A longitudinal qualitative evaluation of an economic and social empowerment intervention to reduce girls’ vulnerability to HIV in rural Mozambique

Evaluation and Program Planning 77 (2019) 101682

Contents lists available at ScienceDirect

Evaluation and Program Planning journal homepage: www.elsevier.com/locate/evalprogplan

A longitudinal qualitative evaluation of an economic and social empowerment intervention to reduce girls’ vulnerability to HIV in rural Mozambique

T

⁎

Holly McClain Burkea, , Catherine Packera, Lázaro González-Calvob, Kathleen Ridgewayc, Rachel Lenzid, Ann F. Greene, Troy D. Moone a

FHI 360, Reproductive, Maternal, Newborn, and Child Health, 359 Blackwell Street, Suite 200, Durham, NC 27701, USA Friends in Global Health, Avenida Maguiguana #32, Maputo, CP604, Mozambique FHI 360, Health Services Research, 359 Blackwell Street, Suite 200, Durham, NC 27701, USA d FHI 360, Behavioral, Epidemiological and Clinical Sciences, 359 Blackwell Street, Suite 200, Durham, NC 27701, USA e Vanderbilt Institute for Global Health, Division of Pediatric Infectious Diseases, 2525 West End Avenue, Suite 725, Nashville, TN 37203, USA b c

A R T I C LE I N FO

A B S T R A C T

Keywords: HIV prevention Economic strengthening Multi-Sector Integrated development Adolescents Girls Mozambique Gender-based violence School attendance Sexual behavior Qualitative

Purpose: An intervention including business training and health education was implemented in Mozambique, where girls are at elevated risk for acquiring HIV. As part of a mixed-methods evaluation, we describe perceived effects of the intervention on girls’ sexual behavior and school attendance. Methods: We conducted 49 in-depth interviews (IDIs) with girl intervention participants (ages 13–19), 24 IDIs with heads of girls’ households, 36 IDIs with influential males identified by girls, and 12 focus group discussions with community members after the intervention ended and one year later. Results: Informants said the primary intervention benefit was realized when girls had money to stay in or return to school and/or to buy necessities for themselves and their households—reducing their need for transactional or intergenerational sex. However, some girls did not make a profit and some businesses were not sustainable. Sometimes the intervention appeared to be implemented in a way to reinforce inequitable gender norms resulting in some girls feeling shame when they reengaged in risky sex after their businesses failed. Conclusions: Earning money enabled girls to potentially reduce their vulnerability to HIV. We offer recommendations for future multi-sector interventions, including the need to address potential harms in programs serving vulnerable girls.

1. Background Mozambique is one of the sub-Saharan African countries most affected by the HIV/AIDS epidemic, with a national HIV prevalence of 15.1 percent in 2015 for adults ages 15–49 (Instituto Nacional Saúde INE, ICF Internacional, 2015). Zambézia Province, Mozambique’s second largest province and home to about 4 million people (Moon, Ossemane, & Green, 2014; PEPFAR, 2019), had the highest estimated number of people living with HIV in the country (˜275,000, or nearly 20 percent of Mozambique’s HIV-infected population) as of 2009 (INSIDA, 2009). As in many places, adolescent girls and young women in Zambézia

are disproportionately affected by the HIV epidemic, where in 2015, the HIV prevalence among females ages 15–24 was estimated to be 14 percent—more than triple the proportion of males the same age (Instituto Nacional Saúde INE, ICF Internacional, 2015). According to the 2011 Mozambique Demographic and Health Survey, 23 percent of girls 15–17 have experienced physical or sexual violence at any time in their life, and 8 percent had sex with at least one man 10 or more years their senior within the last 12 months (Ministerio da Saude INdE, ICF International, 2011). Many factors increase girls’ vulnerability to HIV. These include high rates of poverty (Eaton, Flisher, & Aarø, 2003; Hallman, 2005; Machel, 2001), and individual behaviors that include early sexual debut,

Abbreviations: ASCAs, Accumulated savings and credit associations; FGD, Focus group discussion; GBV, Gender-based violence; HIV, Human Immunodeficiency Virus; IDI, In-depth interview; SEM, Social ecological model; SCIP, Strengthening Communities through Integrated Programming; USAID, United States Agency for International Development ⁎ Corresponding author. E-mail address: [email protected] (H.M. Burke). https://doi.org/10.1016/j.evalprogplan.2019.101682 Received 12 December 2018; Received in revised form 12 July 2019; Accepted 22 July 2019 Available online 23 July 2019 0149-7189/ © 2019 Published by Elsevier Ltd.

Evaluation and Program Planning 77 (2019) 101682

H.M. Burke, et al.

Programming (SCIP) program, locally called Ogumaniha, was a U.S. Agency for International Development (USAID)-funded project that aimed to improve the health and livelihood of children, women, and families in Zambézia Province. SCIP scaled up Women First, a combined economic and social empowerment intervention that trained women to sell products door-to-door in their communities. The business component included business education followed by the sale of items included in a business “kit”. Participants were expected to re-pay the program for the kits with a portion of their sales to receive the next kit, with the remainder of their revenue considered profit to be spent or saved. Generally, the business kits increased in value; once a girl sold her third kit, she was considered a “graduate” and was eligible to receive a bicycle. Soap and ingredients for baking cakes (flour, sugar, oil, etc.) were the most common products in the kits, but participants also sold batteries, rice, biscuits, and spaghetti. In some of the communities, girls also had access to loans and/or savings accounts through accumulated savings and credit associations (ASCAs), which SCIP introduced into the Women First communities to encourage individual saving and to provide access to loans, especially as business capital for Women First participants. ASCA group members make regular contributions and the group fund is used to make loans to community members that are paid back with interest. After a period of time the group fund and its proceeds from interest are paid back to the original members. Adult women were the initial primary focus of Women First and the ASCAs. However, over time, SCIP encouraged adolescent girls ages 13–17 to join the intervention, and targeted vulnerable girls defined as having lost one or both parents, living in a child-headed household, and engaging in transactional sex or other HIV risk behaviors. The Women First intervention was implemented in communities from 2010 through 2015, although the start date of the intervention varied according to the community. However, once initiated in a particular community, the intervention operated continuously until the end of the SCIP program (September 2015). The Women First intervention with adolescent girls used the Go Girls! curriculum from the Go Girls! Initiative, which was designed to reduce adolescent girls’ vulnerability to HIV transmission in select communities in Botswana, Malawi, and Mozambique (HealthCompass, 2014). Women First implemented the full Go Girls! curriculum and a locally-tailored GBV curriculum to encourage social empowerment and reduce adolescent girl participants’ vulnerability to HIV. The intervention also had the goal of encouraging girls to stay in school, since school attendance has been shown to reduce girls’ vulnerability to HIV (Jukes et al., 2008; Stoner et al., 2017). The curriculum included facilitator-led group education sessions covering such topics as gender norms for boys and girls; how to communicate with adults and partners; puberty; pregnancy and HIV prevention; staying in and returning to school; preventing unwanted sexual advances; planning goals; and assessing values, money, and gifts. As such, with its various components, Women First is a combined economic and social empowerment intervention. Adolescent girls were invited to participate in the intervention by community leaders, in conjunction with SCIP program staff and current Women First adult participants. According to SCIP implementers, once invited, individual refusal rates of the intervention were near zero. Indeed, more girls wanted to participate than there were spots available.

multiple partners, and low condom use (Kaponda et al., 2007; McCreary, Kaponda, & Norr, 2008; Underwood, Skinner, Osman, & Schwandt, 2011). Disparities are driven by prevailing gender norms and inequalities (Dunkle, Brown, Gray, & McIntryre, 2004) that increase girls’ vulnerability to HIV, such as limited ability to negotiate condoms, transactional sex, and fear and experiences of violence and abandonment (Organization WH, 2013). A parent’s inability or unwillingness to provide emotional support, material goods, and information about sexual and reproductive health also influences adolescents’ sexual behaviors (Underwood, Skinner, Osman, & Schwandt, 2011). School attendance has also been shown to be protective against HIV (Jukes, Simmons, & Bundy, 2008; Stoner, Pettifor, & Edwards, 2017) and GBV (Kilburn, Pettifor, & Edwards, 2018), yet in Mozambique the school dropout rate is high for girls, especially when they get married and/or have children (Mozambique U. Mozambique Education, 2019). Economic interventions—microfinance, cash transfers, incomegenerating activities, and vocational or business training—can help address the structural factors that increase vulnerability to HIV among girls. Economic interventions in Uganda (The Hunger Project, 2009), South Africa (Pronyk, Hargreaves, & Kim, 2006), and Tanzania (UNFPA, 2006) demonstrated decreases in domestic violence, and increases in youth’s condom use and refusal of sex. Social-level interventions, such as those that focus on gender roles and power relations between sexual partners, recognize the influence of social forces on HIV transmission by influencing practices that can lead to infection. Such interventions identify points of entry at the societal level by altering practices that increase HIV risk in the context where they occur. Examples from Botswana, Malawi, and Mozambique (Kim, Watts, & Hargreaves, 2007; Padian et al., 2011; UNFPA, 2006), South Africa (Pronyk, Kim, & Abramsky, 2008; Pronyk, Harpham, & Busza, 2008), and Tanzania (Jewkes, Levin, & Penn-Kekana, 2003; Maman, Mbwambo, & Hogan, 2002) have demonstrated increases in HIV knowledge and rates of HIV testing and a reduction in gender-based violence (GBV). Given that economic and social empowerment interventions independently reduce HIV vulnerability, some research results suggest that combining the two builds skills to improve financial well-being, women’s empowerment, and gender equity, and thus reduces vulnerabilities to HIV (Kim et al., 2007; Padian et al., 2011; Pronyk et al., 2006). Research from South Africa, for example, found that communities that received an economic intervention and a combined economic–behavioral intervention showed improvement in economic indicators compared to control communities, as well as improvement in indicators of women’s empowerment, reduced levels of intimate partner violence in the previous year, and reduced HIV risk behavior, compared to communities receiving the benefits of either intervention alone (Kim, Ferrari, & Abramsky, 2009). A randomized controlled trial of an intervention package that combined life-skills and health education with vocational training and micro-grants found a lower risk of transactional sex, and a higher likelihood of condom use among adolescent girls in Zimbabwe (Dunbar et al., 2014). 1.1. Study goal We conducted a mixed-methods study to build the global evidence on whether the combined economic and social empowerment intervention called Women First being implemented in Zambézia Province, Mozambique had the desired effect of reducing adolescent girls’ vulnerability to HIV. We also sought to describe potential pathways or processes that contribute to the intervention’s effect on girls’ HIV vulnerability.

2. Methods 2.1. Study design This paper reports on a qualitative assessment that was conducted as part of a longitudinal, mixed-methods evaluation of the Women First intervention implemented with adolescent girls in communities located within six districts of Zambézia Province, Mozambique: Gurué, Alto Mólocuè, Ilê, Mopeia, Mocuba, and Morrumbala. The quantitative

1.2. Overview of the intervention World Vision’s Strengthening Communities through Integrated 2

Evaluation and Program Planning 77 (2019) 101682

H.M. Burke, et al.

Table 1 Total number of respondents or groups and eligibility criteria. Populations

Sample Size

Method

Round 1

Round 2

Girls

49

47

IDI

Household heads

24

24

IDI

Influential males

36

36

IDI

Community members

12 groups

12 groups

FGD

Eligibility Criteria

years at Round 1 • 13-19 in Women First intervention • Participated years • 18+ by girl respondent as household head • Identified years • 13+ by girl respondent as an influential male in her life • Identified not be head of the girl participant’s household • Could years • 18+ • Knowledgeable about Women First intervention and able to discuss community-level effects

Additionally, we conducted 12 FGDs with community members (one per community). Community members familiar with the Women First intervention in their community were recruited for the Round 1 FGDs through government and local leaders and Women First staff. Approximately one year later, from August to October 2016, we reinterviewed 47 of the original 49 girl participants interviewed in Round 1. One girl who was mistakenly interviewed in Round 1 was not reinterviewed at Round 2, and the husband of another girl refused to let her participate in Round 2. At Round 2, the same girls whose household heads and influential males were interviewed at Round 1 were asked to invite their current head of household and top three influential males to participate. As this could change over time, the heads of household and influential males could be different between the two rounds. In Round 2, we conducted 12 FGDs again, with priority participation given to community members who had participated in the Round 1 FGD, and any additional participants being recruited in the same manner as for Round 1. The sample size for the qualitative data collection activities are based on purposive, non-probabilistic sampling, where the size of the sample relies on the concept of saturation—the point at which no new information or themes are observed in the data. Saturation can occur within the first 12 interviews when conducted in a relatively homogeneous group and when the research objective research is to understand common perceptions and experiences (Guest, Bunce, & Johnson, 2006). We selected additional participants to account for potential variation in the implementation of the intervention across communities.

