Research methods in family medicine: an exploratory study of eleven years of congress programs using GPT-5

Background: Scientific conferences reflect trends in Family Medicine research and education. In Germany, the annual congress program of the German Society of General Practice and Family Medicine (DEGAM) encompasses a wide range of topics and is publicly accessible. However, little is known about how research methodologies and topics evolve over time. Methods: All program items from the DEGAM conferences from 2014 to 2024 were analyzed. Using the Large Language Model GPT-5, each item was automatically categorized by research methodology and topic. Descriptive statistics were used to summarize trends. Results: A total of 2,869 program items were identified. Quantitative and interventional studies constituted 53.5% of all methodologies, while qualitative and mixed-methods accounted for 33.1%. The relative proportions remained largely unchanged over the eleven-year period, however future-oriented topics such as eHealth or sustainability do emerge. Although diverse topics were represented, they showed no clear methodological evolution in every topic. Conclusions: This first exploratory analysis of a national Family Medicine conference series shows that, while the thematic range is broad, research methodologies remain stable. Therefore, as in patient care, academic Family Medicine is the ‘decathlon’ of health service research, too.