Keeping the Essentials in Place: Lessons Learned from a Qualitative Study of DREAMS in Northern Uganda

: Peer-facilitated curriculum-based programs, including Stepping Stones, have been shown to be effective in preventing HIV and reducing gender-based violence (GBV). We conducted a qualitative study in early 2017 to hear perspectives of adolescent girls and young women (AGYW) participating in the USAID-funded Determined, Resilient, Empowered, AIDS-Free, Mentored, and Safe Women (DREAMS) intervention (administered by the President’s Emergency Fund for AIDS Relief (PEPFAR) in northern Uganda that featured 10 peer-facilitated sessions of a Stepping Stones curriculum. The study focused on asking AGYW how the initiative had affected their lives and on identifying lessons learned that could support future initiatives. A total of 56 AGYW were interviewed, including the peer facilitator and 6–7 randomly selected participants of nine DREAMS groups in Northern Uganda. Overwhelmingly, participants indicated that regular HIV testing and knowing their status, knowledge and an increased use of family planning, and knowing how to respond to GBV were among the results of their participation. However, a problematic ﬁnding was that peer group discussions relating to reducing GBV included advising AGYW about how to adjust their own behavior in ways that would reduce tension with their male partners, rather than shifting harmful gender norms. This is not consistent with the Stepping Stones program and prompted a retrospective review of factors related to how the program was implemented to better understand this result. Our study points to the important role facilitators play in shifting challenging gender norms, and the importance of ﬁdelity to original program designs, as well as appropriate adaptations for different contexts. Our ﬁndings also signal the need for funders to allow sufﬁcient time to pilot and adapt models.


