Structure and agency in HIV prevention: Multi-level approaches to theory, intervention and research.

摘要

Multi-level interventions are warranted in the prevention of HIV/AIDS globally. This dissertation begins with an overview of theory, interventions, and empirical research that is useful in understanding multi-level or ecological approaches to HIV prevention. Three independent papers using data from three studies provide windows onto several model multi-level HIV prevention modalities. Paper 1 examines the impact of a multi-level 100% Condom Use Program (CUP) replication and Popular Opinion Leader (POL) follow up on condom use reports by male clients of sex workers (CoSW) in the Philippines. Condom availability was more strongly associated with condom use compared to sex work establishment condom rules, but reports of the two policies were highly correlated; results suggest a synergistic effect on condom use. The POL intervention results were mixed for the sub-sample of high risk CoSWs in this analysis. Paper 2 examines multi-level "empowerment" outcomes from a replication of the Durbar (Sonagachi) model Community-Led Structural Intervention (CLSI) with sex workers in West Bengal, India. Outcome indicators included twenty five items from eight domains including: knowledge, self-efficacy, workplace autonomy, finance management and security, mobility and association, and sex worker consciousness and solidarity. Items from all domains had statistically significant intervention effects, as did a summary outcome index. Younger age, self-employed working contract, formal education, and having a live-in male partner were all predictive of empowerment outcomes and/or intervention effects. Paper 3 examines multi-level predictors of Partner Notification (PN) discussion and agreements among HIV-positive test cases reported by providers in Los Angeles County. Discussing PN was common but agreements were rare. Predictors of PN discussion included providers' HIV-test training and knowledge, and client presentation of risk and symptoms. No injection drug using client test-cases agreed to PN. Other predictors of PN agreement included provider or client African-American ethnicity, referral cases to specialists, and test-cases occurring in county public health department operated sexually transmitted disease clinics and mobile units. These studies highlight the interaction of factors operating at multiple levels in the context of HIV prevention. The broad results of the dissertation support multi-level approaches to theory, intervention and research for HIV/AIDS prevention.