As a Whole
Understanding dAIbetes
As a Whole
The federated dAIbetes technology will harmonise existing data of ca. 800,000 type 2 diabetes patients of 6 cohorts distributed across the globe, and learn prognostic virtual twin models. Those will be validated for their clinical relevance and applied in clinical practice through a dedicated software.
We aim to demonstrate that our personalised predictions have a prediction error that is at least 10% lower than that of population average-based models.
This federated virtual twin technology will enable personalised disease management and act as a blueprint for other complex diseases. Our consortium combines expertise in artificial intelligence, software development, privacy protection, cyber security, and diabetes and its treatment.
Ultimately, we aim to resolve the antagonism of privacy and big data in cross-national diabetes research. See the dAIbetes project vision diagram below.
Background
Virtual twins in medicine have been described as “virtual representations” of patients and have the potential to fundamentally impact future advances in precision medicine. The development of multi-scale, multi-organ dynamic, interoperable virtual twin models for personalised disease management requires the integration of a variety of data from diverse data sources.
Yet, this multi-source data integration is hampered due to important patient data protection laws, such as the General Data Protection Regulation (GDPR).
We will utilise these privacy-preserving techniques to develop dAIbetes, a highly integrated health data platform for the development and application of the first internationally unified and validated virtual twin models for personalised outcome prediction of type 2 diabetes treatment.
The developed dAIbetes platform will enable the utilisation of the full potential of the available data while honouring data protection laws and regulations.
Breakthrough & Innovation
The current state of the art in multi-centric big data and statistical data analysis requires data centralization, often using a cloud, which threatens the privacy of patient-derived data, hindering its broad application in health care.
dAIbetes will harmonise and connect such data silos eventually covering type 2 diabetes data sets from ca. 800,000 patients in a federated database network (dAIbetes-Net) that registers the existence of data but does not store the data itself. Federated statistical learning tools for virtual twin model training apps (dAIbetes virtual twin apps) only exchange intermediate parameters computed on local data through the dAIbetes-Net network and learn virtual twins from the distributed data in a privacy-preserving fashion.
In dAIbetes, we leverage federated learning to build a federated health data platform, creating the first internationally trained virtual twin models for type 2 diabetes. Our models integrate big data across various sources while ensuring privacy compliance. This innovative approach aims to personalize treatment outcome predictions, which currently lack precise guidelines, using data from about 800,000 patients.
