台灣版WHOQOL-OLD問卷之發展與驗證

Development and validation of the WHOQOL-OLD in Taiwan

目標：發展及驗證台灣版世界衛生組織老人生活品質問卷（WHOQOL-OLD）。方法：研究之初先以標準化程序進行問卷的翻譯，然後於2012年六至八月，以橫斷面研究設計在大台北地區與嘉義地區收集438名60歲以上社區老人的資料，包括WHOQOL-OLD、WHOQOLBREF、短式台灣老年憂鬱量表（Geriatric Depression Scale-15, GDS-15）、巴氏量表（Barthel Index, BI）、簡短式智能評估（Mini-Mental State Examination, MMSE）。心理計量分析包括內部一致性信度、內容效度、建構效度、同時效標關聯效度、預測效度與區辨效度，採用的統計方法有相關、迴歸、兩獨立樣本t檢定、驗證性因素分析等。結果：WHOQOL-OLD問卷之內部一致性係數介於0.71~0.86。問卷題目與自己所屬的向度之相關（r= 0.62~0.88）都高於跟其他向度之相關。驗證性因素分析（confirmatory factor analysis）大略支持WHOQOL-OLD為兩階層因素之理論模型。WHOQOL-OLD六個層面跟多個同時效標之相關雖然幾乎都達到0.01的顯著水準，但部分相關仍不如預期假設的強。在預測數個整體生活品質指標方面（除WHOQOLBREF的整體自評題G1、G2外），WHOWQOL-OLD六個層面的解釋力約達到四成，而且是一個可接受的外加模組。最後，WHOQOL-OLD可區別健康與較不健康的老人。結論：研究結果支持WHOQOL-OLD台灣版問卷有良好的信、效度，可以使用在如同本研究樣本較為健康的台灣六十歲以上的社區老人。

Objectives: To develop and validate the WHOQOL-OLD-Taiwan version. Methods: This was a cross-sectional study. First, we translated this questionnaire by following standard procedure. After a pilot study, we collected data from 438 seniors over 60 years of age from communities in greater Taipei and Chiayi from June to October, 2012. In addition to filling out the WHOQOL-OLD, the participants also completed the WHOQOL-BREF, Geriatric Depression Scale-15, Barthel Index, and Mini-Mental State Examination as the criterion variables. We examined the following psychometric properties: internal consistency reliability, content validity, construct validity, concurrent validity, predictive validity, and discriminant validity by using correlation analysis, regression analysis, two independent group t-tests, and confirmatory factor analysis. Results: The internal consistency coefficients were between 0.71~0.86. The correlation coefficients between each item (r=0.62~0.88) and the facet of belonging were higher than those with other facets. Confirmatory factor analysis generally supported the idea that the WHOQOLOLD was a second-order factor model, which indicated that six factors were subsumed under an overall ＂quality of life＂ factor. Most of the correlation coefficients between the facets of WHOQOL-OLD and several concurrent criteria were statistically significant (p ＜ 0.01), but some correlations were not as high as expected. Except for predicting G1 and G2, the WHOQOL-OLD explained about 40% of the variation in the overall quality of life indices. This study also showed that the WHOQOL-OLD was a good add-on module. Elderly people with better and worse health conditions could be discriminated by the WHOQOL-OLD through independent group t-tests (p ＜ 0.01). Conclusions: The WHOQOL-OLD-Taiwan version has good psychometric properties. The WHOQOL-OLD module appears to be a useful instrument for use with community-dwelling Taiwanese seniors.