study design—a clustered, non-equivalent (two-stage) cohort trial—and results are described in detail elsewhere (Burke, Field, Calvo, Eichleay, & Moon, 2019). As part of the mixed-methods evaluation, we conducted two rounds of qualitative data collection to understand respondents’ perceptions of the intervention’s effects (Table 1). We conducted longitudinal in-depth interviews (IDIs) with a subsample of girls who participated in the Women First intervention. In each round we also interviewed girls’ heads of households and men whom girls identified as important in their lives at that time point, which was not necessarily the same person. Community members familiar with the intervention participated in focus group discussions (FGDs). We applied the social ecological model (SEM) to the design of the qualitative component of the evaluation, and to the analysis of the data. The SEM is a theory-based framework for understanding the multifaceted and interacting effects of personal and environmental factors that determine behaviors, and for identifying leverage points for health promotion (McLeroy, Bibeau, Steckler, & Glanz, 1988). The most effective approaches to public health prevention use a combination of interventions at all levels of the model, including individual, interpersonal, household, and community levels (UNICEF, 2004). We applied SEM by collecting data from multiple categories of respondents (i.e., girl participants, influential males, household heads, community members) to triangulate findings across the levels. 2.2. Sample selection We randomly selected two intervention communities per district for the qualitative interviews for a total of 12 communities. In each community (Fig. 1), we randomly selected four girls who had participated in the first quantitative survey to participate in a qualitative interview for Round 1. Most girls had finished the intervention by the time of the Round 1 interviews. From August to October 2015, we interviewed 49 adolescent girl intervention participants. We planned to interview a total of 48 girl participants (4 per community) but ended up interviewing one extra girl due to miscommunication during data collection. Five originally selected girls were replaced. In each community, among the girls interviewed, we randomly selected two girls whose heads of household we would interview, and one girl for whom we would interview three influential males in her life (e.g. uncle, boyfriend). Due to random selection, some girls were selected to invite both their head of household and influential males. Girls selected to invite influential males were asked to invite the top three influential males in their lives to contact study staff if the men were interested in participating in an interview. For Round 1, we interviewed 24 of their household heads (two per community) and 36 influential males identified by 12 of these girls (one per community). Two girls originally selected to invite their households’ heads, participated in an interview themselves, but were not able to invite their household heads. One girl originally selected to invite her influential males was interviewed herself, but she could not identify three males to invite.

2.3. Data collection All respondents provided written informed consent prior to participating in the study. We obtained written parental/guardian consent for respondents younger than the age of 18 years. Respondents did not receive reimbursement for participating in any study activity per the guidance of the local research ethics review board who approved the study. Trained interviewers fluent in Portuguese and Lomué, Sena, and/or Chuabo local languages conducted the IDIs and FGDs using semistructured interview guides. IDIs were conducted one-on-one in a private location and lasted approximately one hour. We matched the gender of the interviewer to that of the respondent. FGDs were conducted by two interviewers and included six to 11 participants. FGDs lasted approximately two hours. We audio recorded IDIs and FGDs with the permission of the respondents. Interviewers simultaneously translated and transcribed the recordings from local languages into Portuguese, and then the transcripts were professionally translated into English. The co-authors developed semi-structured interview guides for girls, heads of household, and influential males, which covered the following topics: perceptions of the intervention, especially what girls learned and if/how the intervention affected them; girls’ experience earning money 3

Evaluation and Program Planning 77 (2019) 101682

H.M. Burke, et al.

Fig. 1. Respondent sample selection in each community.

Table 2 Self-reported sociodemographic characteristics of respondents. Girl Participants

Round 1 (n = 49)

Round 2 (n = 47)

Mean age in years, (range) Married, n Has boyfriend, n Has living child(ren), n In school, n

16.0, (13-19) 6 8 7 At least 24

16.5, (13-25) 14 7 13 15

Household Heads

Round 1 (n = 24)

Round 2 (n = 24)

Mean age in years, (range) Household head relationship to girl, n Mother Father Aunt Grandmother Husband Sister Uncle

45.1 (25-79)

47.9 (26-80)

12 3 5 1 1 0 2

10 5 5 1 1 1 1

Influential Males

Round 1 (n = 36)

Round 2 (n = 36)

Mean age in years, (range) Lives with girl, n Influential male relationship to girl, n Uncle Brother Brother-in-law Cousin Father Grandfather Friend/Neighbor Boyfriend

34.8 (16–66) NA

35.4 (18–65) 9

17 6 0 7 2 2 1 1

18 7 2 3 2 1 3 0

Community Members

Round 1 (n = 118)

Round 2 (n = 85)

Male Female Range of number of respondents per FGD

62 56 8–11

43 42 6–10

4

Evaluation and Program Planning 77 (2019) 101682

H.M. Burke, et al.

spending their money on clothing for themselves, products for their business, school-related expenses (supplies, uniforms, fees), and savings. Girls also spent money on household-level expenses. Slightly less than one-quarter of girls in Round 1 stated that they did not earn sufficient money or were not satisfied with the money that they earned. At Round 2, slightly more than half of the girls reported earning money in the past year. Among these, about half were still earning money at the time of the interview, most often stating that they were selling products or working on a farm and selling produce. The other half stated that they were no longer earning money. Some FGD respondents stated that “very few” girls continued to do business, while other FGD respondents described girls struggling to make a profit given higher commodity prices. One girl described the situation in her community:

and what they spent it on; whether respondents thought girls’ relationships had changed, including with household and family members, community members, friends, and sexual partners; changes in girls’ involvement in the community; girls’ thoughts on dating, premarital sex, marriage, the ideal age of their partner, and condom use; and girls’ future goals. FGD guides covered how the intervention was implemented in their community, thoughts on girls earning money, and changes among the girls and other community members resulting from the intervention. Based on Round 1 data, we modified the interview guides for Round 2 to ask explicitly about pathways or processes that may have influenced HIV risk among girls, variations in community norms and intervention implementation, and sustainability of intervention effects. 2.4. Data analysis

Business is the only activity that continued [after the intervention ended], but not all girls are doing business. Some stopped because they spent the money and don't have anywhere to get money to continue with their businesses. I do it to this day, because I usually take some of the money and save it at the warehouse worker’s house. - Girl, Round 2

We developed initial codebooks for each respondent type and data collection round based on the topic interview guides. Three qualitative analysts coded the data using NVivo Version 11 software (QSR International Pty Ltd, 2015). To assess intercoder reliability, teams of two analysts independently coded 12.5 percent of the transcripts and resolved coding discrepancies through discussion. The analysts updated the codebooks based on these meetings and re-coded transcripts when necessary. We took an iterative approach to codebook development for the second round of analysis, developing codebooks based on existing codebooks and making additional revisions to reduce redundancies and harmonize codebooks with the SEM (individual, interpersonal, household, community levels). We generated coding reports and summarized data according to SEM levels and other overarching themes. We use the term “respondents” when reporting results that were consistent across all respondent types (i.e., girls, household heads, influential males, and community members). We note instances when there was contradiction or clear differences in types of responses across respondent types.

Some girls described being unable to sell products or, if profits were low, owing money to the intervention. Most frequently, respondents described that the amount of money girls had to repay to the intervention for products initially given was so high, they were unable to continue having their businesses, as most of the money they earned went to repaying the intervention. Respondents thought that this issue, coupled with the business component being too short and the lack of diversity of the products sold within communities, was contributing to girls’ failed businesses. Additional external factors, such as drought or “hunger,” the need to travel long distances to obtain products, and the high cost of unsubsidized products, also contributed to girls’ financial problems.

3. Findings

3.3. Sexual behaviors

3.1. Respondent characteristics

In the study communities, the HIV risk factors most commonly described by the respondents were sexual risk behaviors (e.g., transactional sex, intergenerational sex, dating and early sexual debut, and multiple concurrent partnerships), early marriage, early pregnancy, and non-use of condoms. Attitudes towards girls’ sexual activity were overwhelmingly negative, as many girls mentioned learning during the intervention that they should not have premarital sex, that it was “wrong” or “bad,” or that girls who engaged in premarital sex garnered less value or respect. Fewer girls mentioned learning about negative health impacts of premarital sex, such as STIs in general, HIV specifically, and unintended pregnancy. Concepts of having sex before age 18, intergenerational sex, and transactional sex were highly conflated, and all treated with similar negative attitudes and general descriptions of health impacts. In both rounds, a few girls and several of their heads of household and influential males reported that girls had ended a formal “dating” relationship or reduced their number of “casual” sexual partners due to the intervention; some said girls were now practicing abstinence or delaying sex until age 18. Although some girls mentioned the potential health risks of early marriage, most mentioned that they delayed or intended to delay marriage because the legal age of marriage was 18, and that they believed that marriage before that age could increase their risk of experiencing GBV. About one-third of girls at Round 2 reported that they were not yet engaged in sexual relationships and either viewed themselves as unaffected and not ready for sex or reported a changed intention to avoid these types of pre-marital relationships due to the intervention. In Round 1, about one-third of girls and respondents in 11 out of the 12 communities mentioned that girls reduced engaging in transactional sex due to earning money through the intervention; some respondents

Of girls who participated in the qualitative interviews the mean age at Round 1 was 16 years old (range 13–25 years). At Round 1, most were single and did not have children; by Round 2, several girls had married and had children (Table 2). The study population had difficulty estimating their ages, and there were sometimes discrepancies between what the girl reported about her age or marital status and what was reported by her household head or influential males. Most household heads interviewed were female, and all were family members except one husband (for the same girl at Rounds 1 and 2). Almost all influential males were family members, with uncle being the most common relationship. At Time 2, most influential males did not live with the girl (this was not asked at Round 1). Only one influential male at Round 1 was described as a boyfriend. In the FGDs, men were slightly more represented than women. About one-third fewer community members participated in FGDs at Round 2. 3.2. Implementation and sustainability of business component Overall, the results suggest that the business component of the intervention increased girls’ financial resources, as nearly all girls reported earning money from this aspect of the intervention; the major drawback being that the businesses did not continue for many girls once the intervention ended. At Round 1, nearly all girls stated that they earned money through the intervention’s business component, and most indicated being satisfied with the money they earned, citing the ways earnings had affected their lives. In both rounds, girls most commonly reported 5

Evaluation and Program Planning 77 (2019) 101682

H.M. Burke, et al.

prevention, pregnancy prevention, and birth spacing; however, fewer girls said that they used condoms compared to those who did not. Girls described that they were unable to use condoms because their husbands refuse, they are married, they trust their husbands, or they simply do not want to use them. During Round 1, girls reported learning about recognizing, preventing, and reporting GBV, including sexual violence. In Round 2, girls similarly described learning about sexual and physical forms of GBV as well as intergenerational sex (sex with older men, including rape/sex with children or child marriage), rape, reporting sexual violence, and physical GBV. At both time points, girls’ knowledge of GBV appeared to be quite superficial, as they often provided very brief descriptions of different examples of GBV. Similarly, girls’ conceptualizations of GBV were linked to other types of violence, and other issues such as division of household labor, verbal harassment or discrimination, general physical violence, and fair treatment of children (e.g., children should not be forced to carry heavy loads). About a quarter of girls in Round 1 and nearly half in Round 2 mentioned learning about reporting GBV to the police, community leaders, and/or parents or seeking medical services at the hospital, without being specifically probed about reporting or seeking services. A few respondents in the other respondent categories also mentioned that girls had learned about reporting and types of GBV; however, they were most likely to mention that girls had learned about equitable gender norms or physical violence. Many respondents described changes in community members’ understanding or knowledge of GBV-related topics, including what constitutes GBV, what victims should do if they experience GBV, and the potential consequences to perpetrators of GBV. Several respondents from all groups in Round 2 said that girls continue to talk about issues related to sexual behavior informally and formally through presentations. While awareness of HIV and GBV remained high among community members, there was limited evidence that this knowledge translated into reduced high-risk sexual practices among girls in the community who had not participated in the intervention. Notably, about equal numbers said transactional sex decreased among girls in the community as those that said it continued, and some respondents described variability in their communities:

similarly described reduced transactional sex in Round 2. Girls and other respondents similarly stated that girls would have been having transactional sex if it were not for the intervention or would have to return to engaging in transactional sex if they were to stop earning money through their businesses. I don’t even know what I would do if it wasn’t for this business that I have. I would be having sex with men in Lua-Lua to earn some [money] for my survival and my family’s … I think I would be a prostitute to help my mother, who is a farmer. What she produces isn’t enough for all of us. - Girl, Round 1 Although intergenerational sex was mentioned less frequently because it was often conflated with transactional sex, many respondents stated that girls took part in less intergenerational sex due to the intervention. For example: R1: This program helps the girls avoid prostituting themselves. They don’t go out at night to whore. R2: Yes, because some stopped going out with married men, older men, and colleagues. They changed now because of the money from the cakes. We regret that it was such a short time for the learning. FGD, Round 1 Positive effects on girls’ sexual risk behaviors were generally sustainable for girls who maintained their businesses and income, but this was not possible for all participants. In Round 2, some respondents stated that girls from the intervention re-engaged in transactional sex or had multiple partners out of financial need; several girls reported getting married in the previous year due to financial pressure or because of pregnancy. For example, one participant described that the program reinforced her desire to wait until she is older to have sex and get married at Round 1, stating that "I stay away from men now to not get pregnant before growing up," and that because of the program, "I grew up a lot. I believe that I am able to face challenges. I already have a more open mind." However, by Round 2, she was no longer reaping the economic benefits of the program. Her household head described that, "she was not able to sell and have profits to continue with the business." The girl, now sixteen, stated that she married a man three years her senior out of financial necessity:

