Plain English summary

Domestic violence against women (VAW) is a well-recognized public health concern and systematic human rights violation. It has a serious negative impact on women’s lives. Domestic VAW is common and still unacceptably high in different parts of the world as observed from several primary studies which have been conducted on the frequency and its associated factors. Additionally, some interventional studies have been conducted in some parts of the world revealed that gender-norms transformation through behavioral change and communication focused program can promote gender-equality norm and avert domestic VAW. Summarized or synthesized evidence is still needed to inform and persuade policy makers and stakeholders, so they can take an evidence based decision making approach. One of the most challenging issue is that most countries’ governments have considered VAW as a minor and socially tricky issue. There is some ambiguity as to whether VAW is a private or public matter. This systematic review and meta-analysis aimed at summarizing existing primary study findings to determine its level and associated factors, identify effective interventions to prevent domestic VAW and make key recommendations. The purpose is to contribute evidence to be used by program planners, policy makers, clinicians and other stakeholders to make an informed decision on the issue of domestic VAW. The study showed that more than half of the women experienced VAW, and almost one-third of the women have experienced current VAW. Intervention strategies should focus on traditional gender role transformation to minimize the relationship power-gap and prevent VAW.

Background

Globally, VAW is a well-recognized public health problem and a gross pervasive violation of human rights. About 35% of women experience VAW [1, 2], and almost two-third of women murders are committed by their intimate partners every year. About five percent of the women’s total health years loss has been attributed by domestic VAW which is also exacerbated by authority inequity in relationship [3, 4]. Additionally, VAW causes ill health and its associated devastating outcome are more than the cumulative problem of cancer, road traffic accidents and malaria which are massive threat and an uncontrollable public health challenge for the upcoming generation [4]. VAW is sturdily interconnected with gender inequality that affects women’s negotiation ability about reproductive health and related issues [5,6,7]. The expenses associated with VAW has been estimated to be 3.7% of the countries’ gross domestic product, which is almost comparable with what several countries devote on primary education [8]. Yet, it has been considered exclusively as private matter and negligible issue by the governments’ of various countries, hence not perceived as a crime [4, 5, 9].

Global and regional commitment to fighting domestic VAW is reflected in various international statues. For example, the United Nations aims to build an enabling household situation in improving women’s right, their political participation, economic empowerment and safety [10]. In addition, provision of comprehensive and universal access to sexual and reproductive health care has been a strategy to avert domestic VAW in the Maputo Plan of Action (2016-2030) [11]. Further, research evidence has revealed that women empowerment and community mobilization are the most recommended interventions to minimize the expenses associated with VAW and its consequences [12]. The problem of domestic VAW is caused and exacerbated by poverty, alcohol consumption and societal receptive attitude towards inequitable gender-norms, which has been exhibited through denying access to education; lack of autonomy and justifying wife-beating by fellow women [9, 13,14,15,16,17,18,19,20].

Furthermore, women’s experience of domestic violence is significantly associated with several and multiple poor physical and mental health outcomes [1, 9, 14, 17, 21,22,23,24,25,26,27]. Likewise, VAW has been associated with various poor reproductive health conditions such as HIV, unintended pregnancy and unsafe abortion [1, 4, 7, 9, 15,16,17,18, 20, 25, 28,29,30,31,32,33,34,35,36,37,38,39,40,41]. Therefore, VAW needs a comprehensive approaches to empower women economically, transform traditional gender-norms in improving their communication and negotiation skills [9, 17, 42]. There is a paucity of summarized evidence on; the level of domestic VAW, its associated factors, proven evidence on the technical approach and key research recommendations in low and lower-middle income countries (LLMICs). However, many primary studies have been conducted in LLMICs. The main purpose of this systematic review and meta-analysis was to summarize existing primary studies in LLMICs to determine the prevalence of domestic VAW and its associated factors; to identify effective and proven interventions and make key recommendations. It will provide an insight to policy makers, program planners, clinicians, researchers and other stakeholders to make an informed decisions on issues related to VAW.

Review question(s)

  • What was the level of domestic VAW in LLMICs?

  • What were the factors associated with domestic VAW in LLMICs?

  • What were the research evidence that should translate into routine action in LLMICs?

