Cluster randomized trial of comprehensive gender-based violence programming delivered through the HIV/AIDS program platform in Mbeya Region, Tanzania: Tathmini GBV study

The Tathmini GBV study was a cluster randomized trial to assess the impact of a comprehensive health facility- and community-based program delivered through the HIV/AIDS program platform on reduction in gender-based violence and improved care for survivors. Twelve health facilities and surrounding communities in the Mbeya Region of Tanzania were randomly assigned to intervention or control arms. Population-level effects were measured through two cross-sectional household surveys of women ages 15–49, at baseline (n = 1,299) and at 28 months following program scale-out (n = 1,250). Delivery of gender-based violence services was assessed through routine recording in health facility registers. Generalized linear mixed effects models and analysis of variance were used to test intervention effects on population and facility outcomes, respectively. At baseline, 52 percent of women reported experience of recent intimate partner violence. The odds of reporting experience of this violence decreased by 29 percent from baseline to follow-up in the absence of the intervention (time effect OR = 0.71, 95% CI: 0.57–0.89). While the intervention contributed an additional 15 percent reduction, the effect was not statistically significant. The program, however, was found to contribute to positive, community-wide changes including less tolerance for certain forms of violence, more gender equitable norms, better knowledge about gender-based violence, and increased community actions to address violence. The program also led to increased utilization of gender-based violence services at health facilities. Nearly three times as many client visits for gender-based violence were recorded at intervention (N = 1,427) compared to control (N = 489) facilities over a 16-month period. These visits were more likely to include provision of an HIV test (55.3% vs. 19.6%, p = .002). The study demonstrated the feasibility and impact of integrating gender-based violence and HIV programming to combat both of these major public health problems. Further opportunities to scale out GBV prevention and response strategies within HIV/AIDS service delivery platforms should be pursued. Trial Registration: Pan African Clinical Trials Registry No. PACTR201802003124149.


Summary of Proposed Research
This research protocol describes an outcome evaluation that the President's Emergency Plan for AIDS Relief (PEPFAR) has funded through USAID to evaluate the comprehensive facility-and community-based gender-based violence (GBV) program that is being rolled out in Tanzania with support from the PEPFAR GBV Initiative. The GBV program and this study focus on intimate partner violence in the forms of sexual, physical, and emotional violence; and on sexual violence among both adults and children, irrespective of the relationship between perpetrator and survivor. The program includes interventions focused both on preventing GBV and on strengthening services for GBV survivors.
The study will employ a pair-matched cluster randomized trial design to compare the effectiveness of this comprehensive programming over time with standard practice in control facilities and communities. The study has two primary aims, or objectives, which are to assess whether the comprehensive GBV program intervention resulted in an expected 1) increase in care for GBV survivors, reflected in delivery and utilization of GBV services at health facilities, and 2) decrease in experience of GBV among females ages 15-49. Secondary study aims include assessment of program effects related to knowledge, attitudes, and norms regarding GBV and gender equality; community-led actions to address GBV; knowledge of and barriers to utilization of GBV services; GBV service components and quality; GBV service and program capacity and coverage; and HIV-related outcomes. Data collection will include a household survey of females ages 15-49 at baseline and endline (i.e., after approximately 24 months of program implementation); extraction of client data from health facility GBV registers; key informant interviews with community representatives and program participants; health facility assessments; and key informant interviews with program managers. The study will take place in 12 clusters across Mbeya region, where each cluster is defined by a Health Center or District Hospital, and communities surrounding these facilities. The program itself will be implemented by local partners who are supported by PEPFAR through the Walter Reed Program in Tanzania.
All studies on GBV involving human subjects pose potential risks for study participants, including risks which may arise from personal distress caused by discussing GBV experiences and from further abuse to the survivor if others learn that s/he has discussed it with the study team. To minimize these risks, the study team will adhere to WHO guidelines regarding the conduct of research on GBV (WHO, 1999) to ensure the safety of study participants and the study team, and confidentiality of participant identity and data. The study will be conducted in a manner that aims to reduce possible distress caused to participants by the study, including ensuring that interviewers are certified in ethical research practice, well trained, and equipped with information and skills to refer participants to GBV services as appropriate.
The proposed study design is methodologically sound and builds upon past research conducted by the study team and others to address key gaps in the evidence base for GBV programming. The study is well timed to coincide with the initial, national roll-out of GBV services by the Ministry of Health and Social Welfare (MOHSW) later this year. Results from the study will inform the scale-up of services and provide critical evidence on public health outcomes. The study team will develop and implement a dissemination plan that will help to ensure that study findings are properly interpreted and used widely to advance GBV policy and programs, both within Tanzania and globally.

Rationale for the study
Global interest and investment in programs to prevent GBV and mitigate its harmful impacts on individuals, families, and communities have accelerated in recent years, in part due to growing understanding of the association between GBV and HIV risk. The United Nations Program on HIV/AIDS (UNAIDS) has identified the elimination of GBV as a core pillar of HIV prevention (UNAIDS, 2009). In 2010, PEPFAR launched an initiative that focuses on countries hardest hit by both GBV and HIV; Tanzania is one of three focus countries in this initiative.
Although HIV program efforts to address GBV have grown significantly in recent years, research on the effectiveness of these interventions is scant. The research that has been done has focused either on interventions at community and individual levels to promote gender equality and change GBV-related attitudes and behavior, or on strengthening response services for GBV survivors. A unique contribution of the proposed study is evaluation of a comprehensive program model that includes a full array of community-and facility-based programs and services that are hypothesized to operate synergistically to address both GBV prevention and care for survivors. The 2010 Tanzania Demographic andHealth Survey (TDHS, 2010) found that 44 percent of ever-married women have experienced physical or sexual violence from their current/most recent husband or partner, and 37 percent of ever-married women experienced intimate partner violence (IPV) in the 12 months prior to the survey; 20 percent of women ages 15-49 have ever experienced sexual violence (SV), usually perpetrated by their partners or former partners. Fiftyfour percent of women and 38 percent of men ages 15-49 believe that a husband is justified in beating his wife for certain reasons. The 2009 Violence Against Children (VAC) Survey (United Nations Children's Fund et al., 2011) found that nearly three in 10 females and one in seven males have experienced violence prior to age 18. Only one in five female and one in 10 male survivors sought any services. Of those who had their first sexual experience prior to age 18, 29 percent of females and 18 percent of males reported that it was not consensual. The prevalence of sex with two or more partners in the last 12 months was significantly higher among those who experienced childhood sexual violence. Studies elsewhere have shown that sexually abused children are more at risk of acquiring HIV as adults (Slonim-Nevo and Mukuka, 2007).
A qualitative study conducted in Dar es Salaam found several factors associated with child sexual offences, including community passivity, weakness in the legal system, inadequacy of the legal framework, and vulnerability of key players (Kisanga, 2010). Specifically, magistrates and legal investigators were found often to live within the communities they serve, thus Gender-based violence: An umbrella term for any act, omission, or conduct that is perpetuated against a person's will and that is based on socially ascribed differences (gender) between males and females. In this context, GBV includes but is not limited to sexual violence, physical violence and harmful traditional practices, and economic and social violence. The term refers to violence that targets individuals or groups on the basis of their being female or male." -Ministry of Health and Social Welfare, 2011 compromising their ability to be objective in handling cases of child sexual abuse. In another study that explored community perceptions about child sexual abuse in Tanzania, focus group discussions with adult community members revealed a lack of trust in the healthcare and legal systems, issues of disclosure related to fear of stigma, myths justifying child sexual abuse, decreased respect for children's rights, and awareness coupled with distress (Kisanga, 2011). The perception gleaned from these studies is that the weakened legal structure, combined with communities' reluctance to get involved in child sexual abuse cases, makes violence against children a particularly complicated issue requiring improved screening and interventions at community and health facility levels.
The following sections outline research conducted by the research team and others on what is known regarding the relationship between GBV and HIV and the effectiveness of various intervention strategies to be implemented as a comprehensive model by PEPFAR-supported partners in Tanzania and evaluated through the study presented in this protocol.

Relationship between GBV and HIV
The gendered nature of the HIV epidemic has received increasing attention, with women now constituting 60 percent of adults living with HIV in sub-Saharan Africa (UNAIDS, 2010). Recent analyses of cohort data regarding incident HIV infection have found IPV to be an independent risk factor for HIV . Baseline results from the cluster randomized control trial of SASA!, a community mobilization intervention developed by Raising Voices in Uganda to change community and individual attitudes and behaviors that support both the perpetration of violence against women and HIV risk behaviors, showed that reports of sexual behavior differed significantly according to whether or not the respondent reported experience of GBV, suggesting associations between sexual risk behaviors and IPV. Women who had experienced and men who had perpetrated IPV in the past year were more likely than those who had not to have had multiple partners and/or extramarital partners and less likely to use condoms (Abramsky et al., n.d.). A cross-sectional study in South Africa found that young men (under age 25) who had been physically violent towards a female partner had twice the odds of having HIV as young men who had not been violent (Jewkes et al., 2011).
While the association between GBV and HIV acquisition is widely accepted, the pathways are myriad and complex. It can be direct, as with physical trauma from SV inflicted by someone who has HIV. For example, coerced sex and other forms of SV may contribute directly to a woman's HIV risk, being a potential route of exposure to the virus as well as increasing the chances of HIV transmission if lacerations or other genital trauma occur (Maman et al., 2000). Women in violent relationships are also less likely to be able to negotiate the frequency or circumstances of sex, thus exacerbating their HIV risk (Campbell et al., 2008;Jewkes and Morrell, 2010).
Evidence also points to several indirect mechanisms of association between GBV and HIV through a chain of effects that include psychological, economic, and gendered conditions and behaviors that lead to a victim's engagement in high-risk sex. For example, women in violent relationships may be more likely to have multiple partners (He et al., 1998;Dunkle et al., 2004), have unprotected sex (Wingood and DiClemente, 1997), and have had an early sexual debut. Sometimes just the threat or fear of violence is sufficient to keep a woman from protecting herself from HIV.
Compounding the situation are barriers that GBV may create to a survivor's access to HIV services. A review for the U.S. Institute of Medicine, based on studies between 1998 and 2007, found that "violence or fear of violence from an intimate partner is an impediment (to) or a consequence of HIV testing" (Campbell et al., 2008). Fears of violent repercussions may prevent women from uptake of HIV testing or disclosing their HIV status (Maman et al., 2001;Medley et al., 2004;WHO, 2000). Individuals typically keep violence hidden to avoid stigma, discrimination, and harmful repercussions. Parents of victimized girls often keep it secret to prevent shame to the family and the perceived ineligibility of their daughters for marriage. In other situations, girls may be forced to marry the perpetrators of rape or violence to maintain social acceptability. Thus, survivors face enormous barriers to accessing most types of services. Interventions that address the stigma and strengthen linkages between services and the community are urgently needed.
Living with HIV also can put a woman at greater risk of violence. Vulnerability may arise from accusations by her partner or family that she has had sex outside the partnership, or as a socially acceptable form of discrimination against persons with HIV. Against a backdrop of gender inequality, an HIV diagnosis and/or its disclosure may put a woman at increased risk of IPV (Maman et al., 2000). GBV also can prevent survivors from realizing the benefits of health services; for example, a controlling partner may keep a woman from adhering to medication.
Inequalities perpetuated by gender-based norms and roles, cultural practices, and laws collectively contribute to power imbalances between women and men and boys and girls. These gendered power imbalances play out at societal, institutional, relationship, and individual levels, and are powerful drivers of GBV and HIV. Female gender roles that assume women's subservience to men, and power imbalances in sexual relations, often sanction the acceptability of violence against women and can increase vulnerability to HIV (Gupta, 2000). Research has shown that women who believe there are justifiable circumstances for a man's violence toward his partner are more likely to experience IPV (Abramsky et al., 2011). Evidence also indicates that women in less equitable relationships are at increased risk of HIV. For example, Jewkes, Dunkle, Nduna et al. (2010) found that inequity in relationship power was an independent risk factor for incident HIV infection among women, even after controlling for partnership duration and other indicators of risk behavior. Some sub-Saharan African countries are characterized by male-dominant cultures, such that violence perpetuated by men against women and children is a socially acceptable means of exerting dominance and control. Given that HIV mainly is acquired through heterosexual relations in sub-Saharan Africa, gender inequality and norms of male dominance are a major force behind the confluence of HIV infection and GBV (Kaye, 2004). For example, rape is sometimes used by men as a means of "punishment" or to exert control over women, including young girls. Such behavior has been described in Namibia and South Africa (Jewkes et al., 2005), as well as in Zimbabwe (Meursing et al., 1995). Further, cultural taboos and social stigma prominent in sub-Saharan Africa severely limit open discussion about sexuality (Ecker, 1994). The combination of a male-dominated cultural landscape and cultural taboos against talking about sex condone and perpetuate GBV.

