Sub-Saharan Africa remains the epicenter of the HIV pandemic, accounting for 59% of new infections in 2021 (UNAIDS, 2022a). In sub-Saharan Africa, 63% of new HIV infections occur among girls and women, demonstrating an overall disparity in risk compared to males (UNAIDS, 2022a). The difference is even more striking for 15–19-year-old youth, given that AGYW accounted for six in seven new HIV infections (UNAIDS, 2022b). While there have been significant reductions in HIV prevalence and incidence in the general population as well as achieving the 90/90/90 goals (UNAIDS, 2014, 2015), targeted strategies for HIV prevention among AGYW are essential to control epidemics (Brown et al., 2018).

The U.S. government’s President’s Emergency Plan for AIDS Relief’s (PEPFAR) Determined, Resilient, Empowered, AIDS-Free, Mentored, and Safe (DREAMS) program addresses the disproportionate incidence of HIV infection among AGYW through evidence-based HIV prevention programming targeting the biological, social, structural, and environmental factors that increase risk of future HIV infection among AGYW (Saul et al., 2018). The DREAMS approach involves layered interventions, including HIV and violence prevention programming, postviolence care services, HIV testing, preexposure prophylaxis, sexual and reproductive health services, parenting/caregiver support, community norms change interventions, and male partner characterization. The program also aims to address economic disparities that predispose AGYW to risk of HIV infection through socioeconomic strengthening approaches that combine financial literacy, entrepreneurship, vocational training, opportunities for internships, and start-up support for entrepreneur activities to vulnerable AGYW. These interventions have been shown to reduce HIV risk behaviors, unwanted pregnancies, exposure to violence, and HIV and violence-related outcomes, such as poor mental health. An analysis in Lesotho found a greater reduction in new HIV diagnoses in antenatal care clinics in DREAMS districts compared to non-DREAMS districts, demonstrating the potential impact of DREAMS (Pelletier et al., 2022). DREAMS targets a key demographic with an aim to improve health equity in HIV service provision and epidemic control through prevention programming focused on the unique structural, cultural, and biological risks that AGYW face, paired with care and treatment for those diagnosed with HIV at intake or seroconvert (Saul et al., 2018).

PEPFAR implements DREAMS in geographic areas in 16 countries that have a high burden of HIV.¹ Participants receive a standard package of primary interventions and services as well as needs-based supplemental services. PEPFAR country programs tailor DREAMS implementation, including the geographic locations in which DREAMS is offered, to address the unique HIV epidemiology among AGYW in the specific country context (PEPFAR Solutions, 2021). In all countries, enrollment requires a screening assessment, based on standardized criteria empirically associated with risk of HIV infection, to determine DREAMS eligibility. In working to reduce risk factors, DREAMS aims to decrease vulnerability to HIV among program participants. DREAMS eligibility criteria include early pregnancy (Christofides et al., 2014); sexual, physical, and emotional violence (Li et al., 2014); sexual risk behaviors (early sexual debut, multiple sex partners, sexually transmitted infections, transactional sex, and no or irregular condom use); and other social risk factors for HIV, including alcohol misuse, being out of school, and orphanhood (Balkus et al., 2016; Santelli et al., 2013) (Table 1). AGYW living with HIV are not excluded from participation in DREAMS; however, those with an unknown status are offered HIV testing and enrolled in care and treatment as needed.

The purpose of the current study was to assess population estimates of DREAMS eligibility criteria using Violence Against Children and Youth Survey (VACS) data from seven sub-Saharan African countries. Although no previous work has examined population estimates of DREAMS program eligibility across multiple countries, such data could provide useful information for program planning to guide efforts to reach the AGYW most at risk for HIV. First, estimates of the presence and prevalence of risk factors highlight which factors are the most substantial drivers of vulnerability, both within a country and at the regional level, to inform prioritization of DREAMS and similar HIV prevention programs among AGYW programming and recruitment efforts. Second, program eligibility estimates can ensure proper alignment and allocation of staffing and other resources toward specific HIV prevention interventions. Third, estimates can inform efforts to tailor interventions to the specific needs of the population a program is intended to serve. For example, a country with a high prevalence of early pregnancy might warrant different prioritization of programming than a country with a high prevalence of alcohol misuse and low rates of early pregnancy. Finally, eligibility estimates can guide outreach and recruitment efforts to ensure efficient use of program resources to reach the eligible population and to promote equity in reach, access, and uptake of services.

This is a descriptive analysis of population-level DREAMS eligibility among AGYW in seven countries using VACS data from Uganda (2015), Zimbabwe (2017), Côte d’Ivoire (2018), Lesotho (2018), Kenya (2019), Namibia (2019), and Mozambique (2019). VACS are interviewer-administered cross-sectional household cluster surveys among youth aged 13–24 years old that are conducted through the leadership of national government institutions with technical assistance from the U.S. Centers for Disease Control and Prevention (CDC), financial support from PEPFAR, and in partnership with UNICEF and multisectoral committees with multiple ministries and key stakeholders. All VACS in this analysis included HIV testing except Uganda, where HIV status was measured only through self-disclosure on the questionnaire. For the six countries that included HIV testing, the age of testing differed based on the national policy at the time of the survey defining when a young person can receive results confidentially.² In each VACS, a three-stage survey sample design was used. The first stage of sampling was selection of the enumeration areas (EAs) using population proportionate to size according to the national frame. The second stage was selection of households within each enumeration area. The third stage consisted of random selection of one eligible household member for survey participation (Nguyen et al., 2019).

In adherence to best practices in protecting participants who may disclose violence in the field (CDC, 2017), a split-sample approach was used, and males and females were interviewed in different EAs. Participants who had experienced recent violence or exploitation received referrals for psychosocial support. Individuals who had tested positive for HIV also received direct referrals to HIV care and treatment services. Each VACS protocol was approved by the CDC Institutional Review Board and one or more country-level ethics committees.

Table 2 provides prevalence estimates and 95% CIs by country for each age group for overall DREAMS eligibility as well as for each of the DREAMS criteria.

Overall, the percentage of HIV-negative AGYW in each country who met eligibility criteria for DREAMS is very high, ranging from 47.0% in Zimbabwe to 81.4% in Uganda among 13–14-year-olds, from 52.1% in Kenya to 78.1% in Mozambique among 15–19-year-olds, and from 28.1% in Zimbabwe to 76.6% in Uganda among 20–24-year-olds. This high prevalence of DREAMS eligibility suggests that PEPFAR overall, DREAMS country programs individually, and other key stakeholders such as country government programs providing HIV prevention programming for AGYW may consider data-driven approaches to determine eligibility targets based on factors that identify the populations and groups most vulnerable to new HIV infection among AGYW who are socially vulnerable. At the same time, programs may consider if, when, and where population-level prevalence of HIV risk factors indicate that community-level interventions are needed to address widespread social and economic vulnerability and harmful norms to complement the current DREAMS package of interventions to fully address AGYW risk.

The prevalence of having two or more risk factors ranged from 11.3% among girls ages 13–14 years in Zimbabwe to 43.7% in Côte d’Ivoire. Among girls ages 15–19 years, having two or more risk factors ranged from 19.2% in Kenya to 44.5% in Mozambique and from 6.4% in Zimbabwe to 34.1% in Uganda for young women ages 20–24 years. That a substantial percentage of AGYW in all countries and across all age groups have multiple HIV risk factors emphasizes the need for the multicomponent interventions targeting multiple risk factors offered in DREAMS. Furthermore, it highlights the need for the incorporation of HIV prevention interventions into sustainable national initiatives within health, child protection, and education systems. The high prevalence of DREAMS eligibility also suggests the potential need for empirical approaches to assess patterns of co-occurrence of risk factors and HIV acquisition among AGYW while taking into consideration the differences in actual risk that AGYW with one risk factor may face in a locality with high HIV prevalence and low viral load suppression compared with AGYW with multiple risk factors in areas of lower HIV burden.

Experiencing physical or emotional violence in the past year represented the most common reason for DREAMS eligibility for girls ages 13–14 years in all countries examined except Zimbabwe. Physical and emotional violence is associated with negative mental and physical health impacts, including increased risk for HIV (Leddy et al., 2022; Shamu et al., 2019). Given the high prevalence of physical and emotional violence and high endorsement of norms and attitudes supportive of violence (Ministry of Women, Family and Children of Côte d’Ivoire et al., 2019; Ministry of Labour and Social Protection of Kenya, 2019; Ministry of Social Development of Lesotho et al., 2020), HIV prevention efforts for AGYW should consider expanding interventions to address structural, economic, and social drivers of violence at the community and societal levels. This approach may in turn result in reduced risk for HIV and create conditions that enable the DREAMS program to more effectively reach AGYW at the highest risk of HIV.

