The Veterans Cardiac Health and AI Model Predictions (V-CHAMPs) Challenge was a first from the VHA Innovation Ecosystem, the FDA Digital Health Center of Excellence, precisionFDA (part of the FDA's Office of Digital Transformation) and the Medicines and Healthcare products Regulatory Agency (MHRA).
Using synthetic datasets based on real US veteran health records, participants explored a range of potential outcomes following heart failure, including the prediction of cardiovascular outcomes, hospital re-admission rates and mortality, over a two phase, nine-month challenge.
This insight is designed to help health systems both in the US and the UK to better understand the cardiovascular risks and specific health outcomes in veterans, as well as increasing understanding of synthetic data and AI/ML models and how they could be used to explore similar topics in the future.
The five winning teams, in rank order, are:
- VHeartAI
- ZS Associates ADS Team
- TeamDSS
- CognizantGoodHeartML
- Team D&D
Puja Myles, Director of the Clinical Practice Research Datalink at the MHRA, said:
This has been a fantastic opportunity for developers to challenge themselves against other experts across the globe, all while contributing towards a worthy goal.
This challenge has not only shown how advances in AI methodology can help improve healthcare, by providing forecasts of certain health outcomes in specific patient populations like veterans, it has also demonstrated the potential of synthetic data for the initial training and testing of AI algorithms.
So, it has been fascinating to see how the best and brightest participants from around the world tested this potential to its limits in this challenge, and the knowledge that our sector has gained from their work will be invaluable.
I'm so pleased to have been able to work with our US partners, bringing together cutting-edge health data expertise from both sides of the Atlantic, and helping to drive forward our combined understanding for application in future projects.
