雖然目前台灣的醫療水平相當高，技術也相當進步，但有不少醫院仍然採用手寫的方式來記錄資料。本篇描述了一套完整的系統概念，可以應用在許多在醫療上需要紀錄資料的場景，除了能夠即時紀錄外，並且利用數位化的優勢，還能進一步使用人工智慧進行分析及預測，達到傳統手寫無法達到的目的。
在本篇中還會將系統應用在兩種場景上，分別是「居家血壓監控」與「血液透析掉壓預測」。「居家血壓監控」中，醫生可以透過此系統針對不同的病人進行血壓監控的規劃，監控的過程中，醫生可以即時接收到患者的血壓資訊，並對規劃做出即時的調整；而「血液透析掉壓預測」則是可以透過許多患者的血液透析資料進行分析，並預測患者進行血液透析的過程中，什麼時候患者的血壓會出現突然下降的狀況，讓醫生可以提前做好應對的準備，同時在本篇中，使用了羅吉斯迴歸對花蓮慈濟醫院提供的血液透析資料進行訓練，得到了大約7至8成的準確率。

Although currently Taiwan has a very advanced and accessible medical environment, many medical facilities are still using pen and paper to keep track of critical information. Here we extensively listed out the concepts of the system that can be used in this situation. In addition to instant and accurate recordings, we can also incorporate machine learning to analyze and predict the data, which is something hard to do with traditional methods.
Here we will demonstrate how the system works in practice on two scenarios, “Home blood pressure monitoring” and “Hypotension in hemodialysis”. During the former, doctors can use this system to accurately monitor various patient’s blood pressure at the same time and plan accordingly. In the later, we analyze the blood data of numerous patients, and predict what type of patients are most likely to suffer from hypotension during hemodialysis. This will allow doctors to have more time to prepare. Later we shall demonstrate how using logistic regression on the hemodialysis data from the patients in Hualien Tzu Chi Hospital to obtain a roughly 70% to 80% accuracy.

序言與誌謝 I
摘要 III
Abstract V
目錄 VII
圖次 IX
表次 XIII
第一章、緒論 1
1.1 研究背景與動機 1
1.2 研究目的與方法 1
第二章、系統的架構與流程 3
2.1 系統架構 3
2.1.1 後端 3
2.1.2 管理系統介面 4
2.1.3 控制接收器 4
2.1.4 裝置群 4
2.2 系統流程 4
2.2.1 設定量測排程 5
2.2.2 量測與接收結果 6
第三章、居家血壓監控系統實作 7
3.1 後端的建置 7
3.1.1 WebSocket 7
3.1.2 關聯式資料庫 8
3.2 管理系統設計 9
3.2.1 漸進式網路應用程式 9
3.2.2 系統權限管理 10
3.2.3 登入與註冊 11
3.2.4 病患新增與管理 13
3.2.5 量測排程管理 15
3.2.5.1 量測計畫 15
3.2.5.2 量測排程 18
3.3 血壓計的控制與接收量測資料 20
3.3.1 血壓計的控制 20
3.3.3.1 藍芽與藍芽低功耗 21
3.3.3.2 接收血壓計的量測資料 21
第四章、包含低血壓預測之血液透析監控系統的概念描述 23
4.1 系統概念 23
4.2 透析低血壓的預測 23
4.2.1 訓練資料集 23
4.2.1.1 資料分析 25
4.2.1.2 判斷是否為透析時低血壓的標準 28
4.2.2 訓練方法 29
4.2.2.1 羅吉斯迴歸 29
4.2.2.2 分層k折交叉驗證 29
4.2.2.3 數字標準化之方法 30
4.2.2.4 訓練步驟 30
4.2.3 訓練結果 31
4.2.3.1 20s_MM的訓練結果 31
4.2.3.2 20s_MM的訓練結果 34
4.2.3.3 所有訓練結果 37
第五章、結論與未來展望 39
5.1 結論 39
5.2 未來展望 39
參考文獻 41
附錄 43

[1] 張維文（2019）擔心洗腎掉血壓，該不該停血壓藥？醫師解答．照護線上．https://www.careonline.com.tw/2019/05/hemodialysis-hypertension-medication.html
[2] 陳亮恭、李威儒（2017）智慧醫療數位轉型與再進化．國土及公共治理季刊，5（4），38-43。https://www.ndc.gov.tw/Content_List.aspx?n=4A6A9E48D08C7042
[3] Fette, I., Google, Inc., Melnikov, A., & Isode Ltd. (2011, December). RFC 6455 - The WebSocket protocol. IETF Tools. https://tools.ietf.org/html/rfc6455
[4] Codd, E. F. (1970). A relational model of data for large shared data banks. Communications of the ACM, 13(6), 377-387. https://doi.org/10.1145/362384.362685
[5] Chen, P. P. (1976). The entity-relationship model—toward a unified view of data. ACM Transactions on Database Systems, 1(1), 9-36. https://doi.org/10.1145/320434.320440
[6] Richard, S., & LePage, P. (2020, February 24). What are progressive web apps? web.dev. Retrieved April 29, 2021, from https://web.dev/what-are-pwas/
[7] Caching files with service worker. (n.d.). Google Developers. Retrieved April 29, 2021, from https://developers.google.com/web/ilt/pwa/caching-files-with-service-worker
[8] O’Connor, J., Aumasson, J., Neves, S., & Wilcox-O’Hearn, Z. (2020, February 21). BLAKE3 - one function, fast everywhere. https://github.com/BLAKE3-team/BLAKE3-specs/raw/master/blake3.pdf
[9] HEM-7600T - Blood pressure monitors (Upper arm). (n.d.). Omron healthcare Asia Pacific. Retrieved May 2, 2021, from https://www.omronhealthcare-ap.com/ap/product/98-hem-7600t
[10] Bluetooth radio versions. (n.d.). Bluetooth® Technology Website. Retrieved May 4, 2021, from https://www.bluetooth.com/learn-about-bluetooth/radio-versions/
[11] 3.1.1 Characteristic behavior. (2020, December 5). In Bluetooth SIG, Blood Pressure Service (v1.1, pp. 11-13). https://www.bluetooth.com/specifications/specs/blood-pressure-service-1-1/
[12] 4.10.1 Notifications. (2019, December 31). In Bluetooth SIG, Bluetooth Core Specification (v5.2, pp. 1584-1585). https://www.bluetooth.org/docman/handlers/downloaddoc.ashx?doc_id=478726
[13] 4.11.1 Indications. (2019, December 31). In Bluetooth SIG, Bluetooth Core Specification (v5.2, pp. 1586-1586). https://www.bluetooth.org/docman/handlers/downloaddoc.ashx?doc_id=478726
[14] Lin, C., Chen, C., Wu, P., Pan, C., Shih, H., Huang, M., Chou, L., Tang, J., & Wu, C. (2018). Intelligent system to predict intradialytic hypotension in chronic hemodialysis. Journal of the Formosan Medical Association, 117(10), 888-893. https://doi.org/10.1016/j.jfma.2018.05.023
[15] The Data Detective. (2019, November 14). A look into feature importance in logistic regression models. Medium. Retrieved June 6, 2021, from https://towardsdatascience.com/a-look-into-feature-importance-in-logistic-regression-models-a4aa970f9b0f
[16] Flythe, J. E., Xue, H., Lynch, K. E., Curhan, G. C., & Brunelli, S. M. (2014). Association of mortality risk with various definitions of Intradialytic hypotension. Journal of the American Society of Nephrology, 26(3), 724-734. https://doi.org/10.1681/asn.2014020222