There are currently no published studies specifically documenting the efficacy of HIV prevention interventions with Hispanic adolescents. This study proposes to test the efficacy of 2 HIV prevention interventions with an urban poor, recent immigrant Hispanic population. All interventions have been constructed to build on Hispanic cultural values by validating the central role of the family as protector and change agent of its youth. 300 families with an eighth grade adolescent will be randomized into 3 conditions: Experimental Condition A = HEPI to improve parenting + PATH for HIV prevention; Experimental Condition B = English classes + PATH for HIV prevention; Condition C = English classes + a cardiovascular prevention intervention. English classes are used as a control for nonspecific factor of the parenting intervention, and the cardiovascular intervention for control of nonspecific factors in HIV prevention. The proposed study investigates if providing HIV prevention in the context of a parenting intervention significantly increases the efficacy of the HIV prevention intervention. The parenting intervention has been shown to significantly increase/improve parental investment, parent-adolescent communication, parental monitoring skills, problem behaviors, and alcohol use in recent immigrant Hispanic adolescents. The HIV prevention intervention has been shown to significantly positively influence parent-adolescent communication about HIV, increase parents' and adolescents' knowledge about HIV, and increase adolescents' intentions to use condoms. Dosage has been equalized across interventions and control. Stratified urn randomization along 7 variables will be used. Efficacy will be assessed in relation to: (H1) ultimate outcomes of risky sexual behavior and drug use; (H2) ecodevelopmental mediators (parental investment, parent-adolescent communication, parental monitoring); and, (H3) social cognitive mediators (attitudes, subjective norms, perceived control, and behavioral intentions). Multi-method, multi-reporter assessments will allow the construction of latent constructs using SEM. HLM will be used for growth curve analysis of the three conditions over the 5 time points (Baseline, 6, 12, 24, and 36 months post baseline).