1. Maximizing Impact and Return on Investments in STD/HIV Prevention Programs David Wilson, Nicole Fraser, Marelize Gorgens and Zukhra Shaabdullaeva Global HIV/AIDS Program The World Bank 15 July, 2013 IUSTI Vienna

2. Estimated national HIV incidence encouraging Measured sub- national HIV incidence worrying The end of AIDS?

3. Estimated national incidence declines 2001 - 2011  Estimated national HIV incidence fell by 20%  39 countries (23 African) - declines > 25% UNAIDS, 2012

4. Concentrated epidemics – MSM Why worry? Proven interventions? Achieve with full implementation? Actual implementation? Investing smarter and implementing better?

5. Concentrated epidemics – MSM Why worry – MSM HIV rates 13.5-fold higher 20% UNAIDS, 2012

6. Concentrated epidemics – MSM Proven interventions? Despite developed world successes, few developing country MSM programs have demonstrably reduced HIV incidence PREP reduced HIV among MSM by 44% (90% among fully adherent) but we don’t even reach MSM in most developing countries with information and condoms In developing countries, scarcely know how to reach hidden MSM, reduce stigma, deliver at scale and change policy Still need to navigate between southern unwillingness to address male-male sexuality and northern temptation to frame response within western constructs

7. Concentrated epidemics – MSM Implementation and coverage limited 70% very low or don’t report 90% of funding international - 19-21 LMIC reliant on external funding 70% very low or no report UNAIDS, 2012

8. Concentrated epidemics – MSM Few reached by HIV prevention services

9. Concentrated epidemics – IDU Implementation and coverage limited 86% low coverage or no report UNAIDS, 2012

10. Concentrated epidemics – IDU Low access to basic services 90% 92% 85% Million IHRA, 2012

11. Generalized epidemics - VMMC Actual implementation is seriously off-target PEPFAR, 2013 Total

12. Generalized epidemics - TAsP Do we have proven approaches? TAsP clinical trial efficacy 96%+ TAsP real world effectiveness lower? – Infection 34% lower in area with 30%-40% ART coverage (the effect saturation point) than area with <10% coverage in KZN (Tanser et al, 2013) – Infection 26% lower in discordant couples in China - for transfusion or sexually infected but not IDU infected indexes (Jia, 2012) – No difference in discordant couples in Uganda (Birungi et al. 2013) – HIV infections continue to rise in highly treated MSM communities in developed countries (Wilson et al, 2012) – With ~85% on ART at CD350, Swaziland has measured HIV incidence of 2.4% on top of 26% adult prevalence (SHIMS, 2013)

13. Generalized epidemics - TAsP Do we have proven approaches? HPTN 071 (PopART) TasPBotswana/ HSPH SEARCH SitesLusaka, CTownSouth AfricaMochudi/BotsKenya, Uganda DesignCluster RCT 24 @ 55,000 3 arm Cluster RCT 34 @ 1,250 2 arm Paired cluster RCT, 30 @ 5,000, 2 arm Paired cluster RCT, 32 cl @ 10,000, 2 arm InterventionImmediate ART if HIV+ HCT, VMMC, condom, risk reduction counselling Immediate ART if HIV+ HCT home-based ART for CD4 10,000 HCT, VMMC, PMTCT-B Immediate ART if HIV+ Combination HIV prevention package Outcome2 y HIV incidence in cohort Cumulat 2 yr HIV incidence in cohort Cumulat HIV incidence 3 + 5 yrs, cross-sectional

14. CD4 ≤ 200 CD4 ≤ 350 + TB/HIV HBV/HIV CD4 ≤ 350 + TB/HIV HBV/HIV + CD4 ≤ 500 “Test and treat” All HIV+ 1 2 34 5 ART regardless of CD4 count for:  HIV-SD couples  Pregnant women ART regardless of CD4 count for:  HIV-SD couples  Pregnant women + TB/HIV HBV/HIV SD couples Pregnant Children < 5 Apollo et al, 2013 11 million 17 million 21 million 26 million 32 million 11 million 17 million 21 million 26 million 32 million Recommended since 2010 Recommended until 2010 9.7 million on ART - 26 million eligible at CD500 and 32 million eligible for “test and treat” Generalized epidemics - TAsP What does full implementation look like?

15. Generalized epidemics - ART What is actual implementation like? US treatment cascade - 28% virally suppressed

16. Generalized epidemics – Financial incentives Do we have proven approaches? Three World Bank RCTs show financial incentives reduce STI/HIV transmission – In Tanzania, people offered up to $60 each annually to stay STI- free had 25 percent lower STI prevalence (De Walque et al 2012) – In Malawi, girls and parents offered up to $15 monthly to stay in school had 60% lower HIV prevalence - whether they stayed in school or not (Ozler et al, 2012) – In Lesotho, adolescents offered a lottery ticket to win up to $50 or $100 every four months if they stayed STI and HIV-free had a 25% lower HIV incidence - 33% lower among girls and 31% in the $100 arm (De Walque et al 2012)

17. TanzaniaMalawiLesotho SitesIfakara, TanzaniaZomba district, Malawi5 rural/periurban districts DesignRCT, 2399 adults 18-30 yrs, 3 arms Cluster RCT, 1,289 never-married females 13-22 yrs in 176 EAs, 3 arms RCT, 3,426 adults 18-32 yrs, 3 arms Intervention and incentive All: STI testing and treatment every 4 mths for 1 yr Low value and higher value CCT for those STI free CCT (for school attendance) and UCT payment (no school attendance required) CCTs and UCTs to student and parent, randomised amounts ($1-10) All: STI testing, counselling and treatment every 4 mths for 2 years High and low lottery tickets for those free of curable STIs EndpointCombined prevalence of 4 STIs (HIV, HSV-2 & syphilis secondary endpoints) Prevalence of HIV and HSV-2 at 18 mths HIV incidence Generalized epidemics – Financial incentives Do we have proven approaches?

19. Watts, 2013 60% reduction in HIV risk

20. Post-MDG High Level Panel Report Word Cloud Wagstaff, 2013

21. Conclusion At best flat financing and competing priorities With existing implementation and rate of expansion, we are likely to prevent less than half of new infections in next decade With full implementation of interventions, we could prevent over half of new infections Bridging the gap between full and actual implementation means moving from age of advocacy to era of implementation - intelligent, inquiring, critical, tailored, targeted, painstaking implementation Ending AIDS requires more and better tools