照護連續性之測量工具分析與應用

Abstract

研究背景與目的： 「照護連續性」在健康照護服務提供中所扮演的角色，自六、七零年代以來，持續的受到國外學術界的探討與重視，且已經發展許多不同面向的測量工具。近年來，國內亦有研究者開始關注照護連續性的相關議題，在國內研究的起步階段，確有必要先行探究適合國內醫療照護制度的測量工具，以利後續進行照護連續性之效益面的研究分析。故本研究之研究目的如下： 一、尋找適用於台灣的照護連續性測量方式或指標。 二、應用上述測量方式，觀察國內個人之照護連續性與健康照護結果是否有顯著的相關。 研究方法和資料來源： 本研究包括照護連續性測量工具評估，以及照護連續性和照護結果之相關性分析兩大部份，研究方法分述如下： 一、照護連續性測量指標評估 設計問卷，調查國內21位熟悉醫療照護與醫療品質等研究和實務的學者專家，並配合研究者對部份指標的統計分析，評估過去研究所採就診記錄與調查型資料的兩類測量方式，共計12項的照護連續性指標中，何者較適合國內之測量應用。 二、照護連續性和照護結果之相關性分析 採橫斷性之分析方式，資料來源為全民健康保險資料庫2005年兩組八萬人之抽樣歸人檔。 三、統計分析方式 應用皮爾森積差相關檢定（Pearson product-moment correlation test）做測量指標評估，卡方檢定（Chi-Square test）進行雙變項分析，並以羅吉斯迴歸（Logistic Regression）和負二項式迴歸（Negative Binomial Regression），進行照護連續性與住院有無暨次數，以及急診有無暨次數之相關性分析。 研究結果： 一、照護連續性測量工具評估 問卷調查結果顯示，國內照護連續性該用何種方式或指標測量未有明顯共識，各類指標有其特殊用途，但於測量操作上，多數專家學者之意見表示研究時應設定疾病類型或科別。同時，配合部分專家建議，採就診記錄測量之照護連續性指標中，最不易受就醫次數干擾的COCI指標，為較佳的照護連續性測量工具。 二、照護連續性與照護結果之相關性分析 這部份研究分析模型係以COCI和MMCI為照護連續性之測量指標。多變項之羅吉斯迴歸模型與負二項式迴歸模型的分析結果顯示，不論是對於一般案件門診之研究對象（n=49803）與慢性病案件門診之研究對象（n=14172），照護連續性高組皆相較照護連續性低組，有顯著較低的住院與急診利用風險，以及顯著較少的住院和急診次數。 結論： 現階段未有明顯共識指出何種照護連續性測量工具最適用於國內之測量應用，各類指標有其特殊用途，但採就診記錄的測量指標中，考量多位醫師且不易受就醫次數干擾的COCI指標，較符合專家對測量研究的建議。同時，本研究亦發現照護連續性高低的確與住院、急診之利用呈顯著的負相關。

Background and Objectives: From 1960s, “Continuity of Care”(COC) has been discussed in the international academia for it’s role in health care service. There are several measurement tools being developed in the past decades. In Taiwan, some researchers are interested in the issues about COC recently. In the preliminary stage of domestic research, it’s necessary to know what kind of COC measure is suitable for research in Taiwan. Thus would facilitate further research concerning COC. The purposes of this study were: 1. To seek for a suitable measure of COC for Taiwan. 2. To apply the suitable measure to examine the relationship between COC and care outcome─hospitalization and emergency department visit. Materials and Methods: The assessment of suitable measures of COC was conducted by survey of experts. We designed a questionnaire and sent out to 21 professors/doctors who were familiar with the research issue or healthcare quality. Using statistic analysis for several of the COC indices, we assessed the two categories of measures, i.e. information from claims data or patient survey data. A total of 12 indices were included in the analysis for COC measures. A cross-sectional analysis was conducted to examine the association between COC and care outcome. Data for the analysis came from two sets of the cohort database from the National Health Insurance research database. Each set included 40,000 persons randomly sampled from the NHI enrollees who were insured in 2005. Pearson product-moment correlation test and Chi-Square test were employed to assess the agreement among COC indicators and bivariate analysis respectively. Logistic Regression and Negative Binomial Regression were used to assess the relationship between COC and hospitalization and emergency department visit. Results: The first part of the thesis was to identify which measures of COC was sutiable for Taiwan’s research. According to the descriptive analysis of the questionnaire, there was no consensus about the best COC measure in Taiwan. But for the application of COC measures, the majority of experts expressed that the researchers should consider the disease pattern while conducting the COC measures. Moreover, among the COC measures based on claims data, COCI was the most stable indicator when taking the number of ambulatory visits into account. The second part of the thesis examined the association between COC and care outcome. COCI and MMCI were used as COC indicators in the regression models. The analysis was carried-out by two separate subsamples: patients with non-chronic disease related visits (n=49803) and with chronic disease related visits (n=14172). Results from the multiple logistic regression and negative-binomial regression showed that people with high COC had lower likelihood and fewer number of utilization in hospitalization and emergency department visit for patients in both sub-sample. Conclusions: There is no consensus on what kind of COC measure is the most suitable for Taiwan. Nevertheless, we suggest that COCI is good for domestic researchers. Furthermore, this study also supports that higher COC is associated with lower risk and frequency in both hospitalization and emergency department visit.