  • What were the studies’ key recommendations on the prevention of domestic VAW in LLMICs?

Method development and protocol registration

The protocol for the systematic review and meta-analysis has been registered in the International Prospective Register of Systematic Reviews (PROSPERO) (ID: CRD42017079977). This systematic review and meta-analysis methods was written according to the preferred reporting items for systematic review and meta-analysis (PRISMA) guideline [43]. The filled PRISMA checklist is attached as Additional file 1.

Searching methods and identifications of studies

Studies were searched using medical subject headings (MeSHs), manual and email methods. Main electronic databases [Medline via PubMed, EMBASE, CINAHL, PopLine and Web of Science], direct Google search and other relevant sources were used to access studies before December 31st, 2017. In addition, emails were sent to authors whose studies were included to request studies. In addition, relevant citations from retrieved studies were searched. Search strings were constructed using keywords and their combinations based on the review questions. However, search strings were modified to suit to the databases interface accordingly. The overall detail of the search strategies are presented as Additional file 2: 2-1 to 2-4. The overall search results were exported to the Endnote citation manager software [44], and duplicate studies were removed.

Eligibility criteria and type of studies included

Observational and interventional studies were eligible to determine the level of, and factors associated with domestic VAW as well as effective interventions and key recommendations to prevent VAW. In addition, published and unpublished studies that have been conducted on women (15–49 years) in LLMICs to assess VAW and were written in English (1994–2017) were eligible for the systematic review. The LLMICs were selected based on the World Bank’s country classification [45]. Case series, editorials, commentaries, life stories and fact sheet reports on VAW were excluded.

Selection of studies

Studies were selected using eligibility criteria and screened through four steps for the systematic review and meta-analysis. Initially, studies were screened and selected for subsequent evaluation based on their titles and abstracts, that is, if studies clearly reported on domestic VAW and its associated factors. Secondly, the two authors (AS and GT) independently screened the studies’ abstract section (aims, methods, results and conclusion) to proceed to the next step. Studies whose abstract section briefly reported the prevalence of domestic VAW and its associated factors were included in the next evaluation process. Thirdly, studies selected by abstract screening were re-assessed independently by authors (AS and GT) with focus on the full-text. Eventually, selected studies were appraised for final inclusion in the systematic review and meta-analysis. In cases where the authors could not reach a consensus on studies, a third person was involved to appraise using same checklist, in the hope of helping to make a final decision. The studies selection process adhered to the PRISMA flow diagram [43] (Fig. 1).

Fig. 1
figure 1

Diagramatic presentation of the selection process of studies for systeamtic review

Full size image

Measurement of outcomes and exposures

The two phrases [domestic violence and intimate partner violence (IPV)] were interchangeably used as an outcome variable in this systematic review process. The outcome was assessed based on the prevalence of domestic/intimate partner VAW (psychological, physical, sexual and concurrent). IPV was presented in two forms (lifetime versus current). The lifetime IPV was assessed using the studies’ report of women’s experience of IPV during their lifetime. Furthermore, women’s experience of IPV over the last 12 months preceding the survey was labeled as current IPV. Exposure variables were classified according to the ecological model (individual, relationship, community and societal level factors).

Quality assurance of the systematic review

Published and unpublished studies were searched for and considered for this systematic review to minimize publication bias. The electronic, manual and email search strategies were carried out to ensure comprehensive retrieval of studies from main databases and other relevant sources. Eligibility, quality assessment criteria, selection process and data extraction templates were properly designed by the authors to assure quality. Methodological quality assessment of the studies was carried out using the Joanna Briggs Institute (JBI) critical appraisal checklist for observational studies [46], for more detail of the critical appraisal is presented as Additional file 3. The authors (AS and GT) performed the selection of studies and data abstractions. Any disagreements were resolved through consensus, and sometimes, other authors who were not involved in data extraction adjudicated to make final decision. Potential publication biases were explored using the funnel plot. Detail of funnel plots for each outcome variables is included as Additional file 4: 4-1. Heterogeneity between included studies was assessed using I2. Sub-group analysis, random effects model and qualitative narration were carried out to minimize the risk of bias. In addition, the risk of bias (ROB) for interventional studies was assessed. The detail of the risk of bias assessment is attached as Additional file 4: 4-2.

