[[alternative]]The Impact of Disease Management on Clinical Indicators and Medical Utilization for Long-term Tracking of Diabetes Patients

[[abstract]]背景:糖尿病盛行率上升，發病年齡層下降，醫療花費居高不下，對於國民健康影響是存在的且必須重視。糖尿病疾病管理在臨床慢性病照護模式中是不可或缺的重要計劃。國健署對於糖尿病照護計畫一直持續推廣及要求增加照護率，健保署對於加入照護網使用支付誘因來執行落實政策。臺灣糖尿病共同照護網已推行18年多，照護計畫推行至此的好壞及優缺點，在研究文獻上多是短期及使用健保資料庫發表，較缺乏單一機構長期加入照護網作追蹤的文獻，期望能有多些長期追蹤文獻來證明糖尿病照護計畫對於病患、機構及醫療花費是否有正面顯著效益。目的:本研究利用某醫學中心門診就診病患資料庫進行分析，討論加入糖尿病照護計畫後，對於糖尿病患之臨床治療指標及醫療利用面在短期與長期追蹤觀察是否具有好處。方法:本研究原先計畫採用三院臨床資料庫數據，但發現實際的資料庫加入時間不一，擔心遺漏值多造成選樣偏差，影響研究結果。故而採用北區某醫學中心門診病患長期就診資料庫進行分析共納入9,967名糖尿病病患。首先使用Excel 2013試算表為研究工具，SAS 9.4進行資料前處理與分析，研究對象基本特質做描述性統計並依傾向分數配對方法(Propensity score matching)進行1:1配對，取得2018年兩組樣本共3,192人（實驗組和控制組各1,096人）進行配對t檢定分析。2017-2018年加入2年共4,811人(實驗組3,764人控制組1,047人)及加入5年2014-2018共1,839人(實驗組1,019人控制組820人)資料皆用獨立樣本t檢定，檢定二組樣本各變項平均值之差異，最後2009-2018年共95人，連續10年加入的糖尿病患做描述性統計進行分析及討論。結果:研究發現加入糖尿病照護計畫1年，糖尿病病患平均之空腹血糖(mg/dl)改善達到統計上顯著差異(有(實驗組)/無加入：平均129.1/132.8, p=0.0078)； 糖化血色素(HbA1c)(%)的改善則未達到統計上顯著差異(6.98/6.96, p=0.6876)，而醫療服務的利用(最後一次門診申報費用(元))也未達到統計上顯著差異(2580.1/2801.6, p=0.2755)。對於加入糖尿病照護計畫2年的病患進行後續討論，可以發現連續加入計畫2年，糖尿病病患在臨床治療指標就可以獲得更好的改善，空腹血糖平均值達到統計上顯著差異(128.3/139.6, p<0.0001) ，糖化血色素平均值亦達到統計上顯著差異(6.93/7.40, p<0.0001)，而醫療服務利用亦是可以發現加入計畫後，病患使用之醫療資源較為精簡(1648.4/2258.5, p<0.0001)。連續加入5年的糖尿病患者，在臨床指標上都可以發現有顯著的改善，無論在空腹血糖（129.4/145.3）、糖化血色素(6.84/7.44)或是低密度膽固醇（81.76/92.78）都達到統計上的顯著差異（p<0.0001）且醫療服務利用亦更為精簡（2604.0/3124.6, p=0.0051）；而連續加入10年的糖尿病病患之臨床指標可以發現，在空腹血糖（133.2/145.6）、糖化血色素（6.79/7.45）及低密度膽固醇（82.15/92.25）都能達到改善，且醫療服務利用也更為精簡（2339.2/3016.9）。最後探討在2018年糖尿病患者是否因為使用口服藥物和胰島素針劑而造成臨床治療結果的差異，進行two-way analysis分析，結果發現使用口服藥病人在加入計畫1年，可達較好的控制(HbA1c:6.69/6.98, p<0.0001)；使用胰島素針劑的組別則未達到統計顯著結果。建議: 明白建議病患在用口服藥初期提早加入糖尿病照護網的好處及優點，在提供及介入衛教技巧也使病患願意參加配合照護網計畫，當然臨床醫師的鼓勵及引導加入計畫方案也使照護網會成功，擴大至跨科別的醫師做慢性病管理是未來重要趨勢。減少合併症產生，降低社會醫療成本的付出，才是國民健康最大的福祉。[[abstract]]Background: The prevalence of diabetes mellitus is increasing while its age onset is declining, and its medical expenses stay high, therefore, there exists an impact on national health and it must be paid attention. Diabetes disease management is an indispensable and important program in the clinical chronic care model. Health Promotion Administration has continued to promote the Diabetes Share Care programs and require increased care rates. National Health Insurance Administration uses payment incentives to join the Diabetes Share Care to implement policies. The Diabetes Share Care in Taiwan has been in operation for more than 18 year, however, the research literatures regarding its implemented quality and advantages and disadvantages are mostly published in the short-term and health insurance database with a lack of joining the Diabetes Share Care for long-term tracking by a single institution. It is expected to have more long-term follow-up literatures to demonstrate whether the Diabetes Share Care programs have a significant positive effect on the patients, institutions, and medical expenses.Objective: The study used the database of outpatient visits in a specific medical center to analyze whether the clinical treatment indicators and medical utilization of diabetic patients were beneficial in short and long-term