慢性疾病之風險評估與遠距照護應用

Abstract

慢性疾病管理已成為全球醫療領域關注的焦點，本論文的研究重點在於慢性疾病之風險評估與遠距照護應用。首先，透過資料探勘技術進行健保資料庫的回溯性世代研究，證實急性胰臟炎病患和胃全切除胃癌病患的葡萄糖代謝功能變化，或者是慢性肺阻塞病患急性惡化之存活情形，合併症的有無及相關用藥皆為影響預後的關鍵因子。另一方面，研究數據也顯示胃癌且合併糖尿病的患者，在接受胃全切除手術後，其葡萄糖代謝功能得以有效地緩解。而控制案例分析的實證則顯示，胃癌且無合併糖尿病的患者，其術後的糖尿病發生率顯著低於一般民眾。 為提供慢性病患連續不間斷的遠距照護，本論文基於網路服務的資通訊技術應用，建立具有生命徵象監測及自動警示機制的整合性遠距照護系統，用以監測分析即時數據，偵測早期病情惡化予以個管師警示提醒，並水平整合院內電子病歷，提供個人化醫療資訊。同時，為提供醫療人員臨床決策輔助，系統配載心電圖自動判讀機制，對於心房顫動的辨識率可達88.4%的準確度、94.1%的敏感度及87.7%的特異度；對於心室早期收縮的辨識率則可達86.1%的準確度、82.7%的敏感度及86.3%的特異度。截至2015年12月為止，本系統已服務2,761名病人、接收並監測1,716,256筆生理數值，及建立268,278筆訪談紀錄。病人量測的生理數值可於三分半鐘內由個管師完成確認，且遠距照護服務獲得使用者95.8%的滿意度支持。此外，隨機控制試驗的實證結果指出，簡易的遠距監測服務便可顯著延緩慢性肺阻塞疾病急性再住院的病程，且服務使用者亦有較低的再住院次數和急診就醫次數。因此，透過整合性遠距照護系統與資通訊加值服務的結合應用，醫療人員及病患皆可在健康照護上獲得許多益處。

The management of chronic diseases has become the most common medical concerns around the world. The aims of this dissertation are to further evaluate the relative risk and conduct the applications of telehealthcare in chronic diseases. First, in terms of the change of glucose metabolism in patients with acute pancreatitis and patients with total gastrectomy and the survival of prognosis in patients with chronic obstructive pulmonary disease exacerbations, population-based cohort retrospective studies based on the Taiwan National Health Insurance Research Database (NHIRD) using data mining technologies showed that the burden of comorbidities has a significant role in patients with chronic diseases, and should be carefully evaluated and managed. On the other hand, the experimental results showed that total gastrectomy contributes to improved glucose metabolism in patients with total gastrectomy. Most importantly, patients without diabetes but who had total gastrectomy had lower rates of newly diagnosed diabetes after total gastrectomy compared with the general population as assessed by control-to-case analysis. In an attempt to provide ubiquitous telehealthcare for patients with chronic disease, a telesurveillance system has been implemented with continuous biometric monitoring and the automatic alarm mechanism using information and communication technologies. The system can analyze synchronous data to detect deterioration at an early stage through long-term health monitoring and inform case managers immediately using instant biometrics monitors and automatic alarm mechanisms. Besides, the system allows users to derive electronic medical records (EMRs) from the hospital information systems using web services. Most importantly, the augmentation service of knowledge-based electrocardiogram interpretation based on the support vector machine is available to assist physicians doing decision-making in clinical practice. The experimental results showed that the recognition classifiers yielded 88.4% accuracy, 94.1% sensitivity, and 87.7% specificity in atrial fibrillation and 86.1% accuracy, 82.7% sensitivity, and 86.3% specificity in ventricular premature contraction, respectively. As of December 31, 2015, the telehealthcare program has served 2,761 patients, derived and monitored 1,716,256 physical data, and performed 268,278 interviews to offer total care to patients. Notably, the program provides feedback within four minutes of the patient measuring biometric data. The telecare service received a high degree of support (95.8 %) from the users. Besides, the randomized controlled trial aimed to investigate the effectiveness of telemonitoring in improving COPD patient outcome. During the follow-up period, time to first re-admission for COPD exacerbation was significantly increased in the telemonitoring group than in the usual care group. Telemonitoring was also associated with a reduced number of all-cause re-admissions and emergency room visits. In conclusion, both of medical staffs and patients could derive benefit health care from the telehealthcare service in many respects.