Skip to main content Skip to main navigation

Project

SmartNTX

Interaktive KI-basierte Entscheidungskomptentenz für die Transplantationsmedizin

Interaktive KI-basierte Entscheidungskomptentenz für die Transplantationsmedizin

  • Duration:

The chances of success of a transplant have steadily improved thanks to medical progress; this also applies to the kidney, the most frequently transplanted organ in Germany. Nevertheless, especially in the first year after the operation, it is important that recipients take their medication regularly and adhere to the prescribed treatment plan. They must first learn what to look out for in everyday life and which physical changes may be alarm signals. Monitoring is crucial to avoid complications. Although the technical requirements exist to record the crucial factors, this information has not yet been systematically collected and used for treatment.

The smartNTX project aims to improve this. To this end, transplanted patients use apps and app-linked measuring devices to document their vital data, well-being and medication intake. This data is digitally transmitted to a telemedicine team, which contacts transplant recipients if their values deteriorate. To evaluate the transplant recipients' data and derive treatment recommendations, artificial intelligence techniques will be used to support the physicians. In weekly case conferences, the telemedicine team and the treating nephrologist will discuss further therapy decisions.

To test the procedure, a study is being conducted with transplanted patients. Half of the participants will be treated for twelve months with the new form of care, the other half according to standard care. A comparison will then be made of which group of patients has a better transplant condition, prognosis, quality of life and disease management.

As part of smartNTX, DFKI staff are working on methods to support telemedicine staff. Specifically, AI components are being developed within clinics that analyze communication between patients and telemedicine staff to identify, for example, inconsistent user input or risk factors. Furthermore, a chatbot component (or rather generative AI) will be developed to answer non-critical and common user queries and increase therapy adherence.

Partners

Friedrich-Alexander-Universität Erlangen-Nürnberg, Charité – Universitätsmedizin Berlin, Universitätsklinikum Essen, Hahn-Schickard-Gesellschaft für angewandte Forschung e. V.

Sponsors

Innovationsfonds des Gemeinsamen Bundesausschusses (G-BA)