R1: Our daughters don't do that anymore; they do their little business and are able to buy their clothes and shoes. R2: To me it happens, but not with all. Some do that because they can't buy soap, clothes and even food, which is the case with fish, oil and beans. R3: This subject still happens, but it decreased because of the Women First program, which gave them product kits to sell and have their own money. The program also taught them that when a person has reckless sex, that person runs several risks, like, for example, a person can catch HIV, syphilis, mule (sexually transmitted disease), and other cases, it depends on each person. - FGD, Round 2

I: Did your thoughts about marriage change since our last interview in August/September of last year, and after the Women First program closed? R: Yes, I had the idea of getting married only after the program closed, because I realized that there was no other way. So when someone showed up and proposed to me, I had no other ideas but to accept it. But when I entered the Women First program the plan was to study. She further noted that this was not her ideal age for marriage, saying she would have preferred to wait until she was at least 18 years old:

Respondents often said intergenerational sex had decreased and that adult men are now less likely to look for teenage girls as sex partners, often because of fear of punishment. Many respondents mentioned that the intervention caused men to fear being caught for having sex with girls under 18 since it was a crime, but they did not mention men changing their attitudes towards girls or becoming more aware of their rights (e.g., having sex with a minor brings about inappropriate power dynamics). In most communities, at least one respondent noted that intergenerational sex still occurs. A few respondents credited the intervention with reducing early marriage and pregnancy in their community:

For a girl to get married she needs to be 18 years old or more, but that depends on the girl’s condition, because you may want to marry at 18 years old, but when her mother doesn't have the means to buy clothes you have to marry so your husband will help you. Respondents from both rounds described girls learning about condoms for HIV and pregnancy prevention and some respondents described that girls now use condoms due to the intervention. For example: I don’t have a husband or a boyfriend. Very few times I have sex, and I demand the man put on a condom. I don’t have sex with many partners like before… some accept to put on a condom for sex, some don’t. For these ones, I say no… When they accept to put on a condom, I say yes. - Girl, Round 2

What we used to see was that some would rush into marriages before growing up, but it was because they didn’t have anywhere to go because of poverty. They had to look for friends to help them and give them soap. That’s what we used to see, there were premature weddings, and with that our country will be ugly.- FGD, Round 1

Many girls in both rounds described wanting or intending to use condoms or other family planning methods in the future for HIV

About two-thirds of the respondents described reduced perpetration 6

Evaluation and Program Planning 77 (2019) 101682

H.M. Burke, et al.

buying my own stuff, paying my school fees, for my notebooks, my uniform and shoes, I started thinking I should continue to study until the 12th grade, then start working…I am still going to school.- Girl, Round 2

of sexual or physical violence against women or girls in the community, more equitable division of labor in homes, respect for children’s rights (such as going to school or child labor), and improved reporting of GBV. Girls were not asked about their own experiences of GBV, but during both rounds a few girls stated that girls in their communities no longer experienced GBV, or, more frequently, that girls were now able to avoid GBV, report their own experiences if assaulted, or help others report GBV, due to what they learned in the intervention. For example:

At the household level, several respondents stated that girls paid for school supplies, uniforms, or school fees for their siblings. A few respondents discussed potential household-level impacts on girls going to school, such as the intervention may have influenced parents’ attitudes towards girls attending school:

[In the past] a lot of men beat people. They would drink and insult others: their wives, neighbors, children. They grabbed breasts and butts. But with the teachings from the program this has reduced, they are afraid of getting arrested by the chief.- Girl, Round 2

After the arrival of the project, we, the parents, also changed our behavior. Before, we didn’t want our girls to go to school because we were afraid they were going to become whores. But with the training they received in the program, they learned they had to go to school and they spread that idea to the parents. That’s why many girls go to school today. - FGD, Round 2 At Round 2, a few respondents said that the intervention had increased school participation among girls who had not participated in the intervention: The community was affected because the girls who are in the program, almost all of them, are going to school. They help their families with that money they earned from the business. Many girls in the community now want to study.- Influential male, Round 1

Although they were not specifically asked, many respondents spontaneously mentioned during Round 1 that girls learned about respect and/or became more “respectful” or “in line” due to the intervention. We explored this concept further during Round 2 and found that most respondents conceptualized girls being “respectful” or “in line” as a set of characteristics that included greeting and respecting elders, obeying parents and contributing to household chores, not being sexually active or interacting with males before marriage (and in a few cases, using condoms), and not “going out” or staying out late. A few respondents in Round 2 credited the business component as having an influential role in reducing girls’ sexual activity and therefore making them appear more “respectful” by no longer engaging in transactional sex. Over half of respondents in Round 2 described whether girls in the intervention continued to be “good,” “in line” or generally respectful and well-behaved after the intervention ended. All respondents except for one stated that girls in the intervention sustained these behaviors after the intervention ended. Although the intervention strived to empower girl participants, respondents’ language surrounding girls and their sexual behavior revealed the continued presence of stigma towards girls’ sexual activity. Also, several respondents volunteered information that implied that the delivery of the Women First intervention in some communities reinforced or perpetuated inequitable gender norms towards girls and women. For example, several girls described learning that girls that were "good" and were valued were those that abstained from sex until married, avoided multiple partners, and/or refused to have transactional or intergenerational sex. Engaging in these activities, according to some respondents, reduced girls' value to their community and their desirability to men.

