ROKAVI Project Logo

ROKAVI

Risk Prediction in Oncology: Development of an AI Algorithm, Insured Persons’ Preferences, and Ethical Implications

 

Cancer is one of the most common diseases worldwide, and the number of people affected is rising steadily. It places a heavy burden on both patients and the healthcare system. Through targeted preventive measures and the earliest possible diagnosis, the chances of successful treatment can be significantly increased. In this field, models that utilize artificial intelligence (AI) offer great potential for early disease detection and for improving patient care in the future. This is where the project “Risk Prediction in Oncology: Development of an AI Algorithm, Insured Persons’ Preferences, and Ethical Implications” (ROKAVI) comes in.

 

Project Duration: 03/2026 – 02/2029

 

Project Sponsor

Innovation Fund of the Joint Federal Committee


The objectives of the project are:

  • the development and testing of algorithms for predicting malignant neoplasms at a specific site
  • to develop an ethical framework for the use of risk prediction algorithms in oncology
  • to survey insured individuals’ preferences regarding whether and under what conditions they wish to receive information about their cancer risk
  • to determine the benefits and risks of the developed prediction model from the perspective of health insurance companies
  • deriving recommendations for the use of risk prediction algorithms in oncology

Based on a systematic literature review and an expert workshop, specific types of cancer for which predictive models appear particularly suitable will be identified. Subsequently, corresponding models will be developed using health insurance data. The predictive accuracy for the risk of developing a selected type of cancer will be compared exploratively with a statistical algorithm developed in parallel.

In addition, the ethical implications arising from the use of such models in everyday clinical practice are analyzed. The preferences of insured individuals are assessed using a discrete-choice experiment. The perspective of statutory health insurance providers is also systematically taken into account through their targeted involvement in the project. Building on these findings, a concluding workshop will develop concrete recommendations for the future use of AI-based predictive models.

MHH team from left to right: Dr. Kathrin Krüger, M.A. Viktoria Wiesner, Prof. Dr. Christian Krauth, M. Sc. Laura Böhme Copyright: Karin Kaiser/MHH
MHH team from left to right: Dr. Kathrin Krüger, M.A. Viktoria Wiesner, Prof. Dr. Christian Krauth, M.Sc. Laura Böhme Copyright: Karin Kaiser/MHH

We are researching

  • the first-ever use of German health insurance data for risk prediction

  • the risk communication regarding cancer to insured individuals

  • the preferences of those with statutory health insurance

  • the use of algorithms for risk prediction in oncology within an ethical framework


Our target group

  • all individuals with statutory health insurance


Contact

Prof. Dr. Christian Krauth

Dr. Kathrin Krüger

M. Sc. Laura Böhme

M.A. Viktoria Wiesner