Introduction
Adolescent girls and young women (AGYW) aged 15-24 in sub-Saharan Africa, and particularly those in Eastern and Southern Africa, have been well-documented as disproportionately affected by HIV and AIDS [1,2]. In 2021, according to UNAIDS, 63% of new HIV infections in sub-Saharan Africa occurred among AGYW, a rate that is three times that of adolescent boys and young men [2]. Causes include intersecting structural factors, including gender norms that drive patriarchal and hierarchical power inequities, multiple forms of gender-based violence (GBV), poverty, lack of access to education and health services, and few economic opportunities [2,3]. Many at-risk AGYW have limited sexual and reproductive health (SRH) knowledge, including an inadequate understanding about how HIV is transmitted and how to prevent infection. They also have poor access to SRH services and HIV prevention methods, including condoms [4] and pre-and post-exposure prophylaxis [5]. Policies, religious beliefs, and social norms often restrict AGYW's access to both knowledge and services [2]. Marginalized AGYW, including those with intersecting identities, for example those who are lesbian, bisexual, transgender, queer, or intersex (LBTQI), AGYW in inter-generational sexual relationships [6], and those that practice transactional sex are particularly at high risk [7].
There is a wide consensus that addressing GBV is a key aspect of HIV prevention programs and that physical and sexual GBV is "pervasive" among adolescent young women across sub-Saharan Africa [8][9][10]. In Uganda, there is an overall high rate of HIV among AGYW despite evidence of reductions in new infections among AGYW in areas targeted by interventions. For example, despite progress made by prevention programming, including large initiatives such as the Determined, Resilient, Empowered, AIDS-Free, Mentored, and Safe (DREAMS) program [11], four out of five (79%) new HIV infections occur among AGYW with ages 10-24 years [12] (p. 15), a rate that is higher than the 63% average across sub-Saharan Africa. The Annual Uganda Joint AIDS Review FY 2021/22 reports that to date only 44 of 62 prioritized districts have comprehensive interventions targeting AGYW [12] (pp. 9,23). Uganda is one of the 28 countries that are part of the Global HIV Prevention Coalition that together account for close to 75% of the global HIV burden.
Today, there is global alarm that the 2030 Sustainable Development Goal 3.3, which envisions ending the AIDS pandemic by 2030, will not be reached with the current efforts [1, [13][14][15]. UNAIDS and the Global HIV Prevention Coalition are among those calling for more urgent progress. The need to implement HIV prevention efforts at scale is central to all strategies [14].
Yet, while there is agreement by both funders and implementers that there is a need to expand (or "scale up") pilot or small-scale interventions that have been proven effective, how to do this successfully is a still nascent area of research [13,[16][17][18]. The World Health Organization and ExpandNet have supported defining scaling up and have provided guidance on what works to organizations [13,16]. The Community for Understanding Scaling Processes (CUSP) is a consortium concerned with the "safe and ethical" scaling of the approaches to evidence-based social-norms changes. CUSP provides a political analysis of efforts by large-scale initiatives to scale some of the member organization's interventions, with implications for both funders and implementers [19]. Among the recommendations are for funders to recognize the need for a sufficient investment of time and up-front money to scale community-based interventions with attention to fidelity of the designs. Implementers are advised to consult with the original project designers and to ensure appropriate engagement and accountability to the communities where the interventions are based [18,19].
Other studies on how evidence-based interventions can be adapted for scaling up, as well as the lessons learned about implementation successes and failures [20], are beginning to create documentation about what is needed to achieve the intended outcomes [19,21].
DREAMS is a public-private partnership that was launched in 2014 by the US President's Emergency Fund for AIDS Relief (PEPFAR) with the objective of reducing new HIV infections in AGYW aged 15-24 [11]. The implementation began in 2015 in 10 countries in Eastern and Southern Africa and expanded to four additional sub-Saharan countries and Haiti in 2017. Close to a billion US dollars had been invested by 2020 [11]. The DREAMS theory of change is based on the layering of evidence based, multi-sectoral interventions to simultaneously address the behavioral, biomedical, and structural drivers of HIV [22,23]. Targets for reaching HIV-negative AGYW have been ambitious from the beginning and have pushed implementers to scale interventions quickly. Analyses of the implementation have identified successes, including the opportunity to meet the complex needs of AGYW. Challenges include expectations for the immediate roll-out of a very complex model that had not been piloted [22], the expectation that ambitious targets would be met within short timelines, and the practical challenges of achieving the "layering" of interventions [24,25].
In 2016, five organizations funded by USAID, in partnership with other private and public groups, began implementing the first iteration of DREAMS in northern Uganda. US-AID Uganda selected Stepping Stones as the core evidence-based mechanism for DREAMS' delivery to peer groups. Stepping Stones is a community-based program implemented by trained facilitators through peer and community groups. A randomized control trial in South Africa (2006)(2007)(2008) showed it reduced risky behaviors among men and reduced men's self-reported perpetration of GBV [26]. Numerous adaptations of this model around the globe have reported positive outcomes relating to both HIV and GBV [27].
This article reports on the findings of a qualitative study conducted in early 2017 by one of the partners, the Strengthening Districts for Sustainability (SDS) Project, implemented by Cardno Emerging Markets USA, Ltd., based in Virginia, USA, now a part of DT Global Inc., at the midpoint of the first two years of DREAMS's implementation in northern Uganda, and as funding for SDS was coming to an end. The study's purpose was to learn about AGYW's perspectives on the differences the DREAMS intervention had made in their lives, as well as the lessons learned that could support future initiatives targeting these populations. The lead author was a program consultant to SDS for DREAMS's implementation from January 2016-March 2017.

Study Design
This was a qualitative study that utilized approximately one-hour semi-structured interviews with a purposive sample of AGYW who had participated in the program. Trained research assistants conducted the interviews in Acholi or Langi, while a note-taker took notes. Interviews were audio-taped and later transcribed into English. Demographic information was collected, including age, education, religion, marital status, children, and income. The interviews used an appreciative inquiry approach [28] with questions that asked participants what they were proud about that had changed in their life as a result of their participation in DREAMS; the role of DREAMS and other factors, including family, community, religious leaders, and other interventions; what they most appreciated about DREAMS; and what they wished was different that might have allowed them to achieve more from their participation. They were also asked to provide details about their specific experiences with DREAMS, including the number of Stepping Stones sessions attended, HIV testing, their participation in workshops about gender norms and genderbased violence, their training on economic strengthening/business development, and plans as part of their peer group for an income-generating activity (IGA), and the roles and engagement of partners, family, and community members. Finally, they were asked about their life's dreams and what they felt it would take to achieve these. Key questions from the interview guide can be found in Appendix A.