Evidence of effective community-based interventions
While the body of evidence on effective interventions to reduce GBV is relatively small, results from the rigorous studies that do exist are compelling. For example, the IMAGE Study conducted in South Africa demonstrated that a microfinance program for poor women, combined with gender and HIV training, resulted in a 55 percent reduction in past-year IPV among intervention participants over the two-year study period (Pronyk et al., 2006). The study also found increases in uptake of HIV testing, HIV-related communication in the household, and a reduction in the prevalence of unprotected sex at last intercourse with a non-spousal partner among younger intervention participants (Pronyk et al., 2008). Increases in female empowerment, measured as autonomy in decisionmaking, increased financial confidence, and increased participation in collective action, also were found. SHAZ! (Shaping the Health of Adolescents in Zimbabwe) combined access to sexual and reproductive health services and life skills education with vocational training and grant support for economic activities for adolescent girls and women in urban Zimbabwe. Using a randomized control trial design, the evaluation compared the SHAZ! intervention package to the life skills intervention alone on economic and social empowerment indicators, sexual behavior, and experience of violence. Both groups saw reductions of more than half in reports of physical and sexual violence over the two-year period, but the magnitude of change was greater among intervention participants (Dunbar, 2009). Stepping Stones, a participatory HIV prevention program that aims to improve sexual health through building more gender-equitable relationships, has been shown to reduce sexual risk behaviors and reported levels of perpetration of IPV against women among male participants (Jewkes et al., 2008). The intervention also was associated with a 33 percent reduction in herpes simplex virus -2 (HSV-2) incidence among women and men.

Response services for GBV survivors
The Tanzania MOHSW recently has prioritized improvements and expansion of health services for GBV survivors. This study is especially timely given the September 2011 MOHSW release of the National Policy Guidelines for the Health Sector Prevention of and Response to GBV and the National Management Guidelines for the Health Sector Response to and Prevention of GBV-and the anticipated training on and roll-out of GBV health services later this year.
GBV has a wide range of health consequences. Apart from physical and emotional harm, and stigma and social ostracism, women suffer sexual and reproductive health implications such as unwanted pregnancy, sexually transmitted infections (STIs), and the risk of HIV infection (WHO, 2006). Health consequences of GBV are summarized in Table 1.  Bott et al., 2005. Survivors thus may present with a wide range of immediate or more chronic complaints. Given that healthcare workers are often the first and sometimes only point of contact for GBV survivors, numerous opportunities exist for the health sector to provide care for these survivors. Further, early identification can help with treatment, limit the consequences, and decrease further violence (Laisser, 2011). Proper care often is not provided, however. For example, many healthcare workers know little about how to manage SV clients, and scant attention is paid to IPV. Additionally, healthcare workers are rarely informed about national legislation on GBV or local services to which women could be referred (Mullick et al., 2010). Many healthcare providers fail to diagnose and register GBV. This is due to sociocultural and traditional barriers, but also to lack of time and resources, inadequate physical facilities, lack of awareness, limited knowledge, and poor clinical practices with minimal direct communication. Inadequate training and awareness often result in the failure of healthcare providers to complete a full examination of the survivor or register and monitor the effectiveness and quality of care.
Given the complexity of the service needs and delivery systems, the provision of GBV services cannot be implemented in isolation, but must be considered as part of a systematic approach that extends beyond the facility to the community. For example, many barriers in access to services exist. Study findings from Kenya show that barriers to GBV services include, for example, community misconceptions about the necessity of reporting GBV to the police prior to accessing healthcare; social stigma against GBV survivors; and the threat of further violence, resulting in survivors' nondisclosure and non-pursuit of services. At the facility level, challenges include staff attitudes and general stigma against survivors, as well as insufficient provider capacity (Njuki et al., 2012).
Integrated health services provided within the context of well-coordinated referral networks of social services is a recognized strategy for meeting the unique health needs of women and children (Ferdinand, 2009;Global Health Initiative, n.d.;Women Won't Wait, 2010). While health services and programs may be fragmented, an individual survivor's health and social needs are comprehensive, including the need for multiple types of care (e.g., primary care, psychosocial counseling and support, family planning and sexual and reproductive health services, antenatal and maternal healthcare, child healthcare, among others) and social services (e.g., education, livelihood programs, legal assistance). Strengthening linkages and integration between and among services can increase access, which is a fundamental priority for individuals already facing barriers due to poverty, low social status, lack of education, stigma, discrimination, and GBV (Keesbury and Askew, 2010).
Emerging evidence regarding integrated and linked programs indicates that access to comprehensive services, whether through one-stop centers, co-location of services, or functional referral systems, among other strategies, can produce better outcomes for GBV survivors (Keesbury and Askew, 2010). Training programs for different cadres of healthcare workers, police, and community leaders have been shown to increase individuals' comfort level in addressing GBV, paving the way for survivor-centered services, community-based violence prevention efforts, increased utilization of HIV testing and counseling services, and better adherence to antiretroviral therapy (ART) (Keesbury et al., 2011). Research studies and program evaluations also point to challenges to integration, related largely to the pressure on already overburdened health systems (Keesbury et al., 2011). Healthcare worker shortages, burnout, poor infrastructure, lack of emergency equipment and supplies, long wait times, and inadequate geographic coverage must be addressed not only within the context of HIV prevention, care, and support, but also in broader attempts to integrate services (Keesbury and Askew, 2010;Keesbury et al., 2011).
GBV screening and referral services are also a key strategy of the MOHSW medical management of GBV. In Tanzania, healthcare staff are not well informed and oriented on how to diagnose and offer appropriate management to GBV survivors (URT, 2011). In this regard, a tool for GBV screening among women was piloted by pre-trained hospital staff in Dar es Salaam for about one month. Among the 102 women screened, 78 percent had encountered emotional, physical, or sexual violence. Among them, 62 percent had experienced IPV, 22 percent were survivors of violence inflicted by a relative, and 9.2 percent by a work mate. Staff acknowledged that the tool was useful in assisting them in detecting and assisting abused women, whom they were unable to assist prior to the training (Laisser et al., 2011).

Specific Aims
The overall purpose of the study is to assess the effectiveness of the comprehensive GBV prevention and response program that will be delivered at facility and community levels through the PEPFAR-supported HIV program platform in Tanzania. The study will compare the effectiveness of this comprehensive programming over time with standard practice in control facilities and communities. The GBV program and this study focus on IPV in the forms of sexual, physical, and emotional violence; and on sexual violence among both adults and children, irrespective of the relationship between perpetrator and survivor.
Intimate partner is defined as including current legal spouses, current common-law spouses, current boyfriends/girlfriends (opposite or same sex), former legal spouses, former common-law spouses, separated spouses, or former boyfriends/girlfriends (opposite or same sex). Intimate partners may or may not be cohabiting. Intimate partners may or may not have an existing sexual relationship.
-From Basile and Saltzman, 2002 Two key objectives of the PEPFAR-supported GBV intervention program in Tanzania are "to increase availability, quality, and utilization of GBV services" and "to reduce societal acceptance of GBV and increase protective factors" in support of the goal "to reduce and respond to GBV by leveraging the PEPFAR platform" for "strengthening the national GBV response." In line with these goals, the primary objectives of the study are to assess whether the comprehensive GBV program intervention resulted in an expected 1) increase in care for GBV survivors, reflected in delivery and utilization of GBV services at health facilities, and 2) decrease in experience of GBV among females ages 15-49.
The study has two primary aims to be measured through three primary outcomes. The research questions representing these aims and the outcomes that will be used to answer them are as follows: Did the comprehensive GBV program lead to increased care for GBV survivors? The primary outcomes are Delivery of comprehensive sexual violence (SV) services at Health Centers and Hospitals and Utilization of GBV services at Health Centers and Hospitals. These outcomes will be measured for adults and children, male and female.
Did the GBV program lead to a decline in GBV? The primary outcome is Recent experience of IPV (sexual, physical, or emotional). This outcome will be measured for females ages 15-49 in heterosexual intimate partnerships.
GBV services, comprehensive SV services, and forms of IPV are defined according to the MOHSW management guidelines (URT, 2011).
The study will address an additional set of associated secondary aims. Secondary research questions include the following: For care of GBV survivors, did the intervention lead to any of the following:  Improved quality of GBV services delivered at health facilities?  Increased GBV screening and referrals at health facilities?  More comprehensive GBV care-that is, did survivors receive more of the services they needed?  Changes in referrals to health facilities? In referrals from health facilities to other services?  Reduced barriers to access to services?  Reduced stigma associated with GBV?  Increased knowledge within communities about health services for GBV? About other GBV services?  Increased community actions (and, if so, what actions) to facilitate survivor access to GBV services?
For a decline in GBV, did the intervention lead to any of the following:  Greater decrease in some forms of IPV compared to others?  Decrease in other forms of GBV?  Decrease within communities in the acceptance of GBV?  Increase in community-led actions to stop GBV perpetration?
 An increase in community-led actions to assist those at risk in preventing experience of GBV?  Shift in community norms toward greater gender equality?  Structural changes aimed at preventing GBV?
For HIV-related outcomes, did the intervention lead to either of the following:  Increased utilization of HIV services?  Reduction in HIV risk behaviors?
The study also includes a process evaluation that will characterize the program and highlight lessons learned including the successes and challenges in rolling out a comprehensive GBV program. Key questions include:  What resources were allocated to the various program components?  Did the program meet its objectives and targets?  What was the coverage and rate of roll out?  Were some components easier to implement than others?  Were there changes in program quality over time?
 What are the critical success factors?  What were the greatest challenges and how were they addressed?
The extent to which key external factors influenced the effects of the program also will be measured. These factors include the following:  Urban compared to rural setting;  Seasonal patterns of GBV arising from agricultural cycles;  Alcohol and substance abuse by survivor or perpetrator;  Distance of survivor's household from nearest Health Center or Hospital; and  Availability of GBV support services (e.g., psychosocial support, legal services, and safe houses) outside of health facilities.
The study includes three data collection phases: baseline assessment (prior to the start of the program), ongoing facility and program data collection through the intervention period, and endline assessment (12-24 months after start of the GBV program). Objectives for these components are as follows: Baseline assessment (within the study cluster communities):  To establish the prevalence of experiencce of GBV among women of reproductive age (i.e., ages 15-49)  To assess the knowledge and attitudes towards GBV of community members  To determine the patterns of GBV among women of reproductive age  To establish the availability and pattern of utilization of GBV services by GBV survivors  To establish the quality of GBV services provided to and received by GBV survivors  To determine the HIV risk behaviours among women of reproductive age  To determine utilization of HIV services by women of reproductive age HIV prevalence in Mbeya is estimated to be 11.0 percent for females and 6.7 percent for males-higher than the national prevalence. Prevalence of GBV also is high. More than 67 percent of women in Mbeya have experienced some form of partner violence since age 15; nearly half have experienced physical violence and 28 percent have experienced sexual violence. Prevalence of all forms of partner violence are higher in Mbeya compared to Tanzania overall. Levels of recent experiences of sexual and physical violence (regardless of perpetrator) also are higher. Compared with the country overall, survivors of GBV in Mbeya are more likely to tell another person about the experience but less likely to seek help. More than half of women in Mbeya agree that a husband is justified in beating his wife for various reasons, which is similar to attitudes among women nationally. More than a third of Mbeya women report that husbands demonstrate controlling behaviors-again, similar to national levels. Statistics on violence among younger ages are not available for Mbeya, but national prevalence estimates indicate that nearly one-third of girls and 13 percent of boys experience sexual violence before age 18. These key indicators are presented in Table 2.