Lifetime sexual violence was the most common DREAMS eligibility criterion among AGYW ages 20–24 in Uganda (50.4%), Zimbabwe, Côte d’Ivoire, Lesotho, and Kenya while it was the second most common (after no or irregular condom use) in Namibia and Mozambique. Sexual violence was also common in the younger age bands (23.2% among 13–14-year-olds and 46.9% among 15–19-year-olds in Uganda). Unsurprisingly, lifetime exposure to sexual violence increased with age in all countries. Exposure to physical or emotional violence in the past 12 months, on the other hand, generally decreased with age, likely due to decreased physical punishment by parents, caregivers and teachers as well as decreased exposure to violence from peers. Although 15–24-year-old AGYW are not eligible for DREAMS due to experiences of emotional and physical violence, exposure remains as high as 50.7% among 15–19-year-olds and 49.2% among 20–24-year-olds in Kenya. Screening young women for recent sexual violence (to gauge ongoing direct risk vs. historical exposure and indirect risk) as well as screening for intimate partner physical and emotional violence could be worth considering given (a) the persistent high prevalence of emotional and physical violence in the past month among 15–24-year-old AGYW; (b) that population-level HIV incidence for girls and women generally increases during adolescence, peaking between the ages of 20 and 24 years (Birdthistle et al., 2019); (c) known associations between intimate partner violence and low condom use and HIV acquisition; and (d) the demonstrable impact of DREAMS gender-based violence programming on lowering HIV risks, including sexual violence, transactional sex, and condom use (Mathur et al., 2022).

No or irregular condom use was also a commonly met DREAMS eligibility criterion in all countries examined in all age groups among those who had ever had sex. DREAMS program planning could consider no or irregular condom use in combination with other sexual risk behaviors among adult AGYW and seek to understand the ecological factors that impact condom use. Considering that the prevalence of no or irregular condom use (as high as 58.2% among 15–19-year-old AGYW in Uganda) far outpaces the prevalence of multiple sexual partners in the past 12 months, it is reasonable to assume that most adult AGYW who report low or irregular condom use are in a monogamous relationship and therefore at low risk of HIV compared to those who report no or irregular condom use and multiple sexual partners. In the oldest age group, young women, whether married or not, may also be trying to get pregnant (UNICEF, 2021). Following that logic, and absent any transactional sexual relationship, low or no condom use may not be a strong marker of HIV risk in adult AGYW and may therefore not represent a strong stand-alone DREAMS criterion. Thus, it may be worth considering an approach that accounts for a combination of eligibility criteria, considering no or irregular condom use in the context of the number of recent sexual partners and whether any of the sexual relationships were transactional in nature. It is worth noting, for those AGYW who are having riskier sex (i.e., nonmonogamous sex without a condom), that condom use has been found to increase among DREAMS participants with school support and complementary programming with male partners (Mathur et al., 2022). Past research has also highlighted the interconnectedness of multiple risk factors, including findings that girls’ experiences of early sex and reduced agency to negotiate condom use increased with their vulnerability scores (Underwood & Schwandt, 2016).

While the current data show strong trends across countries and age groups, they also highlight differences between countries that suggest tailoring country programs’ intervention offerings and targets toward the most contextually important risk factors. For example, more than one in five girls ages 13–14 years in Mozambique had ever had sex. The prevalence of early sexual debut in Mozambique is nearly twice that of the next highest country, and the prevalence of ever had sex rises to more than two in three girls ages 15–19 years. The prevalence of early pregnancy is also high in Mozambique, where 3.5% of girls ages 13–14 had ever been pregnant, compared to all other countries examined in which 1% or fewer had ever been pregnant among the youngest AGYW. The Mozambique program could use these data to help target the youngest girls who have already experienced or are at risk of early sexual debut and pregnancy. In the same way, programs in Côte d’Ivoire and Namibia may use these data to ensure that programs are addressing the high prevalence of alcohol misuse among AGYW, while programs in contexts of lower occurrence of alcohol misuse may focus on other risk factors.

This study focuses on DREAMS eligibility criteria at the national level across seven countries to provide a high-level perspective on AGYW risk. However, VACS sample design has included oversampling of AGYW in HIV high-burden, DREAMS priority areas in several countries, including in five of the seven countries in this analysis. A supplementary table, Table 3, is included to provide overall DREAMS eligibility in the DREAMS subregions for the five countries with subnational sampling (Uganda, Zimbabwe, Lesotho, Namibia, and Mozambique). As an example, in Zimbabwe, the overall DREAMS eligibility at the national level is 28.1% for 20–24-year-old AGYW whereas the eligibility ranges from 17.2% in Chipinge to 40.4% in Bulawayo. Future studies could focus on country-specific examination of DREAMS eligibility by subnational strata, taking into consideration the HIV burden, including population estimates of HIV prevalence, incidence, and viral load suppression, and considering which HIV risk factors may put AGYW at greatest risk in that context. Optimal programming may be different in rural Chipinge, Zimbabwe, compared with urban Bulawayo based on the specific risk factors that put AGYW at risk of HIV in the local context.