Sample
At the time this study was conducted, 280 peer groups of out-of-school AGYW were operating as part of DREAMS in four districts of northern Uganda [29] (p. 11). Each group was led by a peer facilitator and intended to have a minimum of 20 members each. All AGYW enrolled in DREAMS were HIV negative and all were referred to family planning. All peer groups were using Stepping Stones as a core approach to achieve their objectives. At the time of this study, 95 of the groups had participated in a minimum of a half-day training on business skills/IGA selection, and 16 peer facilitators had participated in a one-week intensive workshop on gender norms and gender-based violence.
The study sample consisted of 56 AGYW, including one peer facilitator (PF) and six or seven AGYW peer-group participants (P1-P7) selected from five peer groups in District 1 (Groups 1-5) and four peer groups in District 2 (Groups 6-9). One peer group in each district provided seven study participants (Groups 5 and 8) while all others provided six study participants.
The selection of peer groups for this study was based on a purposive sample with efforts to select those that had the highest levels of exposure to available interventions: 10 Stepping Stones sessions, training on business skills, and an in-depth workshop on gender norms/GBV for peer facilitators. From this purposive sample of groups of AGYW, the researchers or the DREAMS officer for the district met with the peer facilitator to explain the study and gain her consent to participate. She was asked to provide the list of her participants with information about the number of Stepping Stones sessions attended. A random sample was selected from those who had attended a minimum of six sessions. In almost all cases, there was a wide variation in the number of sessions participants had attended.

Ethics
Prior to each person's admittance to this study, there was a one-on-one meeting to gain consent. All participants were either above age 18, or had lived on their own, and/or had children of their own, and, therefore, were considered "emancipated" and able to sign a consent form themselves. One formal interview with two local implementers was conducted and additional program data was consulted. Participant confidentiality and anonymity was protected by using a code for each participant. Only senior researchers had access to the original questionnaires that included participant names and transcripts with identifying information. Because the communities are small, districts where the study was conducted are not named to further protect participants' identities.
The study was approved by the TASO Research Ethics Committee in Uganda and by the Uganda National Council on Science and Technology.

Data Analysis
Qualitative data analysis was iterative and collaborative [30] and used both deductive and inductive content analysis processes [31]. Analysis began as part of the data collection process. Interviewers and notetakers answered post-interview questions at the end of interviews to document initial thoughts on key themes, and data collection teams met where possible at the end of each day to discuss these. As transcripts were translated into English, they were uploaded to Dedoose (https://www.dedoose.com, accessed on 6 January 2023), an encrypted data analysis software platform, useful for collaborative qualitative analysis. In addition, a smaller team began to analyze the data set. Two coders analyzed the initial interviews to ensure consistent interpretations, and an ongoing dialogue was held among the coders, including the lead author, as new codes were identified. The emerging findings were triangulated with Dedoose exports, repeated reviews of the English transcripts, consultations among the analysis team, and further consultations with implementers. The lead author led the drafting of a 2017 unpublished report, with input from the second author. A dissemination-of-findings meeting was held with key stakeholders in 2017. For this article, the lead author revisited the Dedoose exports, interview transcripts, and project documents.

