Center for Education and Drug Abuse Research, University of Pittsburgh, Pittsburgh, Pennsylvania 15261, USA.
Am J Drug Alcohol Abuse. 2009;35(4):260-6. doi: 10.1080/00952990903005916.
BACKGROUND/OBJECTIVES: To date, research on substance abuse prevention relied extensively on large sample randomized clinical trials to evaluate intervention programs. These designs are appropriate for certain types of randomized prevention trials (e.g., efficacy or effectiveness for broad populations) but are unfeasible for other prevention science scenarios (e.g., rare pathologies, pilot studies, or replication tests at specific locales).
An alternative randomized clinical trial is described that relies on much smaller samples, less resources than the large sample designs, randomization, N-of-1 designs for the intervention group, and mixed model analysis.
This methodology is illustrated using a small sample prevention study, which demonstrates its statistical power, flexibility, and sophistication for experimental testing of prevention-oriented research questions.
This methodology can be applied to many existing prevention datasets to facilitate secondary analyses of existing datasets as well as novel studies. It is hoped that such efforts will include further development of the small sample design in substance abuse prevention contexts.
背景/目的：迄今为止，药物滥用预防研究广泛依赖于大规模随机临床试验来评估干预项目。这些设计适用于某些类型的随机预防试验（例如，广泛人群的疗效或有效性），但不适用于其他预防科学场景（例如，罕见的病理学、试点研究或特定地点的复制测试）。
本文描述了一种替代的随机临床试验，它依赖于比大规模设计更小的样本量和更少的资源，对干预组进行随机分组、N-of-1 设计和混合模型分析。
该方法使用一个小样本预防研究进行了说明，展示了其对于预防导向研究问题的实验测试的统计能力、灵活性和复杂性。
该方法可应用于许多现有的预防数据集，以促进对现有数据集的二次分析以及新的研究。希望这些努力将包括在物质滥用预防背景下进一步开发小样本设计。
Zotero 插件