納入糖尿病共同照護網後影響病患遵醫囑行為、治療成效與醫療資源耗用之研究-以某區域教學醫院為例

The Impact of Patients’ Adherence、Outcome and Utilization under the Shared Care Disease Management Program for the Diabetes- An Example from A Regional Teaching Hospital

蔡政麟(Tsai Cheng-Lin)

指導教授 ： 葉玲玲

亞洲大學/健康學院/健康管理研究所/碩士(2006年)

摘要

前言：除了適當的藥物治療，長期的糖尿病防治必需有完善的衛教計畫。為使各醫療院所加強對糖尿病患的衛教照護，健保局於2001年11月開始推出糖尿病醫療給付改善方案，期盼透過財務誘因最終能減緩糖尿病患病情並控制醫療耗用。
目的：探討納入糖尿病共同照護網後，病患的遵醫囑行為、治療成效與醫療資源耗用。並進一步探討衛教是否能改善遵醫囑行為，也比較接受標準照護與不標準照護的個案在治療成效與醫療資源耗用有無差異。
方法：本研究利用回溯性世代研究法進行資料之收集及分析。從2002年5月開始至2005年3月底，累計加入研究醫院糖尿病共同照護網已25個月之個案共1243人，再經過兩階段以篩檢高比率取藥日，即25個月之總取藥日大於或等於684日(90%)，並至少接受4次衛教之研究對象共487人。以SPSS 10.0統計軟體來進行資料分析。
結果：本研究個案的平均年齡略低、教育程度較低、罹病時間於5年以下較多、注射胰島素者較少、有嚴重併發症者為少數。遵醫囑行為的評量，以藥物治療配合的平均分數最高，其次依序是護照使用、飲食學習、飲食改變、運動，得分百分比分別是85.5%、82%、67.2%、67%、56.5%。藥物治療配合、運動配合、護照使用配合及整體配合為依變項之複迴歸分析分別達到統計意義。糖化血色素治療成效為依變項之複迴歸模式中，自變項解釋力(Beta係數)大小依序為初次糖化血色素、有注射胰島素、罹病6至10年。
糖尿病相關的醫療耗用部分，門診費用為依變項之複迴歸模式中，自變項解釋力大小依序為初次糖化血色素、有嚴重合併症、罹病6至10年、罹病11至20年、過重或肥胖、罹病大於或等於21年、照護總次數；急診費用為依變項之複迴歸模式中，自變項以照護總次數之解釋力大於有注射胰島素；住院費用為依變項之複迴歸模式中，自變項解釋力大小依序為罹病大於或等於21年、照護總次數、有高膽固醇血症、初次糖化血色素；總費用為依變項之複迴歸模式中，自變項解釋力大小依序為初次糖化血色素、罹病大於或等於21年、有嚴重合併症、罹病6至10年、過重或肥胖、罹病11至20年、有高膽固醇血症；門診次數為依變項之複迴歸模式中，僅有嚴重合併症變項達統計意義；急診次數為依變項之複迴歸模式中，只有照護總次數變項達統計意義；住院次數為依變項之複迴歸模式中，照護總次數的解釋力大於有嚴重合併症；住院天數為依變項之複迴歸模式中，自變項解釋力大小依序為有嚴重合併症、照護總次數、罹病大於或等於21年、有高膽固醇血症。
非糖尿病相關的醫療耗用部分，急診費用為依變項之複迴歸模式及住院次數為依變項之複迴歸模式中，自變項都以照護總次數的解釋力大於有嚴重合併症；住院費用為依變項之複迴歸模式、總費用為依變項之複迴歸模式及住院天數為依變項之複迴歸模式中，自變項都以有嚴重合併症的解釋力大於照護總次數。門診降血糖藥物及相關耗材費用為依變項之複迴歸模式中，自變項解釋力大小依序為初次糖化血色素、罹病11至20年、罹病6至10年、罹病大於或等於21年、年齡、男性。住院降血糖藥物及相關耗材費用為依變項之複迴歸模式中，自變項以有嚴重合併症的解釋力大於照護總次數。門住降血糖藥物及相關耗材總費用為依變項之複迴歸模式中，自變項解釋力大小排序同門診部分。25個月全部加總費用為依變項之複迴歸模式中，自變項解釋力大小依序為有嚴重合併症、照護總次數、罹病11至20年。
結論：在納入糖尿病共同照護網經過25個月後，本研究發現年齡、初次糖化血色素、有嚴重合併症及照護總次數是影響病患多項遵醫囑行為最重要的因素，年齡高者較能配合，照護愈多次者配合的愈好；病患的初次糖化血色素偏高及有嚴重合併症則配合的不好。初次糖化血色素偏高者其治療成效較好，有注射胰島素者及罹病時間較久者其治療成效較差，照護次數對糖化血色素治療成效的影響在本研究中並未發現統計意義。若病患的罹病時間較長及有嚴重合併症是增加醫療利用與費用最重要的因素。至於增加衛教照護次數雖然會增加糖尿病相關的門診費用，但會減少糖尿病相關的急診費用、住院費用、急診次數、住院次數、住院天數，也會減少非糖尿病相關的急診費用、住院費用、總費用、住院次數、住院天數，並減少住院降血糖藥物及相關耗材費用以及25個月全部加總費用。

