Answers are grounded in this paper and cite their sources.
This study aimed to analyze the main issues patients’ experience while managing their chronic illness by using a large language model (LLM), an artificial intelligence system trained to process human-like text to examine qualitative interview data. This LLM-assisted approach allows for more efficient and nuanced identification of central themes, complementing traditional qualitative approaches. Thirty participants with various chronic conditions were interviewed, and the transcripts were analyzed using an LLM to identify key themes. A correlation analysis was conducted to select weakly correlated themes, indicating limited overlap between themes within the final set, for in-depth exploration. Content domains were then created by grouping related themes based on predefined criteria. The LLM assigned intensity scores for all 46 themes. Variation was observed in how themes were expressed within this exploratory sample of older adults with multimorbidity. Correlation analysis identified patterns of co-occurrence, from which six content domains were created. These domains represent model-derived groupings rather than validated constructs. This study outlines a structured approach for AI-assisted qualitative analysis. LLM-generated scores reflect model interpretation rather than direct patient-reported burden and should be interpreted descriptively. Further validation, including comparison with human coding, is needed before clinical use.
No takes yet. Share an insight, caveat, or question.
Kivity et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: