Behavior Change Interventions to Prevent HIV among Women

Behavior Change Interventions to Prevent HIV among Women Living in Low and Middle Income Countries Sandra I. Mc. Coy, MPH Ph. D 1 R. Abigail Kangwende, MPH MD 2 Nancy S. Padian, MPH Ph. D 1, 3 1 RTI International, San Francisco, California, USA 2 Africa University, Mutare, Zimbabwe 3 Center of Evaluation for Global Action, University of California, Berkeley, California, USA Perspectives on Impact Evaluation Conference ● April 2, 2009 ● Cairo, Egypt

Women and HIV Infection n Globally, women and girls are especially vulnerable to HIV infection n Women represent ~50% of all people with HIV infection, but 59% in Sub-Saharan Africa n Structural and social factors drive transmission in women and decrease women’s ability to avoid HIV infection n Gender inequities ■ Violence n Harmful gender norms ■ Poverty n Lack of education

HIV Prevalence by Age and Sex among 15 -24 year olds, Zimbabwe, 2001 Source: Ministry of Health and Child Welfare (Zimbabwe), Zimbabwe National Family Planning Council, National AIDS Council (Zimbabwe), and U. S. Centers for Disease Control and Prevention. 2004. The Zimbabwe Young Adult Survey 2001 -2002.

HIV Prevention Approaches HIV prevention for women has been discouraging n n Not (yet) effective: female barrier methods, microbicides, and vaccines Male circumcision may not directly lower women’s risk Male and female condoms require male partner knowledge and consent Sexually transmitted disease (STD) services may not be available to all women, esp. those in the developing world

HIV Prevention Approaches HIV prevention for women has been discouraging n n Not (yet) effective: female barrier methods, microbicides, and vaccines Male circumcision may not directly lower women’s risk Male and female condoms require male partner knowledge and consent Sexually transmitted disease (STD) services may not be available to all women, esp. those in the developing world HIV prevention efforts in women (and men) must c to focus on behavior change

Behavior Change for HIV Prevention Behavior change interventions aim to: n n n Delay age of sexual debut Decrease the number of sexual partners or timing of partnerships (e. g. , concurrency) Increase protected sex acts Increase voluntary counseling and testing (VCT) Improve adherence to other successful strategies (e. g. , condoms) Source: Coates, T. Lancet 2008; 372: 669 -84

Behavior Change for HIV Prevention n Focus on individuals, peers, couples, groups, families, institutions, or communities n Need to incite change in enough people for enough time to impact population transmission dynamics n Behavioral change interventions are effective at reducing reported HIV-related risk behaviors in: n MSM (Herbst 2005) n STD clinic patients (Crepaz 2007) n U. S. adolescents (Mullen 2002) n Heterosexual African Americans (Darbes 2008) n People living with HIV (Crepaz 2006)

Impact on HIV, Behavior, and STIs Untangling this (population-level) relationship is critical: n ↓ HIV Intervention ? Behavior change ? ? (how much? ) ↓ Other STIs

Behavioral Intervention Trials n Few studies of behavioral interventions for HIV prevention are evaluated with rigorous designs and with HIV infection as an outcome n Evaluation of HIV incidence is critical: n Reported behavior can be biased n Behavior is inconsistently related to STIs, including HIV n The ultimate objective is to prevent HIV infection

Goal n Systematically review and summarize the effect of behavioral change interventions for HIV prevention in women and girls living in low and middle income countries

Study Inclusion Criteria Dates: 1990 - February 28, 2009 Designs: Randomized controlled trials, quasi-experimental, or prospective designs with a control group Settings: Low- and middle-income countries Outcome: Incident HIV infection Language: No restrictions Grey Literature: Yes (meeting abstracts) Population: Stratified, female-only estimate OR combined estimate where women were at least 50% of population

Study Inclusion Criteria Eligible interventions included, but not limited to: n n n n n Individual or group counseling Conditional and unconditional cash transfers Targeted messages and social marketing campaigns School-based HIV prevention education Voluntary counseling and testing Vocational training Empowerment training Non-cash incentives (like school uniforms) Social support programs Condom promotion Control interventions could be inactive (e. g. , no treatment, waiting list control, or standard of care) or active (e. g. , a diluted or different variant of the intervention being investigated) Per-protocol analyses acceptable if community RCT

Systematic Review Methods Systematic search of: n n Pub. Med/MEDLINE Cochrane Library, including: n n n n n Cochrane Central Register of Controlled Trials (CENTRAL) Database of Abstracts of Reviews of Effects (DARE) Psyc. Info Sociological Abstracts Web of Science African Index Medicus Regional Index for Latin America and the Caribbean: Virtual Health Library Ind. Med NLM Gateway

Systematic Review Methods n Communication with investigators for other studies and unpublished trials n n Current Controlled Trials Register International Clinical Trials Registry Platform Search Portal clinical trials. gov Computer Retrieval of Information on Scientific Projects (CRISP) n Cited Reference Search (Web of Science) of key articles n Reviewed reference lists of included papers and other reviews n Searched recent conference websites n n n Conference on Retroviruses and Opportunistic Infections (CROI) International AIDS Society (IAS) International Society for STD Research (ISSTDR)

3, 864 potentially relevant articles retrieved from electronic databases Literature Search 3, 265 articles excluded after title-level review (e. g. , excluding studies in men, those in the USA, other types of interventions) 599 articles eligible for abstract-level review 551 articles excluded after abstract level review (e. g. , studies without biological outcomes or uncontrolled designs) • 195 studies of behavioral interventions in women and girls without HIV infection as an outcome 48 articles for detailed review 8 met inclusion criteria 3 new eligible articles identified via reference lists 11 articles included in review (8 unique study populations)

Results: 11 Included Studies Author (year) Type Population Follow-up Location Bhave (1995) Quasiexperimental Female sex workers and madams 1 year Mumbai, India Kamali (2003) C-RCT Residents of study communities 3 -4 years Masaka District, Uganda Quigley (2004) C-RCT (PP) Sexually active adults 3 -4 years Masaka District, Uganda Matovu (2005) Cohort Residents of study communities 1 year Rakai District, Uganda Pronyk (2006) C-RCT Residents of study communities 3 years Limpopo Province, South Africa Gregson (2007 a) C-RCT Residents of study communities 3 years Manicaland Province, Zimbabwe Gregson (2007 b) C-RCT (PP) Residents of study communities 3 years Manicaland Province, Zimbabwe Ross (2007) C-RCT Adolescents (years 4 -6 of school) 3 years Mwanza Region, Tanzania Jewkes (2008) C-RCT Residents of study communities 2 years Eastern Cape Province, South Africa Patterson (2008) I-RCT Female sex workers 6 months Tijuana & Ciudad Juarez, Mexico Doyle (2009) C-RCT (LT) Adolescents (years 4 -6 of school) 6 -8 years Mwanza Region, Tanzania C-RCT=Community randomized controlled trial, I-RCT=individual randomized controlled trial, PP=per protocol, LT=Long term

Intervention Characteristics Intervention Target Delivery Duration Group educational and motivational sessions Sex workers G 6 months; 3 -4 groups discussions and 3 -4 video viewing groups Information, education, and communication Adult men and women I, G Activities throughout 3 -4 year study HIV VCT Adult men and women I, C One post-test counseling session IMAGE – microfinance + SFL behavior program Community residents I, G, CM 15 -21 months; loan center mtgs. & 12 -15 month curriculum meetings every 2 weeks Peer education, condoms, Sex workers, adult and STI services residents I, G, CM Activities throughout 3 year study Activities throughout the 3 year period; annually, 12 40 -min school sessions, youth health weeks & days, video shows MEMA kwa Vijana Adolescents Stepping Stones Men and women 15 -26 G, CM 6 -8 weeks; 13 3 -hr single sex groups, 3 peer groups, and a community meeting Mujer Segura Sex workers I 35 -min counseling session G=Group, I=individual, C=couple, CM=community

Impact on HIV incidence * Note: Quigley (2004) and Gregson (2007 b) are per-protocol analyses of C-RCTs described in Kamali (2003) and Gregson (2007 a). Doyle (2009) is the 6 -8 year follow-up of the MKV study described in Ross (2007).