However, some respondents noted that some girls were still not in school, and thus the intervention effects were either not widespread throughout the communities or had not been sustained over time. 4. Discussion Our findings suggest that combined economic and social interventions such as the Women First intervention are a potentially promising strategy for reducing girls’ HIV vulnerability. We observed two main pathways through which the intervention may have reduced participants’ HIV risk: (1) providing girls access to financial resources and thereby reducing girls’ engagement in sexual behaviors that exposed them to HIV risk, and (2) changing household and community social norms with regards to how they perceive and treat girls. Earning money through the intervention affected girls’ HIV risk through reducing transactional (and often intergenerational) sex, reducing early marriage for financial reasons, and reducing early pregnancy to obtain financial support and/or get married. Consistent with evidence that new HIV infections among young women and girls are largely driven by heterosexual transmission (Kharsany & Karim, 2016), the HIV risk factors most commonly described in our study communities were sexual risk behaviors, early marriage, early pregnancy, and non-use of condoms. Despite the presence of these risk behaviors, respondents reported that the Women First intervention reduced some girls’ engagement in sexual risk behaviors. Although our findings are qualitative, evaluations of other integrated economic and social empowerment interventions have also been found to reduce girls’ HIV risk (Kim et al., 2007, Kim et al., 2009; Padian et al., 2011; Pronyk et al., 2006). The intervention appeared to enable some girls to continue attending school and increased some girls’ goals for obtaining higher education. Decreases in girls’ stigmatized behaviors (particularly transactional sex and multiple partnerships) were highly valued by community members, heads of household, and influential males; however, the extent to which girls could enact these changes was, for the most part, realized when girls had increased financial resources. Notably, some changes occurred through a combination of access to financial resources and increased education and empowerment. These changes speak to the complimentary nature of the two intervention components and their combined effects. However, it also reveals the inherent weaknesses when one component is not sustained (i.e., business fails), as has also been found with other multi-sector interventions (Dunbar

3.4. School attendance About half of girls spontaneously mentioned that they were in school during the Round 1 interview. At Round 2, all girls were asked about their school enrolment, and only about one-third of girls said they were still in school and an additional third stated that they had dropped out of school since the intervention ended or in the previous year. Some respondents reported that more girls could continue with their education due to their participation in the intervention. About one-fifth of respondents stated that the intervention did not impact girls’ school attendance, or they had to drop out of school after the intervention ended. At Round 1, some girls stated that they placed higher value on their own education due to the intervention’s emphasis on educational attainment. Furthermore, some girls mentioned that the intervention influenced their goals of attending higher education (secondary education and beyond). Earning money through the intervention also may have facilitated girls’ school enrollment, as some respondents stated that girls used their business profits on school supplies or enrollment. When I joined the Women First program, I learned many things. For example, about business, HIV, hygiene, the right age for a girl to get married, many things! And when I started doing my business, 7

Evaluation and Program Planning 77 (2019) 101682

H.M. Burke, et al.

respondents and describe themselves as separate from the intervention, it is possible that social desirability bias may have influenced the qualitative findings in at least two ways. The first could be related to respondents’ desire for the researchers to bring back the Women First intervention to their community which may have resulted in them inflating the effect of the intervention or describing their situation as worse off than it actually was. Secondly, strong negative attitudes towards girls’ sexuality in this context may have also influenced responses (i.e., respondents may have been less forthcoming in describing girls’ sexual behavior). Another potential threat to the findings is recall bias, since many respondents were recalling experiences from a year before we interviewed them the first time. We also cannot rule out that selection bias may have influenced our findings, especially for the male respondents who were identified by the girl respondents. That is, girls did not often invite their sexual partners to be part of our evaluation; rather, most girls chose uncles or other male family members who may have been less aware of their sexual activity or other behaviors. Inconsistent probing across interviewers, especially in Round 1, affects the precision of the numbers shown in Table 2. For example, we did not directly ask girls whether they were attending school at Round 1, so there are 14 girls for whom it is unclear whether they were in school or not.

et al., 2014). 4.1. Recommendations for strengthening the Women First Intervention The qualitative assessment yielded valuable information to strengthen future iterations of Women First and other interventions like it. We recommend that future interventions address factors along both pathways we identified. The economic component could be strengthened by diversifying the products girls sell within a community so that girls do not compete and drive down their profits. Also, girls’ businesses may be more sustainable if program staff strengthen the connections between the input suppliers and girl participants so that after the intervention ends, girls can negotiate inputs at low costs on their own. This would involve implementing the business component for a longer period, ensuring that the curriculum addresses negotiation skills and strategies (such as bulk purchasing inputs), and strengthening the savings component so that girls would have savings to draw upon when their profits are low or their businesses fail. Girls may also benefit if the curriculum taught resilience skills to prepare them for potential business failure. To avoid potential harms and ensure that interventions are being implemented as designed, it is critical that programs take proactive steps to ensure that staff are not introducing their own biases, especially reinforcing inequitable gender norms, during implementation. We recommend accurate and detailed documentation of actual program implementation down to the site level to facilitate and enhance the quality of monitoring and evaluation efforts. Furthermore, the educational components of the intervention, especially the GBV component, should be expanded to others in the community, especially men. Recent studies have shown that culturally adapted, gender transformative interventions that include men are possible and can reduce attitudes that accept GBV and perpetration of GBV (Doyle, Levtov, & Barker, 2018; Raj, Ghule, & Ritter, 2016). Moreover, the content of the educational components should be enhanced beyond providing superficial information, such as listing types of violence, to provide participants with a deeper understanding of the concepts within a human rights-based approach. Critical reflection around violence, i.e., challenging social norms that are accepting of violence, is essential to counter the prevalent inequitable gender norms.

4.3. Conclusion Our findings provide program implementers with concrete recommendations to improve future programs attempting to reduce girls’ vulnerability to HIV in low-resource settings. Given that adolescent girls disproportionately bear the burden of the HIV epidemic it is critical that we deliver effective interventions that empower, and do not harm, this vulnerable population. Lessons learned The findings from this longitudinal qualitative evaluation revealed some promising effects of the intervention on outcomes and illustrated potential pathways leading from intervention effects to reductions in HIV vulnerability among participant girls. We also demonstrate the value in using more than one research method when conducting programmatic evaluations.

