Reliability and Validity of Assessing User Satisfaction With Web-Based Health Interventions
Reliability and Validity of Assessing User Satisfaction With Web-Based Health Interventions
复制标题
DOI：
10.2196/jmir.5952
复制
发表时间：
2016-08-01
影响因子：
7.4
通讯作者：
Ebert, David Daniel
中科院分区：
文献类型：
--
作者：
Boss, Leif;Lehr, Dirk;Ebert, David Daniel
背景：在评价基于网络的卫生干预措施时，应考虑到用户的观点。评估使用者对其所接受的干预的满意度，可以增强干预效果的证据。因此，需要有效和可靠的措施来评估对基于网络的卫生干预措施的满意度。目的：本研究的目的是分析网络干预客户满意度问卷（CSQ-I）的信度、析因结构和构效度。方法：对来自2个独立的评估网络健康干预措施的随机对照试验的用户样本进行csq - 1的心理测量质量分析，其中一个来自抑郁预防干预（样本1，N=174），另一个来自压力管理干预（样本2，N=111）。首先，分析CSQ-I的底层测量模型，确定其内部一致性。采用多组验证性因子分析检验量表的因子结构和各组间的测量不变性。此外，通过比较满意度得分与主要临床结果来检验量表的结构效度。结果：量表上的多组验证性分析得出了一个拟合良好的单因子结构（近似均方根误差= 0.09，比较拟合指数= 0.96，标准化均方根残差= 0.05），在两个样本之间显示出部分强不变性。量表显示出非常好的可靠性，由McDonald ω表示。样本1和。样本2中是93。与抑郁症状变化显著相关(r=- 0.35， P
Background: The perspective of users should be taken into account in the evaluation of Web-based health interventions. Assessing the users' satisfaction with the intervention they receive could enhance the evidence for the intervention effects. Thus, there is a need for valid and reliable measures to assess satisfaction with Web-based health interventions.Objective: The objective of this study was to analyze the reliability, factorial structure, and construct validity of the Client Satisfaction Questionnaire adapted to Internet-based interventions (CSQ-I).Methods: The psychometric quality of the CSQ-I was analyzed in user samples from 2 separate randomized controlled trials evaluating Web-based health interventions, one from a depression prevention intervention (sample 1, N=174) and the other from a stress management intervention (sample 2, N=111). At first, the underlying measurement model of the CSQ-I was analyzed to determine the internal consistency. The factorial structure of the scale and the measurement invariance across groups were tested by multigroup confirmatory factor analyses. Additionally, the construct validity of the scale was examined by comparing satisfaction scores with the primary clinical outcome.Results: Multigroup confirmatory analyses on the scale yielded a one-factorial structure with a good fit (root-mean-square error of approximation =.09, comparative fit index =.96, standardized root-mean-square residual =.05) that showed partial strong invariance across the 2 samples. The scale showed very good reliability, indicated by McDonald omegas of .95 in sample 1 and .93 in sample 2. Significant correlations with change in depressive symptoms (r=-.35, P