The association between intimate partner violence and unintended pregnancy among married young women in Ethiopia

Background Adolescent and young women aged 15–24 years experience higher rates of intimate partner violence (IPV). Even though IPV has been a common challenge in low-and middle-income countries, there is a paucity of research that explored the effect of intimate partner violence on unintended pregnancy in young women, despite their vulnerability. Therefore, investigating the association between IPV and unintended pregnancy in younger population is critical for developing a multifaceted intervention to reduce the rate of unintended pregnancy and its advese consequences. Methods Data were obtained from the 2016 Ethiopian Demographic and Health Survey (EDHS). A total of 706 ever married or in sexual union aged 15–24 who gave birth at least once in the last ve years or became pregnant at the time of the survey were included in the nal analysis. Bivariate and multiple logistic regression were carried out. SPSS version 20.0 was used for data analysis.

experience at least one form of IPV (physical, sexual, emotional or marital control). Marital control (59.5%) and sexual violence (8.8%) were the most and least prevalent forms of reported IPV respectively. After adjusting for possible confounders, women who ever experienced physical violence (AOR: 1.54; 95% CI = 1.02-2.34), marital control (AOR: 1.68; 95% CI = 1.13-2.48) and emotional violence (AOR: 1.90; 95% CI = 1.25-2.89) had increased odds of unintended pregnancy, compared to those who had no history of the respective form of violence.

Conclusion
Physical violence, marital control and emotional violence were an important predictors of unintended pregnancy among adolescent and young girls. Therefore, reducing unintended pregnancy among adolescent and young girls requires interventions beyond increasing access to contraceptive information and services; including empowering women, promoting men's involvement in fertility control and optimizing violent relationship is critical. Including intimate partner violence screening and treatment in reproductive health services is also necessary.

Background
Violence against women is a signi cant public health problem as well as a fundamental violation of women's human rights (1). United Nations de nes violence against women as "any act of gender based violence that results in, or is likely to result in, physical, sexual, or psychological harm or suffering to women, including threats of such acts, coercion or arbitrary deprivation of liberty, whether occurring in public or private life" (2) p89. It occurs and is classi ed in various ways. It can be de ned depending on the relationship between the perpetrator and victim (intimate partner and non-intimate partner), or by type of the act of violence, such as sexual, physical or emotional violence (3). Intimate partner violence (IPV) is the most common forms of violence against women and major public health and human rights issues that cause the most devastating problem on both the mother and the fetus (4). World Health Organization (WHO) reported a global lifetime prevalence of IPV among ever-partnered women as 30%. The life time prevalence was highest in the Africa (37%), Eastern Mediterranean (37.0%) and South-East Asia regions (38%) which compared with 23-25% in high-income and western Paci c regions (5). IPV affects all spheres of women's lives such as self-esteem, productivity, autonomy, capacity to care for themselves and their children, ability to participate in social activities, and even death (6, 7). Evidence indicated that exposure to IPV is associated with a risk of unintended pregnancy (8) and acquiring HIV or sexually transmitted infections (5). It is also associated with alcohol use, depression and suicide, injuries, and death from homicide (5). Although the mechanism for how IPV is asscoiated with unintended pregnancy is not clear; indirect mechanism through which the climate of fear and control surrounding violent relationships could limit women's ability to control their fertility which could lead to unintended pregnancies (9). Evidence also indicates that women are barely able to prevent pregnancies when sexually assaulted (10,11).
Unintended pregnancies are generally referred to as unwanted or mistimed pregnancies (12). It is a major global concern due to its association with adverse physical, mental, social and economic outcomes. Despite it affects all segments of the community and contributes greatly to maternal and infant mortality (13), young women are highly vulnerable for IPV and potential consequencies including unintended pregnancy (14,15). There were 21 million pregnancies among adolescent girls aged 15-19 years in LMICs (Lower and Middle Income Countries) in 2016; nearly half (49%) were unintended (16). While women aged [15][16][17][18][19][20][21][22][23][24] year is associated with a highest risk of unintended pregnancy the WHO report indicated that life time prevalence of IPV is higher among young women aged 15-19 years which can cause long-lasting harm (14,15).
Adolescent and young women aged 15-24 years experience higher rates of IPV, where high risk of unintended pregnancy is reported (14,15). Although a substantial body of literature explores the adverse health consequences and associated factors of IPV and unintended pregnancy respectively, only a limited body of research has explored the effect of IPV on unintended pregnancy in this vulnerable population (9). While Ethiopia has put in place appropriate and effective legal policies to promote the rights of women (17), IPV continues to be a major challenge where the national rate of IPV is 34% of which the rate increases to 65.7% in young and adolescent women aged 15-24 years. Although Ethiopia is a country with high total fertility rate (4.6 children per woman), low contraceptive use (36%), and high rates of unintended pregnancy (25%) (18), there is limited evidence investigating the association between IPV and unintended pregnancies in younger population (19,20). Hence, this study investigated the association between IPV and unintended pregnancy in Ethiopia using nationally representative data. The ndings will provide evidence for developing a multifaceted intervention to reduce unintended pregnancy and its consequences.

