Trustworthy chronic kidney disease screening in smart health systems using internet-of-things-integrated data processing and verifiable convolutional neural networks
Health Informatics Journal · 2026-10 · 32 (4)
DOI: 10.1177/14604582261508405 ↗Kurz gefasst
Fachartikel in der Zeitschrift Health Informatics Journal, online erschienen am 20.11.2026. Verfasst von S. Mahaboob Basha und 4 weiteren Autor:innen. Die Arbeit nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Objective
To develop and internally evaluate a trustworthy chronic kidney disease (CKD) screening framework emphasizing leakage-controlled preprocessing, feature selection, predictive reliability, calibration, and computational efficiency.
Methods
A retrospective computational study used 172 patient records from a public CKD dataset. The framework integrates High Accuracy Distributed Kalman Filter (HADKF) data repair, Nutcracker Optimizer Algorithm (NOA) feature selection, a lightweight Verifiable Convolutional Neural Network (VCNN), and Snow Geese Optimization Algorithm (SGOA) tuning. Stratified nested cross-validation ensured leakage-controlled evaluation of discrimination, calibration, robustness, interpretability, feature stability, and inference latency.
Results
The framework achieved an AUC of approximately 0.998 and 98.84% accuracy, with 108 true positives, 62 true negatives, one false positive, and one false negative. Mean expected calibration error was 0.015 ± 0.007, Verified Prediction Ratio reached 98.86 ± 1.57%, and inference latency averaged 4.82 ± 0.38 ms/sample. However, the small, single-source cohort limits generalizability.
Conclusion
The framework demonstrates promising computational feasibility for accurate, efficient, and auditable CKD screening. Nevertheless, clinical applicability remains unestablished. Larger multicentre datasets, external validation, prospective clinical evaluation, hardware profiling, and real-world edge-device testing are essential before deployment.
Zitieren
APA
Basha, S. M., Shieh, C. S., Murugan, V. S., Harrison, A., & Mbasso, W. F. (2026). Trustworthy chronic kidney disease screening in smart health systems using internet-of-things-integrated data processing and verifiable convolutional neural networks. Health Informatics Journal, 32(4). https://doi.org/10.1177/14604582261508405
Vancouver
Basha SM, Shieh CS, Murugan VS, Harrison A, Mbasso WF. Trustworthy chronic kidney disease screening in smart health systems using internet-of-things-integrated data processing and verifiable convolutional neural networks. Health Informatics Journal. 2026;32(4). doi:10.1177/14604582261508405
BibTeX
@article{Basha2026,
author = {Basha, S. Mahaboob and Shieh, Chin Shiuh and Murugan, V. Senthil and Harrison, Ambe and Mbasso, Wulfran Fendzi},
title = {Trustworthy chronic kidney disease screening in smart health systems using internet-of-things-integrated data processing and verifiable convolutional neural networks},
journal = {Health Informatics Journal},
year = {2026},
volume = {32},
number = {4},
doi = {10.1177/14604582261508405}
}
EUVIMED