Linking cancer registry data and health insurance data: Investigation of the acceptance by cancer patients and of the potential to describe courses of therapy and to use AI for quality control


In order to be able to use clinical data from cancer registries, it is important to be to assess the completeness of therapy information in the cancer registry data. Established indicators do not yet exist. Health insurance data can provide valuable information regarding the expected frequencies of therapy and types of therapy overall and in relation to other factors, such as age.

Goals and approach:
The KUKUK project therefore develops and applies methods and indicators for assessing the completeness of therapy reports. In the project, therapy information in health insurance data will first be evaluated - exemplarily for lung, colorectal, breast and prostate cancer. For this purpose, the research database GePaRD is used, which contains health insurance data of about 20% of the German population. Based on the findings, an algorithm will be developed that will be applied and optimized step by step to assess the completeness of therapy information in datasets of the Center for Cancer Registry Data and the Bremen Cancer Registry. In order to be able to assess whether this type of data linkage would also meet with acceptance in routine practice, approx. 400 cancer patients will be asked in the course of the information on reporting to the cancer registry whether they would agree to the reported data being linked with their case data.

Perspectives for practice:
The results of the KUKUK project should help to better assess the completeness of therapy information in cancer registries. If it also turns out that the linking of cancer registry and health insurance data would meet with acceptance among cancer patients, this could significantly simplify the collection of complete therapy information in the future.

Funding period

Begin:   October 2022
End:   September 2025


  • Federal Ministry of Health


Dr. med. Sabine Luttmann