Project CANAIRI Correspondence Published
Project CANAIRI: Collaboration for trANslational Artificial Intelligence tRIals. Established in 2024 as an initiative focused on developing guidance for the silent evaluation phase of AI translation efforts at local health settings. This initiative is co-lead by Dr. Melissa McCradden and Dr. Xiao Liu in partnership with 24 academic, regulatory and industry organizations across 4 different countries.
Project CANAIRI’s mission is to provide actionable, ethics-forward evidence-based guidance for health settings to conduct these silent phase evaluations. We will work to ensure this guidance is accessible and practical for anyone - from the largest academic health centre to the smallest community hospital.
To achieve our mission, we are mapping out how we are currently conducting these silent phase evaluations to identify what’s going well, what could go better, and what are some gaps in the current landscape.
We will connect with experts, stakeholders, and consumers all across the world to hear from them about what they think is important, what barriers they’re facing, and where we need to go as a global community to facilitate better translation through local evaluations.
We will work with multi-disciplinary collaborators to consider the literature, the feedback from global stakeholders, values of relevant authorities, and on-the-ground realities to put together a toolkit for local evaluations of AI tools.
We will develop knowledge resources, co-designed with consumers to support health AI literacy knowledge for this critical stage in evaluation of Health AI.
We will test our toolkit through a diverse set of use cases to make sure we’ve got it right!
Project CANAIRI is grateful for the support of the Australian Institute for Machine Learning’s Centre for Augmented Reasoning. This project is funded by a Project Grant from the Canadian Institute for Health Research. Dr. McCradden is grateful to The Hospital Research Foundation Group for their support of her research.
Correspondence
CANAIRI: the Collaboration for Translational Artificial Intelligence Trials in healthcare
Melissa McCradden, Xiaoxuan Liu, and the Project CANAIRI Working Group, Nature Medicine, January 2025.
The full paper can be found here.
Article
AI's canary in the coal mine: Translational trials for accountable AI integration
Melissa McCradden, Xiaoxuan Liu, and the Project CANAIRI Team, Medium, August 2024.
The full paper can be found here.
Goverance Paper
A silent trial is critical to accountable and justice-promoting implementation of artificial intelligence in healthcare
Melissa McCradden, Global Forum on Bioethics in Research, November 2022.
The full paper can be found here.
Poster Presentation
Collaboration for trANslational Artificial Intelligence tRIals: Project CANAIRI
Melissa McCradden, Xiaoxuan Liu; Judy Gichoya, et. al., ML4HC , August 2024.
The full paper can be found here, Paper ID: 114.
Check out work from CANAIRI members!
Article
A framework for understanding label leakage in machine learning for health care
Sharon E Davis, Michael E Matheny, Suresh Balu, Mark P Sendak, JAMIA, September 2023.
The full paper can be found here.
Article
Implementation frameworks for end-to-end clinical AI: derivation of the SALIENT framework
Anton H. van der Vegt, Ian A. Scott, et. al., JAMIA, May 2023.
The full paper can be found here.
Article
How to validate Machine Learning Models Prior to Deployment: Silent trial protocol for evaluation of real-time models at ICU
Sana Tonekaboni, Gabriela Morgenshtern, et. al., PMLR, 2022.
The full paper can be found here.
Article
The silent trial - the bridge between bench-to-bedside clinical AI applications
Jethro CC Kwong, Lauren Erdman, et. al., Front. Digit Health, August 2022.
The full paper can be found here.
Article
The medical algorithmic audit
Xiaoxuan Liu, Ben Glocker, et. al., Lancet Digit Health, May 2022.
The full paper can be found here.
Article
Do no harm: a roadmap for responsible machine learning for health care
Jenna Wiens, Suchi Saria, Mark Sendak et. al., Nature Medicine, September 2019.
The full paper can be found here.
To learn more check out this presentation!