Study description and data extraction

The studies’ characteristics (authors-date, study area/country), its aims, designs, sample size, sampling procedure, response rate, key findings and recommendations are described on a template. More details of the studies characteristics description are presented on tables [see Tables 1 and 2]. Two authors (AS and GT) abstracted the data from selected studies and labeled the data extraction template using Microsoft Excel sheet. Quantitative data, number of women who had experienced VAW (labeled = Yes), who had not experienced VAW (labeled = No) and total participants (n) were separately recorded in a Microsoft Excel sheet. The more detail of abstracted data is stored in Excel Sheet (see Additional file 5).

Table 1 description of observational studies included in the systematic review and meta-analysis
Full size table
Table 2 description of the interventional studies
Full size table

Data synthesis

The pooled estimate of domestic VAW was computed using the Comprehensive Meta-Analysis (CMA) software [91]. Substantial heterogeneity was assumed to be I2 (> 75%) [92, 93]. Potential publication bias was checked through visual assessment of the funnel plot [94, 95]. The random effects model [96] was used to moderate the sample size variation which might have had an influence on the pooled estimate. In addition, sub-group meta-analysis was performed by study settings and types of domestic VAW (lifetime, current, psychological, physical and sexual violence). Furthermore, the associated factors with VAW were qualitatively synthesized according to the ecological framework model (individual, relationship, community and societal factors) [97].

Results

Intimate partner violence in LLMICs

Lifetime intimate partner violence

Nineteen studies with a sample of 35,974 women (15–49 years), the pooled estimate of lifetime IPV was 55% (95% CI: 52, 59%). IPV in sub-Saharan Africa (SSA) (14 studies) and Asian countries (4 studies) were 59% (95% CI: 52, 65%) and 46% (95% CI: 28, 65%), respectively (Fig. 2).

Fig. 2
figure 2

Forest plot of the lifetime intimate partner violence in LLMICs (n = 19)

Full size image

Physical intimate partner violence

From 18 studies with a sample of 44,664 women (15–49 years, the pooled prevalence of lifetime physical IPV was 39% (95% CI: 33, 45%). Furthermore, lifetime IPV in SSA was 43% (95% CI: 35, 50%) (Fig. 3).

Fig. 3
figure 3

Forest plot of the Lifetime physical intimate partner violence against women (n = 18)

Full size image

Psychological intimate partner violence

From 15 studies with a sample of 42,600 women (15–49 years), the pooled prevalence of lifetime psychological IPV was 46% (95% CI: 40, 52%). The sub-group analysis shows consistent findings with the overall pool prevalence across the regions (Fig. 4).

Fig. 4
figure 4

Forest plot of the lifetime psychological intimate partner violence against women (n = 15)

Full size image

Sexual intimate partner violence

From 15 studies with a sample of 29,127 women (15–49 years), the pooled prevalence of sexual IPV was 20% (95% CI: 17, 23%). The lifetime sexual IPV was the highest in SSA [42% (95% CI: 32, 54%)] (Fig. 5).

Fig. 5
figure 5

Forest plot of the lifetime sexual intimate partner violence against women (n = 15)

Full size image

Current intimate partner violence

Thirty three studies with a sample of 216,043 women (15–49 years), the pooled prevalence of IPV was 38% (95% CI: 34, 43%). The prevalence in SSA is almost similar with the pooled prevalence (Fig. 6).

Fig. 6
figure 6

Forest plot of the current intimate partner violence against women (n = 33)

Full size image

Types of current intimate partner violence

Physical intimate partner violence

From thirty one studies with a sample of 141,820 women (15–49 years), the pooled prevalence of physical IPV during the past 12 months was 25% (95% CI: 21, 28%). The subgroup analysis of seven studies in Asian countries was 31% (95% CI: 22, 41%) (Fig. 7).

Fig. 7
figure 7

Forest plot of the current physical intimate partner violence against women (n = 31)

Full size image

Psychological intimate partner violence

From 20 studies with a sample of 115,798 women (15–49 years), the pooled prevalence of psychological IPV was 30% (95% CI: 24.0, 36%) (Fig. 8).

