Yesterday, the first SCOR online seminar was held. For this inaugural session, we had the pleasure to hear Janna Hastings (University of Zurich, University of St. Gallen). You can find the abstract and recording of her talk below.
If you would like to stay informed about the next sessions, make sure to subscribe to our mailing list or to follow us on LinkedIn.
Abstract. While the volume of clinical research is rapidly increasing, it is difficult to gain an overview or use the overall body of evidence to answer specific questions for policy and practice due to fragmentation of the evidence base into different silos. Integrating data from evidence across such silos and building up consensus to inform policy and practice is challenging and, despite the large promises often repeated for AI, can only partially be supported by algorithmic approaches at present, due to subtle context-specific information and the range of different ways that such data is described and reported in the literature. Ontologies are semantic standards that allow data to be more widely integrated and shared, and they support consensus-building through fostering a common understanding of research outputs. However, they are also time-consuming to build and apply manually. Modern ‘large language models’ offer a flexible and powerful technology that can be applied to a wide range of text and data processing tasks, however, they can also introduce errors and biases. In this talk, through a case study in smoking cessation. I will discuss how these technologies can be used together to transform data and evidence into integrated, actionable insights.