人工智慧在初診醫療服務中的應用與效益

許富順、許菡倪¹、朱文儀²

台北市立聯合醫院陽明院區 泌尿科，¹清華大學 計量財金系，²台灣大學 商學研究所

Applications and benefits of artificial intelligence in initial medical services

Fu-Shun Hsu, Han-Ni Hsu¹, Wen-Yi Chu²

Department of Urology, Yangming Branch, Taipei City Hospital, Taipei, Taiwan;
Department of Quantitative Finance, National Tsing Hua University, Hsinchu, Taiwan¹; Business Administration, National Taiwan University, Taipei, Taiwan²

Purpose: Despite growing interest in AI applications in healthcare, existing research has largely focused on diagnostic accuracy, imaging analysis, or specific clinical decision support functions. Empirical studies examining the comprehensive impact of AI implementation on the initial consultation process—particularly in terms of patient waiting time, healthcare worker workload, and hospital management performance—remain limited. This study addresses this research gap by investigating the application and effectiveness of an AI-assisted initial consultation service platform from both clinical workflow and hospital management perspectives.

Materials and Methods: This study adopts a retrospective and comparative research design to assess system effectiveness. We establish an AI-assisted initial medical consultation platform, which based on combination of Large Language Models (LLMs) and enterprise-specific knowledge. Operational and workflow data were collected and analyzed before (100 patients) and after (100 patients) the implementation of the AI-assisted initial consultation platform. We analyzed average patient waiting time, duration of initial consultation completion, frontline healthcare worker working hours, accuracy of initial diagnosis and patient satisfaction.

Results: Research data showed that patient waiting times decreased by 31.7 minutes, while physician consultation time was reduced by 6.3 minutes per visit. Regarding clinical quality, AI provided structured medical histories, enabling faster differential diagnosis. Diagnostic consistency improved from 95% to 98%, demonstrating enhanced stability in clinical decision-making. Patient satisfaction scores rose from 9.28 to 9.92, reflecting strong approval of the improved flow, transparency, and convenience.

Conclusions: Implementation of an AI-assisted consultation system significantly optimized clinical workflows. AI integration not only boosts operational efficiency but also elevates diagnostic precision and the overall patient experience in urological outpatient settings.