Abstract

Background. Intimate partner violence (IPV) against pregnant women is a recognized global public health problem affecting all spheres of women and unborn infants. In Ethiopia, although inconsistent, individual studies avail; there is a dearth of systematic reviews and meta-analysis about the prevalence and associated factors of intimate partner violence. Thus, the present study was aimed at determining the pooled prevalence of IPV and its determinant factors during pregnancy. Methods. The report of meta-analysis follows the Preferred Reporting Items for Systematic Review and Meta-Analysis 20 guideline (PRISMA 20). Databases including PubMed/Medline, CINAHL, SCOPUS, HINARI (research4life), AJOL, IRIS, and AIM were searched. Heterogeneity test was assessed by the Cochrane chi-square () and quantified by statistics test. Publication bias was tested by funnel plots and Egger’s test. Sensitivity test and subgroup analysis were also performed. Effect size was calculated by random effects model. Results. A total of 26 studies, including data from 13, 912 participants, were included in the analysis. The prevalence of IPV ranged from 7% to 81% with overall estimated pooled prevalence of 37% (30% -44%, %, ). Of this, the prevalence of physical, sexual, and psychological violence was 24% (95% CI; 19%-30%), 21% (95% CI; 16%-26%), and 27% (95% CI; 22%-32%), respectively. Factors such as lack of formal education, childhood violence, rural residency, low decision-making power, family history of violence, attitude, unplanned and unwanted pregnancy by women and partners, late initiation of antenatal care, partner alcohol, and khat use were associated with IPV. Conclusion. More than one-third of pregnant women experienced IPV. The most prevalent form of IPV was psychological violence followed by physical and sexual violence. The identified risk for IPV including victim, pregnancy, and perpetrator-related factors indicated the need of a holistic approach in the promotion, prevention, and treatment of IPV. The finding of this study suggests the need of strengthening women empowerments (capacity building) against traditional beliefs, attitudes, and practices. This study also suggests the need of evaluation and strengthening the collaborative work among different sectors such as policy-makers, service providers, administrative personnel, and community leaders, including the engagement of men partners.

1. Background

Intimate partner violence (IPV) against pregnant women is a recognized global public health problem [1, 2] affecting psychological, social, physical, and reproductive health including the following: unintended pregnancy, induced abortion, bleeding, HIV, and other sexual transmitted infections [15]. Moreover, IPV affects the life of unborn infants by damaging the placenta, rupture of the uterus, fetal trauma [2, 6, 7], spontaneous abortion, preterm labor, preterm delivery, and low birth weight [4, 8, 9].

Globally, one in every three women experienced IPV at some point in their life time [6]. In Africa, the prevalence of IPV was found to be 37% [6]. In Ethiopia, the prevalence of IPV was ranging from 50% to 81% [1, 4, 713]. The high proportion of IPV in Ethiopia was associated with the deep-rooted perception of the communities’ and women’s acceptance of IPV [14]. That is why up to 93% of women who have experienced IPV were not disclosed to anyone [2, 12, 13, 1517]. According to the 2016 Ethiopian Demographic and Health Survey (EDHS2016), out of 34% of women who have experienced IPV, 66% of them were not told to any one [18]. During pregnancy, the prevalence of physical, sexual, psychological, and overall prevalence of IPV was ranging from 2% to 35%, 9% to 40%, 22% to 65% [19], and 0.9% to 57% [5, 20, 21], respectively.

Although many of the risk factors were similar among pregnant and nonpregnant women, various studies have shown that pregnancy could be one of the risk factors for the initiation or escalation of IPV for certain women. Generally, factors including (i) pregnancy-related such as unplanned pregnancy, unwanted pregnancies, and elective termination of pregnancy; (ii) victim-related characteristics such as age, marital status, ethnicity, education, employment, substance use; and (iii) perpetrator-related characteristics such as dependency, jealousy, and possessiveness towards female partner were factors associated with IPV during pregnancy [5, 20, 21]. In spite of the high prevalence of IPV and its associated adverse impact, there is a dearth of epidemiological evidence in Ethiopia and only a systematic review and meta-analysis were reported from eight studies [22]. Since, the publications of this systematic review and meta-analysis (2018), several studies have been published. Therefore, the current study could benefit the field by addressing the identified gap such as limited number of studies (8 studies), survey period, study sites, and the inconsistent individual reports with ranging prevalence of 7% to 81% [4, 7, 2346]. This discrepancy between studies could be due to the variation in survey periods, sampling, study sites, diagnostic criteria, and demographic characteristics. Evidence-based strategies such as systematic reviews and meta-analyses are ideal to address such issues. Therefore, the present study was aimed at determining the estimate pooled prevalence of IPV and associated factors against pregnant women in Ethiopia.

