![[ Visit Client Website ]](images/banner.gif)
Schedule:
Friday, January 15, 2016: 4:30 PM
Meeting Room Level-Meeting Room 14 (Renaissance Washington, DC Downtown Hotel)
* noted as presenting author
Background and Purpose: School-based drug prevention programming can be found in all grade levels but previous meta-analyses have indicated that middle school is the ideal period for intervention to be effective. The present review seeks to assess the empirical status of programs that have evolved during the last decade within the context of evidence-based policy making efforts in Canada and the United States. This meta-analysis evaluates the effectiveness of interactive school-based substance use prevention programs among middle school students in North America. Additionally, the analysis includes information on programmatic characteristics and whether they moderate effectiveness.
Methods: A set of predetermined search terms was used systematically in the following databases: Social Work Abstracts; Social Services Abstracts; Sociological Abstracts; PsycINFO; ERIC; Academic Search Complete; PubMed-Medline; Cochrane Library. Grey literature searches were conducted in Dissertations and Theses; ERIC; OpenGrey; and Google Scholar. Inclusion criteria included (1) English studies conducted in North America after 1998; (2) Programs delivered on the school premises during regular school hours or after school hours; (3) middle school-aged adolescents aged 12-14 years or grades 6-8; (4) interactive delivery techniques used by the program; (5) reported outcome measures of actual use of alcohol or any other illicit drugs; (6) random control trial (RCT) or quasi-experimental studies. The analysis estimated an average effect across studies. Final effect sizes (Cohen’s d) were calculated from continuous measures and effect sizes from dichotomous outcome variables were converted to the d index. Studies with multiple treatment conditions or measurement times for each outcome were combined to produce a single effect size per sample per outcome. Random effects models were used to conduct the final pooled effect size calculations.
Results: The initial search yielded 8,314 references. Screening at abstract and title level resulted in 643 studies for full text review. A final sample of 50 studies remained for the analysis, contributing information on 58 separate programs and 82 separate treatment/comparison conditions yielding 394 effect sizes. Of the 82 study conditions, 95.12% were RCTs. The outcomes included use of alcohol, binge drinking/getting drunk, cannabis, other illicit drugs (excluding cannabis), inhalants, and polydrug use. For alcohol-related outcomes, the pooled effect sizes (Cohen’s d) were -0.091 [C.I. -0.134, -0.048] for alcohol use and -0.136 [C.I. -0.202, -0.070] for binge drinking/getting drunk. The pooled effect sizes for substance use were -0.104 [C.I. -0.161, -0.047] for cannabis use; -0.075 [C.I. -0.105, -0.046] for inhalant use, and -0.042 [C.I. -0.077, -0.007] for using other illicit drugs. The pooled effect size for polydrug use was -0.183 [C.I. -0.269, -0.097]. These findings indicated that school-based prevention programs generally have small positive effects on substance use by adolescent program participants in North American middle schools. Moderator analysis revealed differences in mean effects on study and intervention characteristics.
Conclusion and Implications: Findings of the meta-analyses provided support for use of interactive programs in middle schools to prevent substance use by adolescents in North America. The results imply that teachers and social workers are effective instructors in the majority of programs available for middle schools.
Methods: A set of predetermined search terms was used systematically in the following databases: Social Work Abstracts; Social Services Abstracts; Sociological Abstracts; PsycINFO; ERIC; Academic Search Complete; PubMed-Medline; Cochrane Library. Grey literature searches were conducted in Dissertations and Theses; ERIC; OpenGrey; and Google Scholar. Inclusion criteria included (1) English studies conducted in North America after 1998; (2) Programs delivered on the school premises during regular school hours or after school hours; (3) middle school-aged adolescents aged 12-14 years or grades 6-8; (4) interactive delivery techniques used by the program; (5) reported outcome measures of actual use of alcohol or any other illicit drugs; (6) random control trial (RCT) or quasi-experimental studies. The analysis estimated an average effect across studies. Final effect sizes (Cohen’s d) were calculated from continuous measures and effect sizes from dichotomous outcome variables were converted to the d index. Studies with multiple treatment conditions or measurement times for each outcome were combined to produce a single effect size per sample per outcome. Random effects models were used to conduct the final pooled effect size calculations.
Results: The initial search yielded 8,314 references. Screening at abstract and title level resulted in 643 studies for full text review. A final sample of 50 studies remained for the analysis, contributing information on 58 separate programs and 82 separate treatment/comparison conditions yielding 394 effect sizes. Of the 82 study conditions, 95.12% were RCTs. The outcomes included use of alcohol, binge drinking/getting drunk, cannabis, other illicit drugs (excluding cannabis), inhalants, and polydrug use. For alcohol-related outcomes, the pooled effect sizes (Cohen’s d) were -0.091 [C.I. -0.134, -0.048] for alcohol use and -0.136 [C.I. -0.202, -0.070] for binge drinking/getting drunk. The pooled effect sizes for substance use were -0.104 [C.I. -0.161, -0.047] for cannabis use; -0.075 [C.I. -0.105, -0.046] for inhalant use, and -0.042 [C.I. -0.077, -0.007] for using other illicit drugs. The pooled effect size for polydrug use was -0.183 [C.I. -0.269, -0.097]. These findings indicated that school-based prevention programs generally have small positive effects on substance use by adolescent program participants in North American middle schools. Moderator analysis revealed differences in mean effects on study and intervention characteristics.
Conclusion and Implications: Findings of the meta-analyses provided support for use of interactive programs in middle schools to prevent substance use by adolescents in North America. The results imply that teachers and social workers are effective instructors in the majority of programs available for middle schools.