Participant Demographics
The peer groups formed through the SDS component of DREAMS were HIV-negative 15-to 24-year-old AGWW who were out-of-school and segmented according to the following categories: Never pregnant-pregnant; have given birth; married with children; or engaged in transactional sex. Some peer groups had a focus on a particular segment of AGYW, while others were mixed.
Participants ranged in age from 15 to one participant who was an outlier at age 25, based on DREAMS guidelines for participants to be ages 15-24. The average age was 20. Thirty participants in the sample (60%) had given birth, with almost half (14) of these births occurring before the age of 18, including three by age 15. Almost a third of the sample (32%) was currently partnered, with half of these married and half cohabiting. Another 13% report they are "separated" or "divorced." Of those currently or previously married, five participants (almost 10%) were early marriages, with force or coercion a theme among these. All but two young women in the sample dropped out of school prior to completing Secondary 6, the final grade level of a secondary school education. Among the sample, the highest level of education attained ranged from Primary 3 (one participant) to Secondary 6 (two participants). The median education level was Primary 7, the last year of primary education. Uganda's education system includes pre-primary/Kindergarten, seven years of primary, and six years of secondary education. Post-secondary/higher education options include technical and vocational schools, colleges, and universities.
Most of the sample reported a limited knowledge on sexual reproductive health and family planning prior to their engagement with DREAMS. This was true despite being engaged in other interventions before DREAMS. They discussed not knowing how to use condoms correctly; not having access to condoms; a lack of understanding about the menstrual cycles and times of fertility; and fear of many family-planning methods because of various myths and misinformation about them. Some said they still harbored some of these feelings. Many did not know their HIV status prior to their engagement with DREAMS. They did not test for HIV because of a general lack of awareness of why one should test frequently, fear of knowing one's status, and stigma. Fears of testing included the idea that if you go for testing you must think you will be positive.
I was scared too to test my blood because I was not sure if my blood was safe, but I was encouraged during the training . . . and when I return from the training, I went and tested my blood and now I know my status and that its safe . . . This is what [organization] told us, that HIV doesn't kill you if you have it, but it's good to know your status early so that you can start an early treatment... PF, Group 9.
. . . so, I have learnt so many things on HIV-there are so many things on HIV I never knew about. Like the issue of those ones who are on ARV, it will lower the virus down, I even never knew like this issue of one partner not having HIV, and the other partner is having HIV also existed. PF, Group 5.
Participants were not asked explicitly about their personal experiences with GBV; however, almost 40% (22) of those interviewed talked about GBV as part of their stories about their lives-either from partners, or family members.
Many of the study participants discussed the challenges they encountered at home as children and adolescents. These stories included being forced to leave school; being subjected to GBV, including abuse by extended family members, forced sex, and early marriage (which some resisted); being forced to leave home; and unplanned pregnancies. Some were orphaned early in life; for others, their families struggled with poverty, and a few were subjected to traditional views that prioritize boys for education.
Several study participants reported that they engaged in risky behavior with boys and men before their engagement with DREAMS. Some of these AGYW recounted escaping difficult home situations. Others talked about using transactional sex as a form of survivala way to earn money for themselves, their children, or other family members.
The thing is that I stopped clubbing. I was very good at that but when I join DREAMS, they told me the disadvantages and this direct sex I stopped it, the most thing which made me proud is I stopped direct sex. Most time I use condoms. P2, Group 4.

Results of DREAMS Reported by Participants
Study participants were asked the following: Can you tell me about something that you are particularly proud of that has changed in your life as a result of your participation in DREAMS? What are the things you have most appreciated about DREAMS?

Addressing Health-Related Drivers of HIV
DREAMS is meant to promote frequent HIV testing; linkages to health services; knowledge and access to sexual and reproductive health services and commodities for contraception and HIV prevention; and to engage partners in HIV testing and either voluntary male medical circumcision (VMMC) or anti-retroviral treatment (ARV) if positive. Overwhelmingly, participants said that regular HIV testing and knowing their status, knowledge and an increased use of family planning, and knowing how to respond to GBV were among the life-changing results of their participation in DREAMS. They also described knowing the benefits and risks of a range of contraceptive choices, including short-term and long-term methods, and how to correctly use family planning commodities, particularly male and female condoms.
Many AGYW reported that they adopted positive behavioral practices that reduced their risk of HIV and other sexually transmitted infections. Some participants indicated that knowing their HIV-negative status promoted a reduction in high-risk sexual behavior and a stronger commitment to remaining faithful to their partners. In addition, some said they deepened their knowledge of how HIV is transmitted.
This study's participants also discussed spreading the information and improved practices relating to HIV prevention to partners, family members, and friends. Some said that talking to their husbands about HIV risk and getting them to test has promoted more fidelity in the marriage, while one said: " . . . I told him [husband] that if he feels like he wants to get involved with another person he should first go the both of them and do a blood test" P4, Group 1.
Some participants referred to not only the training received through Stepping Stones, but also the training done through DREAMS by volunteer village health teams (VHTs), nurses, and other local implementing partners. Several participants discussed how the myths that various forms of family planning would cause birth defects were dispelled.
Some of the peer educators described how their relationships with health centers and VHTs supported an improved access to services for peer-group members. According to one peer facilitator: "In the health facilities, according to us, or for our group, the moment we go the health center they help us faster than other people . . . " PF, Group 1 At the same time, at least one participant talked about the poor services provided when going for HIV counseling and testing (HCT). She said: " . . . sometimes you go, and you don't get, people are many, like for us they collected our books, but we went there for two days . . . without getting service . . . ." P2, Group 6. Furthermore, according to the local partner's staff that is responsible for mobilizing HCT, HIV test kit stock-outs curtailed testing in some locations.
DREAMS increased awareness and knowledge about GBV, as well as the prevention and response services. Some of the study participants did not know about non-physical forms of GBV. They only realized that they had been subjected to emotional, psychological, or economic GBV when the topic was explored within the peer groups. A participant said: " . . . all those things my father was doing to me-GBV, I didn't know about it . . . " PF, Group 4.
They were also not aware of the available response services and systems where one could seek support if they experience GBV. Several said they did not know about post-exposure prophylaxis or the emergency contraception available in response to rape. DREAMS made them aware of these services and referral points.