2. Methods

This systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Review and Meta-Analysis 20 (PRISMA 20) [47].

2.1. Search Strategy

Initially, the Database of Abstracts of Reviews of Effects (DARE) and the Cochrane Database of Systematic Reviews (CDSR) were searched to confirm the absence of similar studies in Ethiopia.

Then, the first search strategy was started using a compressive search of electronic databases including the following: PubMed/Medline, CINAHL, Scopus, HINARI (research4life), AJOL, IRIS, and AIM. Using the following key words: epidemiology, prevalence, incidence, gender based violence, domestic violence, intimate partner violence, spouses violence, physical abuse, physical violence, emotions violence, emotions abuse, psychological violence, psychological abuse, sex violence, sex abuse, sexual coercion, rape, associated, risk, pregnancy, pregnant, maternal, pregnant mothers, antenatal mothers, pregnant women, antenatal, during pregnancy, and Ethiopia; Medical Subject Headings (MeSH) terms and Boolean operators (“AND” and/“OR”) were established for each databases. A combination of MeSH thesaurus and text words combined with appropriate Boolean operators (AND and OR) was formulated for PubMed/Medline (full search query for the database PubMed/MEDLINE at annexed). Similarly, CINAHL, Scopus, HINARI (research4life), IRIS, AJOL, and AIM were searched using search terms tailored to each database. The second stage of the search phase involved scanning relevant websites, including national and international institutions and research centers, and web of science (Google and Google Scholars) for grey literatures. Finally, the reference lists of the included articles were searched. All searches were performed until January, 07, 2021.

2.2. Selection of Studies

All retrieved studies were imported in EndNote X7 (Thomson Reuters, New York, USA) and duplicates were removed. The study selection process had two stages. The first stage was screening the title and abstract of the retrieved studies. The second stage was a perusal of their full text to determine eligibility. Two reviewers (YDG and BBB) independently screened the studies. Disagreements were solved by discussion.

3. Definition of Concepts

In this study, IPV was defined as any violence whether physical, psychological, and sexual, or any combination of them, regardless of the legal status of the relationship. Physical violence is defined as one or more intentional acts of physical aggression such as pushing, slapping, throwing, hair pulling, punching, hitting, kicking, or burning, perpetrated with the potential to cause harm, injury, or death. Psychological violence is defined as one or more acts or threats of acts including shouting, controlling, intimidating, humiliating, and threatening the victim [1, 2]. Sexual violence is defined as the use of force, coercion, or psychological intimidation to force woman to engage in a sex act against her will, whether or not it is completed [1, 2, 48].

3.1. Eligibility Criteria

Participants: this review targeted studies that were carried out during pregnancy

Outcome measure: this systematic review and meta-analysis has two main outcomes which are the (i) estimated pooled prevalence of IPV and (ii) factors associated with IPV

Study design: observational studies (cross-sectional and cohort/longitudinal)

Setting: this review included studies that were carried out in Ethiopia.

Studies that focused on case reports, conference, and abstracts were excluded.

3.2. Data Extraction

Data were extracted from the eligible studies using a preconceived and piloted data extraction Microsoft Excel spreadsheet. Data were simultaneously extracted by two independent reviewers (BBB and YDG). The extracted data items include the following: name of the first author, year of publication, study setting/region, study design, data collection tool, sample size, and number of cases/prevalence. For the analysis of associated factors, adjusted odds ratio at 95% confidence interval was extracted to address confounders.

3.3. Quality Assessment

The Newcastle-Ottawa quality assessment tool, adapted for cross-sectional studies [49], was used for quality assessment. This tool has three main parts (selection, comparability, and outcome). The first part (selection) has five stars and assesses the methodological quality of the study. The second part of the tool evaluates the comparability of the study. The third part of the tool assesses the quality of the original article’s outcome with respect to the statistical analysis. Individual paper was graded with a score ranging from zero to ten stars. The overall quality of each article was determined using the sum of each star of the three parts and defined as high quality for a out of 10, medium (fulfilling 50% of the quality assessment criteria), and poor .

