postprandial Blood Glucose Variability Using Machine Learning in Frequent Insulin Injection Therapy with a Simplified Carbohydrate Counting Model},
  journal = {Nutrients},
  year = {2025},
  volume = {17},
  number = {24},
  pages = {3832},
  doi = {10.3390/nu17243832}
}
Kostenlos starten →