Building Self-Confidence and Feelings of Safety of AGYW
DREAMS intends to build the social assets of AGYW, including providing them with safe spaces, peer networks, and peer support networks through strengthened families and communities. These strategies are meant to promote positive health-related behaviors and to address some of the structural drivers of HIV. Interview questions included: "Do you feel safer after being part of the DREAMS initiative? How and why?" Many study participants conveyed that they feel more self-confident and safer because of their participation in DREAMS. They talked about having a group of friends to share with; knowing their HIV status; having control over reproduction; knowing where to get health services; being able to help friends and family access services; and improved relationships with partners, family members, and community members.
Some participants reflected on the importance of the peer group in making their lives better. For some, having friends to share fun, as well as fears, with was new and clearly played a role in reducing their social isolation and sadness. For others, DREAMS provided a positive and resourceful social network to replace a negative one.

Addressing the Structural Drivers of HIV Risk
DREAMS intends to transform gender norms and reduce GBV, to promote the income generation of AGYW in order to address poverty, to promote access to education, and to strengthen families and communities as ways of addressing the structural drivers of HIV.
Many participants explained how their participation in the Stepping Stones sessions had given them the awareness and skills to reduce incidences of GBV-either in their own relationships or through help they provide to others. This was attributed by most to improved communication skills, with at least one peer facilitator who talked about how she and her partner could now discuss issues more equitably. However, several peer-group participants said that GBV is reduced because they learned how to be better partners-this is explained with examples of not responding when a partner argues with you and ensuring that the roles expected of a wife are done well.
. . . they [peer facilitators] said if your husband went somewhere and has come back when quarrelling, you should not reply, if you reply it will be increasing the annoyance every time. You should keep quiet if he tells you to do something you should just accept, you should not compete with him. P2, Group 6.
This happened to my cousin, the husband had started an extra marital affair with a younger girl, and they have two children . . . that day I had gone to pay her a visit and I found her packing her things that she is going back home . . . . Specifically, he said that the wife doesn't wash his clothes, doesn't iron them, and doesn't fulfil his sexual needs. I took the wife aside, talked to her alone, and told her what to do. We joined the husband and I told her to ask him for forgiveness and she did so. After one week the husband sent for me . . . he appreciated me more for what I did, and he said there is a change now. P6, Group 7.
I used to experience violence, when I was still together with my husband, whenever he delayed somewhere and came back home, I would just welcome him back with quarrelling, so they taught me how to welcome him back. P3, Group 5.
They taught us how we can welcome our husbands when they come back. . . . they taught us that when our husbands come back, we should receive them, get their bags, give them a chair, greet, and serve food and water for bathing and also, they told us that we can learn to play with our husbands too to show love. That is what Stepping Stones taught us with. P6, Group 7.

