OBJECTIVES: Understanding patient preferences is crucial for informing various aspects of a disease, including the condition, unmet needs, and treatment preferences. However, capturing these is challenging due to extensive logistics, limiting direct patient engagement. This research explores how GPT can promote more conversations with patients (which is what matters the most), while reducing time for conducting patient-preference research(PPR).

METHODS: We reviewed the recent literature and trends in PPR, followed by expert discussions on current challenges and how GPT can mitigate them to allow more time for meaningful patient interactions.

RESULTS: GPT can support several aspects of PPR, including review, conceptualization, design, instrument development, implementation, analysis, and interpretation. GPT can help conceptualize the study by identifying themes of existing data to inform study design and indicate unmet needs, followed by streamlining logistical planning, including detailed plans and establishing timelines that often utilize most resources in PPR. In the design phase, GPT can help create tailored questionnaires and interview guides for diverse populations, including translations and simplified contexts for better patient interpretation. In analysis, GPT can supplement researcher-analyzed data by uncovering additional insights, understanding the tone of transcripts, sentiment analysis, and creating visualizations using the DALL-E extension (though its current applicability is limited), improving communication and efficient dissemination of findings. The use of GPT requires careful considerations, including verification, validation, and ethical aspects (i.e., data security, privacy, autonomy).

CONCLUSIONS: The current emphasis is on utilizing GPT to support PPR, offering researchers more time for direct interactions with patients and robust analysis while optimizing resources. Looking ahead, AI-driven chatbots can play a crucial role in personalized patient engagement by gathering preferences and real-time feedback, ensuring anonymization and careful interpretation. Virtual chatbots can also aid in interview preparation, providing diverse learning opportunities for researchers. This not only expedites time sensitive PPR but also creates opportunities for wider utilization.