Fig. 8
figure 8

Forest plot of the current psychological intimate partner violence against women (n = 20)

Full size image

Sexual intimate partner violence

From 27 studies with a sample of 124,739 women (15–49 years), the pooled prevalence of current sexual IPV was 7% (95% CI: 7, 8%). However, the subgroup pooled prevalence was a bit higher than the overall pooled prevalence. It was 19% (95% CI: 13, 27%) in SSA countries (Fig. 9).

Fig. 9
figure 9

Forest plot of the current sexual intimate partner violence against women (n = 27)

Full size image

Concurrent intimate partner violence

From eleven studies with a sample of 8315 women (15–49 years), the pooled prevalence of women’s experience of IPV concurrently was 13% (95% CI: 12, 15%). In the meantime, the prevalence of concurrent IPV in SSA was 27% (95% CI: 16, 42%) which is two times higher than the overall pooled IPV prevalence in LLMICs (Fig. 10).

Fig. 10
figure 10

Forest plot of the pooled prevalence of concurrent intimate partner violence against women (n = 11)

Full size image

Contributing factors of domestic violence against women

Individual level factors

Socio-demographic factors

The socio-demographic characteristics of the couples were identified as factors associated with IPV. In six studies, place of residence was one of the factor associated with IPV whereby women living in rural area were more likely to experience IPV than urban dwellers [25, 52, 61, 62, 65, 79]. Similarly, in two studies, women’s religion was a factor associated with IPV, such that women belonging to Islam were more likely to experience IPV than non-Muslim [65, 70]. In addition to Muslim women, women belonging to the traditional religion (worshipers) and being in other faith(s) were more likely to accept physical IPV than women belonging to Christianity [52].

Age

Differences between women’s and their husbands’ age had an inconsistent relationship with the experience of IPV. In three studies, physical IPV was significantly associated with women’s age and age at first marriage [40, 48, 56]. Women’s age 20 years and younger was a risk factor associated with IPV [82]. In one study, women who married at age 15 or young were four times more likely to experience IPV than women got married older than 15 years [64]. While women (15–19 years) were less likely to report emotional IPV, women (25–29 years) and women (30–34 years) were more likely to report being physical and sexual IPV, respectively than women (45–49 years) [76]. Nevertheless, in two studies, older age women (35–49 years) were three times more likely to report lifetime and current IPV than women (15-19 years) [53, 79]. In addition, the age of husbands was a positive predictor of IPV [40, 49], but sexual IPV was associated with husbands younger than 35 years [51].

Education

Women’s lifetime IPV remained significantly associated with women’s level of education. In six studies, uneducated or primary educated women had almost double the prevalence rate of any form of IPV who attended secondary or higher education [6, 48, 52, 53, 60, 65]. In five studies, the higher the women’s educational level, the lesser the likelihood of experiencing physical IPV. Women with higher education than their husbands were less likely to experience IPV than women with equal or less education than their husbands [40, 55, 56, 62, 71]. In two studies, women in equally high educated marriages revealed the lowest likelihood of experiencing IPV. Education was found to be significantly protective from IPV, for both women as victim and men as perpetrator [55, 79]. Furthermore, the educational status of husbands was associated with women’s experience of less IPV and less violent behavior to their wives [27, 50, 53, 54, 58]. On the other hand, women with no education were about three times more likely to approve (accept) IPV than women with secondary or higher education [52]. IPV remained higher and even increased in the case of women with secondary to vocational levels of education as compared to those with a higher level of education [57, 66].

Occupation

In two studies, women who were engaged in manual labor (farming) were three times more likely to be exposed to IPV than women with non-farming occupations [66, 70]. In one study, women who has been in the poorest microfinance group under supported by the microfinance programmes in Bangladesh have no significant change on IPV exposure, except some improvement on economic empowerment [65]. Likewise, the women whose partners are employed (earned in cash) were 93% less likely to experience physical IPV than those women whose partners were farmers (earned in kind) [70]. Likewise, in three studies, women were economically dependent, and or did not have their own means of earnings and controlling their earnings were more likely to report sexual IPV than their counterparts [49, 51, 72]. Similarly, women heading business or engaged in different types of jobs were 50% less likely to report lifetime IPV than women have not heading business (almost jobless women) [79]. On the contrary, housewives were less likely to be experienced sexual IPV than women who have a job different from housewives [68].