Data sources
This study used data from the 2016 Ethiopian Demographic and Health Survey (EDHS), which is nationally representative household surveys assessing various measures of population health and nutrition. The 2016 EDHS is the fourth survey conducted in Ethiopia which was designed to provide a representative data on overview of population, maternal, and child health issues for the country as a whole and for nine regional states and two city administrations of Ethiopia. Datails of the survey published elsewhere (18).

Sampling and sample size
The EDHS 2016 used two stages sampling for participant recruitment. In the rst stage, each region was strati ed into urban and rural areas and clusters were selected from both rural and urban areas based on the 2007 Ethiopian population and housing census using a probability proportional to size selection. A list of all the households was prepared in all the selected clusters. The second stage of selection used the list of households as a sampling frame and systematically selected a xed number of households per cluster. Then, all women age 15-49 who were either permanent residents of the selected households or visitors who stayed in the household the night before the survey were included. Only one eligible woman per household was randomly selected for interviewing, and the module was not implemented if privacy could not be obtained (18).
The 2016 EDHS implemented a module of questions on domestic violence for the rst time. This module targeted women aged 15-49 years on their experience of different forms of violence committed by their current or former husbands/partner for ever married women and by others for women who had never married (18). For this study, those ever-married adolescent and young women aged 15-24 years who had been requested on their experience of different forms of IPV was considered. From a total of 5,860 women who has been invited to respond for questions on the domestic violence module, a total of 852 ever married women aged 15-24 who gave birth at least once in the last ve years or currently pregnant at time of survey were selected and nally 706 (weighted) sample were included in the analysis (Fig. 1).

Outcome variable
Pregnancy intention for current pregnancy or recent birth was the outcome variable. Those women responding their current pregnancy or recent birth was 'wanted later' or 'not wanted at all' were regarded as an unintended pregnancy and those who responded 'wanted then' were considered as an intended pregnancy.

Independent variable
Ever experience of any form of violence and the speci ed acts of physical, sexual, emotional violence and marital control was the primary exposures of interest. Violence committed by the current/recent husband/partner for currently married or partnered women around the time of current pregnancy was measured by asking all ever-married women if their husband/partner ever did the following: Physical intimate partner violence push you, shake you, or throw something at you; slap you; twist your arm or pull your hair; punch you with his st or with something that could hurt you; kick you, drag you, or beat you up; try to choke you or burn you on purpose; or threaten or attack you with a knife, gun, or any other weapon.
Sexual intimate partner violence physically force you to have sexual intercourse with him even when you did not want to; physically force you to perform any other sexual acts you did not want to; or force you with threats or in any other way to perform sexual acts you did not want to.
Emotional intimate partner violence say or do something to humiliate you in front of others; threaten to hurt or harm you or someone close to you; or insult you or make you feel bad about yourself.

Marital control
is jealous or angry if she talks to other men; frequently accuses her of being unfaithful; does not permit her to meet her female friends; tries to limit her contact with her family; and insists on knowing where she is at all times.
Ever experience of any form of violence was derived from the response to above listed questions. In each component of violence, if a women answered 'yes' to at least one of the above questions, then the respective component of violence was considered (any violence, physical, sexual, emotional violence and marital control). If a woman answered 'No' to all questions in each component of violence, then it was considered as no experience of IPV.
Other covariates we used for adjustment included socio-demographic characteristics of women (age, residence, education, wealth, and employement status), fertility preference and control measures (ever use of contraceptive, ideal family size, descision maker for using contraceptives and maternal health care), and husbands/partners sociodemographic and behavioural characterstics (education, employement status and alchol consumption) were used in the analyses which were drawn from the literature.

