使用TAM模式探討急診室醫師對AI腦出血判讀系統的接受度

Adopting Technology Acceptance Model to Explore Emergency Department Physicians' Acceptance of toward Cerebral Hemorrhage AI Interpretation System

目的：本研究主要以Davis(1989) 所提出的科技接受模式為架構，探討急診室醫師對於AI 腦出血判讀系統的接受度。方法：本研究以使用過腦出血判讀系統（DeepCT，為愛因斯坦人工智能Deep01 所推出）合作醫院之急診室醫師作為研究對象。此系統可在30 秒判斷腦部CT 影像是否有出血，提供解決方案，並通知和幫助急診科醫生優先處理較危急的病患。有效問卷數為83 份，利用描述性統計、獨立樣本T 檢定、單因子變異數分析等進行各構面的相關性分析。結果：本研究在知覺易用性、知覺有用性、行為意向等構面都得到中高度的相關性，由各構面所得到的分數得知急診室醫師普遍對於此系統的評價頗高。結論：本研究受試者對於腦出血判讀系統之接受度高，並認為可提高工作效率，建議將此系統推廣到所有醫院。

Objective: Based on the technology acceptance model proposed by Davis(1989), this study explored the emergency physicians’ acceptance toward the AI interpretation system for cerebral hemorrhage. Methods: In this study, emergency physicians who had used the cerebral hemorrhage AI interpretation system (DeepCT, developed by Einstein ARTIFICIAL Intelligence Deep01) were enrolled. The system can detect bleeding on CT images of the brain within 30 seconds, provide solutions, and inform and help emergency physicians prioritize more critical patients. The number of valid questionnaires was 83. Descriptive statistics, independent sample T-test and univariate ANOVA, were used to analyze the correlation of each dimension. Results: In terms of perceived level of user-friendly, perceived usefulness, behavioral intention and other aspects, the results of this study indicated relatively high correlation, and the scores obtained from each aspect showed that emergency physicians generally think highly of this system. Conclusion: The acceptance toward the cerebral hemorrhage interpretation system was high among the subjects in this study. They believe it can improve work efficiency. Therefore, it is suggested that the system be adopted by all hospitals.