Generative Artificial Intelligence for Clinical Communication: Implications for Non-Pharmacological Interventions in Health Care

Abstract

Abstract Objectives The objective of this study was to evaluate the potential of Generative Artificial Intelligence (AI) in facilitating clinical communication, particularly in addressing sexual health concerns, which are often challenging for patients to discuss. Methods We employed the Generative Pre-trained Transformer-3.5 (GPT) as the generative AI platform and utilized DocsBot for citation retrieval (June 2023). A structured prompt was devised to generate 100 questions from the AI, based on epidemiological survey data regarding sexual difficulties among cancer survivors. These questions were submitted to Bot1 (standard GPT) and Bot2 (sourced from two clinical guidelines). The responses from both bots were compared to assess consistency and adherence to clinical guidelines. Results Our analysis revealed no censorship of sexual expressions or medical terms. The most common themes among the generated questions were cancer treatment, sexual health, and advice. The similarity rate between responses from Bot1 and Bot2 averaged 92.5% (range 77.0% to 98.4%), with notably lower similarity for items not covered in the guidelines. Despite the lack of reflection on guideline recommendations, counseling and other non-pharmacological interventions were significantly more prevalent in both bots’ responses compared to drug interventions, with odds ratios of 4.8 (p=0.04) in Bot1 and 14.9 (p<0.001) in Bot2. Discussion Generative AI can serve for providing health information on sensitive topics such as sexual health, despite the potential for policy-restricted content. There was a significant skew towards non-pharmacological interventions in responses, possibly due to the prohibitive nature of medical topics. This shift warrants attention as it could potentially trigger patients’ expectations for non-pharmacological interventions.

Ähnliche Arbeiten