Data analysis
Data cleaning and management were carried out using SPSS version 20. Variables were re-coded to meet the desired classi cation. Descriptive statistics were used to describe the characteristics of respondents with regard to IPV and unintended pregnancy.
Binary logistic regression analysis was used to assess the association of different form of IPV and other covariates with unintended pregnancy. Multivariate logistic regression analysis was used to derive the adjusted effects of IPV on unintended pregnancy. Initially, different form of IPV (physical, sexual, emotional violence, marital control and any of IPV) were included to estimate the association between different form of IPV and unintended pregnancy. In the subsequent models, each components of IPV in separate model was adjusted for covariates that showed signi cance (p < 0.2) in binary analysis including sociodemographic, fertility control and preference, and autonomy related variables. P-values less or equal to 0.05 were employed to declare statistical signi cance. Odds ratios and 95% con dence intervals were presented in the results. Multicollinearity was checked using variance in ation factors (VIF).
To account non response rate and sampling design, we used sample weight in all the analysis.

Socio-demographic characteristics
The mean age of participants were 21.2 year (SD = 0.37) and most (77.6%) were between the age of 20 and 24 year.
Most (88.4%) of the study participants were rural residents. Almost half (48.8%) were classi ed in the lower wealth status (Table 1).  The association between intimate partner violence and unintended pregnancy In the bivariate analysis age, parity, residence, education, wealth, employment status, ever use of contraceptive, ideal family size, and decision maker for using maternal health care were signi cantly associated with unintended pregnancy. The nal model of analysis showed the adjusted effects of IPV on unintended pregnancy. As a result physical violence, marital control and emotional violence was strongly associated with unintended pregnancy, while any form of violence and sexual violence have no signi cant relation with unintended pregnancy. After adjusting for possible confounders (age, parity, residence, education, wealth, employment status, ever use of contraceptive, ideal family size, and decision maker for using maternal health care), women who ever experienced physical violence  (Table 3).