關鍵字

糖尿病醫療給付改善方案、糖尿病共同照護網、遵醫囑行為、糖化血色素、醫療耗用

並列摘要

Background: Besides adequate medications, long-term diabetes mellitus control depends on comprehensive education programs. To enhance delivery of education program to diabetic subjects in various medical setting, Bureau of National Health Insurance proposed a Share Care Disease Management Program for the Diabetes on November 2001 and expected to slow down diabetes progression and control medical utilization via financial incentive.
Purpose: To investigate diabetic patients’ adherence, treatment efficacy, and medical utilization after they participated the program. To study whether the program improved adherence. And further to compare the treatment efficacy and medical utilization between patients who received standard education and patients who did not.
Methods: A total of 487 subjects were selected via 2-steps from 1243 patients who had joined this program for at least 25 months between May 2002 and March 2005. These patients had received at least 684 days (90% of the observe duration) medication and at least 4 times of education from nurse and dietitian. All these data were collected retrospectively and analyzed with SPSS 10.0 software.
Results: The baseline characteristics of study cases were lower average age, lower education background, majority with DM history less than 5 years, less with insulin injection, and less with severe complications. Adherence to medication (85.5%) was scored the highest in the adherence evaluation result. DM-passport usage (82%), diet-behavior learning (67.2%), diet-modification (67%), and exercise (56.5%) were scored items in sequence. Medication, exercise, DM-passport usage and total scores as dependent variable were significant via multiple regression analysis. Regression of HbA1c efficacy showed that beta coefficient was highest in baseline HbA1c, then insulin injection, duration between 6 years and 10 years in sequence. In DM relevant utilization, regression of outpatient expense showed that beta coefficient was highest in baseline Hba1c, then severe complications, duration between 6 years and 10 years, duration between 11 years and 20 years, overweigh or obese, duration more than 21 years, and education-frequency in sequence. Regression of emergency expense showed that beta coefficient of education-frequency was higher than insulin injection. Regression of admission expense showed that beta coefficient was highest in duration more than 21 years, then education-frequency, hypercholesterolemia, and baseline Hba1c in sequence. Regression of total DM relevant expense showed that beta coefficient was highest in baseline Hba1c, then duration more than 21 years, severe complication, duration between 6 years and 10 years, overweigh or obese, duration between 11 years and 20 years, and hypercholesterolemia in sequence. Regression of outpatient visit showed that severe complication was the only significant variable. Regression of emergency frequency showed that education-frequency was the only significant variable. Regression of admission frequency showed that beta coefficient of education-frequency was higher than severe complication. Regression of admission days showed that beta coefficient of independent variable was highest in severe complication, then education-frequency, duration more than 21 years, and hypercholesterolemia in sequence. In non-DM relevant utilization, Regression of the emergency expense and admission frequency showed that beta coefficient of education-frequency was higher than severe complication. Regression of admission expense, total expense, and admission days showed that beta coefficient of severe complication was higher than education-frequency. Regression of diabetic medication expense of outpatient and both outpatient and admission showed that beta coefficient was highest in baseline Hba1c, then duration between 11 years and 20 years, duration between 6 years and 10 years, duration more than 21 years, age, and sex in sequence. Regression of diabetic medication expense of admission showed that beta coefficient of severe complication was higher than education-frequency. Regression of total 25 months expense showed that beta coefficient was highest in severe complication, then education-frequency, duration between 11 years and 20 years in sequence.
Conclusion: Patients’25-months data after enrolled into Shared Care Disease Management Program for the Diabetes had been analyzed. This research found that age, baseline HbA1c, severe complication and education-frequency were very important factors for adherence. The older ones and those received more education-frequency showed better adherence. Patients with higher baseline HbA1c and with severe complication showed worse adherence. The treatment efficacy was better in ones with higher baseline HbA1c. Patients with insulin treatment and with longer history had worse treatment efficacy. But education-frequency didn’t show significance in HbA1c treatment efficacy. The most important factors to increase utilization were long duration and severe complication. Although more education-frequency would spend more DM-relevant outpatient expense, it did decrease emergency expense, admission expense, emergency frequency, admission frequency, and admission days. Besides, it decreased non-DM relevant emergency expense, admission expense, emergency frequency, and admission frequency. And it also decreased diabetes-medication expense of admission and total 25 months expense.

並列關鍵字

Shared Care Disease Management Program for the Diabetes、Shared Care Net Work for the Diabetes、adherence、Hemoglobin A1c、utilization

參考文獻

行政院衛生署國民健康局(2003)：糖尿病防治手冊－糖尿病預防、診斷與控制流程指引。

陳榮興、張淳堆、傅茂祖(2004):中台灣某醫學中心執行糖尿病個案管理方案於代謝指標之成效。中華民國內分泌暨糖尿病秋季研討會報告OA-8。

黃三桂、王悅萍、錢慶文(2002)：疾病管理對糖尿病患資源耗用之影響。台灣醫務管理雜誌，3(2)：35-47。

翁瑞亨、徐瑞祥、謝玉娟(2002):台灣地區糖尿病共同照護網現況。台灣醫學，6卷4期，569-573。

周碧瑟(2002)：台灣地區糖尿病流行病學。台灣衛誌，21（2）：83-96。