Statistics & Psychometrics, Mapi Values, Lyon, France.
Clin Trials. 2009 Dec;6(6):574-84. doi: 10.1177/1740774509352515. Epub 2009 Nov 23.
Multinational clinical trials commonly include different language versions of patient-reported outcomes (PRO) instruments without considering the question of their cross-cultural validity. The inclusion of a PRO instrument, the Adult AIDS Clinical Trial Group Symptom Distress Module (SDM), in an multinational clinical trial in HIV-1 antiretroviral-naive patients offered an opportunity to explore the methods to assess cross-cultural validity of PRO instruments in the context of clinical trials.
To assess the cross-cultural validity of the SDM across seven cultural groups in the setting of a multinational HIV clinical trial.
Twenty-five language versions of the SDM were included in a Phase IIb/III trial comparing maraviroc with efavirenz (each in combination with zidovudine/ lamivudine) conducted in 12 countries to assess symptoms perceived by HIV-1-infected antiretroviral-naive patients. Differential item functioning (DIF) detection and the STATIS method were combined in a pragmatic approach to assess the cross-cultural validity of the SDM using pre-antiretroviral treatment data from 759 patients.
Statistically significant DIF between cultural groups was observed for four items: fatigue; fevers; anxiety; and headache. However, examination of these items by linguists did not lead to meaningful explanations for the statistical differences. With the STATIS approach, the Bantu and European Germanic groups were the furthest from the Occidental English group.
The assessment of cross-cultural validity had to be performed on some very small samples and on data aggregated by cultural groups, which suggests the need for a cautious interpretation of the results.
Given the heterogeneity of cultures considered, the absence of meaningful explanations for statistically significant differences between cultural groups supports the cross-cultural validity of the SDM versions included in this trial. Thus, this study demonstrated that it is feasible to conduct assessment of crosscultural validity of PRO instruments using data collected in the setting of multinational clinical trials.
多国临床试验通常包括患者报告结局（PRO）工具的不同语言版本，而不考虑其跨文化效度问题。在一项针对 HIV-1 抗病毒初治患者的多国临床试验中，包含 PRO 工具——成人艾滋病临床研究组症状困扰模块（SDM），这为探索在临床试验背景下评估 PRO 工具跨文化效度的方法提供了机会。
在一项多国 HIV 临床研究中，评估 SDM 在七个文化群体中的跨文化效度。
在一项比较马拉维若韦与依非韦伦（均与齐多夫定/拉米夫定联合使用）的 2 期 b/3 期试验中，纳入了 SDM 的 25 种语言版本，该试验在 12 个国家进行，旨在评估 HIV-1 感染、抗病毒初治患者感知的症状。采用预抗病毒治疗数据，采用差异项目功能（DIF）检测和 STATIS 方法相结合的实用方法，对来自 759 例患者的数据进行分析，以评估 SDM 的跨文化效度。
在四个项目中观察到文化群体间存在统计学显著的 DIF：疲劳、发热、焦虑和头痛。然而，语言学家对这些项目的检查并未为统计学差异提供有意义的解释。使用 STATIS 方法，班图语和欧洲日耳曼语群体与西方英语群体相差最远。
跨文化效度的评估必须在一些非常小的样本和按文化群体汇总的数据上进行，这表明需要谨慎解释结果。
考虑到所考虑的文化的异质性，文化群体间统计学显著差异无有意义的解释，支持了本试验中纳入的 SDM 版本的跨文化效度。因此，本研究表明，在多国临床试验环境中使用收集的数据评估 PRO 工具的跨文化效度是可行的。
