社區整合式照護之成效:系統性文獻回顧

整合式照護的發展，是為了減少醫療照護的片斷化、增進照護品質、提升效率，以及避免不必要的醫療浪費。以社區為基礎的基層醫療是一個重要且良好的整合式照護模式，可以提供連續性、周全性和以病人為中心的照護。本文之研究目的在於評估社區中整合式照護模式相較於一般性照護，在醫療照護品質上的成效。以PubMed （1996 年至2010 年12 月）資料庫做搜尋，使用Delivery of Health Care, integrated, Quality of Health Care, Community Health Services的醫學標題表檢索詞做為關鍵字搜尋。收錄有比較整合式照護和一般照護之成效的隨機對照試驗（randomized controlled trials, RCTs）。出版的語言則限制為英文文獻。由兩位作者分別獨立篩選收錄適當文獻，評估研究品質，並摘錄研究資料。所摘錄的資料經過重複審查，並將其分為二分變項結果（dichotomous outcomes）和連續變項結果（continuous outcomes）。統合的結果，以勝算比（odds ratio, OR）表示二分變項結果；以標準平均差（standardized difference in mean）表示連續變項結果，使用隨機效果模型（random-effect model）來進行統合分析（meta-analysis）。使用漏斗圖（funnel plot）來檢視出版偏差。本篇收集了十篇隨機對照試驗文獻，其中六篇為二分變項結果，四篇為連續變項結果。研究間存在有異質性。統合結果顯示，整合式照護模式相較於一般性照護對於增進照護品質，在二分變項結果與連續變項結果兩者皆有顯著的效果。二分變項與連續變項結果的漏斗圖皆顯示出不對稱形狀，表示有潛在的出版偏差。雖然本研究有異質性及出版偏差的可能性，結果顯示社區中的整合式照護比一般性照護對照護品質有較佳的效果。然而，整合式照護在其他向度的成效則需進一步研究分析。

Integrated health care programs have been developed to reduce fragmentation of care, to improve quality of care, to increase efficacy, and to prevent unnecessary medical cost. Community-based primary care is an important model for integrated health care, which can provide continuous, comprehensive and patient-centered care. Our aim is to assess the effects of integrated care programs versus usual care on quality of care in the community. We searched the PubMed (1996 to December 2010) and used MeSH term of "Delivery of Health Care, integrated", "Quality of Health Care", and "Community Health Services" as keywords. Randomized controlled trials (RCTs) that assessed the effects of integrated care compared with usual care were included. The publication language was limited in English. Two authors independently reviewed and selected the trials for inclusion, assessed trial quality and extracted the data. The data were double-checked and were categorized into dichotomous outcomes and continuous outcomes. The overall effects were estimated by the pooled odds ratio (OR) for dichotomous outcomes and by the pooled standardized difference in mean for continuous outcomes respectively, using the random-effect model for meta-analysis. Publication bias was examined by the funnel plot. Ten RCTs were included in this review. Among these studies, 6 trials were categorized into dichotomous outcomes while 4 trials were categorized into continuous outcomes. Heterogeneity existed among trials. Compared with usual care, integrated care had a significant effect on improving quality of care both in dichotomous outcomes (pooled OR 1.83, 95% CI 1.49 to 2.24; p<0.001) and continuous outcomes (pooled mean effect size 0.31, 95% CI 0.14 to 0.48; p<0.001). Funnel plots showed asymmetry both in dichotomous and continuous outcomes, indicating the existence of potential publication bias. Integrated care programs in the community have a better effect on quality of care than usual care in spite of the existence of heterogeneity and potential publication bias. Further research of integrated care on other dimensions may be required.