Discussion
This study examined the effect of IPV on unintended pregnancy among young women aged 15-24 years in Ethiopia using 2016 Demographic and Health Survey data. We found that 21.1% of participants reported ethier their current pregnancies or recent births were an unintended; which is nearly close to the national rate of unintended pregnancy among women of reproductive age (25%) (18) indicating that uintended pregnancy is common among younger population. In the same age group (15)(16)(17)(18)(19)(20)(21)(22)(23)(24), relatively consistent ndings were reported in Nepal (22.7%) (21), Bangladesh (23%) and India (18%) (22). Higher rate of unintended pregnancies among women of reproductive age was reported in sub-Saharan Africa countries (25.9%) (23), and among adolescent and young women in South Africa (41.9%) (24). Underuse of contraceptives could be attributed to these high rate of unintended pregnancy (25), and the involvement of only unmarried adolescent girls and young women in a study conducted in South Africa may not result a generalizable nding (24).
Most literature considers IPV as a composite measure of physical, sexual and emotional violence only, and exclude marital control. However, considering the in uence of partner controlling behaviour on women's decision-making power, health service utilization (26), and fertility control (27, 28); we have adopted the WHO de nition of IPV (29) p89 and included marital control as one form of IPV in our study. We found that 64.2% of women aged 15-24 experince at least one form of IPV (physical, sexual, emotional or marital control). It was similar with the national overall prevalence of any form of IPV (64%) among women of reproductive age (30), suggesting that adolescent and young women aged 15-24 year takes the greater proportion of overall IPV. On the other hand, the overall prevalence of IPV in this study was much greater than sub-Saharan Africa women (44%) (31), which could be related to study participants age difference and the later study investigated any form of IPV as a composite measure of physical, sexual, and emotional violence only. When we compare with a similar age group (15-24 year), the overall prevalence of IPV in this study was higher than IPV (physical or sexual) reported in India (38%), Bangladesh (52%) and Nepal (28%) (22), keeping a difference in estimating the overall prevalence of IPV.
In the present study the most prevalent form of IPV was marital control (59.5%) and the least prevalent was sexual IPV (8.8%). It was consistent with a study conducted using similar data among women of reproductive age that reported marital control (56.4%) as the most prevalent and sexual IVP (11.9%) as the least prevalent form IPV (30), which implies the occurence of similar form of IPV across different age groups. On the other side, a systematic review and meta-analysis of cross-sectional studies in sub-Saharan Africa indicated emotional IPV as the most prevalent and sexual IPV as least prevaent form of IPV (31). The lower incidence of sexual IPV might be related to underporting by victimized women due to fear of discrimination and feeling shame.
The current study investigated unintended pregnancy in relation to different forms of IPV in Ethiopia. We found that physical violence, marital control and emotional violence by intimate partner have a signi cant effect on unintended pregnancy. After adjusting for potential confounding factors, unintended pregnancy was signi cantly associated with reporting physical IPV. Other studies have shown a similar association (22,32). This could be because women who were physically mistreated by their husbands were less likely to use contraceptives (33), their husbands refused to use condoms or tried to stop them from using a contraceptive (34), or discontinued the use of spacing methods because of fear.
In this study those who ever experienced marital control behaviour had increased odds of unintended pregnancy, after adjusting for possible confounders. Similar nding was reported by a study done among Nicaraguan women (35) which could be due to a reason that controlling behavior by a partner can interfere with women's reproductive freedom through hindering contraceptive use by limiting their autonomy, access to health care, or by facilitating contraceptive failure (36). In Ethiopia, where patriarchal views are common, the decision to seek care is usually made by male partners and they are the chief providers which often determining women's access to economic resources (37).
After adjusting for possible confounders, women who ever experienced emotional violence had increased odds of unintended pregnancy. It was consistent with a study conducted using similar data among women of reproductive age in Ethiopia (30) and population-based study in southern Spain (38). Experincing emotional intimate violence could be related with a limited autonomy in controlling their fertility and are most likely to be discouraged for using contraceptive methods by their husband/partners. Therefore, it sugesst the need to involve male partners and efforts should be made to improve awareness of the male partner's on fertility control through community based health education.
Furthermore, in this study any form of IPV and sexual IPV had no signi cant effect on unintended pregnancy. The lack of signi cant association between any form of IPV with unintended pregnancy was supported by a similar study conducted among women of reproductive age in Ethiopia (30) but contradicts with other study ndings (39,40). Contrary to the general argument, the association between sexual IPV and unintended pregnancy was not signi cant.
Although a study conducted in Nepal (22) indicated no signi cant effect of sexual IPV on unintended regnancy, the reasons for this is uncertain. We assume that this could have been due to small sample size that may affect the results. Moreover, further investigation is needed to explore the nature of these associations.
Previous studies have focused mostly on women of reproductive age group , whereas this study presents the relationship between IPV and unintended pregnancy among adolescents and young women. Our ndings further add to the growing body of knowledge by showing a higher prevalence of unintended pregnancy and IPV among adolescent and young women, despite improved contracetive use among women of reproductive age group in Ethiopia. This study also reinforces the ndings on the effect of different form of IPV on unintended pregnancy indicating physical violence, marital control and emotional violence has been linked to unintended pregnancies. The mechanisms for the occurrence of these relation merit further investigation.
The ndings of this study need to be interpreted in the light of some limitations. It was di cult to determine direct relationship between IPV and unintended pregnancy due to cross-sectional nature of the data used for analysis. The sensitive nature of IPV and recall bias may lead to under-reporting while sharing experiences of IPV. Additionally, unintended births may have been under-reported after a child is born because of the joy associated with having a child.

Conclusion
This study found a higher prevalence of unintended pregnancy and different form of IPV among adolescent and young women in Ethiopia; with marital control being the most prevalent. Physical violence, marital control and emotional violence by intimate partner were signi cantly associated with unintended pregnancy, suggesting that young women who report an unintended pregnancy may experince physical violence, marital control or emotional violence. These ndings indicate the need to include IPV screening and treatment in different reproductive health services.
Therefore, reducing unintended pregnancy among adolescent and young girls requires interventions beyond increasing access to contraceptive information and services; empowering women, promoting men's involvement in fertility control measures, and optimizing initimate partners relationship is critical. Physical violence, marital control and emotional violence by intimate partner appeared to be more useful for predicting unintended pregnancy among young women.  Flow chart showing a sampling procedure for selection of study sample.