4.2. Strengths and limitations Funding This longitudinal, qualitative assessment provided detailed information about the respondents’ perceived effectiveness of the intervention, sustainability of the intervention, and potential pathways of effect. The longitudinal design is an important strength of this evaluation. This is because multiple qualitative rounds allowed us to analyze the data from the first round, adapt the interview guides for the second round, and ask detailed follow-up questions about our interpretation of Round 1 data. Furthermore, having two rounds of data collection that used the same qualitative interviewers allowed us to gain a deeper understanding of individuals’ behaviors, and may have helped build rapport to gather increased richness and quality of data. Our study had other strengths. The use of the SEM as our theoretical approach and interviewing multiple categories of respondents allowed us to triangulate findings across the individual, interpersonal, household, and community levels, which resulted in a more comprehensive understanding of potential intervention effects than we found during the quantitative evaluation of the Women First intervention (Burke et al., 2019). We also achieved thematic saturation in all the main themes due to our design. A high retention rate was another strength. Our data collectors kept in contact with respondents and their communities over the course of the study, even when we were not collecting data, and this facilitated our high retention rate in a traditionally hardto-track population. Despite our data collectors’ best efforts to establish rapport with

This paper was produced under United States Agency for International Development (USAID) Cooperative Agreement No. AIDOAA-LA-13-00001 and was made possible by the generous support of the American people through the United States Department of State, USAID, and the United States President’s Emergency Plan for AIDS Relief. The contents are the responsibility of FHI 360 and do not necessarily reflect the views of USAID or the United States Government. Ethics approval This study was reviewed and approved by the Mozambican National Bioethics Committee for Health (Comité Nacional de Bioética para a Saúde, CNBS), FHI 360’s Protection of Human Subjects Committee, and Vanderbilt University’s Institutional Review Board. Contributors HMB conceived and designed the study. LGC, AFG, and TDM collected the data. CP, KR, RL, and HMB analyzed the data. HMB, CP, KR, and RL wrote the first draft of the article. HMB, LGC, CP, KR, RL, AFG, and TDM made substantial intellectual contributions to the article. Authors do not have any potential, perceived, or real conflict of interest to declare. The study sponsor (USAID) was involved in the 8

Evaluation and Program Planning 77 (2019) 101682

H.M. Burke, et al.

study design by requesting the authors evaluate the Women First intervention being implemented in Zambézia Province, Mozambique and they also reviewed the protocol. The study sponsor was not involved in the collection, analysis, or interpretation of data; the writing of the report; or the decision to submit the manuscript for publication. The named authors wrote the first draft of the manuscript. No honorarium, grant, or other form of payment was given to anyone else to produce the manuscript.

Stoner, M. C. D., Pettifor, A., Edwards, J. K., et al. (2017). The effect of school attendance and school dropout on incident HIV and HSV-2 among young women in rural South Africa enrolled in HPTN 068. AIDS, 31(15), 2127–2134. Kilburn, K. N., Pettifor, A., Edwards, J. K., et al. (2018). Conditional cash transfers and the reduction in partner violence for young women: An investigation of causal pathways using evidence from a randomized experiment in South Africa (HPTN 068). Journal of the International AIDS Society, 21(Suppl 1). The Hunger Project (2009). Change to believe in: THP Uganda’s impact. Pronyk, P. M., Hargreaves, J. R., Kim, J. C., et al. (2006). Effect of a structural intervention for the prevention of intimate-partner violence and HIV in rural South Africa: A cluster randomised trial. Lancet, 368(9551), 1973–1983. UNFPA. I. Change, choice, and power: Young women, livelihoodsand HIV preventionliterature review and case study analysis 2006; http://www.unfpa.org/webdav/site/ global/shared/documents/publications/2007/change.pdf. Accessed May 19, 2014. Pronyk, P. M., Kim, J. C., Abramsky, T., et al. (2008). A combined microfinance and training intervention can reduce HIV risk behaviour in young female participants. Aids. 22(13), 1659–1665. Pronyk, P. M., Harpham, T., Busza, J., et al. (2008). Can social capital be intentionally generated? A randomized trial from rural South Africa. Social Science & Medicine, 67(10), 1559–1570. Maman, S., Mbwambo, J. K., Hogan, N. M., et al. (2002). HIV-positive women report more lifetime partner violence: findings from a voluntary counseling and testing clinic in Dar es Salaam, Tanzania. American Journal of Public Health, 92(8), 1331–1337. Jewkes, R. K., Levin, J. B., & Penn-Kekana, L. A. (2003). Gender inequalities, intimate partner violence and HIV preventive practices: Findings of a South African crosssectional study. Social Science & Medicine, 56(1), 125–134. Kim, J. C., Watts, C. H., Hargreaves, J. R., et al. (2007). Understanding the impact of a microfinance-based intervention on women’s empowerment and the reduction of intimate partner violence in South Africa. American Journal of Public Health, 97(10), 1794–1802. Padian, N. S., McCoy, S. I., Manian, S., Wilson, D., Schwartlander, B., & Bertozzi, S. M. (2011). Evaluation of large-scale combination HIV prevention programs: Essential issues. Journal of Acquired Immune Deficiency Syndromes, 58(2), e23–28. Kim, J., Ferrari, G., Abramsky, T., et al. (2009). Assessing the incremental effects of combining economic and health interventions: The IMAGE study in South Africa. Bulletin of the World Health Organization, 87(11), 824–832. Dunbar, M. S., Kang Dufour, M. S., Lambdin, B., Mudekunye-Mahaka, I., Nhamo, D., & Padian, N. S. (2014). The SHAZ! project: Results from a pilot randomized trial of a structural intervention to prevent HIV among adolescent women in Zimbabwe. PloS One, 9(11), e113621. HealthCompass (2014). Go Girls! Toolkit. [Toolkit]. Accessed August 3 2017 http://www. thehealthcompass.org/project-examples/go-girls-toolkit. Burke HM, Field S, Calvo LC, Eichleay MA, Moon TD. A quasi-experimental evaluation of an economic and social empowerment intervention to reduce girls’ vulnerability to HIV in Mozambique. McLeroy, K. R., Bibeau, D., Steckler, A., & Glanz, K. (1988). An ecological perspective on health promotion programs. Health Education Quarterly, 15(4), 351–377. UNICEF (2004). Module 1: Understanding the the social ecological model (SEM) and communication for development (C4D). https://www.unicef.org/cbsc/files/Module_1_ SEM-C4D.docx. Guest, G., Bunce, A., & Johnson, L. (2006). How many interviews are enough?: An experiment with data saturation and variability. Field Methods, 18(1), 59–82. QSR International Pty Ltd (2015). Nvivo qualitative data analysis software [computer program]. Version Version 11. Doncaster, Australia: QSR International Pty Ltd. Kharsany, A., & Karim, Q. A. (2016). HIV Infection and AIDS in Sub-Saharan Africa: Current Status, Challenges and Opportunities. The Open AIDS Journal, 10, 34–48. Doyle, K., Levtov, R. G., Barker, G., et al. (2018). Gender-transformative Bandebereho couples’ intervention to promote male engagement in reproductive and maternal health and violence prevention in Rwanda: Findings from a randomized controlled trial. PloS One, 13(4) e0192756. Raj, A., Ghule, M., Ritter, J., et al. (2016). Cluster randomized controlled trial evaluation of a gender equity and family planning intervention for married men and couples in Rural India. PloS One, 11(5) e0153190.

