Strong Fidelity-Implementation Results, Post-Test Only Two Groups Design, Propensity Score Matching, Multi-Level Analysis at 12-Month Follow-up Demonstrate Positive Impacts on Co-Occurring Risk Behaviors in the Horizon 7th Grade Program

Background: There is a need for evidence-based multi-component health promotion and risk behavior prevention programs.  Public health youth prevention and health promotion programs need to move away from single a outcome goal.

Purpose: Demonstrate a four-level evaluation of a new program: fidelity, implementation indicated by student responsiveness, psychosocial pre-post outcomes, and, 12-mo. post multiple behavior impacts, quasi-experimental design. Horizon is an innovative combination of three curricula for high risk minority urban adolescents, 14 sessions: 8, abstinence-only (evidence-based, Promoting Health Among Teens); 5, healthy relationship (evidence promising Dibble Institute Love U 2 Relationship Skills); and, 1, contraception/risk reduction (CDC/OPA).

Methods: Total ample includes over 3000 students; 2 trained health educators; 11 schools for fidelity-implementation-outcome analysis; 5 intervention and 3 comparison schools for impact analysis. Attendance tracking provided fidelity data. Post 14-session student survey with Likert-scale items measured 3 domains: teaching; relationships curriculum; sexual activity curriculum. Quasi-experimental design, 737 Horizon and 248 usual instruction 8^thgrade students after propensity score matching examined 12-month impacts.

Mixed effects repeated pre-post) regression models with demographic characteristics as control variables were performed for composite scales. Composite scores yielded means and ordinal scales that represented graded responses for behavior outcome measures. For example, sexual activity was measured in two ways, “ever sex”, yes/no; and ordinal scale, 0, no activity; 1, holding hands; 2, kissing/hugging; 3, touching above/below the waist; 4, other sex; 5, sexual intercourse; and, 6, sexual intercourse, with risky behavior(s) (without a condom). SAS Proc Mixed modeling and Supermix ordinal proportional and non-proportional analyses were used.

Results: 51% female, 51% African American, 34% Hispanic, 85% 13-14 years of age. 84% attended 75% or more of 14 sessions; core activities were completed 93% of the time. Student stated they strongly agree/agree more than: 80% of the time about educator teaching; 94% of the time about learning about peer responsible relationships; and, more than 89% of the time on learning about sexual activity. Results below are significant at p <.05 or less.  Horizon program positive outcome mean score increases were observed: sexual activity attitudes, knowledge and risk reduction and risk avoidance intentions (6.8% to 9.4%); and positive peer relationship attitude, knowledge, and skills (5.3% to 8.0%).  Horizon group had 6.2% lower prevalence on ‘ever’ engaged in sexual intercourse; lower prevalence on emergence of pre-sexual activity (holding hands, touching above/below waist); and, lower scores on juvenile delinquency risk, 4.1%; substance use, 5.4%; and, peer violence, 7.7%.       

Conclusion: The Horizon program had several positive imeediate outcomes and several long term impacts on 7th grade students. A multi-level approach to evaluation reinforces the value of the program.