應用人工智慧建構醫療服務即時衛教機器人

Apply Artificial Intelligence to Develop a Medical Service and Health Education Chatbot

陳瑞仁(Jui-Jen Chen) ； 賴鈺婷(Yu-Ting Lai) ； 顏宏旗(Hung-Chi Yen)

《醫務管理期刊》 22卷2期 (2021/06) Pp. 136-150

https://doi.org/10.6174/JHM.202106_22(2).136

摘要

目的：在臨床檢查或治療常因衛教不足或衛教不當，影響造影影像品質與治療效果，且聊天機器人現況多為就診諮詢，本研究導入微軟QnA Maker平台，建置一醫療服務即時衛教機器人，以彌補臨床衛教不足的一環。方法：本研究將聊天機器人技術，整合至醫病共享決策平台，建構醫療服務即時衛教機器人，系統架構為檢查衛教指引萃取代理人、檢查衛教指引諮詢、檢查衛教指引準則、自然語言機器學習系統、核醫衛教知識庫。結果：2018年11月至2019年3月，以未使用與使用後介面親近性與內容完整性進行問卷調查，平均值由1.99大幅提升至4.15，認知度提升確實解答了病人對核醫檢查治療的相關疑慮與恐懼，另醫師衛教時間減少約125,580分鐘，則可增加病人服務量增加約4,186人。結論：運用機器學習進行自然語言處理，即時衛教機器人可以以幾近人類口吻與病人進行互動式的衛教，增加病人對檢查項目的瞭解程度。

關鍵字

醫病共享 ； 衛生教育 ； 聊天機器人 ； 病人安全 ； 核子醫學

並列摘要

Objectives: In clinical examinations or treatments, the lack of health education or improper health education often affects the quality of images and treatment, and the current status of a chatbot is mostly consultation. This study introduced the Microsoft QnA Maker platform and develop Medical Service and Health Education Chatbot to supplement clinical health education. Methods: In this study, chatbot technology is integrated into the Sharing Decision Making platform to develop Medical Service and Health Education Chatbot. The system architecture is the Examination Report Extraction Agent, Health Education Consultant Agent, Health Education Guidelines Ontology, Natural Language Machine Learning, Nuclear Medicine Health Education Repository. Results: From November 2018 to March 2019, we conducted a questionnaire survey on the proximity and content integrity of the unused and used interfaces. The average value increased from 1.99 to 4.15. Awareness improvement answers the patients' doubts and fears about nuclear medicine examination and treatment. In addition, the time for doctors' health education reduced by about 125,580 minutes. It can increase the patient service volume by about 4,186 people. Conclusions: Using machine learning for natural language processing, real-time health robots can interact with patients in a near-human voice, increasing patients' understanding of the examination program.

並列關鍵字

Shared Decision Making ； Health Education ； Chatbot ； Patient Safety ； Nuclear Medicine

參考文獻

陳瑞仁、顏宏旗、黃淑華、賴鈺婷、張雁翔（2018）‧運用團隊資源管理概念改善核醫檢查病人報到臨床作業流程‧核子醫學暨分子影像雜誌，31(1)，12-21。https://doi.org/10.6332/anmmi.201803_31(1).0002。

Giannoula, E., Iakovou, I., Katsikavelas, I., Antoniou, P., Raftopoulos, V., Chatzipavlidou, V., Papadopoulos, N., & Bamidis, P. (2020). A Mobile App for Thyroid Cancer Patients Aiming to Enhance Their Quality of Life: Protocol for a Quasiexperimental Interventional Pilot Study. JMIR Res Protoc, 9(3), e13409. https://doi.org/10.2196/13409。

Goldenthal, S. B., Portney, D., Steppe, E., Ghani, K., & Ellimoottil, C. (2019). Assessing the feasibility of a chatbot after ureteroscopy. Mhealth, 5, 8. https://doi.org/10.21037/mhealth.2019.03.01。

Ishii, M., Okumura, Y., Sugiyama, N., Hasegawa, H., Noda, T., Hirayasu, Y., & Ito, H. (2017). Feasibility and efficacy of shared decision making for first-admission schizophrenia: a randomized clinical trial. BMC Psychiatry, 17(1), 52. https://doi.org/10.1186/s12888-017-1218-1。

Maron, B. J., Nishimura, R. A., & Maron, M. S. (2017). Shared decision-making in HCM. Nat Rev Cardiol, 14(3), 125-126. https://doi.org/10.1038/nrcardio.2017.6。