Emerging Issues in Analyzing Group-Based Treatment Data under Open Enrollment

描述（由申请人提供）：毫无疑问，大多数用于治疗酒精滥用和酒精依赖的心理社会干预措施都是以团体治疗的形式提供的。在许多情况下，治疗组通过开放注册或“滚动”入院来维持和补充，即患者在不同时间进入正在进行的治疗组，而其他成员则离开（由于毕业或辍学）。随着时间的推移，治疗组成员资格的每次会话变化过程（通常以非随机方式发生）给希望解释和理解开放注册治疗组所产生的依赖性以及在这种情况下对治疗效果进行正确建模的研究人员带来了独特的分析挑战。在酒精治疗试验数据分析中，处理小组成员的逐次变化的困难被认为是许多研究人员避免公开入组数据的分析挑战的主要原因之一。为了避免这些困难，研究人员（a）忽略了滚动治疗组产生的依赖性（导致参数估计和/或标准误差有偏差），（b）设计了以基于个体的干预措施（而不是群体）为中心的研究，或（c）设计了在一定时期后限制入组的研究（即“封闭”组）。这三种方法的主要成本是关于治疗效果的错误推论（即 I 型和/或 II 型错误）（在条件 [a] 下）以及如何进行治疗效果试验与如何在社区环境中进行治疗之间的脱节（在条件 [b] 和 [c] 下）。 面对许多这些分析和设计挑战，迫切需要进一步开发和传播在酒精治疗试验的背景下处理开放注册数据的方法（即多会员模型、模式混合模型）。这个探索性开发 R21 项目的目的是探索分析框架的实用性，该框架可以作为处理三个特定设计和/或分析问题的开放注册数据的基础： (a) 纳入常见的酒精治疗试验设计，(b) 开放入组试验中治疗作用机制的建模，以及 (c) 开放入组下各种基于组的酒精治疗设计（例如，基于组的 tx A 与基于组的 tx B、组 tx 与个人 tx）的统计功效分析。这项探索性/开发性资助申请将结合统计模拟模型和正在进行的酒精治疗试验的数据，最终目标是为酒精治疗研究人员提供清晰的教学示例，说明如何分析此类数据，并提供专门针对公开招募试验设计的样本量估计指南。

DESCRIPTION (provided by applicant): There is little question that most psychosocial interventions used to treat alcohol abuse and dependence are delivered in a group therapy format. In many settings, therapy groups are sustained and replenished using open enrollment, or "rolling" admission, whereby patients enter ongoing therapy groups at various times, while other members leave (due to graduation or dropout). This process of session-to-session change in therapy group membership over time (usually occurring in a nonrandom fashion) creates unique analytic challenges for investigators who wish to account for and understand the dependencies created by open enrollment into therapy groups and for the proper modeling of treatment effects in this context. Difficulties in handling session-to-session changes in group membership in the analysis of alcohol treatment trial data have been cited as one of the major reasons that many investigators avoid the analytic challenges of open enrollment data. To avoid these difficulties, investigators have (a) ignored dependencies created by rolling treatment groups (leading to biased parameter estimation and/or standard errors), (b) designed studies that center on individual-based interventions (as opposed to groups), or (c) designed studies that restrict enrollment after a certain period (i.e., "closed" groups). The major costs of each of these three approaches are incorrect inferences concerning treatment effects (i.e., Type I and/or Type II errors) (under condition [a]) and a disconnect between how treatment efficacy trials are conducted and how treatment takes place in community settings (under conditions [b] and [c]). In the face of many of these analytic and design challenges, further development and dissemination of approaches to handle open enrollment data in the context of alcohol treatment trials (i.e., multiple membership models, pattern mixture models) are sorely needed. The purpose of this exploratory- developmental R21 project is to explore the utility of analytic frameworks that may serve as the foundation for handling open enrollment data for three specific design and/or analytic problems: (a) incorporation into common alcohol treatment trial designs, (b) modeling of mechanisms of treatment action in trials with open enrollment, and (c) statistical power analysis for various group-based alcohol treatment designs (e.g., group-based tx A versus group-based tx B, group tx versus individual tx) under open enrollment. This exploratory/developmental grant application will use a combination of statistical simulation modeling and data from an ongoing alcohol treatment trial with the ultimate goal of providing alcohol treatment researchers with clear pedagogical examples on how to analyze such data and providing guidelines for sample size estimation specifically for open enrollment trial designs.