文章基本信息

标题：Detecting Selection Effects in Community Implementations of Family-Based Substance Abuse Prevention Programs
本地全文：下载
作者：Laura G. Hill ; Scott G. Goates ; Robert Rosenman 等
期刊名称：American journal of public health
印刷版ISSN：0090-0036
出版年度：2010
卷号：100
期号：4
页码：623-630
DOI：10.2105/AJPH.2008.154112
语种：English
出版社：American Public Health Association
摘要：To calculate valid estimates of the costs and benefits of substance abuse prevention programs, selection effects must be identified and corrected. A supplemental comparison sample is typically used for this purpose, but in community-based program implementations, such a sample is often not available. We present an evaluation design and analytic approach that can be used in program evaluations of real-world implementations to identify selection effects, which in turn can help inform recruitment strategies, pinpoint possible selection influences on measured program outcomes, and refine estimates of program costs and benefits. We illustrate our approach with data from a multisite implementation of a popular substance abuse prevention program. Our results indicate that the program's participants differed significantly from the population at large. Federal, state, and local agencies increasingly require that funds for substance abuse prevention be spent on programs that have been found efficacious in randomized controlled trials (RCTs). However, community-based implementations may reach a population that is noncomparable to that of the RCT for a variety of reasons, including differences in recruiting, self-selection among program participants, differential attrition, or a different base population. The present study was designed to determine the nature of selection effects among participants attending a universal family-strengthening substance abuse prevention program called the Strengthening Families Program for Parents and Youth 10–14 (SFP). First, we briefly review the literature on selection effects in community-based prevention programs. Such programs are often delivered to families and focus on enhancing family strengths (e.g., parents' involvement with children and parental monitoring). Family interventions of this type have been shown to prevent substance abuse, delinquency, and other risky behaviors.¹ We then describe an evaluation design and analytic method that together enable detection of selection effects, even when the primary evaluation sample contains observations only on program participants, as is the case in many real-world implementations. Finally, we apply the technique to a community-driven multisite evaluation of SFP, a successful and widely disseminated prevention program,² and we discuss the results. Specifically, we use pretest data to examine the risk status of adolescent children of families who self-selected into SFP, to determine whether the probability of program participation was related to risk and protective factors. We then discuss the evaluation design in terms of its implications for assessment of program benefits, and we provide recommendations for translational research on preventive interventions in community settings.