生成式AI與醫療衛教：探討慢性病病患之科技介入與焦慮處理策略

Abstract

台灣在2025年即將邁入超高齡化社會，老年化及慢性病的人口逐年上升，但台灣目前醫護人力正面臨緊缺的狀態，慢性病病患的照護的需求只會日益增加。然而生成式AI(Generative AI, GenAI)在ChatGPT問世之後，該技術的應用受到廣大的討論。此現象讓研究者感到好奇，GenAI技術性的突破，是否符合慢性病病患的臨床需求。因此，本研究從慢性病病患必須面對長期心理的壓力與挫折的角度出發，針對心理學的因應策略審視過往的文獻，提出一個結合因應策略中兩個主要的策略：「問題聚焦因應策略」及「情緒聚焦因應策略」與相關變數之模型，藉此探討慢性病病患是否會因為GenAI提供因應策略中的兩種面向的相關因子，進而提升對於AI健康聊天機器人的偏好與使用意願。 本研究於 2024年 6 月 10 日至 6 月 24 日期間，透過便利抽樣法於社群平台發放 SurveyCake 平台問卷連結以蒐集相關樣本，最終有效樣本數為 187 筆(回應率 64.71%)。研究中使用描述統計分析、結構方程式模型分析進行資料分析，並發現：若AI健康聊天機器人可提供「問題聚焦因應策略」及「情緒聚焦因應策略」，慢性病病患對其偏好程度將有正向的影響，而這樣的偏好程度會進而影響其使用意願。此外，情緒聚焦因應策略的影響程度略高於問題聚焦因應策略。另外，研究中也發現：年齡、教育與焦慮程度並不會影響患者對於AI健康聊天機器人的偏好與使用意願。 整體而言，本研究透過病患因應壓力與焦慮的策略，來進一步探討GenAI是否符合慢性病病患的需求，而研究中也證實無論是問題聚焦因應策略或情緒聚焦因應策略，確實都能提升病患的偏好程度，且不受其對於疾病之焦慮程度而有所影響。本研究試圖提供GenAI於醫學衛教上可行之應用方向，同時也強調GenAI可透過有效的應對策略來滿足慢性病病患需求的潛力。

In 2025, Taiwan is approaching a super-aged society with a growing elderly population and an increasing prevalence of chronic diseases. However, Taiwan’s healthcare system is facing severe shortages in personnel, and the demand for care for chronic disease patients is only set to rise. Following the emergence of Generative AI (GenAI), particularly ChatGPT, extensive discussion has arisen regarding its applications. This phenomenon has sparked curiosity among researchers about whether the technological breakthroughs of GenAI meet the clinical needs of chronic disease patients. Therefore, this study begins from the perspective of the long-term psychological stress and challenges faced by chronic disease patients. It reviews the literature on psychological coping strategies and proposes a model that integrates two primary coping strategies: “problem-focused coping” and “emotion-focused coping,” along with related variables. The study explores whether chronic disease patients’ preferences for AI health chatbots and their willingness to use them can be enhanced by GenAI’s provision of these coping strategies. From June 10 to June 24, 2024, the study employed convenience sampling via SurveyCake platform links distributed on social media to collect relevant samples. The final dataset included 187 valid responses (with a response rate of 64.71%). Data analysis utilized descriptive statistics and structural equation modeling. The findings indicate that the provision of “problem-focused coping” and “emotion-focused coping” strategies by AI health chatbots positively influences chronic disease patients’ preferences for these technologies, affecting their willingness to use them. Additionally, the impact of emotion-focused coping strategies is slightly more substantial than that of problem-focused coping strategies. Furthermore, the study found that age, education, and anxiety level do not significantly influence patients’ preferences for or willingness to use AI health chatbots.