Impact on Behavior and STIs Reduced HIV Incidence? Reduced Risk Behavior? Reduced Other STIs? Bhave (1995) Yes: ↑ Condom use with clients Yes: ↓ Syphilis, HBV Kamali (2003) No No Yes: ↓ HSV-2* Quigley (2004) Yes No NR Matovu (2005) No No NR Pronyk (2006) No No NR Gregson (2007 a) No No NR Gregson (2007 b) No NR NR Ross (2007) No No No Jewkes (2008) No No Possibly ↓ HSV-2** Patterson (2008) No Yes: ↑ Condom use, total protected sex Yes: ↓ “Any STI” Doyle (2009) No Yes: ↑ Condom use with non -regular partner No Author (year) NR=not reported, HBV=Hepatitis B Virus, HSV=Herpes Simplex Virus * No effect found for active syphilis, Chlamydia, or gonorrhea ** Incidence ratio for women: 0. 69, 95% CI (0. 47, 1. 03)

Conclusions Only two of 11 studies in women and girls showed an effect on HIV incidence n 3 of 10 studies reduced any risk behavior n 4 of 6 studies reduced STI incidence (non-HIV) n n Important research and prevention gaps remain n How and whether to incorporate behavioral change programs into existing prevention packages in the absence of clear data on effectiveness

Limitations One study not peer-reviewed (meeting abstract) n Per-protocol analyses included & 1 follow-up n Potential to miss studies if biological outcomes not reported in abstract n Potential to miss studies from other databases n Variable statistical power to detect effects n

Ongoing Behavioral Intervention Trials for HIV Prevention Name Type Location Intervention Results Community C-RCT Popular Opinion Leader (CPOL) China, India, Peru, Russia, Zimbabwe 15% of population trained as CPOLs to diffuse HIV prevention messages Regai Dzive Shiri C-RCT 30 communities in Zimbabwe Youth, community, and clinic 2009 -2010 based prevention intervention for adolescents Project ACCEPT C-RCT 34 communities in South Africa, Tanzania, and Zimbabwe and 14 in Thailand Community-based HIV VCT intervention plus standard clinicbased VCT (SVCT), or SVCT alone 2009 2011

Acknowledgements n International Initiative for Impact Evaluation (3 ie) n Center of Evaluation for Global Action (CEGA), University of California, Berkeley n Women’s Global Health Imperative (RTI International)

Questions?

HIV Prevalence by Age and Sex among 15 -24 year olds, South Africa, 2003 Source: Pettifor, A. AIDS 2005; 19: 1525 -34.

Data Abstraction n Data were abstracted from each eligible study by one investigator (SM) and reviewed for accuracy n Most adjusted measure of effect on HIV incidence (e. g. incidence ratio or risk ratio) in women n Overall incidence ratio if women-only not presented n Measure of effect and/or 95% CI calculated if necessary Trial years, location, population, intervention n Impact on knowledge, behavior and other sexually transmitted infections, when available, in women n

Why don’t behavioral interventions translate to reductions in HIV incidence? n Greater than observed behavioral change required n Follow-up times too short n Behavior changes misreported n Traditionally “low-risk” women may be at high-risk independent of their personal behavior n We are measuring the wrong behavioral intermediates n n n Sexual networks or concurrency, or Which behavior is important depends on the STI in question Structural factors (e. g. , poverty) further up the causal chain drive transmission and individual behavior

Impact on HIV, Behavior, and STIs n Despite variability in the intensity, delivery, and target populations, only 2 interventions reduced HIV incidence n 6 -month program for female sex workers in India n↑ n condom use with clients, ↓ syphilis and HBV infection Per-protocol analysis of sexually active women in Uganda who attended at least one intervention activity n No reduction in risk behavior, STIs not measured

Intervention Characteristics Intervention Study(s) Target Group educational and motivational sessions Bhave (1995) Sex workers Group 6 months; 3 -4 video viewing groups, 3 -4 group discussions Information, education, and communication Kamali (2003) Quigley (2004) Adult men and women Individual and group Activities implemented throughout the 3 -4 year study period Voluntary HIV counseling Matovu (2005) and testing Adult men and women Individual and couples One session IMAGE – microfinance + behavior program Community residents Individual, group, and community 15 -21 months; loan center mtgs. and 12 -15 month curriculum meetings every 2 weeks Peer education, condoms, Gregson (2007 a) Sex workers, and STI services Gregson (2007 b) adult residents Individual, group, and community Activities implemented throughout the 3 year study period MEMA kwa Vijana Ross (2007) Doyle (2009) Adolescents Individual, group, and community Activities implemented throughout the 3 year period; annually, 12 40 min school sessions, youth health weeks & days, and video shows Stepping Stones Jewkes (2008) Men and women 15 -26 Group and community 6 -8 weeks; 13 3 -hr single sex groups, 3 meetings of peer groups, and a community meeting Mujer Segura Patterson (2008) Sex workers Individual 35 -min counseling session Pronyk (2006) Delivery Duration