Prevalence of intimate partner violence and abuse and associated factors among women enrolled into a cluster randomised trial in northwestern Tanzania

Background Intimate partner violence (IPV) is recognised as an important public health and social problem, with far reaching consequences for women’s physical and emotional health and social well-being. Furthermore, controlling behaviour by a partner has a similar impact on women’s well-being, yet little is known about the prevalence of this type of behaviour and other related abuses in Tanzania and in other sub-Saharan African countries. Methods We conducted a cross-sectional study to determine the lifetime and past 12-month prevalence of physical and sexual IPV, economic abuse, emotional abuse and controlling behaviour among ever-partnered women in Mwanza, Tanzania. Women (N = 1049) were enrolled in an ongoing trial (Maisha study) to assess the impact of microfinance combined with gender training on participants’ experience IPV, and other related outcomes. Interviews were conducted by same sex interviewers to collect information about socio-demographic characteristics, experiences of specific acts of IPV and abuse, and symptoms of poor mental health status. Results Overall, about 61% of women reported ever experiencing physical and/or sexual IPV (95% CI: 58–64%) and 27% (95% CI: 24–29%) experienced it in the past 12 months. Partner controlling behaviour was the most prevalent type of abuse with 82% experiencing it in their lifetime and 63% during the past 12 months. Other types of abuses were also common, with 34% of women reporting economic abuse and 39% reporting emotional abuse during the past 12 months. The prevalence of IPV and abuses varied by socio-demographic characteristics, showing much higher prevalence rates among younger women, women with young partners and less educated women. After we adjusted for age and socio-economic status, physical violence (OR = 1.8; 95% CI: 1.3–2.7) and sexual violence (OR = 2.8; 95% CI: 1.9–4.1) were associated with increased reporting of symptoms of poor mental health. Similarly, experience of abuse during the past 12 months was associated with increased reporting of symptoms of poor mental health. Conclusions The high prevalence of IPV and abuses and its strong links with symptoms of poor mental health underline the urgent need for developing and testing appropriate interventions in settings like Tanzania to tackle both violence and abusive behaviours among intimate partners. Trial registration ClinicalTrials.gov – ID NCT02592252, registered retrospectively on 13 August 2015.


Background
Intimate partner violence (IPV) is recognised as an important public health problem, development issue and human rights concern. Globally, it is estimated that about 30% of women will experience physical and/or sexual violence from an intimate partner during their lifetime [1], and that one in three homicides among women are by an intimate partner [2]. Lifetime rates of physical and/or sexual IPV have been found to be highest in south-East Asia, the Mediterranean region and sub-Saharan Africa [1]. In Tanzania, almost 2 in 5 women aged 15 to 49 years have experienced physical violence at some point in their lives, 44% of ever-married women have experienced physical and/or sexual violence by their current or most recent husband or partner, and 37% of ever-married women experienced such spousal violence in the past 12 months [3]. Results from the baseline survey of an ongoing clusterrandomized HIV and gender-based violence prevention trial in Dar es Salaam, Tanzania showed that within the last 12 months, 35.8% of women reported experiencing physical, sexual and/or psychological IPV [4].
Experience of IPV has far reaching consequences for women's physical and emotional health and social wellbeing [5][6][7]. Women who experience IPV show more physical symptoms of poor health, and more days out of work and injuries than women who have not been abused [5,7]. IPV has also been associated with mental health problems, including depression, anxiety, phobias, posttraumatic stress disorder, suicide, and alcohol and drug abuse [6][7][8][9][10]. Furthermore, recent evidence suggests that controlling behaviour by a partner has a similar impact on women's well-being [11], yet little is known about the prevalence of this type of behaviour and other related abuses in Tanzania. Several intervention studies have been implemented in low and middle income countries to address IPV [12][13][14][15]. One example is the Intervention with Microfinance for AIDS and Gender Equity (IMAGE) trial implemented in rural South Africa [13], which found that a gender empowerment intervention linked to microfinance loans was associated with 55% reduced IPV incidence in the past year. We are currently conducting a trial (MAISHA study), to adapt and evaluate the impact of the IMAGE model in Mwanza city, northwestern Tanzania. In this paper, we describe the design of the MAISHA study and present the baseline lifetime and past 12-month prevalence and severity of physical and sexual IPV and other abuses, including economic abuse, emotional abuse and controlling behaviour. Additionally, we assessed the co-occurrence (or overlap) of different forms of violence and abuses, and the associations between experience of violence and abuses, and socio-demographic characteristics and women's self-reported mental health status.

