In August, Janna, Lotti, and Marie attended the Medical Informatics Europe Conference “MIE2024” held in August in Athens, Greece. Marie had the privilege of presenting our research paper, “Applying Large Language Models to Interpret Qualitative Interviews in Healthcare.”
In our study, we explored how Large Language Models (LLMs) can enhance and streamline the process of analyzing qualitative data in healthcare research. Our findings suggest that LLMs have the potential to significantly complement traditional human-driven methods, such as coding and annotation, by providing an additional layer of automation and efficiency. This opens up exciting new possibilities for scaling up qualitative research while maintaining the depth and nuance required in healthcare contexts.
By leveraging LLMs, researchers can accelerate the often time-consuming process of qualitative analysis, freeing up time for deeper interpretation and allowing for a more iterative and expansive exploration of healthcare data.
If you are interested in our findings and the potential applications of LLMs in this field, you can access our full conference paper here, and full interview study here.
Medical Knowledge and Decision Support
We advance “Medical Knowledge and Decision Support” with data science and artificial intelligence, and qualitatively explore digital transformation in the clinic.
The overarching objectives of the research of the Medical Knowledge and Decision Support group are to implement innovative data science approaches, accelerate discovery, translation and integration of evidence into clinics, and to align technological progress with the needs of clinicians.
Methods
Our research draws methodologically from data science, bioinformatics, computer science, and psychology. We build ontologies and an associated ecosystem for logic-based automation. We apply and develop predictive models using AI/ML, particularly LLMs and neuro-symbolic approaches. We apply mathematical approaches including constraint-based modelling. In addition, we use qualitative approaches to study human perspectives.
Use Cases
We address a wide range of application use cases across our projects, including the discovery and prediction of functions and regulatory factors, biomedical and clinical evidence synthesis and automation, pre-trained large language and multi-modal model evaluation, extraction of data from clinical notes for predictive modelling, decision support, medical question answering, and interpretability and explanations for AI applications.
Clinical Areas
Our projects span a range of biomedical and clinical domains, including fundamental metabolism and metabolomics; drug discovery and molecular property prediction; RNA biology and regulatory interactions; oncology and personalised cancer biology; behavioural science and health psychology; mental health including anxiety, depression and psychosis; ophthalmology; infectious diseases; and physiotherapy.
Contact and Collaboration
If you are interested in partnering with us, have an idea you’d like to discuss, or simply want to get in touch, we would love to hear from you!
Please reach out via email to: janna.hastings@unisg.ch or janna.hastings@uzh.ch
School of Medicine
University of St.Gallen (HSG)
St. Jakob-Strasse 21
9000 St.Gallen, Switzerland
Institut for Implementation Science in Health Care (IfIS)
Universität Zürich
Universitätsstrasse 84
8006 Zürich, Switzerland
Applying LLMs to Interpret Qualitative Interviews in Healthcare: Insights from MIE2024