3.4. Data Analysis

The extracted data were exported to Stata version 14. Metaprop command in Stata was used to calculate the pooled prevalence estimate of DV with 95% confidence intervals (CIs) using a random-effects model via the DerSimonian and Laird transformed inverse-variance method [50]. Heterogeneity was assessed using the Cochrane chi-square () and statistics. value greater than 50% was considered indicative of substantial heterogeneity [51]. Sensitivity analyses were carried out to assess the contribution of each individual study to the overall effect size. In addition to this, subgroup analyses were conducted by study region, sample size, quality of each study, and publication year. Furthermore, publication bias was tested by the funnel plot [52] and Egger’s test [53]. A was considered indicative of statistically significant publication bias. The meta-analysis of associated factors was conducted when at least two studies reported the same associated factors based on the overall IPV [54]. Results were presented using texts, tables, forest plot, and summary of descriptive statistics.

4. Results

The initial database search resulted in 3747 publications. Of these records, 2327/1474 were excluded because of duplication. Of the remaining 2300, 1009 studies were excluded as their title and abstract reading were not related to the outcome. From the remaining 1291 studies, 1190 articles were excluded based on the full text/information and found unclear results. Again, from the remaining 101 studies, 76 publications were excluded based on the eligibility criteria. Moreover, from the other source, 6 studies were located and following the removal of 5 studies because the focus were not pregnant women IPV. The remaining 26 studies were included in the systematic review and meta-analysis (Figure 1).

4.1. Study Characteristics

In this study (detail Table 1), a total of 26 studies, including data from 13, 912 study participants, were included in the analysis for the overall IPV. Regarding the types of IPV, the most common type of reported IPV was physical violence (), followed by psychological (), and sexual violence (). Moreover, for the analysis of associated factors, a total of 17 studies, including data from 10, 940 participants, were included. These studies were carried out in five different regions: Southern Nation and Nationalities of People (), Amhara (), Oromia (), Addis Ababa (), and Tigray (). Majority (62%) of the included studies used institution-based cross-sectional study design. Eighty-five percent of the included studies used probability (e.g., random sampling) sampling, whereas 15% relied on nonprobability (convenience) sampling. Most (77%) studies were published between 2016 and 2020. The sample sizes of included studies ranged from 183 to 3015. For the assessment of IPV, the WHO questionnaires () and AAS () were used, whereas the remaining three studies used self-developed assessment tools.

4.2. Quality of Included Studies

The quality of the included studies was ranging from 4 to 9. Majority (88%) had good quality, while the remaining 12% had fair quality (Table 2).

4.3. Test of Heterogeneity, Sensitivity, Publication Bias, and Subgroup Analysis

We found evidence of significant heterogeneity for the overall estimate pooled prevalence of IPV (% and ), physical violence (% and ), sexual violence (% and ), and psychological violence (% and ) (Figures 2, 3, 4, and 5, respectively); yet, the sensitivity analysis showed that none of the point estimates was outside of the overall 95% confidence. The subgroup analyses of overall IPV by study setting, study design, year of publication, sample size, and quality of study were not identified the source of heterogeneity for the overall IPV; yet, the highest pooled prevalence was reported from studies carried out in Oromia region (50.7% (95% CI: 35%-66%). However, the subgroup analysis of physical violence indicated heterogeneity by the study setting (Tigray: 17% (95% CI: 12%-21%, , and )). Moreover, the subgroup analysis of sexual violence by study setting (SNNP: 12% (95% CI: 8%-15%, , and )) and study design (community based 12% (95% CI: 9%-15%, %, and ) also indicated as source of heterogeneity. Moreover, the subgroup analysis of psychological violence by year of publication (2003-2015 : 25% (95% CI: 21%-28%, %, and )) could be a source of heterogeneity (Table 3). Regarding the publication bias test, there was no evidence of publication bias from the visual inspection of the funnel plot and Egger’s test for the overall IPV [(Figure 6) and Egger’s test ()], sexual violence [(Figure 7) and Egger’s test ()], psychological violence [(Figure 8) and Egger’s test ()], and physical violence [(Figure 9) and Egger’s test ()].

4.4. Prevalence of Intimate Partner Violence

In this systematic review, the lowest (7%) and the highest (81%) prevalence of IPV were reported from Tigray and Oromia Region, respectively. The prevalence of IPV ranged from 7% to 81% with overall estimated pooled prevalence of 37% (30% -44%, %, and ). Of this, the prevalence of physical, sexual, and psychological violence was 24% (95% CI; 19%-30%), 21% (95% CI; 16%-26%), and 27% (95% CI; 22%-32%), respectively.

