Exploring large language models for dermatology inbox management: A solution for provider burnout
Need to claim your poster? Find the KiKo table at the conference and they'll help
you get set up.
Presented at: Society for Investigative Dermatology 2025
Date: 2025-05-07 00:00:00
Views: 2
Summary: Abstract Body: Background: Studies have shown that dermatologists spend a significant portion of their workday responding to patient messages, with recent data identifying inbox management as a leading contributor to provider burnout and decreased job satisfaction. Large Language Models (LLMs) have demonstrated capabilities in natural language processing and content generation across various industries, suggesting potential applications in clinical communication management. Methods: This paper explores the theoretical framework and practical considerations for implementing LLM assistance in dermatology inbox management. We review current challenges in message handling, examine available LLM technologies, and analyze potential integration strategies within existing electronic health record systems. Key considerations include message categorization, response automation capabilities, provider oversight requirements, and patient privacy protection. Results: Our analysis suggests that LLM implementation could streamline common inbox tasks such as medication inquiries, appointment scheduling clarifications, and basic skincare questions. Critical factors for successful integration include maintaining appropriate provider oversight, ensuring HIPAA compliance, and developing dermatology-specific training datasets. Potential benefits include reduced provider time burden, standardized response quality, and improved work satisfaction while maintaining patient care quality. Conclusion: LLM technology presents a promising solution for dermatology inbox management challenges. Future research should focus on developing specialized dermatology language models, establishing safety protocols, and conducting controlled trials to measure the impact on provider workload and patient satisfaction. Faranak Kamangar<sup>2</sup>, Andrea Leung<sup>1</sup>, Georgia Marquez-Grap<sup>1</sup>, Allison Kranyak<sup>1</sup>, Wilson Liao<sup>1</sup>, Tina Bhutani<sup>1</sup> 1. Dermatology, University of California San Francisco, San Francisco, CA, United States. 2. Dermatology, Palo Alto Medical Foundation, Palo Alto, CA, United States. Bioinformatics, Computational Biology, and Imaging