Exclusion Criteria in Alcohol-Related Services Research

几乎所有的酒精治疗结果研究都使用排除标准来确定哪些患者亚样本将参与研究方案。尽管出于科学、实践或伦理原因，这种排除有时是必要的，但它们可能会通过创建不具代表性的研究样本而在未知程度上降低外部效度。本项目的目的是对现代酒精治疗结果研究历史中采用的排除标准进行全面评估。具体而言，我们将检查在治疗结果研究中使用了哪些排除标准，它们如何影响结果的普遍性，它们是否不成比例地将弱势群体（例如，非洲裔美国人，妇女）排除在研究参与之外，以及它们如何影响结果。为了实现这些目标，将使用自1970年以来进行的701项酒精治疗结果研究的荟萃分析数据集来评估排除标准的性质和流行程度。该数据中的排除标准将被应用于六个大型“现实世界”数据集，其中包括在美国不同地区寻求酒精问题治疗的约20 000人。通过这一分析程序，我们将能够确定现实世界中有多少比例的酒精患者没有通过典型的研究排除标准，以及在各种排除标准下，被排除的患者的基线和随访特征（例如，结果）是否与未被排除在结果研究中的患者不同。

Almost all alcohol treatment outcome studies use exclusion criteria to determine which subsample of patients will participate in research protocol. Although such exclusions are sometimes necessary for scientific, practical or ethical reasons, they may lower external validity to an unknown extent by creating an unrepresentative research sample. The purpose of this project is to provide a comprehensive assessment of exclusion criteria employed during the modern history of alcohol treatment outcome research. Specifically, we will examine which exclusion criteria have been utilized in treatment outcome research, how they have affected the generalizability of results, whether they have disproportionately excluded members of vulnerable populations (e.g., African-Americans, women) from research participation, and how they may have affected outcome results. To pursue these objectives, the nature and prevalence of exclusion criteria will be assessed using a meta-analytic data set comprising 701 alcohol treatment outcome studies conducted since 1970. The exclusion criteria in this data will be operationalized and applied to six large "real world" data sets comprising about 20,000 individuals seeking treatment for alcohol problems in different parts of the United States. Through this analytic procedure, we will be able to determine what proportion of real-world alcohol patients do not pass typical research exclusion criteria, and whether, under various exclusion criteria, the baseline and follow-up characteristics (e.g., outcomes) of excluded patients differ from those not excluded from outcome research studies.