運用生成式AI分析醫療費行動支付程式使用者之品質體驗意見

Applying Generative AI to Analyze User Quality Experience Feedback for a Medical Payment Mobile Application

目的：本研究旨在分析台灣醫院專用行動支付應用程式「醫指付」的使用者回饋，利用生成式AI工具，精確分析使用者留言中的關鍵詞彙與情緒值，期能為行動支付應用程式開發者提供具體且實用的參考依據，以優化產品設計與使用者體驗。
方法：蒐集了2023年1月至2024年7月期間，於Google Play與Apple Store平台上「醫指付」應用程式的840則有效使用者評價留言。透過生成式AI工具，進行情緒值測定與文字探勘，從中發現使用者意見中的核心關鍵詞彙。
結果：研究結果顯示，「醫指付」使用者的留言情緒值與其滿意度呈現顯著的相關性。以生成式AI分析出，正向情緒留言僅占10.8%，平均情緒值為0.69；相對地，負向情緒留言高達88.8%，平均情緒值為-0.75。負評內容主要集中於「感知表現／期望不一致」這一構面，關鍵詞彙包括「資安問題」、「個資外洩」、「閃退／當機」及「盜刷信用卡」。這些詞彙明確反映出使用者對於應用程式資安保護不足，進而導致盜刷問題的強烈不滿。結論：使用者對「醫指付」的留言明確指出，資安問題是最為關注的焦點。由於資安漏洞未能及時得到修復，導致使用者留下了「不推薦」、「刪除安裝」等負面評價，嚴重影響了潛在用戶的下載意願。因此，支付型應用程式開發者應建立暢通的客服管道，並即時發布更新，以有效增強使用者信心，減少負面評論的產生。

Purpose: This study aimed to analyze user feedback on E-Health-Pay, a widely used mobile payment application specifically designed for hospitals in Taiwan. Using generative AI, the research analyzed keywords and sentiment responses in user comments, aiming to provide developers with concrete and practical insights to optimize product design and enhance user experience.
Methods: A total of 840 user comments on the E-Health-Pay app were collected from the Google Play and Apple App Store between January 2023 and July 2024. Generative AI was applied to analyze sentiment responses and linguistic expressions, to identify key terms and recurring themes emphasized by users.
Results: The findings revealed a significant correlation between the sentiment responses expressed in user comments on the E-Health-Pay app and overall user satisfaction. The generative AI analysis showed that only 10.8% of comments conveyed positive sentiment (average sentiment score: 0.69), while a substantial 88.8% reflected negative sentiment (average sentiment score: -0.75). Negative feedback was primarily linked to a“perceived performance–expectation discrepancy,”with recurring key terms such as information security, personal information leakage, system crashes/shutdowns, and unauthorized credit card charges. These concerns highlighted strong user dissatisfaction, particularly regarding information security vulnerabilities that resulted in unauthorized charges.
Conclusion: User comments on the E-Health-Pay app consistently identified information security as the primary concern. The failure to promptly address these vulnerabilities led to negative feedback, including remarks such as“not recommended”and“deleted app installation,”which in turn diminished the willingness of potential users to adopt the app. To mitigate these issues, developers of mobile payment applications should provide readily accessible customer service and release timely updates, thereby enhancing user confidence and reducing negative feedback.