AI4HF aims to develop the first trustworthy AI solutions for personalised risk assessment and management of HF patients. The project will build on a unique set of big data repositories, trustworthy AI methods, computational tools and clinical results from major EU-funded projects in cardiology. To test robustness, fairness, transparency, usability and transferability, the validation with take place in eight clinical centres in both high- and low-to-middle-income countries in the EU and internationally. AI4HF will develop a comprehensive and standardised methodological framework for trustworthy and ethical AI development and evaluation based on the FUTURE-AI guidelines developed by the consortium members. The ambitious objective is to codesign, develop, evaluate and exploit an integrative and trustworthy artificial intelligence (AI) solution for personalised HF risk assessment, trained from large, complementary multi-source cardiovascular data, including (1) cardiac imaging, (2) cardiac biomarkers, (2) electrocardiography (ECG) and (4) information from clinical reports. AI4HF will be implemented through continuous multi-stakeholder engagement, taking into account clinical needs and patient preferences, as well as socio-ethical and regulatory perspectives.
SRDC is the leader of a work package responsible for designing and developing the traceability and monitoring tools for post-deployment of AI models. In addition, SRDC is leading the tasks working on health data extraction, harmonization and management to build and deploy a health data interoperability pipeline for AI model training and online execution. Under these roles, SRDC will specifically be working on the following items:
- Utilize available health data modeling and management standards like HL7 FHIR (https://www.hl7.org/fhir/) and OMOP CDM (https://ohdsi.github.io/CommonDataModel/)
- Harmonize the available datasets into standard formats addressing both syntactic and semantic interoperability
- Guide the clinical partners while preparing Data Management Plans
- Design and develop a data quality controller component to assess the quality of harmonised datasets w.r.t many conformance, completeness and plausibility metrics
- Design and implement an AI product passport that will provide key information about the AI’s production and maintenance, including model properties and hyperparameters, training and testing datasets, evaluation metrics and results, biases and other limitations, ethical approvals and data governance, as well as monitoring and continuous evaluations.
- Build human feedback interfaces to ease human-AI interaction loop while feeding the AI processes with human input,
- Design and develop monitoring dashboards to present live information to the end-users, including time-series statistics and visualisations that will inform on potential deviations, anomalies or biases of the AI model performances.
|1.||Netherlands Heart Institute||Netherlands|
|2.||University of Barcelona||Spain|
|3.||Centre for Research & Technology||Greece|
|4.||Barcelona Supercomputing Centre||Spain|
|6.||Vall d’Hebron Hospital Research Institute||Spain|
|7.||St. Anne's University Hospital Brno||Czechia|
|8.||Muhimbili University of Health and Allied Sciences||Tanzania|
|9.||Cardiovascular National Institute||Peru|
|12.||European Society of Cardiology||France|
|13.||European Heart Network||Belgium|
|14.||University Medical Centre Utrecht||Netherlands|
|15.||Academic Medical Centre Amsterdam||Netherlands|
|16.||University of Oxford||United Kingdom|