From Ethical Principles to Auditable Systems: A Regulatory Operational Assurance Framework for Trustworthy Healthcare AI
International Journal of Drug Delivery Technology · 30.04.2026 · 16 (23s)
DOI: 10.25258/ijddt.16.23s.3 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Drug Delivery Technology, online erschienen am 30.04.2026. Verfasst von Amna Akram und 5 weiteren Autor:innen.
Abstract (Originalsprache)
Artificial intelligence (AI) has proven significant capabilities of enhancing the accuracy of diagnoses, clinical decision making, and efficiency of care. Nevertheless, the successful application of AI in safety critical clinical contexts has not been well implemented in practice resulting in persistent concerns of fairness, accountability, transparency, and regulatory compliance. Current ethical frameworks define desirable principles of trustful AI, whereas technical approaches offer metrics and interpretability means, but neither offers a single, operational channel of regulatory assurance. This paper introduces the Trustworthy AI Assurance Framework (TAIAF), a regulatory operational model that redefines trustworthiness as a system assurance property, similar to safety or reliability. TAIAF incorporates fairness, accountability and transparency as verifiable and auditable properties throughout the AI lifecycle to align system design, implementation, and monitoring to U.S. regulatory expectations. The framework consists of five assurance layers integrated: Governance & Compliance, Data and Model Development, Decision and Control, Deployment and Runtime Monitoring and Audit and Accountability producing inspection read artifacts that can be certified, audited, and monitored in the after market. The empirical evidence of TAIAF is based on publicly available healthcare datasets on the U.S. such as NIH ChestX-ray14, MIMIC-III, and OpenFDA. Instead of benchmarking predictive performance, the case studies demonstrate how the three components of demographic risk stratification, data governance signals, provenance records, and post market transparency artifacts are operationalized as the components of trustworthiness. Findings demonstrate that fairness may be viewed as equitable exposure to diagnostic risk, accountability as reconstructions of traceability, and transparency as auditability other than explainability. This work bridges the gap between ethical intent, technical implementation, and regulatory oversight by actually mapping TAIAF to FDA guidance, NIST AI Risk Management Framework, and ISO 14971, as well as U.S. quality system regulations. TAIAF offers a viable, lifecycle grounded roadmap to the development, assessment, and governance of secure, disproportionate, and answerable healthcare AI systems within current regulatory frameworks
Zitieren
APA
Akram, A., Riaz, K., Singh, P. K., Rahim, S. A., Smerat, A., & Abbas, S. (2026). From Ethical Principles to Auditable Systems: A Regulatory Operational Assurance Framework for Trustworthy Healthcare AI. International Journal of Drug Delivery Technology, 16(23s). https://doi.org/10.25258/ijddt.16.23s.3
Vancouver
Akram A, Riaz K, Singh PK, Rahim SA, Smerat A, Abbas S. From Ethical Principles to Auditable Systems: A Regulatory Operational Assurance Framework for Trustworthy Healthcare AI. International Journal of Drug Delivery Technology. 2026;16(23s). doi:10.25258/ijddt.16.23s.3
BibTeX
@article{Akram2026,
author = {Akram, Amna and Riaz, Kinza and Singh, Prakash K. and Rahim, Safwan Abdul and Smerat, Aseel and Abbas, Salbia},
title = {From Ethical Principles to Auditable Systems: A Regulatory Operational Assurance Framework for Trustworthy Healthcare AI},
journal = {International Journal of Drug Delivery Technology},
year = {2026},
volume = {16},
number = {23s},
doi = {10.25258/ijddt.16.23s.3}
}
EUVIMED