Childhood witness

Domestic IPV was significantly associated with women and their husband childhood wittiness of parental violence. Although the strength of association varies (ranged from four to thirteen times), those women whose mothers were hit by their fathers during their childhood were more likely to report lifetime and current IPV than their counterparts [25, 53, 61, 72, 79]. Likewise, in some studies, women whose husbands were beaten by someone in their family during their childhood were two times more likely to report experiences of IPV than those who were not beaten during childhood [53, 72, 79]. In addition, women’s exposure to physical IPV was 5 to 6 times more likely to be higher on women whose husband having witnessed that their mothers being beaten by their fathers than women whose husband had no history (witness) of maternal IPV [51, 66, 70]. Furthermore, in one study, IPV was associated with women who had a history of sexual abuse during childhood, adolescence or even an early age of first sex [60].

Husbands’ controlling behaviors and mental health condition

IPV was higher among women who were afraid of their husbands [53, 80]. Women whose husbands had controlling, hostile and or rude behaviors were almost three to four times more likely exposed to any type of IPV than their counterparts [47, 53, 68, 77, 79]. Women who had high levels of emotional distress were associated with IPV [82]. Furthermore, women whose husbands previously engaged in physical fight were 3.5 times more likely to experience IPV than others. In a similar way, women with an unhealthy mental status were two times more likely to experience IPV than women with a healthy mental status [47].

Husband alcohol consumption

Women whose husbands’ drank alcohol were more likely to experience IPV than women whose husbands did not drink alcohol. As we found from several studies, husbands’ alcohol consumption is the most commonly reported factor associated with IPV [6, 25, 49,50,51, 56, 60, 62,63,64, 66, 70, 79, 80]. Furthermore, in one study, IPV was significantly associated with husband’s use of psychoactive substances [66].

Pregnancy status

In three studies, pregnant women with high parities had a higher probability of experiencing lifetime IPV than non-pregnant women [48, 49, 66]. Furthermore, in one study, women whose pregnancies were undesired by their partners was six times more likely to have risk of current IPV than pregnancy desired by partner [64]. Likewise, in two studies, the likelihood of women’s experience of lifetime IPV was consistently higher as the number of children increased [48, 54].

Relationship factors

Women’s decision making status

Decision making power was a predictor of IPV [49]. Generally, the odds of IPV was less by 50% for women who had an equal say in household decision-making. Sexual violence was 35% less likely to occur among women who had a share in household decision-making [25]. Interestingly, the probability of women being physically abused decreased by 8.2% for those women whose husbands dominated household decision-making, whereas wife-dominance in household decision-making had a marginal effect on physical violence [54]. However, women who decided on spending their own or husband’s earning with a joint decision-making approach in view of their own healthcare utilization, large household purchase, or contraceptive use were less likely to be victims of IPV than women who made decisions by themselves [53].

Infidelity and lack of satisfaction in marriage

In four studies, women whose husbands had engaged in extramarital sex or had multiple sexual partners (unfaithful) were two times more likely to be at risk of IPV than their counterparts [51, 62, 72, 80]. Yet, exposure to physical IPV was associated with being unfaithful. Women who agreed that a woman was obliged to have sex with her husband had a lower risk of exposure to physical IPV than those who disagreed [51]. Women who had poor relationships with their husbands were 2.6 times more likely to experience physical IPV [58]. Furthermore, women who had a satisfaction in their marriage were noted to face a low risk of IPV [62]. Sexual IPV was higher among women whose partner were jealous if they talked with other men, suspected them of unfaithfulness, did not permit them to meet even female friends, limit their contact with family [80]. Those women who did not believe a wife could do anything if a husband wanted a girlfriend were three times more likely to be exposed to physical IPV [68]. In the same view, women who could refuse sex with their partners or ask their husbands to use condoms were two times more likely to be victims of IPV than their peers who could not [53]. Women who had worries on issues about their daily activities and did not discuss them with their partners were more likely to experience IPV than women who did discuss their issues [66].

