以混合研究法探討樂智AI桌遊對輕度認知障礙長者之成效

Abstract

背景：具有交互功能的技術輔助培訓工具可以改善患有輕度認知障礙(MCI)長者的健康結果。本研究探討基於人工智慧的桌遊針對患有MCI的長者，以改善他們的認知功能、憂鬱程度、一般自我效能和對生活的滿意度的有效性；及這套桌遊的易用性與未來MCI長者的使用意圖。方法/設計：採取準實驗設計。台灣北部八家日照中心或長照據點輕度認知障礙長者的109位參與者，分為實驗組(n=52)與對照組(n=57)，實驗組接受樂智AI桌遊，對照組則為維持一般常規活動。將兩組參與者分別完成前後測之問卷調查，包括認知功能(MMSE、ADAS-Cog)、憂鬱程度、一般自我效能與生活滿意度的影響，使用廣義估計方程式(GEE)分析資料。樂智AI桌遊的可用性通過系統易用性量表(SUS)進行評估與分析未來使用意圖。結果：GEE分析顯示，實驗組在 MMSE（P<0.001）、ADAS-Cog（P<0.001）、憂鬱程度（P=0.018）、一般自我效能（P<0.001）與生活滿意度方面(P=0.001)均有顯著改善，這些影響持續了三個月。另外，SUS的平均分數為87.50，表明參與者的評價非常好，且未來使用意圖高。 結論：這項研究的結果支持基於交互式人工智慧的桌遊在改善機構化長者的認知功能、憂鬱、一般自我效能與生活滿意度方面的有效性。參加者對課程的評價非常好。建議長期護理機構的管理人員可以採用基於人工智慧的桌遊作為培訓工具，以協助健康促進計劃，以減少機構老年人認知和心理健康惡化的可能性。

Background: A technology-assisted training tool with interactive features may improve the health outcomes of older adults with mild cognitive impairment (MCI). This study explores the effectiveness of artificial intelligence-based board game targeting institutionalized older adults with MCI to improve their cognitive function, depression, general self-efficacy and satisfaction with life.Materials and Methods: We used a quasi-experimental design. 109 institutionalized older adults from ten long-term care facilities were recruited and assigned to the experimental (n=52) and comparison groups (n=57). The participants of the experimental group received 12-week cognitive training program. Participants of both groups completed before-and-after and three-month follow-up measurements for outcome variables that included cognitive function (scores of MMSE and ADAS-Cog), depression, general self-efficacy, and satisfaction with life. Data were analyzed using the generalized estimating equation (GEE). System Usability Scale (SUS) assessed the program's usability, and decomposed theory of planned behavior (DTPB) analyzed future usage intentions. Results: GEE analyses indicated that the experimental group showed significant post-intervention improvements on MMSE (P<0.001), ADAS-Cog(P<0.001), depression(P=0.018), general self-efficacy (P<0.001) and satisfaction with life(P=0.001). These effects persisted for three-month follow up. The mean score of SUS was 87.50 which indicated the participants’ evaluation was excellent. The behavioral intention to use AI-based board games in the future is high.Conclusion: The findings of this study support the effectiveness of the interactive artificial intelligence-based board game to improve the cognitive function, depression, general self-efficacy and satisfaction with life of institutionalized older adults. Participants’ evaluation of the program is excellent. It is suggested that the administration of long-term care facilities can adopt artificial-based board game as a training tool to assist in health promotion programs to reduce the likelihood of cognitive and mental health deterioration among institutionalized older adults.