While these data provide new insights into the population-level drivers of DREAMS eligibility and HIV risk, additional research is needed to inform identification of specific combinations of risk factors. Future studies could leverage person-driven approaches such as Latent Class Analysis (Lanza et al., 2013) to identify patterns of heterogeneity within and between groups of AGYW. Improved data on the complex interplay among risk factors for HIV acquisition in AGYW, including the frequency, severity, and co-occurrence of risks as well as the cultural and geographic context in which AGYW live, might also allow for better prediction and prevention of potential future HIV acquisition.

Some DREAMS programs have already incorporated strategic approaches to assess AGYW’s screening results for eligibility. In Malawi, for example, AGYW are automatically eligible if they have certain specific risk factors (e.g., 10–14-year-olds who have ever had sex, experienced sexual violence, or contracted an STI), while for other risk factors eligibility is based on an overall weighted score rather than on a single risk factor. Programs interested in a weighted screening approach could consider using VACS data to determine priority risk factors for their country and score those accordingly. Similarly, in Uganda, recognizing the high prevalence of emotional violence, DREAMS-implementing partners suggested further screening to identify high-risk impacts (e.g., suicidal ideation, self-harm behaviors, and poor mental health) of emotional violence on AGYW. VACS data could be assessed to determine the prevalence of such nuanced risk factors and support programs to inform more responsive screening criteria. These data indicate that such adaptations may be increasingly important as data provide further evidence of trends and differences in vulnerability across diverse contexts. Previous work has highlighted the successes and challenges in screening for vulnerability among AGYW and can inform development and refinement of such tools (UNICEF, 2021).

These data are subject to several limitations. These population-level estimates of DREAMS eligibility may differ significantly from actual DREAMS program enrollment due to differences in the geographic area served and populations targeted through programmatic efforts, as well as due to changes over time since VACS data collection occurred. Furthermore, the VACS lower age limit is 13 years, so the lowest DREAMS age band of 10–14-year-olds is truncated in this analysis. The VACS surveys are subject to recall bias, especially questions that cover the lifetime. The data analyzed here were collected between 2015 and 2019 and may not reflect more recent changes in the context in each country. There are some limitations of using VACS data to approximate DREAMS eligibility. In VACS, data on alcohol misuse cover only the past 30 days, while DREAMS eligibility may be based on any lifetime alcohol use. Thus, alcohol misuse is likely underestimated compared to actual population-level DREAMS eligibility in the countries examined. Finally, the COVID-19 pandemic has introduced myriad challenges for AGYW, including global increased prevalence of orphanhood (Hillis et al., 2021). In Kenya, a study using qualitative phone interviews found disproportionate increases in both economic hardship and transactional partnership among AGYW during the COVID-19 pandemic compared to young men in Kenya (Decker et al., 2021). Another study in Kenya found that the risk of pregnancy doubled and the risk of school dropout tripled among AGYW affected by lockdown compared with those who were not affected (Zulaika, 2022). Such exacerbations in risk for AGYW may be especially relevant in Uganda, where schools were closed for nearly 2 years and in which the current study shows a vast majority of girls already had risk factors for HIV acquisition. There is a need for more current population data to understand how COVID-19 may have impacted HIV risk for AGYW.

AGYW face complex social, biological, and environmental challenges to avoid HIV infection. Through a greater understanding of vulnerability and risk, the DREAMS program can better customize prevention programming by targeting and tailoring interventions based on the different needs of AGYW in each age band and in the specific country context. VACS data provide critically important population data on the prevalence of key factors and patterns of vulnerability. These data are collected following rigorous ethical standards that ensure privacy and confidentiality and enable AGYW to disclose risk behaviors safely and without stigma. VACS with subnational-level data can also provide insight into regional variations of risk factors and vulnerabilities. Overall, VACS data can be considered a valuable resource to understand DREAMS eligibility in the population and to sharpen programming to ensure that the most vulnerable AGYW are supported with the necessary services to prevent HIV.