tracking after the Diabetes Share Care programs.Methods: The original plan used the data from the clinical databases of the three hospitals; however, it was found that some errors in the actual databases were likely to cause an increase in missing values, which worried about the impact of the research results. Therefore, a long-term medical database of outpatients in a medical center in the North District was used for analysis. A total of 9,967 diabetic patients were included. The Excel 2013 spread sheet were used for the research tool and SAS 9.4 were performed for data management and statistical analysis. The basic characteristics of the subjects were summarized by descriptive statistics. According to the propensity score matching methods 1:1 pairing, a total of 3192 subjects from the two groups in 2018 (1096 subjects in each of the experimental and control groups) were selected for paired t-test analysis. In the 2 years of 2017-2018, a total of 4811 subjects (experimental group 3764 subjects and control group 1047 subjects) data were analyzed by independent sample t test, and the difference between the mean values of the two groups were tested. Finally, during 2009-2018 period, a total of 95 diabetes patients who joined the Diabetes Share Care programs continuously for 10-year were analyzed and discussed by using descriptive statistics.Results: One year of joining the Diabetes Share Care, the mean fasting blood glucose (mg/dl) improvement in diabetic patients reached a statistically significant difference (to join, yes (experimental group)/no: mean 129.1/132.8, p = 0.0078), the improvement in HbA1C did not reach a statistically significant difference (6.98%/6.96%, p = 0.6876), and medical utilization (NT dollars) (last outpatient declaration fee) did not reach statistically significant difference (2580.1/2801.6, p = 0.2755). For 2 years of continuously joining the Diabetes Share Care, diabetic patients got better improvement in clinical treatment indicators, the average of fasting blood glucose was statistically significant (128.3/139.6, p<0.0001), statistically significant difference of mean HbA1C were also observed (6.92%/7.39%, p<0.0001), and medical utilization was also found to be more streamlined after joining the Diabetes Share Care (1648.4/2258.5, p<0.0001). Patients who joined the program for 5 years received more qualified diabetes care, and we found fasting glucose (129.4/145.3, p<0.0001), HbA1c( Glycated Hemoglobin) (6.84/7.44%) and also LDL(low-density lipoprotein,) (81.76/92.78 mg/dl) were significantly improved (p<0.0001). As the medical indicators of diabetic patients who continuously joined the Diabetes Share Care for 10 years, the fasting glucose (133.2/145.6 mg/dl), HbA1c (6.79/7.45%) and LDL (82.15/92.25 mg/dl) were significantly reduced. Finally, two-way analysis of variance was conducted to evaluate whether the clinical treatment indicators were different due to the use of oral drugs and insulin injections for diabetic patients in 2018. It was found that patients who used oral drugs and insulin injections could achieve better results when joining the program 1 year (HbA1c: 6.69/7.0%, p