One peer facilitator narrated what causes violence:
I will give as an example, we are in the home, you as the wife don't listen to what the husband says, take for example the man says don't do this, but you do it; won't this bring about violence? And yet he told you not to do it. And also, there is one thing that is making men violent or what is causing violence in the home is sex. If the man insists, he wants to have sex and yet the woman doesn't feel like, even that will cause violence, right? . . . And you as a woman, you should respect your husband in every way, when the man comes back home show him love, bring food, and put it on the table, welcome him in all ways. This will make it that even when you ask for anything from the man, he will accept, there will not be violence but love only. Even it means kneeling down before the man, do it. PF, Group 3.
DREAMS is intended to improve the capacity of AGYW to participate in saving, loaning, and business activities, although these efforts remained nascent at the time of this study. Many of the Stepping Stones' groups had started Village Savings and Loans Associations (VSLA) where members could make contributions to a savings pool and access loans at a low interest rate. Some also cited training on business management as part of DREAMS, which improved their capacity to plan for both the group IGA and their own micro businesses.
Almost all participants interviewed referred to the IGA that they anticipated starting with financial support from DREAMS. Many expressed frustrations with the long process of obtaining this support, which for some was the motivation for remaining in DREAMS. Overall, most believed they would eventually receive the support though some were not sure.
About a third of the AGYW interviewed indicated they have some cash income from small business ventures. While some of these were started prior to DREAMS, others said they started these as a result of DREAMS.

Participants' Visions for Their Futures
One of the interview questions asked AGYW to "imagine it is 10 years from now and that in these 10 years you have been able to achieve whatever you want to in life. Describe yourself and your life to me. The main themes included: Some participants had simple hopes, while others had larger visions for themselves, demonstrating some degree of self-efficacy and an inroad towards agency [32]. One participant said: "If I get much money, I will also open my hotel" P4, Group 7. Another said: "I need to be having a business different from doing pan cakes to a big shop . . . " P2, Group 8. A 17-year-old mother said: "I want to go back to school and become a medical worker . . . " P7, Group 5. When asked what they need to achieve their dreams, many participants said they needed to work hard, and they needed the support of their partners and families. Almost all participants said that they needed to go back to formal school or vocational or technical training. Most participants said they would need financial support to go back to school to start their own businesses, which would then support them going back to school.

Peer Facilitators' Perspectives on DREAMS's Implementation
Most of the nine peer facilitators interviewed related that they were proud of their roles, what they had learned, and that they had opportunities to support their peers. However, most also reported challenging jealousies from peers. One facilitator said: "Seven people left the group. They say that for us . . . the peer facilitators, we are eating their money . . . " P1, Group 2. Others also noted that while facilitators received some compensation, participants received nothing to support their participation.
While questions of facilitators did not explicitly ask about their training, one commented that it was not enough if you did not already have some knowledge and skills: " . . . we went for the training together but they did not go deep into this Stepping Stones-they just give guidelines and if you don't have knowledge you will not transfer it to fellow girls . . . " PF, Group 1. While confident in her own abilities, she said another participant could not even say what Stepping Stones was after the training. She and others commented that the Stepping Stones manual was in English, and they were expected to keep records and report in English.
A key theme among peer-group facilitators and participants was that community members and participants had expectations of DREAMS for their assistance in creating an IGA, as well as with their help in getting into school or technical training that had not yet materialized.
. . . when they introduced us to DREAMS, they told us if you want to go back to school, they will take you back to school; if you want to go for vocational training-you choose what you want to do, like catering, depending on what you want to do. But, up to now we have waited in vain, so if they can fulfill their promise, it will be good so that our colleagues who have given up can be motivated. P2, Group 2.
Many facilitators reported that their groups had grown smaller because some participants were disenchanted. They also talked about their work in recruiting new members to take the place of those who left and advocating with implementers for promises to be kept. Most said they thought they would eventually get their group IGA funded by DREAMS. One facilitator lamented that without this component some AGYW would go back to transactional sex to survive, despite all they had achieved in DREAMS.
Finally, a challenge that facilitators spoke about was the lack of supplies, such as flip charts, and demonstration dildos for facilitating groups. Participants talked about condom shortages, commonly known as "stock outs".