MAISHA study design and procedures
MAISHA study is a cluster-randomised controlled trial to assess the impact of a combined microfinance and gender training programme for women on participants' experience of physical and/or sexual violence, as well as other gender-empowerment, economic, and health related outcomes. The trial is a collaboration with an established local microfinance provider in Tanzania (Bangladesh Rural Advancement Committee or BRAC). As part of BRAC microfinance programme standard procedures, loans are provided to women in groups, with about 10-30 women per group. We recruited into the trial 66 existing BRAC microfinance loan groups from low socio-economic communities in Mwanza city. Each woman in the groups approached for this study met with a staff member to go through the participant information sheet and those who agreed to participate and demonstrated understanding of the study procedures were invited to sign the consent form. Following completion of baseline data collection, loan groups were randomly allocated in a 1:1 ratio to either continue with microfinance only (N = 33 groups) or to receive a combined microfinance and gender training intervention (N = 33 groups). The gender training intervention included 10 sessions delivered by female facilitators over 20 weeks at a venue convenient to the participants. The training was developed drawing on six curricula including the Sisters For Life curriculum originally developed for the IMAGE trial [13].

Baseline survey
During the baseline survey, consenting women were interviewed, face-to-face, using a structured questionnaire, with responses entered directly onto a tablet computer programmed to check for accuracy and consistency of information entered during the interview. Interviews were conducted in a private location by female interviewers trained on interviewing techniques, gender issues, violence, and ethical issues related to research on IPV. We used a standardised structured questionnaire which was translated into a local language (Swahili) and then independently back-translated into English to check for accuracy. The questionnaire included information about household details, income, relationships, health, childhood and experiences of specific acts of IPV and abuse. Symptoms of poor mental health were assessed using a standardised tool (SRQ-20), integrated into the health section of the questionnaire, which was developed by WHO [16] and has been validated in a range of populations [17]. Questions asking about each type of violence and abuse were adapted from the WHO Violence Against Women instrument (Table 1) [18].
In order to ensure support for participants who reported acts of violence or abuse perpetrated against them, we identified service providers in the communities where we recruited study participants. Our study team met with representatives of each service provider to inform them about the study and to establish effective mechanisms for referring women for ongoing help and support.

Data analysis
Intimate partner violence was defined as one or more acts of physical and/or sexual violence and abuse was defined as one or more acts of either controlling behaviour, emotional abuse and/or economic abuse (Table 1). We defined the prevalence of ever experiencing IPV as the proportion of ever-partnered women (i.e. women who had ever been in a relationship including those who had ever been married or lived with a partner) reporting one or more acts of physical and/or sexual violence by an intimate partner at any point in their lives. Current prevalence of IPV was defined as the proportion of everpartnered women reporting one or more acts of violence during the 12 months before the interview. The same Table 1 Questions used in the study to document violence, controlling behaviours, emotional abuse and economic abuse by an intimate partner in Mwanza, Tanzania approach was used to define prevalence of the different forms of abusecontrolling behaviour, emotional abuse and economic abuse.
Physical violence was considered severe if a participant reported having been hit, kicked, chocked or threatened with a weapon; and less severe if they reported having been pushed or slapped. For emotional abuse, economic abuse and controlling behaviours, severity was defined by the number of yes responses to the component questions of each abuse experienced by participant (Table 1).
We summarised the prevalence (with corresponding 95% confidence intervals (CI)) of each type of violent and abusive act experienced, overall and by severity. We also summarised the prevalence of current violent and abusive acts for each stratum of the following sociodemographic variables: women's age, education, marital status, number of children, monthly earnings and overall socio-economic status; and male partner's age and education. We used structural equation modelling [19] to generate a summary socio-economic status indicator from 19 variables collected in the questionnaire. Symptoms of poor mental health were assessed by using SRQ-20 cutoff point of 7/8 (yes to 8 or more questions is a case, 7 or less is a non-case). The associations between symptoms of poor mental health and different types of violence and abuses were explored by cross tabulations and logistic regression models. We summarised the associations firstly in terms of crude odds ratios and corresponding 95% CIs, and secondly after controlling for age, socio-economic status, violence and abuses. Analyses were done using Stata version 13.1 software (StataCorp, College Station, TX, USA).

