Department of Biostatistics, Shiraz University of Medical Sciences, Shiraz, Iran.
Qual Life Res. 2013 Aug;22(6):1255-63. doi: 10.1007/s11136-012-0250-3. Epub 2012 Aug 18.
This article is a report of using seemingly unrelated regression (SUR) models to examine the determinants of different dimensions of quality of life (QoL) among childbearing age women. There are a limited number of studies on QoL and its associated factors among women in developing countries such as Iran. Therefore, more attention should be focused on identifying these issues.
We administered the Persian's abbreviated version of the World Health Organization Quality of Life (WHOQOL-BREF) questionnaire to 1,067 married women aged between 15 and 49 years. The women were chosen via a multistage research design from the rural region of Shiraz, the center of Fars Province in Iran in 2008. Clinical and socio-demographic characteristics as well as their reproductive health-related characteristics were investigated. To identify associated factors of QoL dimensions, ordinary least squares (OLS) regression and SUR were used and their findings were compared.
The WHOQOL-BREF showed acceptable consistency (Cronbach's alpha range: 0.62-0.75 across domains). Lower age, absence of long-term illness, economic status satisfaction, higher level of education, lower number of pregnancies, and higher body mass index were important associated factors of different dimensions of the QoL among these women. The estimated parameters for these factors were in close agreement in both OLS and SUR estimation methods. However, the SUR estimator provided the higher precision of the estimates than the OLS estimator, as the parameters obtained by SUR are characterized by lower standard errors. Women's age, income satisfaction, and level of education were common for all domains.
This study presents a novel approach to simultaneously predict QoL domains using the SUR estimators and the results are relevant for implementing objective QoL. SUR estimators performed consistently better than the OLS estimators, since SUR takes the correlation between error terms into account. Thus, the SUR method could be a useful methodology for predicting QoL domains.
本文报告了使用似乎不相关的回归（SUR）模型来检验生育年龄妇女生活质量（QoL）不同维度的决定因素。在伊朗等发展中国家，关于 QoL 及其相关因素的研究数量有限。因此，应该更加关注这些问题。
我们对 1067 名年龄在 15 至 49 岁之间的已婚妇女进行了世界卫生组织生活质量（WHOQOL-BREF）问卷的波斯语缩写版调查。这些妇女是通过 2008 年在伊朗法尔斯省首府设拉子农村地区的多阶段研究设计选择的。调查了她们的临床和社会人口统计学特征以及与生殖健康相关的特征。为了确定 QoL 维度的相关因素，我们使用了普通最小二乘法（OLS）回归和 SUR，并比较了它们的发现。
WHOQOL-BREF 显示出可接受的一致性（各领域的 Cronbach's alpha 范围为 0.62-0.75）。较低的年龄、没有长期疾病、经济状况满意度、较高的教育水平、较低的怀孕次数和较高的体重指数是这些妇女生活质量不同维度的重要相关因素。在 OLS 和 SUR 估计方法中，这些因素的估计参数非常一致。然而，SUR 估计器提供了比 OLS 估计器更高的估计精度，因为 SUR 获得的参数具有较低的标准误差。妇女的年龄、收入满意度和教育水平与所有领域都有关。
本研究提出了一种使用 SUR 估计器同时预测 QoL 领域的新方法，其结果对于实施客观的 QoL 具有重要意义。SUR 估计器的表现始终优于 OLS 估计器，因为 SUR 考虑了误差项之间的相关性。因此，SUR 方法可能是预测 QoL 领域的有用方法。
