BRIDGE stands for Bias Reduction for Implementation, Deployment, Governance, and Equity. BRIDGE-AI is a new, cross-design reporting guideline for evaluation studies of artificial intelligence (AI) in digital health, particularly diagnosis, prognosis, clinical decision support, workflow integration, and post-deployment monitoring. It is not conceived as an extension of a single existing guideline. Instead, it addresses reporting items that cut across study designs and evaluation phases and are currently dispersed across the AI literature.
The guideline is organised around eight workgroups: WG1 fairness and data representativeness; WG2 robustness and stress-testing; WG3 traceability, documentation and accountability; WG4 explainability and transparency; WG5 usability, human-AI interaction and automation bias; WG6 governance, ethical oversight and regulatory alignment; WG7 clinical and public health integration/workflow evaluation; and WG8 post-deployment monitoring, drift detection and lifecycle quality.
Candidate reporting items are also being grounded in ISO/IEC standards and technical specifications, including ISO/IEC TR 24027 and ISO/IEC TS 12791 for bias and unwanted bias treatment, ISO/IEC TR 24028 for trustworthiness and transparency, ISO/IEC TR 24029-1 and ISO/IEC 24029-2 for robustness assessment, ISO/IEC 23894 for AI risk management, and ISO/IEC TS 4213 for machine learning classification performance assessment.
A new guideline is needed because existing AI guidance on EQUATOR is largely design-specific or phase-specific—for example trials, protocols, diagnostic accuracy studies, prediction models, or early-stage clinical evaluation—whereas trustworthy digital health evaluation also requires consistent reporting of subgroup performance, oversight, lifecycle monitoring, workflow fit, and post-deployment drift. BRIDGE-AI will therefore complement existing guidance while remaining a standalone new reporting guideline entry.
Project website: https://bridge-a-i.com/
Registered: 01 April 2026
Last updated:
Contact: Hamid Reza Marateb Email: [email protected]