Implementor's Perspectives on Challenges
DREAMS was launched in northern Uganda through five USAID-funded implementing partners (IPs) with relevant ongoing projects that were asked to add DREAMS's objectives to their current statements of work. While all were community-based, their target populations were different. SDS, which is responsible for the coordination of IPs and HIV-prevention services, placed peer groups of targeted out-of-school AGYW at the center of its approach; the IP responsible for economic strengthening had a focus on families of orphans and vulnerable children; and another IP targeted in-school AGYW. Additionally, the northern Uganda DREAMS model did not include cash assets or cash subsidies to address poverty and lack of education, identified as two of the structural drivers of HIV risk [33,34] that DREAMS is meant to focus on [35].
Early expectations were to pilot interventions in a limited number of communities in three targeted districts. One district was later split, creating four target districts. IPs were challenged to identify overlapping communities; however, funders quickly expanded targets and required IPs to scale DREAMS to all hotspot areas in all districts.
USAID Uganda had selected Stepping Stones, an evidence-based workshop series that addresses issues relating to HIV, gender, sexual and reproductive health, communication, and relationship skills, as the central approach and entry point for delivering HIV-prevention interventions to out-of-school AGYW in northern Uganda. However, as peer groups were scheduled to begin, USAID put a hold on the use of the original Stepping Stones curriculum. Groups formed early in the process had to wait to begin activities, creating some confusion and frustration. Later, USAID instructed partners to move forward with a version of Stepping Stones adapted by the Medical Research Council in South Africa [36].
The implementation differed in several ways from what is recommended in the Stepping Stones manual [36], attributed by some to the rush to meet ambitious targets. Facilitators were trained over one week, rather than the recommended 3 weeks; peer groups were comprised primarily of girls and young women aged 15-24, with a few groups of adolescent boys and young men, rather than peer groups of various segments of the community, including boys and men, but also other age and social groups that should work in parallel and occasionally meet together; and peer groups for DREAMS strived to have a minimum of 20 members, but ideally more, rather than the maximum of 20 members recommended by Stepping Stones.

Discussion
This qualitative study looked at perspectives from AGYW on how participating in the DREAMS initiative in northern Uganda from mid-to late-2016 to early 2017 changed their lives and sought to identify lessons for future initiatives targeting this population. The results showed that overall, AGYW felt they benefited from their participation, and they pointed to many positive results in their lives. The findings, however, pointed to more positive results around addressing the health-related drivers of HIV than in addressing the structural drivers of HIV, including poverty, a lack of education, unequal gender norms, and GBV.
The positive changes included knowing their HIV status through regular HIV testing; an increased knowledge on, access to, and use of condoms and a range of family planning methods; and improved communication skills, self-confidence, self-efficacy, and feelings of safety. However, the results relating to reducing GBV, a key objective of DREAMS [22], showed that at least some of the peer-group discussions on reducing GBV included suggestions that women should avoid intimate partner violence (IPV) by adopting more traditional roles as women and being better wives. This is not an intended outcome of the Stepping Stones program.
This finding prompted a retrospective review of factors related to how the program was implemented to better understand this result and what recommendations can be made to improve outcomes. We found issues that have implications for discussions related to how to scale and/or adapt evidence-based programs.
As noted above, implementers for this initiative did not adhere to the recommendations in the Stepping Stones manual for training of facilitators or for the composition of peer groups [36].
The centrality of facilitators or change-agents in norms-shifting and social behavioral change interventions such as DREAMS is well-documented [37]. Peer facilitators for DREAMS must be able to negotiate complicated relationships in the community, both among their peers and with community members, and most importantly must motivate AGYW to adopt the behaviors being promoted. As change agents, peer facilitators must internalize the changes they seek to effect in others.
The peer facilitators in our sample talked about shifting their own attitudes and practices around HIV testing, condom use, and FP use, and acting as role models to their peers and others to adopt the desired behaviors. However, when it came to GBV, our data suggests that facilitators had not all undergone the more challenging process of shifting their own beliefs about the roles of women in Ugandan society that support gender inequality and drive GBV.
This finding adds to the mounting evidence demonstrating that when evidence-based initiatives are scaled or adapted, as with the use of Stepping Stones in DREAMS, achieving the same results as the original initiative requires a combination of fidelity to the original design and adaptation that is responsive to a new context [37,38].
Where programs rely on the use of peer-or community-based facilitators, providing sufficient training, mentorship, and follow-up monitoring and support is critical to success [19,37,39]. This is similar to the lessons learnt from related programs that use facilitators, for example SASA!, a community mobilization approach that uses "community activists" to prevent HIV and GBV [40]. In addition, the Transforming Masculinities project [41] used a model of cascading transformations to shift social norms underpinning sexual GBV. In this model, the Gender Champions and Faith Leaders first underwent their own process of training, critical reflection, and transformation before leading efforts to facilitate the reflection and transformation of other members of their communities. The program was shifted from the rural, conflict-affected eastern Democratic Republic of Congo to urban and peri-urban Kinshasa as a new initiative, Masculinite, Famille, et Foi, that included a focus on both family planning and SGBV. The adapted program showed success in shifting norms to support family planning but did not have results on reported rates of sexual GBV.
A retrospective analysis conducted by the Passages Project suggested that the complexity of generating changes in social norms relating to GBV requires longer time frames to affect change, as well as the need for monitoring and support to change agents to ensure that messaging remains consistent with what is intended [41]. Our study results resonate with these.
A recent study looked at the efficacy of religious leaders as change agents to reduce IPV and shift norms around men's power in relationships in the Becoming One (B1) intervention in northern Uganda. One finding was that the religious leaders who were most successful were those who held more gender equitable norms and attitudes from the outset [20]. When timeframes are necessarily short, as with the DREAMS implementation, selecting facilitators that already embody norms that are closer to what is needed to create change may support program objectives.
In addition to issues related to facilitators of Stepping Stones' peer groups, there was not an opportunity to pilot the DREAMS model to test what happened when individual interventions were "layered" or how to effect layering with IPs working with different approaches and target populations. Additionally, while the DREAMS model conceptualized by PEPFAR included addressing poverty and the lack of education [11], consistent with research about the importance of structural interventions [33], these components were non-existent in the DREAMS model implemented in northern Uganda in 2016-2017.
Others have also raised ethical concerns about programs addressing GBV and HIV that are scaled to new locations without closely adhering to quality standards for implementations [42]. Analysis of the relationships between funder, international implementing partners, and local peer facilitators during the implementation of DREAMS in northern Uganda also highlights power dynamics that have an impact on adaptation, contextualization, and the implementation of programs similar to DREAMS. The current movement towards localized and decolonized development funding provides important reflections for consideration by large, complex, top-down initiatives such as DREAMS [43]. Recent projects such as the Agency for All project that has a strong focus on locally led and equitable partnerships between funder (USAID), the prime institution (University of California San Diego, Centre on Gender Equity and Health) and local partners in West Africa, East Africa and South Asia seem to demonstrate a step in the right direction in being intentional about equalizing these power dynamics. How these strides towards addressing power dynamics are playing out during the implementation of the project have not yet been assessed. For details see: https://geh.ucsd.edu/agency-for-all/ (accessed on 30 March 2023).