Results
Between September 2014 and June 2015, 66 out of 110 microfinance loan groups approached were recruited into the trial. Of the 1154 women in the 66 groups who received information about the trial, 1049 provided consent to join the study. Baseline interviews were conducted among 1021 women who were available at the time of the survey and reported to have ever been in a relationship. Women's age ranged from 19 to 70 years, with a median age of 39 years (interquartile range [IQR] 33-46), and 28.7% were less than 35 years of age. Most participants were married or living with a man as if married (72.8%), and had completed primary education (72.3%). Table 2 presents the lifetime and current prevalence of each type of violent and abusive act. Overall, about 61% of women reported ever experiencing physical and/or sexual IPV (95% CI: 58-64%) and 27% (95% CI: 24-29%) experienced it in the past 12 months. When examined separately, the prevalence of lifetime physical and sexual violence was 53 and 35%, respectively. The prevalence of physical and sexual violence during the past 12 months before the interviews was 19 and 17%, respectively. Overall, most women who had experienced physical violence in their lifetime (68%) or during the past 12 months (68%) reported severe forms of violence.

Prevalence of violence and different forms of abuses
Among other abuses, partner controlling behaviour was the most prevalent type of abuse with 82% experiencing it in their lifetime and 63% during the past 12 months. Other types of abuses were also common, with 34% of women reporting economic abuse and 39% reporting emotional abuse during the past 12 months. Most women who reported economic (51%) and emotional (54%) abuse during the past 12 months had responded 'yes' to at least two different acts of abuse.

Co-occurrence of past 12-month violence and abuse
In Fig. 1, we present the co-occurrence (or overlaps) of reported violence and abuse during the past 12 months. Overall, there was considerable overlap between physical and sexual violence and the different forms of abuses, with only 276 (27%) women not reporting any violence or abuse during the past 12 months. Among 271 women who reported physical and/or sexual violence by their current partners in the last 12 months, 94 (35%) women experienced both physical and sexual violence; and almost all (98%) report at least one form of abuse within the same time period. Out of 739 women who experienced at least one form of abuse, 200 (27.1%) experienced all abuses (i.e. controlling behaviour, economic abuse and emotional abuse) in the last 12 months.

Associations with socio-demographic characteristics
In Tables 3 and 4 we present the associations between experience of violence and different forms of abuse in the past 12 months and selected socio-demographic characteristics. The prevalence of physical and/or sexual violence was highest among women below 30 years of age, and among those with primary education or less. Overall, the prevalence of violence decreased with increasing women's age and level of education. Similarly, the prevalence of violence decreased with increasing male partner's age and level of education. For controlling behaviour, higher prevalence rates were also found among younger women. Economic abuse showed no distinct patterns, although the prevalence of economic and emotional abuses was relatively low among women ≥ 50 years and among those with post-secondary education. Women who were widowed and those without children below the age of 18 years were relatively less likely to experience violence or abuses when compared with other women. Women's monthly earnings and socio-economic status was not related to experience of violence or abuses.

Associations with symptoms of poor mental health
Overall, about 40% (95% CI: 37-43%) of women reported symptoms of poor mental health based on SRQ-20 cut-off point of 7/8. In Table 5, we present the associations between experience of IPV and abuse during the past 12 months and symptoms of poor mental health. In general, the proportion of women with symptoms of poor mental health was relatively higher among women experiencing IPV and abuse. After we adjusted for age and socio-economic status, women who experienced physical violence (OR = 1.8; 95% CI: 1.3-2.7) and sexual violence (OR = 2.8; 95% CI: 1.9-4.1) were significantly more likely to report symptoms of poor mental health when compared to other women. Similarly, women reporting to have experienced emotional abuse (OR = 1.5; 95% CI: 1.1-2.2), and economic abuse (OR = 1.9; 95% CI: 1.4-2.6) were more likely to report symptoms of poor mental health after we adjusted for age, socio-economic status and experience of violence.

