Bridging Disciplinary Boundaries (January 11 - 14, 2007)

This paper presents findings from Gateway-Longview's Families United program that provided preventive interventions to reduce the incidence of substance abuse and sexually transmitted infections, such as HIV/AIDS and hepatitis, among pregnant and parenting youth ages 9 – 19. Additional outcomes included increase in family interaction, improvement in parenting and knowledge of child development, increase in social support and advocacy skills, and participation in educational or vocational services. Families United offered a continuum of care, including home and community-based mentoring and case management, education and prevention activities, referral services, social and emotional support, recreational activities and transportation. Services were offered to pregnant and parenting adolescents through Family Mentors recruited from the neighborhoods served by the program and trained in counseling and case management skills.

Methodology

The main method was pretest posttest design using the above outcomes based on the self-report GPRA as well as repeated risk assessments from the mentors' perspective. The data analysis was based on realist evaluation (Kazi, 2003) to investigate what interventions work and in what circumstances, including the use of binary logistic regression (BLR) models to control for the effects of key covariates in relation to each outcome. The BLR was selected as the variables that are actually influencing the outcome are retained in the equation and can provide a prediction of the intervention achieving the outcome where the significant factor(s) may or may not be present.

Findings

The results were positive for the 65 teenage mothers included in the sample over a three-year period. For example, family interactions improved in 69%, and this outcome was correlated with lower risky behaviors of peers assessed at baseline, e.g. girlfriends' use of alcohol and drugs, and the number of friends that get suspended from school. Binary logistic regression indicated that those who did not have friends suspended at school were 8 times more likely to improve than those who did. It was also found that 69% improvement their knowledge of parenting and child development, and this improvement was significantly correlated with the baseline stage of change assessment for sexual risk behavior. Clients who were assessed to be at advanced stages of change for this target behavior (Ready for Action, Action, Maintenance) improved at higher rates than those at lower stages (Pre-contemplative, Contemplative). Clients at these higher stages displayed less high-risk behavior and/or more involvement in mutually monogamous relationships, consistent use of safe sex, or were sexually inactive. The combination of readiness for change and relationship status provided a supportive context for improvement in this outcome.

Implications for practice

The results indicate that the use of trained mentors in HIV/AIDS prevention programs can help to reduce risk factors and increase protective factors for high-risk pregnant and parenting adolescents. This study also demonstrates that advanced non-parametric methods such as binary logistic regression can help programs to investigate the circumstances in which particular components of the intervention may or may not be effective, and help to better target the programs to meet the needs of clients.