paulorades / mi-claim

A community repository for MI_CLAIM (Minimum Information for CLinical AI Modeling) reporting standards

MI-CLAIM

A community repository for MI_CLAIM (Minimum Information for CLinical AI Modeling) reporting standards

Authors and Original Manuscript

Beau Norgeot, Giorgio Quer, Brett K. Beaulieu-Jones, Ali Torkamani, Raquel Dias, Milena Gianfrancesco, Rima Arnaout, Isaac S. Kohane, Suchi Saria, Eric Topol, Ziad Obermeyer, Bin Yu & Atul J. Butte

MI-CLAIM was published in Nature Medicine on September 9th, 2020 (https://www.nature.com/articles/s41591-020-1041-y)

About MI-CLAIM

MI-CLAIM has two purposes:

1. to enable a direct assessment of clinical impact, including fairness
2. to allow rapid replication of the technical design process of any legitimate clinical AI study

About This Repo

Our goal is to develop a documentation standard that can serve clinical scientists, data scientists, and the clinicians of the future who will be using these tools. To that end, a checklist is provided as a part of MI-CLAIM that should be included along with each clinical AI model or manuscript. Additionally, we hope that this description will stimulate discussion of the proposed MI-CLAIM standards, and we encourage the clinical community, as well as the AI community, to provide us with their views on how this standard can be improved. This repository has been set up to coincide with the release of the paper and will allow the community to comment on existing sections and suggest additions. Please contribute by generating Issues.

As we gather feedback, and as the field evolves, we plan to publish comprehensive updates to the MI-CLAIM standards

The MI-CLAIM Checklist

A copy of the checklist can be obtained in this repo for easy access (MI_CLAIM_checklist_repo.docx). The checklist is in MS Word Table format, but otherwise exactly matches the table attached to the original manuscript