Data Integration for the Cancer Registry Using Artificial Intelligence

Project »EPIK: Decision Support for the Allocation of Individuals in the Rhineland-Palatinate Cancer Registry«

The cancer registries of the German federal states are important institutions in the fight against cancer. They collect and collate data on cancer cases and treatments. Based on this, they compile statistics on incidence and survival rates and provide comparative data for medical research. However, compiling this information from various sources is often still done manually – a laborious process that requires a great deal of care and time.

In the »EPIK« project, we are working on a method and software solution for the Rhineland-Palatinate epidemiological cancer registry that supports correct and efficient data integration.

Challenging Integration of Data From Different Sources

The cancer registry brings together data from patients diagnosed with cancer by doctors, registration data from residents' registration offices and information on deaths. After checking the incoming data for relevance, it is imported into the cancer registry database and, where possible, assigned to the cancer patients already recorded.

Our solution proposes customized procedures for dealing with gaps, inconsistencies or errors in the data.

AI Methods to Support Data Consolidation

The software solution being developed in the project uses various Artificial Intelligence (AI) methods to support data integration with suitable suggestions for action. With the help of unsupervised learning, similar data records are recognized and suggested for possible merging. A knowledge-based system maps established manual procedures and automatically provides suitable suggestions for action. The decisions to accept or reject the suggestions are saved and evaluated using statistical and supervised learning methods – creating opportunities for targeted process optimization.

Project Partner

IDG Institute for Digital Health Data Rhineland-Palatinate gGmbH (IDG Institut für digitale Gesundheitsdaten Rheinland-Pfalz gGmbH)

Process of the Project »EPIK: Decision Support for the Allocation of Individuals in the Rhineland-Palatinate Cancer Registry«
© Fraunhofer ITWM / freepik
Process of the Project »EPIK: Decision Support for the Allocation of Individuals in the Rhineland-Palatinate Cancer Registry«