Associated Projects

Within the KI-AIM ("AI-based anonymization in medicine") project, researchers develop an anonymization platform to provide large volumes of realistic data. This data has no personal reference and can therefore be used more easily under data protection law. This will facilitate access to data for the research and commercial development of data-based medical solutions. The project's research focuses on procedures for the flexible combination of anonymization methods for sensitive patient data and synthesis methods for the generation of data sets. The developers also focus on possible transferability to different medical specialties. In addition, researchers assess privacy risks using innovative models and evaluate the realism of the output data for anonymized and synthesized data.

Innovations and perspectives

As part of KI-AIM, innovative methods for anonymizing personal data are being developed and evaluated on the basis of a specific application in oncology, taking into account residual risks. The availability of verifiably anonymous data will make it easier both to conduct research in the field of AI and to develop AI-based commercial products for medical institutions. This will enable further innovations in the field of AI applications in the future. By providing the developed platform as open source software and establishing a user community, it is possible to create a location advantage for Germany and Europe in the medium to long term through the improved availability of data.

Project coordinator: University of Münster (Press release in German)

The project Medinym ("AI-based anonymization of personal patient data in clinical text and voice datasets") investigates the possibility of further reusing sensitive data by removing sensitive information through anonymization. Two medical use cases, text-based data from electronic patient records and voice data from diagnostic doctor-patient consultations, are being implemented as examples in the project. To this end, open technologies for anonymization are being investigated, further developed and applied to real data. The researchers are also investigating how the informative value of such anonymized data can be preserved for further use. In addition, methods that prevent or hinder misuse of the technology outside of the intended use case will be considered.

Innovations and perspectives

Information-preserving anonymization should make it possible to further process clinical data, as de-anonymization is no longer possible. These data sets can then be used to train AI models on clinical data in compliance with data protection regulations or be extended to other cohorts. This would make it possible for small and medium-sized companies to collect corresponding amounts of data cumulatively. This would allow sensitive data to be pooled across multiple applications and used for AI training routines, provided it is always anonymized accordingly. The desired anonymization should also increase the willingness of patients to consent to participation in studies, data analyses and general donations of health data. Ultimately, information-preserving anonymization allows the technology to be integrated into current development methods and diagnostic systems, thereby strengthening Germany as a location for science and business in the fields of diagnostics, treatment and therefore healthcare in general.

Project coordinator: Otto von Guericke University of Magdeburg (Further information in German)

© Fraunhofer IDMT

Project NEMO ("Non-identifiability of data from electroencephalography for Open Science") provides relevant insights into specific risk scenarios and the development and testing of anonymization procedures in the sensitive field of health data collection. In addition, the project lays the foundation for solid data protection concepts and anonymization procedures. In this way, a tried-and-tested technical infrastructure can be created that enables the comprehensive use of data in various applications while safeguarding data protection. The envisaged technological solutions offer a highly unique selling point in terms of data protection in the healthcare market. There is a realistic possibility of adapting the procedures for other time-dependent sensor data with manageable effort, which opens up a multitude of further application and utilization possibilities.

Innovations and perspectives

The NEMO project provides relevant insights into specific risk scenarios and the development and testing of anonymization procedures in the sensitive field of health data collection. In addition, the project lays the foundation for solid data protection concepts and anonymization procedures. In this way, a tried-and-tested technical infrastructure can be developed that enables the comprehensive use of data in various applications while safeguarding data protection. The envisaged technological solutions offer a highly unique selling point in terms of data protection in the healthcare market. There is a realistic possibility of adapting the procedures for other time-dependent sensor data with manageable effort, which opens up a multitude of further application and utilization possibilities.

Project coordinator: Fraunhofer IDMT, Oldenburg (Further information)

In the "Personal Mastery Health & Wellness Data" (PATH) project, the researchers are pursuing the goal of creating a data protection-compliant platform to link personal health data from patient records with data from home healthcare systems. One aspect is the simple and user-friendly access to health data in a graphical overview, which creates the necessary transparency for effective control by users. This enables citizens to make individual decisions about the sharing of their data and to give specific consent based on easily understandable and accessible platform interfaces. The researchers are developing open source modules for obtaining consent, managing it and monitoring the use of data in research. As a demonstration model, the resulting "data hub" will be tested, validated and subsequently evaluated in realistic case studies with diabetes and psychiatric patients. 

Innovations and perspectives

The PATH project uses the opportunities offered by digitalization to make use of the full potential of medical data in the service of health. For the first time, the "data hub" integrates individually generated medical data and traditional paper-based hospital cards into a data protection-compliant platform. The platform enables medical data to be shared in compliance with data protection regulations. The data collected can be used for market surveillance to demonstrate the safety, performance and clinical benefit of marketed medical devices. In the future, the project will open up further potential for research and development of medical products and services through the provision of anonymized data.

Project coordinator: Dresden University of Technology (Press release)