Program Interventions
The comprehensive GBV program that this study will evaluate includes a set of five interrelated intervention components, some of which will be delivered at the health facility level and others that will be implemented within communities. The five components are the following:  Facility-based services for GBV survivors;  Facility-based screening and referral in clinical settings;  Clinic and community outreach;  Community-based GBV prevention activities; and  Referrals to psychosocial support, legal services, and safe houses.
It is important to note that this study and its protocol include the evaluation of program outcomes only and not implementation of the program itself. The intervention program will be carried out by PEPFAR-supported governmental and nongovernmental partners, which are not directly involved in the study except through the collection of routine project monitoring and evaluation (M&E) data that will be included in the study analysis. These partners include the Tanzania MOHSW; the Mbeya Regional Medical Office (RMO); and nongovernmental organizations (NGOs) provided with financial and technical support from PEPFAR through the U.S. Department of Defense (DoD)-specifically, the Walter Reed Army Institute of Research (WRAIR) and the Walter Reed Program (WRP) Tanzania.
The Mbeya RMO will be the primary implementer of GBV facility-based services. The facilitybased interventions will be standardized in line with the Tanzanian National GBV guidelines launched in 2011; efforts currently are underway to develop the guidelines training package and accompanying materials, which are directed to doctors, professional nurses, and social welfare officers. The guidelines cover provision of comprehensive care for survivors of GBV, screening for GBV, and the provision of referrals.
The WRP is currently in the process of selecting nongovernmental partners for implementing community-based GBV interventions, including GBV prevention, community outreach, and community-based referrals to services. The intention is to standardize community-based interventions across study clusters and employ an adaptation of the SASA! community mobilization methodology, which focuses on the benefits of non-violence and gender equity to both men and women, promotes discussion of community norms around relationship power and gender inequality, and mobilizes communities to take and promote action and change (Michau, 2008). The program intervention at the community level includes the six components, as follows: 1. Group workshops on gender and GBV for women and men. The program will convene group workshops with women and men of the intervention communities to increase their understanding of how existing gender norms negatively affect their lives, their partners and their families. The evidence-based Men As Partners® methodology and curriculum will be used. This intervention will address male and female attitudes in tandem to promote a more synchronized support for gender equity and other healthy behavior changes. These workshops are expected to lead to raised individual awareness on GBV, reduction in norms of acceptability surrounding GBV, and promotion of respectful, non-violent conflict resolution among workshop participants.

Study Design
The study will employ a pair-matched cluster randomized trial design to determine the effectiveness of the comprehensive GBV program (Hayes and Moulton, 2009). Randomization at the community rather than individual level is appropriate, since the intervention will be delivered at both community and facility levels.
In this study, a cluster is defined as a Health Center or District Hospital and the communities (defined by administrative ward or wards) that the Health Center or District Hospital serves. Clusters will be selected such that sufficient geographical distance exists between them to minimize cross-cluster contamination that might arise if the population within a given cluster utilizes health facilities within a nearby cluster.
A total of 12 clusters will be defined from locations within the Mbeya Region where WRP currently supports HIV activities. These include 20 Health Centers and 17 Hospitals. These facilities are located in all seven administrative districts and Mbeya City. The facility-based intervention also will be implemented at Mbeya Referral Hospital and Mbeya Regional Hospital in Mbeya City because these hospitals serve as referral facilities for the entire region.
The 12 clusters will be grouped into six matched pairs based on matching factors hypothesized to affect the study outcomes. Key factors to be considered include population density (urban, periurban, or rural), client volume at the health facility (i.e., Health Center or Hospital), availability of GBV services outside of the health facility, and geographic size of the health facility's catchment area. Within each pair, one cluster will be randomized to the intervention group and the other will be assigned to the control group. Randomization within each pair will be implemented at the data management center at Muhimbili University of Health and Allied Sciences (MUHAS). Randomization and notification of sites will occur as close to the time of the GBV program roll-out as is feasible.
The intervention will consist of the immediate roll-out of the GBV program at the health facilities and communities within the six intervention clusters. Roll-out of the GBV program for the six control clusters will be delayed by approximately 20 to 24 months.
Primary outcome measures will be obtained over time from both intervention and control clusters. A cross-sectional household survey will be conducted within all 12 clusters at baseline (prior to initiation of the intervention) and at endline (24 months later). Utilization of health facility services will be tallied from health facility GBV registers on a monthly basis and captured quarterly during the course of the study. Secondary outcomes also will be assessed through these and other data collection methodologies and sources, as described in Section 3.7.

Study Outcomes
The study will include primary and secondary outcomes hypothesized to form a causal pathway from program interventions to shorter-and longer-term GBV and HIV-related outcomes. The conceptual framework summarizing these outcomes and pathways is presented in Exhibit 1.

Exhibit 1. Conceptual framework
In brief, strengthening of facility-based services for GBV survivors and screening and referral are expected to lead to an increase in the availability and quality of GBV services at health facilities. Increased availability and quality of services, in combination with outreach and referrals to/from service entry points outside of health facilities, are expected to improve access to GBV services. At the same time, all program components are expected to lead to increases in community knowledge of GBV and community actions to reduce GBV. These intermediate outcomes then collectively are hypothesized to result in improved utilization of GBV health services, decrease in community acceptance of GBV, shift in community norms toward greater gender equality, and decline in GBV experience and perpetration. These changes in GBV outcomes also are hypothesized to lead to positive HIV-related outcomes, including increases in utilization of HIV services and reduction in HIV risk behaviors, both in the near and longer terms.
Below are the three primary outcomes on which the study will be powered:  Recent experience of IPV (sexual, physical, or emotional) among females ages 15-49 years;  Utilization of GBV services at Health Centers and Hospitals (by females and males, all ages); and  Delivery of comprehensive SV services at Health Centers and Hospitals.
A summary of measurements for these three primary outcomes, together with data collection methods, target population or data source, and timing of data collection, are presented in Table 3. Outcomes and measurements for secondary outcomes of the study are presented in Table 4.

Household surveys of females ages 15-49 years
The study team will conduct a cross-sectional survey of 100 households within each of the 12 clusters at baseline (prior to initiation of the intervention) and approximately 24 months later (at endline). For each household, the team will identify eligible members and select a single member at random to complete the household survey through a one-on-one interview. An eligible member is defined as a female ages 15 through 49 years at the time of the survey. This age group has been selected to represent those at high risk of GBV and to facilitate comparison of results with previous studies, such as the TDHS.
The study team will use a multistage stratified random sampling process to sample households. Within each of the 12 study clusters, it will obtain a list of households from the most recent census and generate a random sample of 130 households. The team will oversample by 30 percent to ensure that approximately 100 households are identified in which at least one female between the ages of 15 and 49 resides and agrees to participate in the survey. A household selection form similar to that employed by the DHS will be used to ascertain whether the selected household has any eligible members (see Annex B-1). When the team identifies more than one eligible household member, one will be chosen randomly for the interview, with no substitution if that person refuses to participate, or if parental consent is not given in the case of minors who have been selected for the interview. If an interview is not completed on the first visit, team members will make further attempts with the sampled household or respondent up to three times over three different days before classifying the case as non-response. The team will schedule subsequent contacts at times when the respondent is more likely to be at home.
The sample size of 100 household surveys per cluster for 12 clusters is based on the formula for a pair-matched cluster randomized trial with a binomial outcome given in Hayes and Bennett (1999). Table 5 shows the minimum percentage reduction in IPV and GBV after 18 months of delivery of the comprehensive GBV program that can be detected with 80 percent power using a two-sided 0.05 level test for varying within cluster pair coefficients of variation (km).
In the table, estimates for the proportion of women experiencing IPV and GBV in the past 12 months in the control clusters are based on the Mbeya findings from the 2010 TDHS. Based on data from trials such as SASA! (Abramsky et al., 2012), the coefficient of variation is expected to be in the range of 0.1 to 0.4. If the coefficient of variation is close to 0.1, then percentage reductions in IPV and GBV on the order of 30 to 40 percent should be detectable. However, if the coefficient is closer to 0.4, then only large reductions on the order of 65 percent will be detectable. Questions on the occurrence of GBV are based on questions used in the 2010 TDHS, in which respondents were asked about their experiences of specific acts from partners and non-partners without reference to leading terms such as 'abuse' or 'violence.' The target population for the household survey was chosen to reflect the population most likely to have experienced genderbased violence-women ages 15-49 living within the defined study clusters. The sampling frame thus comprises households situated in all enumeration areas in which participants (in intervention clusters) or passive volunteers (in control clusters) live.
While the study would benefit from the inclusion of household survey information from men, they are not included for the following reasons. Specifically,  A separate survey instrument would need to be developed for males as the questions that would be asked would differ from those asked of females. This would require additional effort and time beyond that which is needed for development of the female questionnaire as most of the GBV-related questions in the female survey are taken from existing, tested survey instruments. These survey instruments do not exist for men. Thus, more extensive development and pre-testing would be required for the male questionnaire.  Separate sampling frames would be needed for women and men. Given the sensitive nature of the topic and ethics and safety considerations, men and women cannot be sampled from the same or nearby households.  Additional numbers of individuals would need to be sampled in order to achieve the current estimated levels of statistical power, which are based among other factors on estimated prevalence of GBV among women. In other words, the current sample could not simply be distributed among males and females and still maintain the same statistical power of the study.  A separate data collection team would need to be employed and trained as female interviewers could not interview males (and vice versa), and the questionnaires would differ.
Given the resources available for the study and the fact that prevalence of GBV in Mbeya is significantly higher among females compared to males among all ages, the experience of females was chosen over that of men as the focus of the household survey. Although the study will not collect data from males in a household survey, the facility-based component will collect data on experience of GBV among both males and femalesand of all ages. Additionally, data on gender norms and attitudes will be collected from male participants in the GBV program through key informant interviews (described in the following section).

Key informant interviews with community leaders/gatekeepers and male participants of the prevention and community programs
The team will conduct all key informant interviews in private settings with no interference, so that participants can communicate freely. Interviews will be audio-recorded, and written notes will capture the information shared. The team will inform participants of the modalities for capturing information before the onset of data collection.
Two groups of key informants will be interviewed. The types of information that the team will collect from these groups are described below.

Community Leaders/Gatekeepers
Key informant interviews (KIIs) will be conducted with government officials and community leaders and gatekeepers, including ward leaders, police, religious leaders, traditional leaders, leaders of women's NGOs, managers/providers of GBV services outside of health facilities, and others. The purpose of the KIIs is to gather opinions about these individuals' experiences in handling GBV cases, their own attitudes toward GBV, and whether they believe attitudes have changed in the community during the study period. They also will be asked about the strengths and weaknesses of the GBV program, their suggestions for ongoing improvements, and any community-led actions to address GBV that came about due to implementation of the program.
The study team will conduct approximately 20 KIIs within the intervention clusters during the second half of the study period. The interview guide is provided in Annex B-2.

