Applying Large Language Models to Interpret Qualitative Interviews in Healthcare

Area of Science:

• Healthcare Systems
• Qualitative Data Analysis
• Artificial Intelligence

Background:

• Healthcare challenges necessitate integrating patient and professional perspectives.
• Current qualitative data analysis is time-consuming and labor-intensive.
• Existing automated methods like topic modeling may reduce data richness.

Purpose of the Study:

• To evaluate Large Language Models (LLMs) for semi-automated interpretation of qualitative interview data.
• To compare LLM-based approaches against traditional topic modeling and manual analysis.
• To explore LLMs' potential in enhancing efficiency and depth in qualitative research for healthcare.

Main Methods:

• A novel approach utilizing LLMs was developed for qualitative data interpretation.
• LLM performance was compared with topic modeling techniques.
• The study used two distinct qualitative interview datasets for validation.

Main Results:

• LLMs demonstrated potential in supporting the interpretation of qualitative interview data.
• The LLM approach showed promise in retaining data richness compared to topic modeling.
• Exploratory findings suggest LLMs can aid in incorporating diverse human perspectives.

Conclusions:

• LLMs offer a viable tool for semi-automated qualitative data analysis in healthcare research.
• This technology can accelerate the interpretation process, making it more efficient.
• LLMs have the potential to significantly contribute to the development of sustainable healthcare systems by better integrating stakeholder insights.