ANALYSIS OF TEAM-BASED SUBSTANCE ABUSE PREVENTION

点击翻译按钮获取中文摘要

DESCRIPTION: (adapted from Investigator's abstract) The purpose of this continuation proposal is to continue to complete and publish the evaluation of a prospective, randomized, study of the effects of the school-based anabolic steroid and other drug use prevention program, entitled ATLAS (Adolescents Training and Learning to Avoid Steroids). ATLAS is a social-influences based primary prevention program aimed at modifying those factors that increase the desire of young athletes to use anabolic steroids. In addition, nutrition and strength training alternatives to anabolic steroids are provided. The program uses peer leaders and coaches in a team-centered approach to exert positive peer pressure and role modeling. After one year, positive effects on self-reported steroid use, intentions to use steroids, nutrition behaviors, and many mediators of program effects were observed. 31 Portland, Oregon-area high school football teams were randomized to receive the ATLAS program (15 high schools) or a standard pamphlet on steroid use. Starting in the summer of 1994 with cohort 1, a total of four cohorts of all football players at these 31 high schools were measured for four years-- the last data collection was November, 1997. Over 4000 subjects are in this multiple wave longitudinal data set that includes questionnaire and anthropometric (e.g., height, weight, body fat, strength) measures. The data for the project were collected, cleaned and merged five months after funding for the grant expired. No new data collection is planned for this continuation proposal. We will apply recently developed statistical methods and software in the evaluation of program effects, mediation analysis to determine how the ATLAS program worked, moderator analysis to identify subgroups where the program is most or least effective, and etiological analysis to identify the precursors to steroid and other drug use. Advanced statistical methods including growth curve models, mixture models, multiple mediator models, hierarchical linear models, missing data models, and models for categorical data will be applied across multiple waves and across each of the four cohorts. The overall purpose of the grant is to use advanced as well as more common statistical methods to extract the maximum important information from these data and publish the results.