Participants (female and male) of the prevention and community programs
The team will conduct qualitative key informant interviews (KIIs) with females and males who have participated in the prevention and/or community-based programs within the six intervention clusters. The purpose of these interviews is to learn about participants' experiences with the program, their personal experiences with GBV, how their program involvement may have influenced their attitudes and behaviors, and their suggestions for improving the GBV program. Results of these interviews will be pooled across clusters and used to help interpret the quantitative findings from the household surveys with women, and the data from the health facilities.
The study team will work with program implementers to select a purposeful sample of females and males aged 18-49 years old who are participating in, or who have been exposed to the community and/or prevention programs. In order to be eligible for selection, they will have participated in at least one program component delivered as part of the community program. Efforts will be made to select interviewees who represent a wide range of women and men in the community, for example, women and men across the selected age range, married and unmarried, and women and men with different educational levels and professional backgrounds.
The study team will conduct the KIIs during months 12-14 the study, before and separate from the when the team conducts of the end-line household survey for women. By interviewing participants at this time point, we will be able to select participants who have been a part of, or exposed to, the program for enough duration that changes could have taken effect. The study team will verify selected individuals' participation in the program(s) with program implementers (following the criteria noted above for selection) and also record the intensity and length of their program participation.
Interview topics will include: • Knowledge of the comprehensive GBV program • Awareness, norms, and attitudes about gender and GBV • Response to GBV • Experience of GBV • Knowledge of and experience with GBV services Questions will focus on those related to the individual's perceptions and self-reported changes in his knowledge, perceptions, attitudes, and behaviors as a result of the program. Interviewees also will be asked to describe changes they have seen in the community that also may be related to the GBV program.
The team will conduct a total of 48 interviews (8 per each intervention cluster). The interview guide is provided in Annex B-3.

Health facility GBV registers: Data extraction and review
The national GBV register, as developed by the MOHSW, will be adapted and implemented at each of the health facilities in both intervention and control clusters. Adaptations to the existing register, which currently collects information on socio-demographic data, type of violence, and some elements of care provided, will include additional information on date and time of GBV act, type of GBV experienced, information on HIV testing, and data reflecting where the client was referred and any referrals provided. The GBV register is provided in Annex B-4.
Healthcare providers at the health facilities will complete the GBV register on an ongoing basis and fill it in for all GBV clients screened (adult and child, male and female) and treated for GBV at the health facility. GBV registers from the various health facility departments will be collected and stored in a central location at each facility. The registers will be printed in duplicate form. One copy will remain at the health facility and the study team will collect the other copy. Registers from all health facilities will be collected on a monthly basis and sent to the study office at MUHAS in Dar es Salaam for data entry. The register forms will be produced so that the study team's copy will exclude client-identifying information.
Training for providers on the use of the registers will occur as part of GBV service delivery training. The study team, in collaboration with the RMO, WRP, and PEPFAR-supported partners providing technical assistance for GBV M&E, will provide additional training in the month prior to the start of the study. Participants will include healthcare providers from facilities in the control clusters. The study team, in collaboration with the M&E officer at the WRP, will provide periodic technical support to ensure completeness and accuracy of data recording in the GBV register.
Data collection using the Tathmini GBV register will continue for 16 months, at which time facilities will transition to the national GBV register. For months 17-24 of the 24-month evaluation period, we will use national register data reports submitted by facilities through the Health Management Information System to triangulate facility and household survey endline data.

Health facility assessments
The study team will conduct a health facility assessment at the Health Centers and Hospitals in the study. This assessment aims to document health facility readiness and capacity to deliver GBV services as stipulated in the 2011 Tanzanian national GBV guidelines. The data will be used in the evaluation to assess service quality and its effects on study outcomes. The team will conduct assessments at baseline and 14-16 months after the start of the comprehensive program.
The health facility assessment will collect information on infrastructure, supplies and equipment, staff training, and provider knowledge and attitudes. Data will be collected through a combination of surveys of facility heads, observation at facilities, and interviews with staff involved in the provision of GBV services at point of first contact.
The assessment will include two components:

1) Audits of health facilities through surveys of facility heads and study team observation
For each health facility, the team will identify facility heads and obtain verbal consent to conduct the assessment and interview him or her. The team also will conduct survey interviews with the facility head at each Health Center and Hospital in the study clusters. The questionnaire is provided in Annex B-5. This survey is based on the tool that the MOHSW and PEPFARsupported clinical partners developed to assess facility readiness to provide GBV services. It also includes a standardized observational checklist widely used by other researchers for similar purposes (Christofides et al., 2006).

2) Key informant interviews with service providers
At each facility, the team will interview one or two key staff involved in the provision of services related to GBV and selected by the facility manager. Interview topics include provider knowledge, skills, attitudes, and working environment. Also included will be general understanding of GBV, perceptions and attitudes toward provision of GBV clinical services, training skills related to STI management, skills on provision of emergency contraception, and other topics. The provider's knowledge of the service delivery guidelines, attitudes toward providing services, and practices will be assessed using a structured questionnaire. The team also will ask open-ended questions to explore challenges and achievements at the facility level. The interview guide is provided in Annex B-6.

Key informant interviews with program managers
The study will include a process evaluation to characterize the roll out and implementation of the comprehensive program over the 12-month study period. Key informant interviews with program managers will be the primary data source. The key informants will be selected from two types of managers: 1) GBV program managers at WRP and 2) community program managers who are sub-partners of WRP. This information will be supplemented with information from the health facility assessments (e.g., interviews with facility managers and service providers) to describe the roll out, ongoing implementation, successes and challenges of all components of the comprehensive program.
Key informants at WRP will include the overall program manager, the clinical services manager, the community program manager, and the M&E officer. Two representatives from each of the sub-partner community program organizations will be interviewed. All will be interviewed at 4-6 months and again at 12 months. Interview topics will include:  Roles and responsibilities  Description of the overall program or component and workplan  Resources allocated/utilized  Progress vis a vis plans  Implementation successes  Implementation challenges  Corrective actions taken  Recommendations for program improvement The interview guide is given in Annex B-7.

Routine M&E Data Collected by Organizations Implementing the Programs
Implementing partners will collect routine monitoring data, including three indicators being piloted by PEPFAR under the GBV initiative. Two of the indicators relate to facility-based services and are compatible with primary and secondary facility outcomes in this study. 3 The third indicator, which captures community coverage, is of particular importance and is as follows:  Number of people reached by an individual, small group, or community-level intervention or service that explicitly addresses GBV and coercion This indicator will be used to determine coverage and confirm and document study assumptions that community interventions are operating at scale (and if they are not, to inform the interpretation of our study outcomes).
The study also will draw from any project evaluation studies to be conducted by implementing partners to evaluate specific components of the comprehensive program model. The results of these evaluation studies will be used to interpret and triangulate the findings of the outcome evaluation.
All data collection activities involving human subjects are described in this protocol. Additional M&E activities carried out by WRP as part of this protocol will not involve human subjects. Rather they will comprise routine program data monitoring (e.g. document the number of workshops carried out, and the number of participants, etc.), but will not include the collection of research data from these participants. The study team will work closely with the WRP to coordinate M&E activities throughout the implementation of the study. At the design and planning stages, routine communications have been established between the two parties and opportunities for M&E collaboration have been identified. In particular, at the facility level, where GBV register data will be reported through RMO (and WRP) structures in addition to being collected by the study team, arrangements are being made for collaborative data collection and capacity building of implementing partners. Similarly, WRP and Tathmini GBV will collaborate in the conduct of facility assessments. Similar arrangements and agreements will be developed for data collection and use on the community intervention side. Specific roles and responsibilities will be detailed in the operations manuals for the study. We will explore options for formalizing these arrangements, for example, through a memorandum of understanding.

Data Analysis
For the baseline household survey, all demographic variables will be summarized by intervention and control clusters. Demographic variables will include age, marital status, number of children, education, recent employment, and alcohol usage. For women who have ever been married or lived with a male partner, the education, employment, and alcohol usage of the most recent husband/partner will also be summarized. Those variables for which a significant imbalance occurs between intervention and control clusters will be identified and used in an adjusted analysis of the primary endpoint drawn from the endline household survey.
A logistic regression analysis of the matching factors on the proportion of women reporting IPV in the past 12 months will be performed with the data from the baseline household survey. The multiple correlation coefficient R 2 , which estimates the percentage variation in the outcome variable explained by the matching factors, will be calculated. If the estimated R 2 exceeds 0.5, a matched paired t-test will be utilized in the analysis of the primary endpoint from the endline household survey. Otherwise, an unmatched t-test will be used to assess the reduction in physical or sexual violence due to the intervention. Breaking the matches when the matching factors are poor or weak predictors of the outcome should increase the statistical power of the study (Diehr et al., 1995). Since the multiple correlation coefficient is based on the baseline survey, not the endline survey, this approach should not inflate the Type 1 error probability.
For the endline household survey, the primary analysis will be based on the intention-to-treat principle, in which all surveys are included in the analysis regardless of compliance with the study protocol. Additional analyses, in which surveys are excluded for reasons of non-compliance, including lack of exposure to the intervention, will be performed as part of a sensitivity analysis.
Depending on the degree to which the matching factors predict the primary outcome, either a paired t-test, in which the matching is preserved, or a two-sample t-test, in which clusters are unmatched, will be used to assess the association between the intervention and report of IPV in the past 12 months. The appropriate t-test will use cluster summaries of the proportion of women reporting IPV and, for the primary analysis, will not adjust for imbalances in demographic variables or other potential confounders. The estimated difference in the proportion of women reporting IPV between the intervention and control groups, along with its two-sided 95 percent confidence interval, will be reported. Also, the Mantel-Hanzel estimate of the common odds ratio and its two-sided 95 percent confidence interval will be provided.
As part of a sensitivity analysis, an analysis in which an adjustment is made for the imbalance in demographic variables will be conducted using the two-stage procedure suggested by Hayes and Moulton (2009). Briefly, a logistic regression model is fit to the individual-level data, in which the explanatory variables are the demographic variables for which the baseline survey found an imbalance. No intervention effect is included in the model. For each individual, her predicted outcome is determined. Then, for each cluster, the predicted number of outcomes is obtained by summing over the individuals in the cluster. The cluster-level difference residual Rd is calculated as the observed number of outcomes minus the predicted number divided by the number of surveys completed in the cluster. The adjusted t-test is then performed on the difference residuals. Hayes and Moulton further provide a formula for calculating a two-sided 95 percent confidence interval for the adjusted difference in the proportion of women reporting GBV between intervention and control groups.
Linear mixed effects (LME) models will be used to evaluate the utilization of GBV services at health facilities over time. From the monthly GBV registers maintained at each facility, the number of encounters at the facility in which a person reports GBV will be tabulated and used as the outcome variable. The number of encounters will be broken down further by sex, age group (i.e., <15, ≥15), and type of violence experienced (i.e., physical, sexual, or emotional) by the person reporting GBV. Fixed explanatory variables will include study month, type of facility, and an indicator variable for whether or not the facility is in an intervention cluster. Study month will be treated as a categorical variable to assess the seasonality of GBV as well as the impact of the availability of GBV services on utilization over time. Other explanatory variables will be explored, such as staffing levels and the number and mix of health facilities in a given cluster. A random effect for each facility will be included in the model to account for correlated outcomes within a facility over time. Suitable linear contrasts will be used to estimate and test the significance of the difference in GBV utilization between intervention and control clusters as well as the change in utilization between baseline and the end of the study.
LME models also will be used to assess the difference between intervention and control clusters as well as the change over time in the percentage of persons reporting SV for whom comprehensive services were delivered according to MOHSW guidelines. Since comprehensive services for SV survivors are not available at Dispensaries, these models will be restricted to GBV register data from Health Centers and Hospitals. The number of visits in which a person reported SV for a given study month and facility will be tabulated and treated as an offset variable. The number of visits in which comprehensive services were delivered according to MOHSW guidelines will be the outcome variable. Such models will allow an assessment of the proportion of SV survivors receiving comprehensive services by the end of the study.
For SV survivors seen at Health Centers or Hospitals, an additional score will be derived for each survivor that reflects the number of services received. This score will be used as an outcome variable in suitable LME models to assess the difference between intervention and control clusters as well as the change over time in the level of services provided to SV survivors.
The analysis of secondary endpoints from the household survey, GBV registers, key informant interviews, facility assessments, and services mapping will be provided in a separate and more detailed data analysis plan.

