Community-level moderators of a school-based childhood sexual assault prevention program.

Childhood sexual abuse (CSA) is highly prevalent and associated with a wide variety of negative mental and physical health outcomes. School-based CSA education and prevention programs have shown promise, but it is unclear to what extent community-level characteristics are related to their effectiveness. The present cluster randomized controlled trial evaluated community-level moderators of the Safe@Lastprogramcomparedtoawaitlistcontrolcondition.(*) Knowledge gains from pre- to post-intervention were assessed in 5 domains: safe versus unsafe people; safe choices; problem-solving; clear disclosure; and assertiveness. Participants were 1177 students (46% White, 26% African American, 15% Hispanic, 4% Asian American, 6% Other) in grades 1 through 6 from 14 public schools in Tennessee. Multilevel models accounting for the nesting of children within schools revealed large effect sizes for the intervention versus control across all knowledge domains (d's ranged from 1.56 to 2.13). The effectiveness of the program was moderated by mean per capita income and rates of substantiated cases of child abuse and neglect in the community. Intervention effects were stronger for youth living in lower as compared to higher income counties, and for youth attending schools in counties with lower as compared to higher abuse/neglect rates. Child characteristics (sex, race) did not moderate intervention effects. This research identified two community-level factors that predicted the effectiveness of a CSA education and prevention program designed to improve children's knowledge of personal safety skills. School-based CSA prevention programs may require modification for communities with higher rates of child abuse and neglect.