Discussion
In this paper, we present baseline results from a study of women attending microfinance loan groups enrolled in the ongoing MAISHA study in Mwanza, Tanzania. We found a high prevalence of IPV, with 61% of women reporting ever experiencing physical and/or sexual IPV and 27% of women reporting it in the past 12 months. A substantial proportion of women reported severe physical violence which happened relatively frequently, suggesting that this is a common experience among women in this population. We also found a high prevalence of abuses, confirming that both violence and abuse from an intimate partner are common social, public health and human rights concerns among women in this study population.
The prevalence of sexual and/or physical violence in this population is consistent with previous studies conducted in Tanzania [3,8,18] and in other countries in sub-Saharan Africa [1]. To our knowledge, this is one of the first studies to examine the prevalence of emotional and economic abuse and controlling behaviour in a relationship in an African setting. A substantial proportion of women reported economic and emotional abuse in the last 12 months and almost two thirds of women reported experiencing at least one form of controlling behaviour during the same time period, suggesting this may constitute a form of normal behaviour in this Tanzanian population. We observed high rates of co-occurrence of reported IPV and abuses, with almost all women who reported experiencing physical and/or sexual violence in the last 12 months also reporting to have experienced one or more forms of abuse during the same time period. This high degree of overlap suggests that women in this population are more likely to experience both violence and multiple forms of abuse from their intimate partners. While more in-depth research is required to understand the impact of the different forms of abuse and controlling behaviour, our findings suggests that comprehensive violence prevention intervention programmes areneededto address IPV and more neglected forms of abuses [20].
We examined the associations between the prevalence of IPV and abuses in the last 12 months and sociodemographic characteristics. In general, IPV (and to some extent controlling behaviour) varied by sociodemographic characteristics, with much higher prevalence rates among younger women, women with young partners and women with less education. This is also in line with prior studies reporting women below the age of 25 having higher rates of violence than women aged 25 or older [21]. Potential reasons for this association are that relationships in younger age groups are more likely to be less  stable, and to be associated with early pregnancy and higher levels of financial stress [22]. It is also possible that the levels of recall and social desirability bias is relatively lower among young women and therefore the likelihood of reporting IPV is higher in this population [23]. We observed relatively lower prevalence of IPV and abuses among women who were widowed and those without children below the age of 18 years. This is likely to be related to other factors, such as age, that are related to IPV in this population. Surprisingly, women's monthly earnings and socio-economic status were not found to be related to experiences of violence or abuses. This finding warrants further investigation in order to examine the role of socioeconomic status as a risk factor for IPV and abuses. We found that women who reported experiencing IPV during the past 12 months were more likely to report symptoms of poor mental health, and this association remained unchanged after we adjusted for age and socioeconomic status. This is consistent with findings from other studies [5,24,25]. We also found that economic and emotional abuses were significantly associated with symptoms of poor mental health. These cross-sectional Table 4 The associations between experience of intimate partner abuse during the past 12 months and socio-demographic characteristics among ever-partnered women in Mwanza, Tanzania   findings suggest a strong association between IPV/abuses and symptoms of poor mental health, although it is not possible to confirm a causal link between them. Women who report symptoms of poor mental health may be more likely to experience violence [5], suggesting that poor mental health may be a risk factor for IPV. It is also possible that experience of severe and prolonged IPV may lead to poor mental health and other adverse health outcomes, as previously reported in prospective studies [26][27][28].
The high prevalence of IPV and abuses and its strong links with symptoms of poor mental health underline the importance of developing and testing appropriate interventions. As part of the MAISHA study, we are conducting a cluster randomised controlled trial to assess the impact of microfinance combined with gender training for women on participants' experience of physical and/or sexual IPV, as well as other gender-empowerment, economic, and health related outcomes in Mwanza, Tanzania. We have enrolled 1049 women receiving group-based microfinance loans and successfully delivered 10 sessions of gender training to women randomised to the intervention arm of the trial. The impact of the intervention will be assessed through a face-to-face interview during a followup survey to be conducted in 2017, approximately 29-30 months after randomisation.
The findings of this cross-sectional survey need to be considered in view of several limitations. Firstly, reporting of IPV may be prone to under-reporting due to social desirability or shame as women are commonly stigmatized and blamed for the abuse they receive. Several measures were implemented to minimise this problem, including training of interviewers to adhere to high ethical standards, and closely following the WHO guidelines on researching violence against women [29]. Secondly, the study was conducted among a sample of women receiving microfinance loans living in Mwanza, Tanzania's second largest city. It may therefore not be representative of the whole of Tanzania. However, this does not affect the validity of the findings presented, especially since the prevalence rates are comparable to previous studies. Thirdly, we analysed cross-sectional data collected at baseline, and this does not allow any inferences to be made regarding causality and temporality with respect to the socio-demographic factors and health outcomes assessed. Lastly, women's mental health status was assessed through self-report of symptoms and does not represent a clinical diagnosis.

Conclusions
In conclusion, these findings show that violence by an intimate partner is a major problem among women in Tanzania. We also found high levels of economic and emotional abuse and very high levels of controlling behaviour, with high rates of co-occurrence of reported IPV and abuses. The prevalence of IPV and abuses varied by socio-demographic characteristics, showing much higher prevalence rates among younger women, women with young partners and less educated women. The high prevalence of IPV and abuses and its strong links with symptoms of poor mental health underline the urgent need for developing and testing appropriate interventions in settings like Tanzania to tackle both violence and abusive behaviours among intimate partners. As part of the ongoing MAISHA study, we aim to assess the impact of microfinance combined with gender training for women on physical and/or sexual IPV in NW Tanzania.