4.5. Factors Associated with Intimate Partner Violence during Pregnancy

A total of 17 studies, including 53 distinct odds ratios, from 13 unique risk factors were included in the meta-analysis. Based on previous similar studies, these factors were broadly categorized as follows: (i) victim-related: lack of formal education, history of childhood violence, rural residency, lack of decision-making power, family history of violence, and attitude toward IPV; (ii) pregnancy-related factors: unplanned pregnancy, unwanted pregnancy, and late initiation of ANC; and (iii) perpetrator-related factors: lack of formal education, partner alcohol use, partner khat chewing, and unwanted pregnancy (Table 4).

4.5.1. Victim Related

In this study, the pooled effect of five studies showed that those participants who have no formal education were around four times more likely experienced IPV compared to those pregnant women who have formal education (; 95% CI: 1.48, 6.27, , and ). Similarly, those participants who had history of childhood violence (; 95% CI: 1.37, 4.90, , and ), family history of violence (; 95% CI: 1.14, 2.22, , and ), attitude toward DV (; 95% CI: 6.58, 19.25, , and ), and rural residency (; 95% CI: 1.29, 3.67, , and ) were associated with DV. On the other hand, the pooled effect of four studies showed that those women's who had no decision making powers were more than 1.3 time (; 95% CI: 1.03, 1.58, , and ) more likely experienced IPV as compared to those who had decision making power; yet, insignificant.

4.5.2. Pregnancy-Related Factors

Unplanned pregnancy (; 95% CI: 1.48, 2.05, , and ), unwanted pregnancy (; 95% CI: 1.00, 2.08, , and ), and late initiation of ANC (; 95% CI: 1.15, 1.44, , and ) were pregnancy-related factors associated with pregnant women’s experience of IPV.

4.5.3. Perpetrator-Related Characteristics

In this study, the pooled effect of eleven studies showed that partner alcohol use was two times (; 95% CI: 1.66, 2.71, , and ) more likely experienced IPV compared to those who did not use alcohol. The result of publication bias from the visually examined funnel plot and Egger’s tests showed that there was no evidence of small study effect using Egger’s test (). The pooled effect of the three studies showed that partner khat use was around two times (; 95% CI: 1.42, 2.05, , and ) more likely experienced IPV compared to those partners who did not chew khat. The pooled effect of two studies showed that partner’s unwanted pregnancy was 1.2 times (; 95% CI: 1.03, 1.36, , and ) more likely experienced IPV. On the other hand, the pooled effect of six studies showed that the partner with no formal education was 1.1 times (; 95% CI: 0.93, 1.28, , and ) experienced IPV compared to those partner who have formal education, yet statistically not significant.

5. Discussion

In this systematic review and meta-analysis, more than one in three (37%) pregnant women experienced IPV. The overall pooled prevalence of psychological violence [27% (95% CI; 22%-32%)] was higher than physical [24% (95% CI; 19%-30%)] and sexual violence [21% (95% CI; 16%-26%)]. Comparing with other studies, the observed range of IPV (7% to 81%) in this systematic review and meta-analysis is consistent with previous systematic review and meta-analysis conducted in Ethiopia (12% to 45%) [23], Nigeria (2.3% to 44.6%) [21], African countries (0.9% to 57%) [5, 19, 21], and the WHO Multi-Country Study (1% to 28%) [55].

The estimated pooled prevalence of IPV (37%) in this study is higher than the reported prevalence of IPV in Ethiopia (26.1% (95% CI: 20, 32.3)), Africa (15%), China (7.7%) [56], and a meta-analysis of 92 studies across 23 countries [19.8% (13.3% for developed countries and 27.7% for developing countries [57]. On the other hand, lower than a systematic review and meta-analysis conducted in Iran (48% (95% CI: 38-58)) [58]. This variation may be attributed to the sociocultural difference of the study population, educational status, lack of access to correct information and reproductive health information, and stressful life events. This is supported by evidence from previous studies [5, 57] as factors such as unplanned pregnancy, low socioeconomic status, lower education status, stressful life events, and lack of social support were associated with IPV during pregnancy [5, 57, 58].

