目標：發展一份適用於台灣、以病患經驗為主軸且具有信、效度之醫院醫療照護品質(本研究中簡稱為醫院品質)測量工具。方法：便利取樣全台灣20間自願參與研究之醫院，選取出院急性病人2,676名為樣本，進行問卷調查，有效回收問卷數為2,005份(回收率75%)。以心理計量項目分析程序選取核心題目，並檢驗核心問卷之信、效度。信度分析使用Cronbach’s α值與以Spearman-Brown formula修正後之折半信度值驗證量表內部一致性信度；效度分析使用皮爾
森相關係數值評估量表內容效度與同時效度；量表之建構效度以因素分析進行驗證。結果：醫
院品質核心問卷計有27題，包含六大向度：醫師照護、護理照護、整體醫病互動、社會心理支持、尊重與醫院環境。信度方面：六個向度之Cronbach’s α值為0.72至0.89、折半信度值為0.64至0.86，顯示內部一致性良好。效度方面：題目與所屬向度總分之皮爾森相關係數值介於0.57至0.88 (p＜0.01)，量表內容效度良好；向度總分與效標題「請問您是否願意推薦這間醫院給您的親友」之皮爾森相關係數值介於0.41至0.58，具有同時效度；建構效度分析中，探索性因素分析顯示整體醫病互動、尊重與社會心理支持三個向度可被整合為單一因素，然而藉由驗證性因素分析所進行之模型比較顯示，本研究所建構之六因素模型與資料間適配程度最佳。結論：本量表具有良好之信、效度，未來將藉由問卷資料的收集，健全國內目前已具有專業面之醫療照護品質指標。

Objectives: The purpose of this study was to develop and test the
reliability and validity of the Patient Reported Hospital Quality (PRHQ) Scale. Methods: 2,675 discharged patients across 20 hospitals in Taiwan were used in the current survey. Core items of the PRHQ were selected by
the process of item analysis. Reliability was then examined by conducting Cronbach’s alpha and split-half reliability tests. Validity analyses including content validity, criterion-related validity,and construct validity were tested by applying Pearson’s correlation coefficient, Exploratory Factor Analysis (EFA), and Confirmatory Factor Analysis(CFA).
Results: Through the process of item analysis, the original questionnaire was revised to include 27 core items, which were defined by six domains: medical care, nursing care, overall interaction, respect, psychosocial support,and hospital environment. The internal consistency (Cronbach’s alpha) coefficients and the splithalf reliability coefficients were 0.72 to 0.89 and 0.64 to 0.86 at the domain levels, showing good reliability of the scale. In addition, the range of the correlations between items and their domain scores was 0.57 to 0.88 (p<0.01), showing good content validity. The range of the correlations between domain scores and the criterion item was 0.41 to 0.58, indicating adequate criterionrelated
validity. Results from EFA indicated that among the six hypothetical domains, overall interaction, respect, and psychosocial support could be integrated into one factor; however, CFA of the six-factor model showed it to be the best fit for the data, indicating good construct validity
for the PRHQ. Conclusions: The first version of the PRHQ was confirmed to be reliable and valid and it can be applied in future studies to enrich the existing system for evaluation of hospital quality.