Type of marriage

In five studies, the type of marriage was significantly associated with IPV. Women who were married or cohabitated by abduction; women married to distant relatives; women with a polygamous partner or in a polygamous marriage; payment of dowry and marriage undergone without ceremony were more likely to experience IPV than their counterparts [27, 48, 50, 61, 79].

Duration of marriage

In two studies, this factor was significantly associated with higher experience of physical abuse. Women who were married for 5–9 years (OR, 3.8) or ten or more years (OR, 3.7) were at higher risk of being abused than women who were married for less than 5 years [48, 60].

Wealth index and economic status

IPV and wealth index had an association but an inconsistent relationship across the wealth quintiles. In six studies, women belonging to the poorest wealth index categories were most likely to be exposed to IPV than women in the rich wealth index [51, 52, 54, 63, 65, 74]. Other studies, microfinance program membership was associated with a two-three-fold increase in exposure to IPV [65], as were family and financial problems associated with IPV [66, 79].

Community level factors

Presence of the traditional gender norms

In two studies, women’s exposure to IPV were more likely to be high in communities who adhered to traditional patriarchal gender norms or beliefs and supported (accepted) attitudes towards wife beating [51, 54]. Meanwhile, women who agreed that “a good wife obeys her husband” and/ or “a man should show who the boss is” were more likely to experience sexual IPV than women who disagreed. In addition, women who agreed that a woman had no reason to refuse sex with her husband were three times more likely to be exposed to IPV as compared to who agreed for some reasons [51]. In addition, women who worked outside the home but whose husbands did not make enough money had an increased risk of IPV by 5.2% than women whose husbands made enough money [54].

Community attitude towards wife-beating

Communities whose attitudes supported IPV [66] by thinking that justified wife-beating is acceptable [60], or encouraged societal gender beliefs or norms [54], reflected relationship control, and relationship inequalities [40, 51] were significantly associated with women’s experience of IPV.

What were the key recommendations from the studies?

Transformation of community’s traditional gender norms

In seven studies, undertaking a massive and intensive information, education and communication (IEC) approach is a recommended strategy on transforming a community’s culture and traditional gender norms in order to enhance gender equality [25, 27, 38, 47, 49, 51, 61]. Similarly, in eight studies, there is a need to employs comprehensive and culturally acceptable approaches including medical (psychiatric) counseling, community mobilization, gender advocacy and effective development of IEC to dispel myths, misconceptions, negative traditional norms and beliefs, gender inequality and to reduce the costs of IPV [6, 40, 54, 68, 70, 72, 75, 79].

Human right based approach

In one study, community level awareness of human rights as well as advocacy for women’s rights is crucial [67]. In four studies, more investments in IPV prevention strategies are needed to address the intergenerational transfer of deeply entrenched cultural-norms which support male dominance and gender inequality [57, 66, 73, 90].

Stakeholder engagement

The urgent attention of policymakers, stakeholders, professionals and other concerned bodies is needed at all levels of society. Stakeholders should design interventions targeting behavioral and social factors which can help to prevent IPV [79]. Likewise, resources should be mobilized by policy-makers, public health experts, researchers and other stakeholders to prevent IPV [48, 59, 63, 72]. Advocacy is very important and can be done as religious institutions, media, government and non-governmental associations encourage gender equality by [66, 67, 70, 73].

Policy formulation and legal framework

The issue of gender equality, women’s rights and legal sanctions need due attention during policy formulation and endorsement of laws to prevent VAW are crucial. In many instance of VAW, the punishment to perpetrators were light and not commensurate with the offence victims filed. The law should be more responsive to VAW to help address this challenge [25, 68]. Perhaps, survivors are encouraged to disclose their experience of IPV to people who are in position or have an autonomy, and implementing the existing law for punishment [70]. Policymakers should take immediate action to break hierarchical barriers between spouses, and promote gender equality while amending the existing laws or formulating new policies [52, 53, 75]. However, evidence based efforts are needed to re-enforce legal rights or existing laws and policies and ensure their effective implementation to prevent and respond to VAW [54, 57, 67].

