Structural economic interventions to address women’s vulnerability to HIV infection in sub-Saharan Africa Charlotte Watts Ph.D. Head, Social and Mathematical Epidemiology Group Research Director, Structural Interventions Research Programme Consortium Department of Global Health and Development London School of Hygiene and Tropical Medicine

Challenge #1: Gender disparities in rates of heterosexual HIV transmission Source: UNAIDS Report on the Global AIDS Epidemic

Challenge #2: Inequalities in relationship power & partner violence a risk factor for incident HIV infection Source: Jewkes et al The Lancet 2010

Challenge #3: Benefits of HIV prevention least accessed by the poor Source: Hargreaves & Howe AIDS 2011 Adjusted Odds of difference in HIV prevalence between and in Tanzania AOR95% CI No education – 1.47 Stable HIV prevalence Primary education – 1.03 Small decline in HIV prevalence Secondary education – 0.84 Larger decline in HIV prevalence

Effective HIV prevention for women needs to address poverty, gender inequality & violence Effective prevention Prevention technologies Poverty Social power Brady, Martha. Population Council, Conceptual Framework Interventions that address women’s economic and social vulnerability Women controlled HIV prevention methods

What role could different forms of economic intervention play? CCT interventions Incentivise increased service use HIV testing or STI treatment services Enrolment in education – knowledge & self-efficacy, delay onset of sex, change sexual networks, increase opportunity cost of pregnancy Reward safer sex behaviours Reduce poverty, and on levels of sex in exchange for resources Economic development / empowerment programmes, eg micro-finance Reduce poverty, so potentially impacting on: Ability to negotiate sex Ability to address violence Need to engage in sex in exchange for resources

Conflicting perspectives on CCTs Potential concerns Negative viewPositive view Social engineering? paying people to act against their wishes offered to achieve outcomes most people desire Paternalistic?undermines individual autonomy facilitates autonomy when makes it more likely that people act in line with preferences Unfair?selective poverty alleviationpotentially potent means of changing behaviour in most socially deprived, reducing health inequalities Inefficient?poor use of resources when many competing demands potentially large health benefits from modest investment Disempowering?top down model of intervention financial benefits lead to greater empowerment Sustainability?continued funding requirements, creating dependency prevention reduces downstream costs Marteau, Ashcroft & Oliver, 2009

Population based CCTs to increase demand to learn or maintain HIV -ve status InterventionLocationDesignResults CCT to learn HIV status after testing 2,812 women and men in rural Malawi RCT – individuals randomly assigned voucher payments (between $0 and $3) to be paid when test results made available Without any incentive, 34% learned HIV results. Even smallest incentive doubled numbers returning Positive linear effect with level of incentive CCT to maintain HIV status 1,300 women and men in rural Malawi RCT –voucher amounts for maintenance of HIV- status for 1 year Rewards ranged from zero to 4 months wage No effect on HIV status or reported sexual behavior during study Following intervention, men with incentive 8.5% more likely and women 7.5% less likely to engage in risky sex

Schooling, Income, and HIV Risk (SIHR) intervention to encourage girls to stay in or return to school InterventionLocationDesignFindings after 18 mths Zomba cash transfer program that provides CCT to current school girls and recent drop outs - average $10/month for 10 months) + payment of school fees - 30% payment to girls, 70% parents Zomba, Malawi 3,796 unmarried girls aged 13 – 22 from 176 enumeration areas CRCT. Randomised payment amount to parents ($4-10) and girls ($1-5) Some offers conditional, some unconditional ITT analysis HIV prevalence 60% lower than control (1.2% vs. 3.0%). o HSV-2 prevalence 75% lower (0.7% vs. 3.0%). o No significant differences between those offered conditional and unconditional payments o Girls reported smaller numbers of partners, and less transactional sex Baird, Chairwa, McIntost, Ozler, 2009,

Alternative economic interventions: Intervention for Micro-finance And Gender Equity (IMAGE intervention) InterventionStudy populationDesignFindings Combined micro-finance programme + participatory training in gender, violence and HIV (Sisters for Life) + community mobilisation Limpopo Province, South Africa 860 women enrolled 1,750 loans disbursed Total value USD $ Repayment rates 99.7% CRCT 4 intervention & 4 control communities Primary outcomes Past year experience of intimate partner violence by recipients Incidence of HIV among adolescents in community Among participants  Past year experience of violence reduced by 55%  Significant reductions in hh poverty  Improved HIV communication  $43 per client in trial / $13 per client at scale up) Among younger women participants (18 – 35) 64% higher uptake HIV testing 25% reported unprotected sex Among adolescents in community (18 – 35) No impact on HIV incidence over 2 years Pronyk et al The Lancet 2004, Pronyk et al AIDS 2008

* All aRR for indicators represented as bar graphs on a logarithmic scale Source: Kim et al 2008 WHO Bulletin Evidence that training essential to impact Indicators of household economic wellbeing Indicators of women’s empowerment Attitudes and past year exposures to violence HIV risk behaviours MF vs Control Attitudes condoning IPV Past year experience of controlling behaviour Past year experience of violence Household communication Collective action Condom use with non-spousal partners HIV risk behaviours IMAGE vs Control IMAGE vs MF HIV RISK BEHAVIOURS VIOLENCE INDICATORS EMPOWERMENT INDICATORS ECONOMIC INDICATORS Microfinance vs. Control IMAGE vs. Control IMAGE vs. Microfinance

households 4500 households (30,000) households (80,000) Ongoing scale-up of intervention alongside roll out of micro-finance Pilot Study Additional cost = US $43/client Scale-up Additional cost = US $13/client Expansion Additional cost = US $??/client

Conclusions  Promising findings from both CCT and micro-finance interventions  Improved uptake of services  Reductions in reported risk behaviours  Impacts on HIV and HSV-2  Evidence that the HIV impact of livelihood programmes enhanced by added training components  Importance of conditionality element of intervention varies between studies:  Evidence of influence on uptake of services  Less evidence of impact on rates of partner chance / maintaining HIV risk  Limited evidence on the costs of implemented conditionality

How do these structural interventions fit into the new AIDS investment framework? Source: Swartlander et al Lancet 2011

Investment needed to realize synergies between HIV & development programmes Basic HIV programme activities Critical enablers Development programmes School and livelihood programmes provide important infrastructures upon which HIV interventions can build Systems of social protection that explicitly seek to achieve HIV benefits important May be possible to achieve synergies at relatively low incremental cost Cross-sectoral communication & collaboration essential to make prevention a reality