To investigate ORL-HNS patients' familiarity with AI/LLMs, their usage patterns, and their trust in LLM-based medical information and recommendations.
Approach:
Questionnaire: Patients completed a 15-item questionnaire assessing their familiarity with AI/LLMs, usage patterns, and trust in LLMs, including specific scenarios of use.
Key Findings:
96% of patients were familiar with AI and 78% with LLMs.
59% reported using LLMs, with one third using them for health information.
LLMs were rated lower for perceived medical accuracy compared to comprehensibility, conciseness, and coherence.
Patients showed significantly lower confidence in LLM recommendations compared to those from doctors.
Interpretation:
Patients are familiar with and frequently use LLMs, but their trust in LLM-generated health information is limited, favoring recommendations from doctors.
Limitations:
The study was conducted at a single center, which may limit generalizability.
The sample size may not represent the broader population of ORL-HNS patients.
Self-reported data may introduce biases.
Conclusion:
While ORL-HNS patients are familiar with and use LLMs, they prefer medical advice from doctors, indicating a potential role for clinician-led LLM integration.
A phase 2 randomized trial suggests reducing nasal commensal viruses with ribavirin spray may improve allergic rhinitis symptoms while supporting a proposed disease mechanism.
