Mechanisms of Intervention Impact for Diverse Dementia Caregiving Populations

描述（由申请人提供）：预计老年人口的增长以及对阿尔茨海默病和类似痴呆症的早期识别的更加重视，对卫生保健系统提出了沉重的要求，包括超过1500万的家庭成员，他们为痴呆症患者提供某种形式的持续非正式护理。大量研究记录了痴呆症护理的主要和次要压力源，提供基于证据的服务以减轻与护理相关的负担和压力仍然是一项主要的公共卫生优先事项。由于家庭照顾者是一个非常多样化的群体，他们在许多人口统计学和情境变量上都存在差异，因此需要仔细而复杂的分析来更好地理解现有干预试验在不同亚群体中取得效果的机制。在本提案中，我们描述了一系列的二次分析，这些分析将对来自第二个增强阿尔茨海默氏症护理者健康资源（REACH II）项目的公开数据进行。REACH II在现有的护理干预随机试验中是独一无二的，它纳入了地理和种族多样化的样本，但迄今为止只有少数出版物研究了干预结果的影响，没有发表的分析研究了干预影响的中介因素，也没有发表的分析研究了这种机制是否因种族或其他重要的人口统计学亚组（例如，性别，配偶与非配偶照顾者）而不同。我们提出了一系列协调的分析，以解决四个相互关联的目标。特别是，我们建议1)对评估REACH II结果的常用工具进行项目层面的心理测量分析，包括差异项目功能（DIF）分析，以确定人口统计亚组的一些发现是否可能部分受到测量偏差的影响；2)对干预可能取得主要结果效益（减轻照顾者抑郁和负担）的潜在机制（改善社会支持，减少压力评估）进行中介分析；3)进行多群体结构方程模型分析，以确定目标2的中介效应和先前发表的调节效应是否随种族或民族而变化；4)检验符合REACH II干预措施的预测因素，并对充分遵守干预措施的参与者子集进行干预效果的平均因果效应（CACE）分析。这些分析将实质性地告知REACH II干预措施的潜在人群影响，并有助于指导特定痴呆症护理亚组的额外转化工作。

DESCRIPTION (provided by applicant): The projected growth of the older adult population and the greater emphasis on early stage identification of Alzheimer's disease and similar dementias are placing heavy demands on the health care system, including the more than 15 million family members who provide some type of ongoing informal care to persons with dementia. A considerable body of research documents the primary and secondary stressors of dementia caregiving, and providing evidence-based services to ease the burden and stress associated with caregiving continues to be a major public health priority. Because family caregivers are an incredibly diverse group who differ along many demographic and situational variables, careful and sophisticated analyses are needed to better understand the mechanisms by which existing intervention trials have achieved their effects among diverse subgroups. In this proposal, we describe a series of secondary analyses to be conducted on the publically-available data from the second Resources for Enhancing Alzheimer's Caregivers Health (REACH II) project. REACH II was unique among the existing randomized trials of caregiver interventions in its enrollment of a geographically and racially diverse sample, but only a few publications to date have examined intervention outcome effects, and no published analyses have examined the mediators of intervention impact or tested whether such mechanisms differ by race or by other important demographic subgroups (e.g., gender, spouse vs. non-spouse caregivers). We propose a coordinated series of analyses that will address four inter-related aims. In particular, we propose 1) to conduct item-level psychometric analyses of the common instruments used to assess REACH II outcomes, including differential item functioning (DIF) analyses, to determine if some findings across demographic subgroups might be partially contaminated by measurement biases; 2) to perform mediation analyses of the potential mechanisms (improved social support, reduced stress appraisals) by which the intervention may have achieved primary outcome benefits (reduced caregiver depression and burden); 3) to conduct multiple group structural equation modeling analyses that will determine if the mediation effects from aim 2 and previously published moderator effects vary as a function of race or ethnicity; and 4) to examine the predictors of compliance with the REACH II intervention and conduct complier-average causal effect (CACE) analyses of the effect of the intervention among the subset of participants who sufficiently complied with it. These analyses will substantially inform the potential population impact of the REACH II intervention and be useful for guiding additional translational work for particular dementia caregiving subgroups.