A Meta-Analysis of Direct and Mediating Effects of Community Coalitions That Implemented Science-Based Substance Abuse Prevention Interventions

Abstract

This article reports results of a meta-analysis of the effects of a set of community coalitions that implemented science-based substance use prevention interventions as part of a State Incentive Grant (SIG) in Kentucky. The analysis included assessment of direct effects on prevalence of substance use among adolescents as well as assessment of what “risk” and “protective” factors mediated the coalition effects. In addition, we tested whether multiple science-based prevention interventions enhanced the effects of coalitions on youth substance use. Short-term results (using 8th-grade data) showed no significant decreases in six prevalence of substance use outcomes—and, in fact, a significant though small increase in prevalence of use of one substance (inhalants). Sustained results (using 10th-grade data), however, showed significant, though small decreases in three of six substance use outcomes—past month prevalence of cigarette use, alcohol use, and binge drinking. We found evidence that the sustained effects on these three prevalence outcomes were mediated by two posited risk factors: friends' drug use and perceived availability of drugs. Finally, we found that the number of science-based prevention interventions implemented in schools within the coalitions did not moderate the effects of the coalitions on the prevalence of drug use. Study limitations are noted.

Keywords:

• community intervention
• science-based prevention
• risk factors
• protective factors

Notes

Notes

1. Regarding risk factors, three items measuring attitudes favorable to drug use asked how wrong the respondent thought it was for someone his or her age to use cigarettes, marijuana, and other drugs; responses used a 4-point scale ranging from “Very wrong” to “Not wrong at all”. Four items measuring friends' drug use asked how many of the respondent's four best friends had used different substances, with a 5-point response scale ranging from “None” to “Four friends.” Family conflict was measured by three items that asked the respondent whether family members have arguments, yell at each other, or insult each other; responses used a 4-point scale ranging from “NO¡” to “YES¡” Three items measuring parental attitudes favorable to drug use asked how wrong the respondent's parents feel it would be for him or her to use substances; responses used a 4-point scale from “Very wrong” to “Not wrong at all.” School days skipped was measured using a single item that asked the respondent how many full days of school he or she had skipped in the past 4 weeks, with responses on a 7-point scale ranging from “None” to “11 or more.” Academic failure was measured using a single item that asked what the respondent's grades were like in the past year, with responses on a 5-point scale ranging from “Mostly A's” to “Mostly F's.” Three items measuring neighborhood adults' attitudes favorable to drug use asked how wrong most adults in the respondent's neighborhood think it is for youth to use marijuana, drink alcohol, and smoke cigarettes, with responses on the same scale used for the youth and parent attitude measures. Perceived low risk of being caught for drug use was measured using three items that asked the respondent whether a youth would be caught by police for using marijuana, alcohol, or carrying a gun in their neighborhood; responses used a 4-point scale ranging from “YES*” to “NO*” Four items measuring perceived availability of drugs asked how easy it would be for someone the respondent's age to get alcohol, cigarettes, marijuana, and other drugs; responses used a 4-point scale from “Very hard” to “Very easy.” Of the protective factors, perceived risk of drug use was measured using three items that asked the respondent how much he or she thought people risk harming themselves physically or otherwise by using marijuana (once or twice, and regularly) and by using alcohol regularly; responses used a 4-point scale from “No risk” to “Great risk.” Family attachment was measured using three items that asked whether the respondent felt close to his mother and shared thought and feelings with father and mother; responses used a 4-point scale from “NO¡” to “YES*” School commitment was measured using three items that asked how important the respondent's school work was and how often he or she enjoyed being in school; responses used a 5-point scale appropriate to each survey item.

2. In testing the moderating effect of whether the non-KIP school in each pair was unique or whether it also appeared in another pair, we compared (for the short-term data) the results for 13 school pairs with unique comparison schools to the 50 school pairs where the school was used twice. In the data involving sustained effects, 7 school pairs with unique controls were compared with 40 pairs using duplicate comparison schools. In the short-term effect data, none of the comparisons of school pairs with unique controls versus twice-paired controls showed significant differences between slopes for the intervention effect. For the sustained effect data, only one outcome showed a significant difference between the pairs using unique controls and those with controls used twice. For smokeless tobacco prevalence, the 7 school pairs with unique controls had a positive slope for the intervention effect (b = 0.066, SEb =. 011), whereas the school pairs with controls paired twice showed a negative slope (b = −0.021, SEb = 0.004). The Z test showed the difference in these slopes to be highly significant (p < 0.0001).

3. We also conducted a separate set of analyses in which the distal outcomes were frequency, rather than prevalence, of use of each substance. We reported in the narrative the prevalence results only, because prevalence is much more commonly reported in evaluations of prevention interventions. For the frequency models, the distal outcomes of smokeless tobacco, cigarette, alcohol, marijuana, and inhalant frequency of use were computed from items measuring lifetime, past year, and past 30-day use through latent-variable scores (i.e., factor scores), again using LISREL 8.54. Response categories for the survey items measuring alcohol, marijuana, and inhalant use asked (for each time period—lifetime, past year, and past 30 days) the number of occasions on which the respondent had used each substance, with categories ranging from 0 (never used in that time period) to 6 (40 or more occasions). Responses for items measuring smokeless tobacco use were as follows: for lifetime use: 0 (never used) through 4 (use regularly now); for past year: 0 (no occasions) through 6 (40 or more occasions); and for past 30-day use, ranging from 0 (not at all) through 4 (more than once a day). Responses for items measuring cigarette use were as follows: for lifetime use: 0 (never used) through 4 (use regularly now); for past year use: 0 (no occasions) through 6 (40 or more occasions); and for past 30-day use: frequency of use, ranging from 0 (not at all) through 4 (two packs or more per day). The binge drinking outcome was measured by a single indicator of frequency of binge drinking (more than five drinks in a row in the past two weeks); response categories ranged from 0 (none) to 5 (10 or more times). Using the same meta-analytic procedures described in the narrative, the results showed the following significant short-term effects: an intervention effect in the desired direction on frequency of alcohol use (beta = −0.012; SEb = 0.011); and an intervention effect on increase in frequency of inhalant use (beta = 0.027; SEb = 0.012). Sustained intervention effects were significant and in the same direction (decreases in use) as for the three prevalence outcomes reported in the narrative: frequency of cigarette use (beta = −0.039; SEb = 0.014); frequency of alcohol use (beta = −0.026; SEb = 0.012); and frequency of binge drinking (beta = −0.043; SEb = 0.015).