Data Management
The data will be housed in the project office in Dar es Salaam. The office has Internet, fax, phone, photocopy, printing, and computer facilities available to staff engaged on the project to allow for project monitoring. The office is also equipped with locked data storage cabinets to ensure the safety and security of the information collected from the field.
Data collection forms from the household survey will be reviewed in the field by the interviewers and a data collection supervisor for inconsistencies, errors, and missing data so as to immediately flag potential data quality issues. Data from the field will be transported in safe boxes to the project office. There, the study team data managers will review and code the data, conduct double entry of all the quantitative data, and conduct initial data cleaning simultaneously with data entry.
Recorded key informant interviews will be transcribed verbatim within two to five days after collection to keep abreast with what occurred during data collection. The transcribed text will be translated from Kiswahili into English by bilingual project staff members skilled in transcript translation. A sample of the translated material will be back-translated to Kiswahili by an independent qualitative skilled translator to check for accuracy and consistency. This information then will be entered and processed for analysis using NVIVO software. 4 Back-ups of audio tapes and transcribed data will be produced and stored separately in the secured storage cabinets.
Backup of electronic data will be done whenever newer entries are completed. Electronic data will be stored both on computer and external hard disk drives. Project computers will have log-in passwords, and external hard disk drives will be password protected and kept in locked cabinets. Data forms will be stored in locked cabinets. Only members of study team will have access to the data. Standard operating procedures and data quality checks will be developed and all data collection and data management team members will be trained on these procedures. The team's study coordinator will ensure compliance with all procedures.

Risk and Benefit
The proposed study will use a robust design to evaluate a comprehensive GBV program model of facility-and community-based interventions to provide information on the validity of the conceptual model and contribute new knowledge to the GBV program field. Results of the study will be used by the Tanzanian MOHSW; the Ministry of Gender, Children and Community Development; and other national policy and program actors to inform and strengthen GBV prevention and response services scale-up in Tanzania. The experience of monitoring community-and facility-based activities and outcomes in study clusters will strengthen M&E capacities at the regional, district, and national levels. This study also is designed to provide evidence to a broader set of global stakeholders, including PEPFAR; the Global Fund to Fight AIDS, TB and Malaria; other national governments; and other multilateral organizations on the impact of programs responding to GBV. These programs ultimately benefit adults and children most at risk of both GBV and HIV.
All studies on GBV involving human subjects pose potential risks for study participants. For example, asking individuals to disclose difficult and painful experiences may create personal distress, and the act of research itself may put individuals at further risk of abuse. A respondent may suffer physical and/or emotional harm if a partner finds out that s/he has been talking to others about their relationship. Because many violent individuals closely control the actions of their partners, even the act of speaking to another person without permission may trigger violence. On the other hand, asking respondents to reveal stories of trauma can be a transforming experience for both the respondents and the interviewer. The experience of the study team and other researchers indicates that many women welcome the opportunity to tell their stories if asked in a sympathetic, nonjudgmental manner. Interviews also provide an opportunity to provide participants with information on GBV. The need to establish rapport with the respondent to ensure confidentiality and privacy during the interview is important for the entire survey, but is also critical for ensuring the validity of the data on GBV. Complete privacy is also essential to ensure the security of the respondent and the interviewer. Asking about or reporting violence, especially in households where the perpetrator may be present at the time of interview, may carry the risk of further violence.
The study team acknowledges its ethical responsibility to minimize these risks and maximize the benefits, and will conduct this study accordingly. Specifically, the study team will adhere to WHO guidelines regarding the conduct of research on violence against women (WHO, 1999). In this regard:  The study team will consider the safety of respondents and the team as its highest priority in all project decisions.  The proposed study design is methodologically sound and builds upon the research experience of the team and lessons learned from other researchers in the GBV field to minimize under-reporting of GBV.  Confidentiality of participant identity and data will be safeguarded to ensure both participants' safety and data quality.  All current research team members are experienced in GBV research and have undergone trainings in research ethics; all additional team members will be carefully selected, certified in research ethics, and receive specialized training and ongoing support.  The study design includes a number of actions aimed at reducing any possible distress caused to the participants by the research.  If for any reason privacy cannot be assured the interviewer will be trained not to proceed with the interview on GBV.  Fieldworkers will be provided with information on available local GBV support services and resources, and trained to refer study participants for assistance.  The study team will develop and implement a dissemination plan that will help to ensure that findings are properly interpreted and used to advance policy and intervention development within Tanzania and globally.
The study team will inform participants that their participation will assist the team in learning how to provide services that can improve the health and well-being of communities in Mbeya and throughout Tanzania. Participants also will be told that they may not benefit directly from their participation in the study and that no monetary compensation will be provided to them. The team also will inform participants that some of the questions will cover intimate topics and may be embarrassing or make them feel uncomfortable. Participants will be advised to respond to questions according to their own comfort level. All study participants will be advised prior to participating in data collection that they have the right to withdraw from the study at any time or refuse to answer any questions with which they may feel uncomfortable.
All interviewers will be of the same sex as the respondent and trained in ethical procedures, the informed consent process, and confidentiality. Training will include specific considerations for research on GBV using existing guidelines (Ellsberg and Heise, 2005). Above all, interviewers will be trained to ensure safety for the participant and the highest level of privacy. Interviewers also will be trained to stop the interview and offer the client a referral to services if the participant is visibly distressed. The study team will comprise interviewers from areas outside of Mbeya to ensure that participants are comfortable in discussing personal issues, and to minimize the possibility of a breach of confidentiality.
Interviewers will be advised that discussion of sensitive issues such as GBV may cause discomfort for them as well. Although GBV-related issues have been addressed widely both in the media and research findings, hearing first-person accounts may evoke strong emotions during some interviews. Members of the study team will be provided with the opportunity to debrief each other regularly during the research process and will be provided with details on how to seek counseling should a team member wish to do so. Two of the study co-investigators are trained and practicing mental health specialists, and are available to provide counseling support to team members.
During the initial design phase of the study, the study team consulted broad groups of stakeholders at the national and Mbeya regional levels regarding the proposed study outcomes, design, and conduct. Stakeholders included among others Tanzanian government officials and leads for the roll-out of GBV services in health facilities, program implementing partners, community and religious leaders, policymakers, health care providers, and U.S. government agency and multilateral donor representatives. The study team will establish a Technical Advisory Group that will serve to assist the team in addressing technical issues that may arise during the conduct of the study and provide comments on study results and interpretation. Approximately six representatives Steps that will be taken to minimize harm to study participants  Only one woman per household will be interviewed in the survey.  The wider community will not be informed that the household survey focuses on GBV.  Male key informants will not come from the same households in which the survey has been conducted and women have been asked about violence.  All interviews will be conducted in complete privacy.  Dummy questionnaires may be used if others enter the room during the interview.  Interviewers will be trained to recognize and deal with a respondent's distress during the interview and provide referrrals, as appropriate.  All interviews will end on a positive note that emphasizes the respondent's strengths.
from among stakeholders consulted earlier will serve on the TAG. The study team also will liaise with Community Advisory Boards (CAB) in Mbeya, established by WRP, to serve as community monitor for the study and to build support for the study within the cluster communities. The study team will meet with CAB representatives during quarterly monitoring visits establish routine communication mechanisms including email for more frequent updates on study progress and feedback from the CAB.

Confidentiality
No individual GBV survivors will be targeted directly in this study. None of the data collected, including GBV registers from health facilities, household survey questionnaires, key informant interview notes, and audio tapes, will include personal identification information. Participants and records of cases will be identified only by record numbers assigned at the facility level or by a study number assigned by the research team.
To minimize the risk of breach of confidentiality, the team will conduct household survey interviews in a private location within the household setting; no partners or other family members will be present. The team will not record names on the questionnaires or other data collection forms. The team member will give each participant a study identification number that will be used in place of his or her name on the interview form. Only study team members will have access to the questionnaires and data collection forms. The questionnaires and forms will be kept in a locked cabinet at the MUHAS offices in Dar es Salaam. The team will inform those participating in key informant interviews that the discussions may be recorded, and that the recordings will be used for research purposes only and destroyed once the information has been transcribed.
Participants will be asked to acknowledge the possibility that an interview may be requested by a representative of the trial's sponsor to determine whether they gave informed consent to the study team. If the sponsor requests an interview, participants will be informed that they have the option to accept or refuse the interview.
Interviewers having at minimum bachelor's degrees will be trained intensively to develop and refine research skills over a period of five days. This training will include (1) understanding the concept of GBV, its relationship with HIV, and its effects; (2) developing and improving interviewing skills, both quantitative and qualitative; (3) understanding special considerations in conducting research on GBV and applying skills; and (4) understanding and practicing on the specific research tools for the study. Furthermore, staff will undergo a research ethics course and certification test. The training will highlight the topic of confidentiality and its importance.
The data collection team also will be trained to provide participants with information about GBV services in their communities and to refer participants, if the participants so choose. A clusterspecific list of GBV support services will be developed from the services mapping activity. This list will be given to all study participants. In accordance with GBV research best practice and approaches previously employed in Tanzania (e.g., in the VAC survey), the list of services will also include services not directly associated with GBV (e.g., malaria, HIV/AIDS, alcohol use, etc.). This will be done to protect the privacy and safety of participants and minimize potential stigmatization and/or harm from others.
This list of services also will be used by the study team to refer participants to services, as needed.
Following the guidelines used in the VAC survey, if a participant becomes upset during the interview the interviewer will offer to place the participant in direct contact with a service provider located near the participant. (Interviewers will be trained to recognize signs of anxiety and/or agitation during the interview process and to follow these steps). Further, for all participants who report experiencing GBV in the past 12 months, interviewers will be instructed to ask participants whether they currently feel safe in their living situation. If a respondent indicates that she does not feel safe, the interviewer will offer to place the participant in direct contact with someone who can help (a service provider). If a participant indicates that she would like a direct referral, the interviewer will ask permission to obtain her contact information, including her name and a safe place (way) where a service provider can find her. The interviewer will record the contact information on a separate form, which will not be linked in any way with the interview form. The interviewer will then give the contact form to the data collection team leader, who will identify an appropriate service provider from the directory list and deliver the form directly to the provider. Information shared by the participant during the interview will not be shared with the service provider.

Compensation
Study participants will not be paid for their participation in the study. It is anticipated that household interviews will be conducted at or near participants' homes. However, participants who may need to travel to participate in key informant interviews may be provided with fare for public transport and refreshments.

Informed Consent
For all data collection methods involving interactions with participants (i.e., the household survey, key informant interviews, and facility assessments), the study team will inform participants about the objectives of the study and ask them for written informed consent prior to participating. For study participants ages 15-17, parental consent also will be required. A participant's autonomy to choose whether or not s/he wishes to be interviewed will be respected. The informed consent process will include a description of the purpose of the study, procedures to be followed, possible discomfort and risks, benefits, compensation, and confidentiality. The team also will tell participants that no monetary compensation will be provided for participation. Team members will obtain informed consent in the local language (Kiswahili).

