The goal of the proposed research is to enhance health status among homeless women by conducting a 2- year longitudinal quasi-experimental study designed to assess the effectiveness of a Peer Mentored Program and a Case Managed Program on subjective and objective measures of risky behaviors, health protecting behaviors, and health outcome of 648 African-American, Latino, and White women and 648 significant others (SOs) at risk for AIDS. Structural equation modeling of a revised theoretical framework to assess its ability to predict risky behaviors, health protecting behaviors, and health outcome will also be conducted using the longitudinal data. The setting, sober living shelters (SLSs), will be selected by randomizing study-eligible SLSs with similar service delivery characteristics into African-American, Latina (English and Spanish Speaking) and white ethnic groups. Eighteen SLSs within each of these ethnic groups will then be randomized into the Peer-Mentored, Case- Managed, or control group. Women and their SOs in the Peer-Mentored and Case-Managed groups will complete baseline questionnaires and will receive the intervention weekly over 6 weeks, and follow-up at 6, 12, and 24 months. The weekly intervention program incorporates information on behaviors that increase risk for AIDS, personal health protecting and maintenance techniques, empowerment in negotiating health protecting behaviors, and assistance with referrals to health and community services. Women and their SOs in the Peer-Mentored Program will receive this content by a peer mentor, assisted by an outreach worker. Women and their SOs in the Case-Managed Program will receive the intervention by a nurse and outreach worker. Women and SOs in the control group will be administered questionnaires similar to subjects in the treatment groups. It is hypothesized that: 1) women in the Peer-Mentored Program will have lower risky behaviors, more health protective behaviors, and improved health outcomes in the posttest measures than the Case-Managed or control group; and, 2) the SOs of women in the Peer-Mentored group will, likewise, have lower risky behaviors, more health protective behaviors, and improved health outcomes than SOs in the Case-Managed or control group.