Declaration of Competing Interest Authors do not have any potential, perceived, or real conflict of interest to declare. The study sponsor (USAID) was involved in the study design by requesting the authors evaluate the Women First intervention being implemented in Zambézia Province, Mozambique and they also reviewed the protocol. The study sponsor was not involved in the collection, analysis, or interpretation of data; the writing of the report; or the decision to submit the manuscript for publication. The named authors wrote the first draft of the manuscript. No honorarium, grant, or other form of payment was given to anyone else to produce the manuscript. Acknowledgements The authors thank the research assistants from Friends in Global Health, Mozambique, and field coordinator, Samuel Martinho, for collecting the data for this evaluation. We appreciate the Ogumaniha project staff for their assistance locating the program beneficiaries. We are grateful to staff at the Vanderbilt Institute for Global Health for research support. We acknowledge Mary Ellen Duke (USAID/ Mozambique), Amelia Peltz (USAID/Washington), and Michael Ferguson (FHI 360) for their input on the study design. References Instituto Nacional Saúde INE, ICF Internacional (2015). Inquérito de Indicadores de Imunização, Malária e HIV/SIDA em Moçambique 2015. Relatório Preliminar de Indicadores de HIV. Maputo, Maputo, Moçambique. Rockville, Maryland, EUA: INS, INE e ICF. Moon, T. D., Ossemane, E. B., Green, A. F., et al. (2014). Antiretroviral therapy program expansion in Zambézia Province, Mozambique: geospatial mapping of communitybased and health facility data for integrated health planning. PloS One, 9(10), e109653. PEPFAR. PEPFAR Zambezia: Partnership for an AIDS-free Mozambique. INSIDA (2009). Inquérito nacional de prevalência, riscos comportamentais e informação sobre o HIV e SIDA em Moçambique Maputo. Mozambique: Ministério de Saúde. Ministerio da Saude INdE, ICF International (2011). Moçambique Inquérito Demográfico e de Saúde 2011. Calverton, MD: MISAU, INE e ICFI. Organization WH (2013). Gender inequalities and HIV. Geneva: WHO. Mozambique U. Mozambique Education: Current Situation http://www.unicef.org.mz/ en/our-work/what-we-do/education/. Accessed April 4, 2019. Eaton, L., Flisher, A. J., & Aarø, L. E. (2003). Unsafe sexual behaviour in South African youth. Social Science & Medicine, 56(1), 149–165. Hallman, K. (2005). Gendered socioeconomic conditions and HIV risk behaviours among young people in South Africa. African Journal of AIDS Research, 4(1), 37–50. Machel, J. Z. (2001). Unsafe sexual behaviour among schoolgirls in Mozambique: A matter of gender and class. Reproductive Health Matters, 9(17), 82–90. Kaponda, C. P., Dancy, B. L., Norr, K. F., Kachingwe, S. I., Mbeba, M. M., & Jere, D. L. (2007). Research brief: Community consultation to develop an acceptable and effective adolescent HIV prevention intervention. Journal of the Association of Nurses in AIDS Care, 18(2), 72–77. McCreary, L. L., Kaponda, C. P., Norr, K. F., et al. (2008). Rural Malawians’ perceptions of HIV risk behaviors and their sociocultural context. AIDS Care, 20(8), 946–957. Underwood, C., Skinner, J., Osman, N., & Schwandt, H. (2011b). Structural determinants of adolescent girls’ vulnerability to HIV: Views from community members in Botswana, Malawi, and Mozambique. Social Science & Medicine, 73(2), 343–350. Dunkle, K. L., Jr, Brown, H. C., Gray, G. E., & McIntryre, J. A. (2004). Harlow SD Genderbased violence, relationship power, and risk of HIV infection in women attending antenatal clinics in South Africa. Lancet, 363(9419), 1415–1421. Underwood, C., Skinner, J., Osman, N., & Schwandt, H. (2011a). Structural determinants of adolescent girls’ vulnerability to HIV: views from community members in Botswana, Malawi, and Mozambique. Social Science & Medicine, 73(2), 343–350. Jukes, M., Simmons, S., & Bundy, D. (2008). Education and vulnerability: The role of schools in protecting young women and girls from HIV in southern Africa. AIDS, 22(Suppl 4), S41–56.

Holly McClain Burke, PhD, MPH is a behavioral scientist in the Reproductive, Maternal, Newborn, and Child Health (RMNCH) Division at FHI 360. She joined FHI 360 in 2002 and has more than a decade of experience designing and leading international research studies. Her expertise is in (1) designing studies that incorporate both quantitative and qualitative research methods to expand understanding of contraceptive and HIV prevention behaviors and (2) evaluating the impact of programs, including health communication campaigns and integrated economic strengthening and health interventions. Catherine Packer, MSPH, is a Research Associate in the Reproductive, Maternal, Newborn and Child Health Division at FHI 360. She has more than a decade of experience working on quantitative and qualitative research studies related to international public health and social science, with a specific focus on adolescents. Lázaro González Calvo, PhD, MSc, MPH, a medical anthropologist with training in nursing and mental health, has spent the last eight years as the Research Director for Friends in Global Health in a rural province in Mozambique. He joined FGH in 2010 and has experience designing and leading international implementation research studies and experience in implementing large population-based surveys as well as evaluations utilizing focus groups and in-depth interviews in resource-limited settings.

9

Evaluation and Program Planning 77 (2019) 101682

H.M. Burke, et al.

Ann F. Green, MPH is a public health professional with over 13 years of international experience. She currently serves as Senior Program Manager at the Vanderbilt Institute for Global Health, managing implementation of program analysis and evaluation at both community and health facility levels.

Kathleen Ridgeway, MSPH, is a mixed-methods researcher with experience in social and behavioral interventions related to youth development and HIV prevention among most at-risk populations. As a Research Associate at FHI 360, she has spent the last three years contributing to intervention design, research program management, and data analysis for intervention-based and implementation science research conducted in resource-constrained settings.

Troy D. Moon, MD, MPH is Associate Professor of Pediatrics and infectious diseases, fluent in Portuguese, who served as the in-country Clinical and Research Director from 2007 to 2012 of the Vanderbilt Rural Mozambique Project supported by PEPFAR. His current focus is on the conduction of implementation science research and program evaluations in sub-Saharan African settings and he has over 10 years of experience implementing both clinical and research capacity building projects in low-resource settings.

Rachel Lenzi, MPH, is a Research Associate in the Behavioral, Epidemiological & Clinical Sciences Division at FHI 360. She has several years of experience implementing social science and public health research projects with a focus on gender and designing and evaluating social and behavior change programming.

10