Household surveys
The sample for the household surveys includes both minors (ages 15-17) and adults (ages 18-24). Parental/guardian consent will be obtained for the minors. The team will not interview nor obtain any other data directly from minors without their prior assent and parental/guardian consent. Three types of consent forms will be used for the household surveys: 1) Informed consent for females ages 18-49, 2) Minor assent for females ages 15-17, and 3) Parental consent for females ages 15-17.
For the household surveys, the consent forms will describe the program that is the focus of the study as aiming "to improve the health and well-being of women and children." They will not include specific reference to GBV. The potential sensitivity of questions that will be asked, however, will be clearly communicated in the consent form. Language will be included that informs participants they are free to not answer the questions or to end the interview without penalty or loss of any benefits to which they are entitled. This approach is based on the following:  The absence of consensus within the global GBV research community on the explicitness of violence language that should be used in the consent process (see Ellsberg and Heise, 2005;WHO, 1999). In particular, many researchers argue that it is sufficient (and advisable) to include only language that communicates the sensitivity of topics that will be raised. Key reasons cited in the literature for excluding explicit GBV language include protection of the participant from potential harm inflicted by others as a result of others more easily becoming aware that questions on GBV will be asked; and, poor validity of the study findings resulting from potentially high rates of refusal and bias in the response sample (Scholar et al, 1995). Specifically, several studies reviewed note both child and adult participants may be vulnerable to further abuse or retribution as a consequence of participating in the research. Therefore, it may not be in the best interests of participants for the researchers to be explicit about the nature of the research. The literature review report concludes that terminology used in the information provided may in itself affect levels of anxiety, and sometimes labeling certain descriptors, for example labeling experiences as "domestic violence" or "abuse," may be stigmatizing. Other researchers have noted that given that a significant proportion of women and children are abused by people they know, the assumption that parents have the best interests of the child at heart cannot be assumed in VAC research as the parent may be abusive or have conflicting allegiances between the child and another abusive adult (Knight et al., 2000).
 Advisement by the study team's locally based co-PIs and MUHAS research ethics advisors to follow the practice previously approved and carried out in Tanzania for similar studies. These include the WHO multi-country violence against women (VAW) study (WHO, 2004), the 2009 Tanzania Violence Against Children (VAC) Study (UNICEF et al, 2011), and the 2010 Tanzania Demographic and Health Survey (Tanzania DHS, 2010). In all three of these studies, which included large scale surveys, the consent forms did not include any reference to violence or GBV. Rather, they utilized an oral consent process that referred to the survey as a study on women's (or children's) health. Additionally, the WHO VAW and UNICEF VAC study consent forms included language to indicate the studies also were interested in learning about "life experiences." The WHO VAW study consent also noted, "Some of the topics discussed may be personal and difficult to talk about, but many women have found it useful to have the opportunity to talk." In all three surveys, participants were told that they could end the interview at any time or skip any question they did not want to answer. Other smaller studies on VAC in Tanzania have followed a similar approach.
The study team will use a two-stage "consent" process for the household surveys. The first stage comprises the consent process described above and signing of consent forms prior to the interview. The second stage involves introducing the concept of GBV immediately prior to asking questions about GBV experience and informing participants again about the sensitive nature of the questions, that they can end the interview at any time, or they can skip any question they do not want to answer. This approach was used in the WHO VAW Study and the Tanzania DHS. Researchers have noted that this approach allowed for rapport to develop before introducing the topic of violence, but still gave the participant the opportunity to opt out of the violence-related questions at the time they were introduced. These studies noted that this consenting process was well received by survey respondents in Tanzania.

Key informant interviews
Key informant interviews will be conducted with adults (ages 18-49) only. Since all key informants (i.e., community leaders/gatekeepers and male participants of the prevention and community programs, and health care providers at the study sites) will have some familiarity with the GBV program, the consent forms include language that conveys that the purpose of the study is to measure the effects of the GBV program. Like the household survey consent forms, the key informant consent forms include language on the potential sensitivity of the issues and the participant's choice to not answer the questions or to end the interview. Within the interview just prior to asking sensitive questions, the interviewer will remind the participant that s/he is free to not respond.
There are two key informant consent forms, one for community leaders/gatekeepers, participants of the prevention and community programs, and program managers; and the other for health care providers who will be interviewed as part of the facility assessment. The forms differ only in the language within the section, "Why have I been invited to take part?" and reflect the different perspectives sought from these groups.

Health facility GBV registers
The team will not request consent from individuals for data collected through the health facility GBV register. The facility-based services provided as part of the intervention are those included in established MOHSW national guidelines. Data for the study will be extracted by the service provider from routine client records that are maintained separate and apart from the study. Data forms from the registers collected by the study team will not include client names or other personal identification information.

Consent forms
Consent forms are provided in Annex A as follows: Annex

Questionnaires
Questionnaires and other data collection tools to be used in the study are included in Annex B.

Ethical Training Certification
All principal investigators and key study personnel have completed a research ethics training course and received certification. These certifications were received within the past three years. Certificates for these individuals are provided in Annex C.

Introduction
You are invited to take part in a research study. Before you decide whether to participate, you need to understand why the research is being done and what it will involve. Please take the time to read or to listen as I read the following information. You may talk to others about the study if you wish. Please ask me if there is anything that is not clear, or if you would like more information. When all of your questions have been answered and you feel that you understand this study, you will be asked if you wish to participate in the study, and if yes, to sign this Informed Consent form. You will be given a signed copy to keep, if you would like one.

Purpose of the Study and Study Requirements
What is the study? The study is being conducted by a team of researchers from Muhimbili University of Health and Allied Sciences (MUHAS) in Dar es Salaam and the Futures Group, Population Council, and Pangaea Global AIDS Foundation in the United States. The overall purpose of the study is to measure the effect of a set of programs and services that will be delivered in health facilities and in the community to improve the health and well-being of women and children. The study will compare data from communities where these programs and services are being delivered with data from communities that aren't yet receiving these services to determine if the program had the desired effect.
The study has been funded by the United States Government.
Why have I been invited to take part? You have been invited to take part in the study because, as part of the data collection, we are visiting a number of households in the area where you live to interview one individual in each selected household. Your household has been selected by chance from a list of all households in your area. The information you give us regarding your knowledge and experiences with the services provided in your community will be important to understanding how these services are delivered and the effect they are having in the community. In addition the information you provide will assist the study team to understand what works and does not work in order to strengthen services in Mbeya and other regions.
What will happen if I take part? If you agree to take part in the study, we will ask you to sign this form indicating that you agree to be part of the study. An interviewer will sit with you and ask you a series of questions. The interviewer will record your responses on a form. The information that you provide will be kept completely confidential. Your responses will be combined with those of other participants in the study. Together this information will be analysed to determine if the programs and services are making a difference.
How long will the interview last? The interview will take between 60 and 90 minutes.
We may contact you again if we need to obtain further information or clarifications regarding the information you provide us with today.

Risks
What are the risks of the study? If you decide to participate, you may be inconvenienced by the time and effort that it will take to complete the interview. You may find one or more questions that we ask to be sensitive. You do not have to respond to any question that makes you uncomfortable. You may end the interview at any time without penalty or loss of any benefits to which you are entitled.
Breach of confidentiality (that is, something you say in confidence during the interview that then accidentally becomes known to others) is a potential risk of your participation in the study. We will take precautions, however, to see that this does not happen.

Benefits
What are the benefits of participating? There may be no direct benefit to you for participating in the study. However, you may find an indirect benefit in knowing you have participated in an important study that could help others in the future. You may also appreciate the opportunity to share your experiences in confidence with another person. If you wish to receive it, we can also provide you with a list of relevant services in your area. You may also ask the interviewer to refer you for services if you wish to seek help. This information will not be shared with anyone.

Confidentiality
Will my participation in the study be kept confidential? The information that is collected during this interview will not include your name or any other information to personally identify you. Only members of the study team will have access to your questionnaire and it will be kept in a locked cabinet at the MUHAS offices in Dar es Salaam. You will not be identified in any reports or publications regarding the study. Your participation in the study will not be divulged to anyone.

Voluntary Participation
What are my rights as a study participant? Your participation in this study is completely voluntary. If you decide not to participate, you will not lose any existing benefits to which you are entitled. If you agree to participate in this study, you may end your participation at any time without penalty or loss of existing benefits to which you are entitled. If you decide to take part, you are free to refrain from answering any questions. You are free to withdraw at any time without affecting your relationship with the services you are entitled to as a community member.

Additional Information
What will I receive for participating? You will not receive any monetary or other incentive for participation.
What will happen to the results of the research study? The results and progress of the study will be presented in research publications and at meetings and conferences in Tanzania and in other countries. The Ministry of Health and Social Welfare (MOHSW), Regional Medical Office (RMO), and community partners will use the results to make improvements to their programs and services. In addition, the study progress and results will be put on a website so that others can learn from this study and strengthen services in their countries.

Who has reviewed the study for ethical issues? This study has been reviewed by research Institutional Review Boards at the Population Council, MUHAS, Tanzanian National Institute for Medical Research (NIMR), and Mbeya Research and Ethics Committee.
What if I need more information? If you have a concern about any aspect of the study, you should ask to speak to the researchers, who will do their best to answer your questions and address your concerns. You can contact Dr. Jessie Mbwambo, a psychiatrist at MUHAS and a principal investigator of the study, telephone number: +255 784 339 747; or Ms. Stella Mujaya, Study Coordinator, at the Futures Group, telephone number: +255 754 307 205.
What if there is a problem? For any concern you may have about the way you have been treated during the study or any possible harm you might suffer, please contact: Professor Mainen Moshi, Director of Research and Publications, MUHAS, Phone number: +255 222 150 302/6. Subject Statement: I have read or listened to the Informed Consent for this study. I have received an explanation of the planned research, procedures, risks and benefits and privacy of my personal information. I agree to take part in this study. I understand that my participation in this study is voluntary.

Your name: ________________________________________________________________
Your signature: ____________________________________________Date:____________ Investigator or person who conducted Informed Consent discussion: I confirm that I have personally explained the nature and extent of the planned research, study procedures, potential risks and benefits and confidentiality of personal information.

Minor Assent Form for the Household Survey for Females ages 15-17 years
You are invited to take part in a research study. Before you decide whether to participate, you need to understand why the research is being done and what it will involve. Please take the time to read or to listen as I read the following information. You may talk to others about the study if you wish. Please ask me if there is anything that is not clear, or if you would like more information. When all of your questions have been answered and you feel that you understand this study, you will be asked if you wish to participate in the study, and if yes, to sign this Informed Consent form. You will be given a signed copy to keep, if you would like one.
Your parent or guardian has already given permission for you to participate in the study. However, you do not have to say yes. We have talked to your parent or guardian and he/she agrees that you do not have to say yes to participate in the study.

Purpose of the Study and Study Requirements
What is the study? The study is being conducted by a team of researchers from Muhimbili University of Health and Allied Sciences (MUHAS) in Dar es Salaam and the Futures Group, Population Council, and Pangaea Global AIDS Foundation in the United States. The overall purpose of the study is to measure the effect of a set of programs and services that will be delivered in health facilities and in the community to improve the health and well-being of women and children. The study will compare data from communities where these programs and services are being delivered with data from communities that aren't yet receiving these services to determine if the program had the desired effect.
The study has been funded by the United States Government.
Why have I been invited to take part? You have been invited to take part in the study because, as part of the data collection, we are visiting a number of households in the area where you live, to interview one individual in each selected household. Your household has been selected by chance from a list of all households in your area. The information you give us regarding your knowledge and experiences with the services provided in your community will be important to understanding how these services are delivered and the effect they are having in the community. In addition the information you provide will assist the study team to understand what works and does not work in order to strengthen services in Mbeya and other regions.
What will happen if I take part? If you agree to take part in the study, we will ask you to sign this form indicating that you agree to be part of the study. An interviewer will sit with you and ask you a series of questions. The interviewer will record your responses on a form. The information that you provide will be kept completely confidential. Your responses will be combined with those of other participants in the study. Together this information will be analysed to determine if the programs and services are making a difference.
How long will the interview last? The interview will take between 60 and 90 minutes.
We may contact you again if we need to obtain further information or clarifications regarding the information you provide us with today.
Household Survey Minor Assent Form

Risks
What are the risks of the study? If you decide to participate, you may be inconvenienced by the time and effort it will take to complete the interview. You may find one or more questions that we ask to be sensitive. You do not have to respond to any question that makes you uncomfortable. You may end the interview at any time without penalty or loss of any benefits to which you are entitled.
Breach of confidentiality (that is, something you say in confidence during the interview that then accidentally becomes known to others) is a potential risk of your participation in the study. We will take precautions, however, to see that this does not happen.