Women empowerment

Six studies have recommended that building women’s capacity through education, employment, income and other economic opportunities, and addressing imbalance of power between men and women are crucial to prevent IPV [25, 47, 52, 53, 56, 79]. Likewise, promising public health strategies (increasing awareness of the consequences of IPV, strengthening the self-esteem of women and girls and promoting equity in marital relationships) are needed to change attitudes towards gender inequality, and are essential to avert IPV [47, 70, 79]. Furthermore, addressing household poverty-wealth comprehensively [74], enhancing the safety of women, promoting fertility control methods and women’s reproductive health service [40] are some other recommended interventions to prevent and control IPV.

Intervention integration

Innovative strategies are needed to provide support and counselling to IPV survivors, who are rarely assisted by health care professionals [82]. Research findings have strongly recommended that IPV prevention interventions be integrated with community health programs; reproductive health and other health services to be more comprehensive, close to household level and accessible to IPV survivors [27, 49, 53, 67, 70]. Integration may help to enhance medical screening of survivors for STI including HIV, provide male partner counseling and other health care support. Furthermore, the authors [38, 48, 67, 78] recommended that reproductive health service providers should be encouraged to advocate for IPV and mental health screening during antenatal care.

Engaging men and other influential persons on IPV prevention

Research findings have proposed that community programs that have a couple-centered approach are needed to promote non-violent masculinity values and norms [47, 51, 58]. Moreover, working with men is a win-win-approach to prevent IPV. Efforts are needed to focus societal, community, relationship and individual level approaches which engage men to promote men’s non-violent behavior and gender equality and to minimize infidelity [51, 62, 67]. In addition, interventions addressing IPV should place more emphasis on reducing partners’ controlling behaviors and to prevent men’s alcohol drinking habits [56, 62, 66, 80]. Furthermore, involving mother-in laws would be a significant move [47]. Also, community health workers could be active players in raising community awareness about IPV [27, 49].

What are the proven evidence?

We systematically selected eight interventional studies that had been conducted in LLMICs [83,84,85,86,87,88,89,90]. The detail of GRADE for the summary of evidence for different outcomes is attached as an Additional file 6. In two studies, the social accepting attitude of IPV was lowered by community mobilization intervention. Women’s attitude towards sex refusal when necessary was 1.3 times more likely higher in the community mobilization intervention than control group. Likewise, current physical, sexual and concurrent IPV were lowered in the intervention group by 52, 24%, and 435 respectively than the control group. In addition, women experiencing IPV in intervention communities were more likely to receive supportive community responses. [83, 90]. In one study, the community-level normative attitude towards physical IPV and IPV acceptability norms were improved in the intervention group than the control. In addition, men’s suspicion of their partner for infidelity and communication about sex were improved [90]. Similarly, in one study, community engagement and group education combined interventions reduced IPV almost by 20% while the CE-only group reduced it by 23%, and enhanced gender equitable norms [85].

In the two interventional studies which is mainly focused on women’s economic empowerment through a village loan and saving association approach, women in the combined groups were significantly less likely to report economic abuse than control group (OR, 0.39, 95% CI: 0.25, 0.60) [84, 86]. In one study, while attitude towards refusal of sex did not significantly change, women in the combined (VLSA and GDG) intervention group reported a lower experience of current IPV and also acceptance of justified wife-beating than VLSA alone, but it was not statistically significant [84]. In one study, women who got married when they were children were 46% less likely to report physical and or sexual violence in the combined intervention group than the control. The reduction was however, not statistically significant in the overall IPV [86].

Women’s attitudes about IPV and power relationships were associated with their IPV experience. In one study, women who report violence was ever justified if a woman refuses sex were two times more likely to experience IPV than control group. Furthermore, women in joint (both partners) sexual decision making relationships were 30% less likely to report IPV as compared to women whose partners controlled sexual decision-making. Notably, women were 57% less likely to report IPV when both partners had equal power [87]. However, women’s economic empowerment was doubled for those women in the safe home and respect for everyone (SHARE) intervention group. Likewise, women in the intervention group were 20% less likely to experience IPV than those in the control group. Nevertheless, SHARE had no significant effect on emotional IPV, men’s behavior (perpetration) [88], overall IPV, women’s autonomy and women’s attitude towards gender norms [89].