Conclusions
Our results add to the growing body of literature around what produces desired results when programs that were successful in one setting are scaled and/or adapted to a new context. They point to the important role facilitators play in shifting challenging gender norms, and the importance of fidelity to program designs as well as the appropriate adaptation of program models proven effective in other contexts. Findings also signal the need for funders to allow sufficient time to pilot and adapt models while needed lessons are learned about how to effectively localize and decolonize development assistance. These efforts have the potential to improve results by increasing opportunities for contextualization, learning, and adaptation of norms shifting interventions.

Limitations
The purposive sample focused on those with the highest level of exposure to the intervention. While meant to capture the successes of the initiative, this excluded perspectives on why some AGYW did not engage as fully or dropped out. This study employed only one form of interview techniques, so information was not triangulated by using other methods, such as focus group discussions. In addition, because of time limitations, our study design included only one meeting with each study participant, limiting the opportunity to build a better rapport with her. A few of the study participants were shy and the depth of the data was limited, although others were quite verbal and expressive. It is possible that more participatory methods would have helped the AGYW to be more open as well.
Author Contributions: Conceptualization, D.G.; methodology, D.G. and P.B.; formal analysis, D.G. and P.B.; investigation, D.G. and P.B.; supervision, D.G. and P.B.; writing-original draft preparation, D.G. and P.B.; writing-review and editing, D.G. and P.B. All authors have read and agreed to the published version of the manuscript. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study, which are qualitative, are unavailable to protect the identities and privacy of study participants. This is consistent with ethical agreements made with the funder.

Conflicts of Interest:
The authors declare no conflict of interest.