Benefits
What are the benefits of participating? There may be no direct benefit to you for participating in the study. However, you may find an indirect benefit in knowing you have participated in an important study that could help others in the future. You may also appreciate the opportunity to share your experiences in confidence with another person. If you wish to receive it, we can also provide you with a list of relevant services in your area. You may also ask the interviewer to refer you for services if you wish to seek help. This information will not be shared with anyone.

Confidentiality
Will my participation in the study be kept confidential? The information that is collected during this interview will not include your name or any other information to personally identify you. Only members of the study team will have access to your questionnaire and it will be kept in a locked cabinet at the MUHAS offices in Dar es Salaam. You will not be identified in any reports or publications regarding the study. Your participation in the study will not be divulged to anyone. We will not share any of your information with your parents.

Voluntary Participation
What are my rights as a study participant? Your participation in this study is completely voluntary. If you decide not to participate, you will not lose any existing benefits to which you are entitled. If you agree to participate in this study, you may end your participation at any time without penalty or loss of existing benefits to which you are entitled. If you decide to take part, you are free to refrain from answering any questions. You are free to withdraw at any time without affecting your relationship with the services you are entitled to as a community member.

Additional Information
What will I receive for participating? You will not receive any monetary or other incentive for participation.
What will happen to the results of the research study? The results and progress of the study will be presented in research publications and at meetings and conferences in Tanzania and in other countries. The Ministry of Health and Social Welfare (MOHSW), Regional Medical Office (RMO), and community partners will use the results to make improvements to their programs and services. In addition, the study progress and results will be put on a website so that others can learn from this study and strengthen services in their countries.
Who has reviewed the study for ethical issues? This study has been reviewed by research Institutional Review Boards at the Population Council, MUHAS, the Tanzanian National Institute for Medical Research (NIMR), and Mbeya Research and Ethics Committee.

Household Survey Minor Assent Form
What if I need more information? If you have a concern about any aspect of the study, you should ask to speak to the researchers, who will do their best to answer your questions and address your concerns. You can contact Dr. Jessie Mbwambo, a psychiatrist at MUHAS and a principal investigator of the study, telephone number: +255 784 339 747; or Ms. Stella Mujaya, Study Coordinator, at the Futures Group, telephone number: +255 754 307205.
What if there is a problem? For any concern you may have about the way you have been treated during the study or any possible harm you might suffer, please contact: Professor Mainen Moshi, Director of Research and Publications, MUHAS, Phone number: +255 222 150 302/6. Subject Statement: I have read or listened to the Assent Form information for this study. I have received an explanation of the planned research, procedures, risks and benefits, and privacy of my personal information. I agree to take part in this study. I understand that my participation in this study is voluntary. Your child is invited to take part in a research study. Before you decide whether to allow your child to participate, you need to understand why the research is being done and what it would involve. Please take the time to read or to listen as I read the following information. You may talk to others about the study if you wish. Please ask me if there is anything that is not clear, or if you would like more information. When all of your questions have been answered and you feel that you understand this study, you will be asked if you would like your child to participate in the study, and if yes to sign this Parental Consent form. You will be given a signed copy to keep, if you would like one. Your child will also be asked whether she agrees to participate in this study and to sign a consent form. We will conduct the interview only if she agrees to participate.

Purpose of the Study and Study Requirements
What is the study? The study is being conducted by a team of researchers from Muhimbili University of Health and Allied Sciences (MUHAS) in Dar es Salaam and the Futures Group, Population Council, and Pangaea Global AIDS Foundation in the United States. The overall purpose of the study is to measure the effect of a set of programs and services that will be delivered in health facilities and in the community to improve the health and well-being of women and children. The study will compare data from communities where these programs and services are being delivered with data from communities that aren't yet receiving these services to determine if the program had the desired effect.
The study has been funded by the United States Government.
Why has my child been invited to take part? Your child has been invited to take part in the study because, as part of the data collection, we are visiting a number of households in the area where you live to interview one individual in each selected household. Your household has been selected by chance from a list of all households in your area. The information your child gives us regarding her knowledge and experiences with the services provided in your community will be important to understand how these services are delivered and the effect they are having in the community. In addition, the information she provides will assist the study team to understand what works and does not work in order to strengthen services in Mbeya and other regions.
What will happen if my child takes part? If you agree to let your child take part in the study, we will ask you to sign this form. An interviewer will sit with your child and ask her a series of questions. The interviewer will record her responses on a form. The information that she provides will be kept completely confidential. Her responses will be combined with those of other participants in the study. Together this information will be analysed to determine if the programs and services are making a difference.
How long will the interview last? The interview with your child will take between 60 and 90 minutes.
We may contact you again if we need to obtain further information or clarifications regarding the information you provide us today.

Risks
What are the risks of the study? If you give permission for your child to participate, she may be inconvenienced by the time and effort it will take to complete the interview. She may find one or more questions that we ask to be sensitive. She does not have to respond to any question that makes her uncomfortable. She may end the interview at any time without penalty or loss of any benefits to which she is entitled.
Breach of confidentiality (that is, something she says in confidence during the interview that then accidentally becomes known to others) is a potential risk of her participation in the study. We will take precautions, however, to see that this does not happen.

Benefits
What are the benefits of participating? There may be no direct benefits to you or your child for participating in the study. You and/or your child may find an indirect benefit in knowing your child has participated in an important study that could help others in the future. If you or your child wishes to receive a list of relevant services in your area, you will be provided with such a list from the interviewer. You may also ask the interviewer to refer you or your child for services if you wish to seek help. This information will not be shared with anyone.

Confidentiality
Will my child's participation in the study be kept confidential? The information that is collected during this interview will not include your name, your child's name or any other information to personally identify her. Only members of the study team will have access to your child's questionnaire and it will be kept in a locked cabinet at the MUHAS offices in Dar es Salaam. Your child will not be identified in any reports or publications regarding the study. Your child's participation in the study will not be divulged to anyone.

Voluntary Participation
What are my child's rights as a research participant? Your child's participation in this study is completely voluntary. If your child decides not to participate, she will not lose any existing benefits to which she is entitled. If you agree to let your child participate in this study, you or she may end her participation at any time without penalty or loss of existing benefits to which she is entitled.

Additional Information
What will my child receive for participating? Neither you nor your child will receive any monetary or other incentive for participation.
What will happen to the results of the research study? The results and progress of the study will be presented in research publications and at meetings and conferences in Tanzania and in other countries. The Ministry Of Health and Social Welfare (MOHSW), Regional Medical Office (RMO), and community partners will use the results to make improvements to their programs and services. In addition, the study progress and results will be put on a website so that others can learn from this study and strengthen services in their countries. What if I need more information? If you have a concern about any aspect of the study, you should ask to speak to the researchers who will do their best to answer your questions and address your concerns. What if there is a problem? For any concern about the way your child has been treated during the study or any possible harm your child might suffer, please contact:

Informed Consent Form for Key Informant Interviews Introduction
You are invited to take part in a research study. Before you decide whether to participate, you need to understand why the research is being done and what it would involve. Please take the time to read or to listen as I read the following information. You may talk to others about the study if you wish. Please ask me if there is anything that is not clear, or if you would like more information. When all of your questions have been answered and you feel that you understand this study, you will be asked if you wish to participate in the study, and if yes to sign this Informed Consent form. You will be given a signed copy to keep.

Purpose of the Study and Study Requirements
What is the study? The study is being conducted by a team of researchers from Muhimbili University of Health and Allied Sciences (MUHAS) in Dar es Salaam and the Futures Group, Population Council, and Pangaea Global AIDS Foundation in the United States. The overall purpose of the study is to measure the effect of a set of services to improve the health and well-being of women and children. These programs and services are intended to address gender-based violence (including physical, sexual, and emotional violence), which most typically takes the form of violence against women and children. The study will compare data from communities where these programs and services are being delivered with data from communities that aren't yet receiving these services to determine if the program had the desired effect.
The study has been funded by the United States Government.
Why have I been invited to take part? You have been invited to take part because the information you give us regarding your knowledge and experiences with the programs and services provided in your community will be important to understanding how these services are delivered and the effect they are having in the community. In addition, the information you provide will assist the study team to understand what works and does not work in order to strengthen services in Mbeya and other regions.
What will happen if I take part? If you agree to take part in the study, we will ask you to sign this form. You will also be asked to answer some questions regarding your knowledge, opinions, and experiences with programs and services that are being provided to address prevention and services for violence (including physical violence, sexual violence and other types of abuse).
How long will the focus group discussion last? This will take between 60-90 minutes.
We may contact you again if we need to obtain further information or clarifications regarding the information you provide us today.

Risks
What are the risks of the study?
An inconvenience may be the time and effort you take to be a participant. You may find one or more questions that we ask to be upsetting or emotionally sensitive particularly as the study focuses on violence against women and children and sexual violence. You do not have to respond to any question that makes you uncomfortable. You may end the interview at any time without penalty or loss of any benefits to which you are entitled.
A risk may be a breach of confidentiality (something you say during the interview) but we will take precautions to see that this does not happen.

Benefits
What are the benefits of participating? There are no direct benefits to you for participating in the study. You may find an indirect benefit in knowing you have participated in an important study that could help others in the future. The study will help the government of Tanzania to strengthen services to prevent and respond to violence against adults and children.

Confidentiality
Will my participation in the study be kept confidential? The information that is collected during this interview will not include your name, only your occupation or designation. Only members of the study team will have access to the information we record form the interview and that will be kept in a locked cabinet at the MUHAS offices in Dar es Salaam. You will not be identified in any reports or publications regarding the study. Your participation in the study will not be divulged to anyone else.

Voluntariness
What are my rights as a research participant/subject? Your participation in this study is completely voluntary. If you decide not to participate, you will not lose any existing benefits to which you are entitled. If you agree to participate in this study, you may end your participation at any time without penalty or loss of existing benefits to which you are entitled. If you decide to take part, you are free to refrain from answering any questions. You are free to withdraw at any time without affecting your relationship with the services you are entitled to as a community member.

Additional Information
What will I receive for participating? You will not receive any monetary or other incentive for participation.
What will happen to the results of the research study? The results and progress of the study will be presented in research publications and at meetings and conferences in Tanzania and in other countries. The MOHSW and community partners will use the results to make improvements to their programs and services. In addition, the study progress and results will be put on a website so that others can learn from this study and strengthen services in their countries.
Who has reviewed the study for ethical issues? This study has been reviewed by Institutional Review Boards at the Population Council, MUHAS, Tanzanian National Institute for Medical Research (NIMR), and Mbeya Research and Ethics Committee.
What if I need more information? If you have a concern about any aspect of the study, you should ask to speak to the researchers, who will do their best to answer your questions and address your concerns.

Introduction
You are invited to take part in a research study. Before you decide whether to participate, you need to understand why the research is being done and what it will involve. Please take the time to read or to listen as I read the following information. You may talk to others about the study if you wish. Please ask me if there is anything that is not clear, or if you would like more information. When all of your questions have been answered and you feel that you understand this study, you will be asked if you wish to participate in the study, and if yes, to sign this Informed Consent form. You will be given a signed copy to keep, if you would like one.

