FHI 360, Durham, North Carolina, United States of America.
Centers for Behavioral & Preventive Medicine, The Miriam Hospital, Providence, Rhode Island, United States of America.
PLoS One. 2018 Apr 12;13(4):e0195499. doi: 10.1371/journal.pone.0195499. eCollection 2018.
Low adherence in recent HIV prevention clinical trials highlights the need to better understand, measure, and support product use within clinical trials. Conventional self-reported adherence instruments within HIV prevention trials, often relying on single-item questions, have proven ineffective. While objective adherence measures are desirable, none currently exist that apply to both active and placebo arms. Scales are composed of multiple items in the form of questions or statements that, when combined, measure a more complex construct that may not be directly observable. When psychometrically validated, such measures may better assess the multiple factors contributing to adherence/non-adherence. This study aimed to develop and psychometrically evaluate tools to screen and monitor trial participants' adherence to HIV prevention products within the context of clinical trial research.
Based on an extensive literature review and conceptual framework, we identified and refined 86 items assessing potential predictors of adherence and 48 items assessing adherence experience. A structured survey, including adherence items and other variables, was administered to former ASPIRE and Ring Study participants and similar non-trial participants (n = 709). We conducted exploratory factor analyses (EFA) to identify a reduced set of constructs and items that could be used at screening to predict potential adherence, and at follow-up to monitor and intervene on adherence. We examined associations with other variables to assess content and construct validity. The EFA of screener items resulted in a 6-factor solution with acceptable to very good internal reliability (α: .62-.84). Similar to our conceptual framework, factors represent trial-related commitment (Distrust of Research and Commitment to Research); alignment with trial requirements (Visit Adherence and Trial Incompatibility); Belief in Trial Benefits and Partner Disclosure. The EFA on monitoring items resulted in 4 Product-specific factors that represent Vaginal Ring Doubts, Vaginal Ring Benefits, Ring Removal, and Side Effects with good to very good internal reliability (α = .71-.82). Evidence of content and construct validity was found; relationship to social desirability bias was examined.
These scales are easy and inexpensive to administer, available in several languages, and are applicable regardless of randomization. Once validated prospectively, they could (1) screen for propensity to adhere, (2) target adherence support/counselling, and (3) complement biomarker measures in determining true efficacy of the experimental product.
最近的艾滋病毒预防临床试验中，低依从性凸显出需要更好地理解、衡量和支持临床试验中的产品使用。艾滋病毒预防试验中常用的传统自我报告依从性工具往往依赖于单项问题，已被证明效果不佳。虽然理想情况下需要使用客观的依从性测量方法，但目前尚无适用于活性药物和安慰剂臂的方法。量表由多个以问题或陈述形式组成的项目组成，这些项目组合在一起可以衡量更复杂的结构，而这些结构可能无法直接观察到。经过心理测量验证后，此类测量方法可能更能评估导致依从性/不依从性的多种因素。本研究旨在开发和心理测量评估工具，以在临床试验研究背景下筛查和监测艾滋病毒预防产品的使用情况。
基于广泛的文献回顾和概念框架，我们确定并完善了 86 项评估依从性潜在预测因素的项目和 48 项评估依从性体验的项目。我们对前 ASPIRE 和 Ring 研究参与者和类似的非试验参与者（n=709）进行了一项包含依从性项目和其他变量的结构化调查。我们进行了探索性因素分析（EFA），以确定一组简化的结构和项目，这些项目可以在筛查时用于预测潜在的依从性，在随访时用于监测和干预依从性。我们还检查了与其他变量的关联，以评估内容和结构效度。筛选项目的 EFA 得出了一个具有可接受至非常好的内部一致性（α值：.62-.84）的 6 因素解决方案。与我们的概念框架类似，因素代表与试验相关的承诺（对研究的不信任和对研究的承诺）；与试验要求的一致性（就诊依从性和试验不相容性）；对试验益处的信念和伴侣披露。监测项目的 EFA 产生了 4 个产品特定的因素，分别代表阴道环疑虑、阴道环益处、环脱落和副作用，具有良好至非常好的内部一致性（α值：.71-.82）。我们发现了内容和结构效度的证据；并对社会期望偏差的关系进行了检查。
这些量表易于管理且经济实惠，有多种语言版本，并且无论随机分组如何都适用。经过前瞻性验证后，它们可以（1）筛查依从性倾向，（2）针对依从性支持/咨询，（3）补充生物标志物测量，以确定实验产品的真实疗效。