Regarding the different form of IPV, in this study, the overall pooled prevalence of psychological violence (27%), physical violence (24%), and sexual violence (21%) is similar with previous systematic review and meta-analysis carried out in Ethiopia with reported prevalence of 16%, 21%, and 16% for physical, psychological, and sexual violence, respectively [23]. Similarly findings from Iran revealed that the pooled prevalence of physical, psychosocial, and sexual violence was found to be 17% (95% CI: 12-32), 41% (95% CI: 33-50), and 21% (95% CI: 16-23), respectively [58]. Another systematic review and meta-analysis of 92 studies across 23 countries [psychological violence (28.4%), physical abuse (13.8%), and sexual abuse was 8%] [57]. Another systematic review and meta-analysis of 73 studies carried out across the worldwide showed that the pooled prevalence of sexual violence during pregnancy was 17% (CI 95%:15%-18%) [59]. In this study, the observed highest prevalence of psychological violence and lowest prevalence of sexual violence may be associated with traditional culture and women’s unwilling to talk about sex-related topics and acceptance IPV. This is supported with other similar studies [15, 6062].

Regarding the associated factors, the association of lack of formal education; history of childhood violence; rural residency; lack of decision-making power; family history of violence; attitudes toward unplanned pregnancy, unwanted pregnancy, and late initiation of ANC; partner alcohol use, partner khat chewing; and unwanted pregnancy with IPV supported previous meta-analysis [14, 5962]. In Ethiopia, a previous systematic review and meta-analysis showed that women’s educational status, partner’s educational status, and partner’s alcohol use were associated with IPV [22]. The findings of meta-analysis in Iran showed that low level of maternal education, partner use of tobacco, and regular visits of ANC were factors associated with IPV. Similarly, a systematic review and meta-analysis in China showed that number of children and unplanned pregnancies were factors significantly associated with IPV. A meta-analysis of 55 studies across the world showed that lower educational level, low socioeconomic status, single, partner alcohol use, unwanted pregnancy, and lifetime adversity were factors associated with IPV [57, 62].

Therefore, the finding of this study implies the following: high prevalence of IPV against pregnant women, which suggests the needs of strengthening the awareness of IPV and women’s empowerment and evaluates the effectiveness of the national IPV prevention strategies. The identified risk for IPV including victim, pregnancy, and perpetrator-related factors indicated the need of a holistic approach in the promotion, prevention, and treatment of IPV. Evidence showed that victim-related interventions such as screening, in conjunction with supportive counseling, and strengthening women’s empowerment; pregnancy-related interventions such as: strengthening community awareness toward family planning and perpetrator-related interventions such as: designing an effective controlling mechanisms for partner substance use were important in the control of IPV [5, 20, 21].

5.1. Strengths and Limitations of the Study

To manage heterogeneity bias and make the findings more meaningful, the use of a random effects model was to control the effects of heterogeneity, sensitivity analysis was to identify influential studies and subgroup analysis, and the funnel plot asymmetry and Egger test were used to help the findings become more meaningful. However, limitations like the use of different measurement tools (self-developed) may affect the different forms of domestic violence. The use of reference lists and Google Scholar to include the available studies may have overlooked some studies.

6. Conclusion

More than one-third of pregnant women experienced domestic violence. The most prevalent form of domestic violence was psychological violence followed by physical and sexual violence. The identified risk for IPV includes victim, pregnancy, and perpetrator-related factors which indicated the need of a holistic approach in the promotion, prevention, and treatment of IPV. The finding of this study suggests the need of strengthening women empowerments (capacity building) against traditional beliefs, attitudes, and practices. This study also suggests the need of evaluation and strengthening the collaborative work among different sectors such as policymakers, service providers, administrative personnel, and community leaders, including the engagement of men partners.

Abbreviations

CBCS:Community-based cross-sectional
DV:Domestic violence
EDHS:Ethiopian Demographic and Health Survey
IBCS:Institution-based cross-sectional
SNNP:Southern Nation and Nationalities of People
PRISMA:Preferred Reporting Items for Systematic Review and Meta-Analysis
SPSS:Statistical Package for Social Sciences.

Data Availability

The data are included in manuscript.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Authors’ Contributions

BBB designed the systematic review and meta-analysis in collaboration with YDG. BBB developed the search strategy and drafted the protocol. YDG improved the drafted systematic review and meta-analysis. BBB and YDG provided their expertise to the section of depression and methodology. BBB and YDG performed the search strategy and conducted the data selection and extraction. Both authors were involved in the data analysis and interpretation of the results and write up the manuscript. Both authors have read and approved the final manuscript.

Acknowledgments

The authors’ gratitude goes to the University of Gondar for the office arrangement and motivational support to conduct this systematic review and meta-analysis.

Supplementary Materials

PRISMA 2020 Checklist. (Supplementary Materials)