Purpose of the Study and Study Requirements
What is the study? The study is being conducted by a team of researchers from Muhimbili University of Health and Allied Sciences (MUHAS) in Dar es Salaam and the Futures Group, Population Council, and Pangaea Global AIDS Foundation in the United States. The overall purpose of the study is to measure the effect of a set of programs and services that will be delivered in health facilities and in the community to improve the health and well-being of women and children. These programs and services are intended to address gender-based violence (including physical, sexual, and emotional violence), which most typically takes the form of violence against women and children. The study will compare data from communities where these programs and services are being delivered with data from communities that aren't yet receiving these services to determine if the program had the desired effect.
The study has been funded by the United States Government.
Why have I been invited to take part? You have been invited to take part in this study because you have been identified as someone who may provide services to survivors of gender-based violence (GBV) at this health facility. The information you give us regarding your knowledge and experiences with the services you provide will assist the study team to understand how health services are being provided and what gaps and challenges may exist. The information you give us will help to strengthen services in Mbeya and other regions.
What will happen if I take part? If you agree to take part in the study, we will ask you to sign this form indicating that you agree to be part of the study. You will also be asked to answer some questions regarding your knowledge of GBV and the types of services for GBV survivors that are being provided at your health facility. An interviewer will sit with you and ask you a series of questions. The interviewer will record your responses on a form. The information that you provide will be kept completely confidential. Your responses will be combined with those of other participants in the study.
How long will the interview last? The interview will take between 60 and 90 minutes.
We may contact you again if we need to obtain further information or clarifications regarding the information you provide us with today.

Risks
What are the risks of the study? If you decide to participate, you may be inconvenienced by the time and effort that it will take to complete the interview. You may find one or more questions that we ask to be sensitive. You do not have to respond to any question that makes you uncomfortable. You may end the interview at any time without penalty or loss of any benefits to which you are entitled.
Breach of confidentiality (that is, something you say in confidence during the interview that then accidentally becomes known to others) is a potential risk of your participation in the study. We will take precautions, however, to see that this does not happen.

Benefits
What are the benefits of participating? There may be no direct benefit to you for participating in the study. However, you may find an indirect benefit in knowing you have participated in an important study that could help others in the future. You may also appreciate the opportunity to share your experiences in confidence with another person.

Confidentiality
Will my participation in the study be kept confidential? The information that is collected during this interview will not include your name or any other information to personally identify you. Only members of the study team will have access to your questionnaire and it will be kept in a locked cabinet at the MUHAS offices in Dar es Salaam. You will not be identified in any reports or publications regarding the study.

Voluntary Participation
What are my rights as a study participant? Your participation in this study is completely voluntary. If you decide not to participate, you will not lose any existing benefits to which you are entitled. If you agree to participate in this study, you may end your participation at any time without penalty or loss of existing benefits to which you are entitled. If you decide to take part, you are free to refrain from answering any questions. You are free to withdraw at any time without affecting your relationship with your employer or the community.

Additional Information
What will I receive for participating? You will not receive any monetary or other incentive for participation.

What will happen to the results of the research study?
The results and progress of the study will be presented in research publications and at meetings and conferences in Tanzania and in other countries. The Ministry of Health and Social Welfare (MOHSW), Regional Medical Office (RMO), and community partners will use the results to make improvements to their programs and services. In addition, the study progress and results will be put on a website so that others can learn from this study and strengthen services in their countries. What if I need more information? If you have a concern about any aspect of the study, you should ask to speak to the researchers, who will do their best to answer your questions and address your concerns. For purposes of this study, a household is defined as: A group of individuals who usually live and eat together, whether or not they are related by blood or marriage, with one person, male or female, acknowledged as the head of the household. A household can consist of one person or many persons.
Several households may reside in one dwelling. If this is the case, randomly select one of the households (writing household labels on pieces of paper, then randomly drawing from a bag).
Upon reaching the house, ask to speak to the household head. If s/he is not available, ask to speak to any (competent) adult who resides in the household. If such individual is not there, ask for a time when s/he will return.
Then proceed with Brief Introduction of the Study. Then ask the following questions of the household head/adult.

Introduction
INTERVIEWER: As you have just learned, the overall purpose of this study is to measure the effect of a set of programs and services delivered in health facilities and in the community to improve the health and well-being of women and children. I am going to ask you a series of questions and then will record your answers on this form. The interview will take from one to one and one half hours. As we noted in the consent form, you may find some questions that I will ask to be sensitive. You do not have to respond to any question that makes you uncomfortable. You may end the interview at any time. We will only begin the interview once we have ensured we are in a private place and you are comfortable. If anyone enters the room, we will pause the interview and look for another private space, if necessary, to ensure your comfort in moving forward with the interview. INTERVIEWER: Next, I would like to ask you some questions about your current (or recent) male partner. Again, please let me assure you that your answers will be kept completely confidential and will not be told to anyone. If we should come to any question that you don't want to answer, just let me know and we will go on to the next question. You told me previously, that you currently have a partner/had a partner in the past year… 13 We have now finished the interview. Do you have any comments, or is there anything else you would like to share? RECORD COMMENTS.

904
FINISH I would like to thank you very much for helping us. I appreciate the time you have taken. I realize that these questions may have been difficult for you to answer, but it is only by hearing from women themselves that we can really understand about their health and experiences in life.
In case you are interested or if you ever hear of another person who needs help related to violence and the things we have been talking about, I'm going to leave you with the name and contact information of a service provider at _________________ Health Center/_________Hospital, who can offer assistance. This person is service provider at this health facility who also is working with us on this study. S/he is aware that we are giving out this information as part of the interview. If you go, you can be assured that s/he will keep anything that you say private. You can go whenever you feel ready to, either soon or later on.
LEAVE PAPER SLIP WITH NAME OF HEALTH FACILITY IN THE STUDY CLUSTER, NAME OF TATHMINI GBV DATA FOCAL PERSON, AND CONTACT INFORMATION.

INTERVIEWER:
We are coming to the end of the interview. There is just one last step.
I'm going to give you a card. On this card are two pictures. No other information is written on the card. One picture is of a sad face, the other is of a happy face.

GIVE RESPONDENT THE CARD AND PEN.
No matter what you have already told me, I would like you to put a mark next to the sad picture if someone hurt you physically, touched you sexually, or made you do something sexual that you didn't want to in the last 12 months.
Please put a mark next to the happy face if this did not happen to you in the last 12 months. Once you have marked the card, please fold it over and put it in this bag. The bag contains similar cards from many other women. This will ensure that I do not know your answer. If you are interested, I can also ask this service provider to contact you directly. If you would like this referral, I will take down your contact information, including your name and a safe place where or private way that s/he or another trained colleague can find you. I will record this information on a separate form, which will not be linked in any way with the interview we have just completed. I will then give this form to my team leader, who will provide the form directly to the service provider. Information you have shared during this interview will not be shared with the service provider.
Would you like me to put you in direct contact with a service provider? YES…………………….

INTERVIEWER OR TEAM LEADER INSTRUCTIONS:
For purposes of this study, a household is defined as: A group of individuals who usually live and eat together, whether or not they are related by blood or marriage, with one person, male or female, acknowledged as the head of the household. A household can consist of one person or many persons.
Several households may reside in one dwelling. If this is the case, randomly select one of the households (writing household labels on pieces of paper, then randomly drawing from a bag).
Upon reaching the house, ask to speak to the household head. If s/he is not available, ask to speak to any (competent) adult who resides in the household. If such individual is not there, ask for a time when s/he will return.
Then proceed with Brief Introduction of the Study. Then ask the following questions of the household head/adult.  Identify the last digit of the ID Number. Go to this row in the table below.  Obtain the total number of eligible women from the listing above. Go to this column number.  Follow the selected row and column to the cell where they meet and circle the number in the cell. This is the line number of the female to be selected for the interview (from table above).  Write the name and line number of the selected female in the space below the table.

AGE VERIFICATION Interviewer instructions:
Ask to speak with the selected girl/woman. Ask her age and date of birth and record: Reported age: _ _ Date of birth: _ _/_ _/_ _ _ _ Calculated age: _ _  If reported or calculated age is 20 or younger:  Ask to see written documentation. This could include birth certificate, clinic card, passport, religion card, ID card, or other official document.  If documentation indicates she is aged of 15-17, continue with Parental Consent process.
If documentation indicates she is aged 18+, continue with Adult Informed Consent.  If no documentation is available to verify age, then remove the name of this person from the list of eligible female members and move up all other names by one row. Repeat the random selection process, and then age verification, as indicated by reported age.
[ ] Tick here if you had to repeat the random selection because of no age documents Introduction INTERVIEWER: As you have just learned, the overall purpose of this study is to measure the effect of a set of programs and services delivered in health facilities and in the community to improve the health and well-being of women and children. I am going to ask you a series of questions and then will record your answers on this form. The interview will take from one to one and one half hours. As we noted in the consent form, you may find some questions that I will ask to be sensitive. You do not have to respond to any question that makes you uncomfortable. You may end the interview at any time. We will only begin the interview once we have ensured we are in a private place and you are comfortable. If anyone enters the room, we will pause the interview and look for another private space, if necessary, to ensure your comfort in moving forward with the interview. INTERVIEWER: Next, I would like to ask you some questions about your current (or recent) male partner. Again, please let me assure you that your answers will be kept completely confidential and will not be told to anyone. If we should come to any question that you don't want to answer, just let me know and we will go on to the next question. You told me previously, that you currently have a partner/had a partner in the past year… Section 4 IPV in the past 12 months INTERVIEWER: Read this only the first time/for first partner. I am now going to ask you about some situations that happen to some women. I'd like you to think back over the past 12 months and please tell me if they happened to you. You may find some of these questions to be sensitive. Please remember if we should come to any question that you do not want to answer, just let me know and we will go on to the next question. I want to remind you that if you prefer, we can stop the interview at any time. (1) Number of staff (Don't know=888)  2) If YES, can you please describe those roles and tasks?

STANDARD OPERATING PROCEDURES AND CONFIDENTIALITY for GBV clients
Do not read. If R mentions the role/task, circle 1. If not mentioned, circle 2.

3.5
An informed consent process involves counseling the client on steps the provider recommends taking, asking for and addressing the client's questions and concerns, and receiving her/his agreement to proceed.

Are providers at this facility required to obtain informed consent from the client before …
Read each item. Circle the response that is given. Read the definition below regardless of the response to Q 5.0.

CHILD SEXUAL VIOLENCE Definition:
Child sexual violence is the involvement of a child (in Tanzania, defined as a person under the age 18) in sexual activity that he or she does not fully comprehend, is unable to give informed consent to, or for which the child is not developmentally prepared and cannot give consent, or that violates the laws or social taboos of society. Child sexual abuse is evidenced by this activity between a child and an adult or another child who by age or development is in a relationship of responsibility, trust, or power, the activity being intended to gratify or satisfy the needs of the other person.
This may include but is not limited to: • the inducement or coercion of a child to engage in any unlawful sexual activity, • the exploitative use of a child in prostitution or other unlawful sexual practices, and

5.2
Imagine that a girl 14 years of age comes into the facility with her mother and reports that she is pregnant. Would you (or another appropriate service provider) conduct an examination to assess whether she may have experienced sexual violence? 1=Yes 2=No 8=Don't know

5.3
If a girl wears a short dress that leaves her body parts half naked, is a man justified in forcing her to have sexual intercourse?

6.2
What are the challenges, if any, that this facility faces in responding to GBV? List all mentioned.

6.3
Do you have any final suggestions for improving this facility's response to GBV? List all mentioned 6.4 Do you have any questions for me? List all mentioned Interviewer: Thank you very much for answering these questions. We appreciate your time and honesty.