• Cohort I: 12 month experience of intimate partner violence
• Cohort II: Condom use with last non-spousal partner
• Cohort III: Condom use with last non-spousal partner
• Cohort III: HIV incidence

The IMAGE Study is a cluster randomised trial of a structural intervention for the prevention of HIV and gender based violence being conducted in South Africa.

Background : While there is a high level of knowledge on HIV/AIDS in rural South Africa, structural factors, including poverty and gender inequalities, have critically impeded prevention efforts based on modifying individual risk. Attempts to rigorously evaluate structural interventions for the prevention of HIV infection in sub-Saharan Africa remain under-developed.

Microfinance is a well-established development tool with nearly 100 million clients world wide. Loans to poor households, particularly to women, facilitate business development opportunities. Prior research has demonstrated significant economic and social benefits. Health benefits have also been documented and include improvements in immunization coverage, better nutrition, and contraceptive use. While poverty and HIV/AIDS frequently co-exist, the impact of microfinance on vulnerability to HIV has not previously been explored.

Intervention Design : The IMAGE intervention combines two components:

1. Microfinance: Implementation of a community-level, poverty-focused microfinance programme exclusively targeting women
2. Gender and HIV training: A two phase participatory learning and action curriculum, has been integrated into bi-weekly microfinance meetings. The first phase consists of 10 structured sessions facilitated by trainers, while the second phase facilitates community mobilsation activities from within the microfinance centre.

Study Design

The IMAGE Study randomised four pairs of communities that were matched on size and accessibility. The intervention communities received the IMAGE intervention from late 2001. Comparison communities will receive the intervention strategy approximately three years later.

The study enrols individuals into three pairs of cohorts (intervention and comparison). Data is collected on the major health-related endpoints as detailed below. Qualitative and quantitative data are also collected to document processes of change and potential pathways of impact. Data analysis will be appropriate to the cluster-randomised study design.

Cohort I : Women recruited to IMAGE intervention and Randomly selected women eligible to join IMAGE from control villages. Sample size per arm - 450. Length of follow up - 2 years. Primary outcomes: 12 month experience of intimate partner violence

Cohort II : Young people aged 14-35 living in the households of women in cohort I (Intervention and Comparison). Project sample size per arm - 750. Duration of follow up - 2 years. Primary outcomes - Condom use with last non-spousal partner

Cohort III : Randomly selected young people aged 14-35 living in Intervention communities and Comparison communities. Project sample size per arm - 1500. Duration of follow up - 3 years. Primary outcomes - Condom use with last non-spousal partner, HIV incidence

• HIV Infections
• Gender Based Violence

Inclusion Criteria:

• Cohort I (Intervention community) : Joining the IMAGE programme
• Cohort I (Comparison community) : Female, Eligible to receive microfinance loan and residing in an IMAGE Comparison community, matched on community pair and age group (18-25, 26-35, 36-45, 46-55, 56+) to the index IMAGE participant.
• Cohort II (Intervention community) : 14-35 years old at time of interview. Current residence in the household of a woman who joins the IMAGE programme.
• Cohort II (Comparison community) : 14-35 years old at time of interview. Current residence in the household of the non-IMAGE participant selected for inclusion in Cohort I.
• Cohort III (Intervention community) : 14-35 years old at time of interview, Household member in a community where IMAGE is operating.
• Cohort III (Comparison community) : 14-35 years old at time of interview. Household member in an IMAGE study community where IMAGE is not operating.

Exclusion Criteria:

• Lack of informed consent

• University of Witwatersrand, South Africa
• Small Enterprise Foundation, Tzaneen, South Africa
• National Department of Health, South Africa
• Henry J.Kaiser Family Foundation, USA
• Ford Foundation
• Department for International Development, United Kingdom