Discussion

This systematic review and meta-analysis determined the prevalence of domestic IPV and its types, associated factors, effective interventions and key recommendations to prevent domestic VAW. We found out that the pooled prevalence of lifetime IPV was 55% (95% CI: 52, 59%). Of these, lifetime physical, psychological and sexual violence were 39, 45 and 20%, respectively. Furthermore, the pooled prevalence of current IPV was 38% (95% CI: 33.0, 43%). Of these, prevalence of current physical, psychological and sexual violence were 25, 30 and 7%, respectively. In addition, the pooled prevalence of women’s experience of concurrent IPV was 13% (95% CI: 12, 15%). This finding is consistent with the WHO’s global estimates and multicounty study whose findings indicated that one-in- three women experienced domestic VAW in their lifetime [1, 98, 99].

As evidence shows that IPV was significantly associated with educational level, place of residence, economic status, having witnessed abuse during childhood, husbands’ having controlling behavior, husbands’ alcohol consumption, pregnancy status and parity. Notwithstanding, sexual violence was common among women who had husbands younger than 35 years. Furthermore, women in unfaithful relationships and unsatisfied marriages as well as women who married early and have experienced forced first sex, short duration of marital life, and poor mental health had a higher risk of experiencing IPV. In addition, the presence of traditional gender-norms and wife-beating accepting attitudes were linked with a high risk of domestic violence. This finding is also consistent with the ecological model formulated in 1998 [97, 100] which is very applicable in low and lower-middle income settings and is targeted at the deep rooted causes.

This systematic review found out that IPV prevention should focus on community culture or traditional gender-norm transformation; stakeholders’ engagement; women empowerment (capacity building); engaging men and other influential people (mother and/ or father in-laws, sister-in-law, neighbors); intervention should focus on service integration with other relevant sectors (mainly health sector for screening and other care and support); policy formulation and provision of legal framework and implementation of human right based approaches. This finding is consistent with other reviews that recommended focus on the structural drivers of unequal power in relationships to prevent VAW [98, 100,101,102].

We included observational and interventional studies. However, we interpreted the findings to meet our review objectives. Nevertheless, we did not mix-up the findings of the observational and interventional studies. Though, this systematic review and meta-analysis used relatively comprehensive search of the major databases; included both quantitative and qualitative studies; published and unpublished studies and sub-group analysis by setting and violence types, it had some limitations. One of the main limitations was the inconsistent definition of IPV across some studies. Hence, we used both IPV and domestic VAW synonymously. The second limitation of this systematic review and meta-analysis is that it did not carry out quantitative synthesis on the factors associated with domestic VAW due to the high heterogeneity between included studies. However, it is difficult to find absolutely homogeneous studies in terms of setting, method, analysis and interpretation.

Implication of the review

Domestic violence against women is a common women’s life experience and mostly perceived as minor and socially tricky by many governments. However, its consequences have a devastating impact on national gross domestic product and costs much higher than the budget allotted for primary education. Hence, generating evidence on the prevalence, associated factors and identifying effective interventions applicable in poorly resourced settings is very crucial. This systematic review and meta-analysis can give critical insight about VAW, associated factors, and effective interventions. Therefore, concerned stakeholders can use the findings of this study as main evidence to inform policymakers, program designers and local planners construct and for implement policies to prevent IPV in LLMICs.

Conclusion

Both lifetime and current domestic IPV are still high in LLMICs. More than half of the women in the studies had experienced lifetime domestic IPV, and almost one-third of the women had experienced current IPV. In addition, almost one-in-ten women experienced more than one type of IPV concurrently. Domestic IPV is a complex public health and human rights violation which is associated with factors at the individual, relationship, community and entire system level. However, most of the associated factors are preventable. Interventions integrating legal framework and programs that focus on transformation of traditional gender-norms are most important to prevent IPV. Community mobilization and awareness creation to transform gender-norm reduced IPV by half. However, the economic empowerment intervention had reduced some influences on women (reduced economic abuse), but it was not statistically significant in the prevention of IPV. Therefore, we suggest that researchers, program planners, policy makers, clinicians and other concerned stakeholders should invest in the implementation of gender-norms focused on community based interventions to prevent IPV.