照護連續性與老年慢性病患照護結果的關係－以個別醫師或醫療院所為測量基礎之比較

Abstract

本研究主要是比較以個別醫師與醫療院所為測量基礎之照護連續性之差異，並探討照護連續性對老年人口中罹患單一慢性病與雙重慢性病患照護結果之影響。 本研究利用國家衛生研究院全民健康保險資料庫，以2005年承保抽樣歸人檔為基礎，選取在2008年1月1日滿65歲以上且就醫三次以上者，分成單純只有糖尿病、單純只有高血壓、以及同時患有糖尿病及高血壓三組為此研究分析主要樣本。自變項為醫師照護連續性指標(UPCP)及醫療院所照護連續性(UPCS)，並分為高、中、低三組，依變項為住院與急診醫療利用，以及總醫療費用，控制變項包含個人特性(年齡、性別、總門診次數、共病指標)，以及最常就醫院所特性(就醫層級、分局別)等。 統計分析納入糖尿病病患9,018人、高血壓病患為24,454人、糖尿病併高血壓病患為10,219人。當年糖尿病病患有住院者佔28.47%，有急診者佔30.82%。高血壓病患有住院者佔21.65%，有急診者27.25%。糖尿病併高血壓病患有住院者佔28.45%，有急診者佔32.63%。當以醫師為測量基礎，糖尿病病患照護連續性指標平均為0.448，高血壓病患照護連續性指標平均為0.443，糖尿病併高血壓病患照護連續性指標平均為0.432。以醫療院所為測量基礎時，三組病患的照護連續性指標值分別為: 0.610, 0.575, 以及0.590。 接著，本研究以負二項式迴歸分析及複回歸搭配gamma distribution來分析醫師照護連續性與住院與急診次數、以及總醫療費用的相關性，結果發現糖尿病病患以醫師照護連續性高者為參考組時，照護連續性為低與中者，其住院次數分別為2.75倍、1.58倍，急診次數為2.87倍和1.53倍，醫療費用為1.54倍和1.23倍，皆達統計顯著意義(P<0.001)。相對的，醫療院所照護連續性的分析結果分別為:住院次數1.90倍與1.29倍，急診為1.97倍與1.39倍，總費用為1.14倍與1.02倍。以高血壓病患為樣本時，分析結果與糖尿病患呈現相同趨勢，差異不大。當以糖尿病併高血壓病患為分析樣本時，結果發現以醫師照護連續性高者為參考組時，照護連續性為低與中者，其住院次數分別為2.53倍、1.70倍，急診次數為2.52倍和1.63倍，醫療費用為1.41倍和1.29倍，皆達統計顯著意義(P<0.001)。相對的，醫療院所照護連續性的分析結果分別為:住院次數1.77倍與1.35倍，急診為1.84倍與1.38倍，總費用為1.12倍與1.06倍。 本研究發現，以老年慢性病患為研究對象時，以醫師或醫療場所為測量基礎的照護連續性愈高者，其住院與急診次數、以及總醫療費用均愈低，而且醫師照護連續性的影響比醫療院所照護連續性來得大，此發現與過去文獻的發現相似。另外，照護連續性與照護結果的相關性，在單一慢性病老年族群或雙重慢性病老年族群之間，似乎沒有明顯的差異。

The study aims to examine the continuity of care comparison between the physician and the institution based measures, and to understand the association between continuity of care and health outcome among chronically illed elderly. The cross-sectional analysis was employed in this study. The database of National Health Insurance in National Health Research Institute was used. This study was using the 2005 National Health Insurance Registry for Beneficiaries Claims Data files. Only collected the people who aged over 65 years old on 2008/01/01, and those people who divided into three groups should visited physicians or institutions for more than 3 times in 2008. The putative index of usual provider continuity in physician (UPCP) and usual provider continuity in site (UPCs) were used as independent variables and divided into three groups. The hospitalization and emergency visits and total medical expense were dependent variables. The controlled variables in this study included personal characterics (age, gender, presence of chronic diseases, and total ambulatory physician visits), and institution characterics (location and hospital level). In this study, a total of 9,018 diabetes patients, 24,454 hypertension patients and 10,219 diabetes and hypertension patients were recruited. First, 2,567 diabetes patients (28.47%) had hospitalized, while 6,451 diabetes patients (71.53%) had not hospitalized. In others words, 2,780 diabetes patients (30.82%) had emerged, while 6,238 diabetes patients (69.17%) had not emerged. Second, 5,294 hypertension patients (21.65%) had hospitalized, while 19,160 hypertension patients (78.35%) had not hospitalized. Futhermore, 6,664 hypertension patients (27.25%) had emerged, while 17,790 hypertension patients (72.75%) had not emerged. Finally, 2,907 diabetes and hypertension patients (28.45%) had hospitalized, while 7,312 diabetes and hypertension patients (71.55%) had not hospitalized. Moreover, 3,334 diabetes and hypertension patients (32.63%) had emerged, while 6,885 diabetes and hypertension patients (67.37%) had not emerged. As examined with a single provider, the mean UPC index in diabetes patients was 0.48. In this study, higher UPC could be seen in those people who aged between 65 and 70, male, those of lower ambulatory physician visits, patients without others chronic diseases. In other words, the mean UPC index in hypertension patients was 0.443. In the study, higher UPC could be seen in those people who aged between 65 and 70, female, those of lower ambulatory physician visits, patients without others chronic diseases. Moreover, the mean UPC index in diabetes and hypertension patients was 0.432. In the study, higher UPC could be seen in those people who aged between 65 and 70, female, those of lower ambulatory physician visits, patients without others chronic diseases. However, as examined with a single site, most of characterics were the same as physician but the one different characteric was aged than 75 years old. In negative binomial regression and gamma distribution, groups of lower and moderate UPCp in diabetes patients showed higher probability in hospitalization, emergency visits, and total medical expense than the group of high UPC, with 2.75 and 1.58 folds in hospitalization ,with 2.87 and 1.53 folds in emergency visits, and with 1.54 and 1.23 folds in total medical expense, respectively. The result of hypertension and diabetes and hypertension patients also revealed similar trend. Furthermore, the similar results showed in UPCs. Higher COC would present with lower frequency in hospitalization and emergency visits, and fewer total medical expense. According to this study, the continuity of care in physician or institution plays a role in hospitalization and emergency visits. Therefore, It is an important issue to facilitate the the continuity of care in physician or institution.