Intimate partner violence (IPV) is one of the most pervasive and under-reported public health problems around the world. According to the World Health Organization, IPV refers to “behavior by an intimate partner or ex-partner that causes physical”, sexual or psychological harm, including physical aggression, sexual coercion, psychological abuse, and controlling behaviors (11). The IPVs could also be different forms within each type; For example, within physical IPV, there could be less severe or more severe form (12). The controlling behavior can be in the form of economic (13) or controlling free movement (i.e., limiting women to see, visit, and interact with others) (14). IPV can also include stalking, i.e., repeated unwanted attention and contact by an intimate partner igniting fear and safety concerns (15).

Globally, one-in-three women experienced some form of violence within their lifetimes (11). Perpetrators of IPV come from diverse socioeconomic, cultural, and religious backgrounds, leaving all women at potential risk (12). In many countries, legal systems and cultural norms often normalize IPV, leading to inadequate protection and support for victims. This is particularly evident in the Middle East and North African (MENA) countries, where studies indicate that between 6% and 59% of women have experienced physical IPV, 3% to 40% have experienced sexual IPV, and 5% to 91% have suffered from emotional IPV (13). Specifically, a previous umbrella review study found that the prevalence of any type of IPV during pregnancy was as high as 44.1% in the MENA countries (16).

Intimate partner violence (IPV) has far-reaching consequences for women's overall health and social well-being (17). The impacts of IPV include serious physical injuries and increased risk of chronic health conditions such as anxiety, depression, post-traumatic stress disorder (PTSD), gastrointestinal problems, cardiovascular diseases, sexually transmitted infections including Human Immunodeficiency Virus (HIV), and unintended pregnancies (18–20). Women experiencing IPV often engage in high-risk behaviors such as smoking and substance abuse, while simultaneously facing social isolation, fear of retaliation, coercion by abusers, financial dependence, and significant barriers to accessing essential healthcare services (21). Additionally, many women who experienced IPV victimization received poor social support (22), despite it was known that social support can lessen the severity of IPV at least (23).

Societal factors contributing to the high prevalence of IPV include social norms that condone violence against women, gender inequality in both political and economic spheres, and legal impunity for perpetrators. Individual risk factors include women's acceptance of violence as a husband's right, young age, and infrequent communication with their original family (24–28). Also, prior literature identified IPV and APO risk factors such as education (29), poverty (30), place of residence, region, use of skill birth attendants, number of ANC visit, having children under age 5 years, parity, smoking (31–33), and presence of sexually transmitted infections (STI) (34, 35). The physical, psychological, emotional, sexual, economic, and social manifestations of IPV significantly affect maternal health and frequently result in APOs (36–39).

In Jordan, IPV remains a substantial issue despite notable progress in gender parity, women's empowerment, and women's health (40). A nationally representative survey in Jordan revealed that the lifetime prevalence of IPV was 25.9%, with emotional violence being the most common (20.6%), followed by physical violence (17.5%) and sexual violence (5.1%) (40). It was also reported that fatal IPV cases contributed to 33% of all the femicide cases in Jordan (41). A prior study from Jordan reported that women living in urban areas of Amman have a higher incidence of IPV, with potential reasons of stress associated with urban lives among the perpetrators (42). A recent World Health Organization (WHO) report also indicated that 24% of ever-partnered women in Jordan have experienced some forms of IPV during their lifetime (11). Additionally, there were studies that identified risk factors for IPV among Jordian women including intergenerational violence (43) or having number of children, but examination on APOs outcome was lacking (44). Also, another study in Jordan examined the disparities negative birth outcomes, however, IPV exposure was lacking (45). However, very rare study focuses on APOs based on IPV exposure.

Although most research concentrates on the adverse effects of IPV during pregnancy, there is substantial limitation of research on a lifetime IPV victimization and APOs in Jordan context. For example, Abujiban and colleagues identified an association between physical IPV and higher risks of APOs (i.e., pregnancy-induced hypertension, cesarean section, more pain killer use during birth, and excessive use of postnatal medication) using a hospital-based sample (46). However, their study did not find any significant association between physical IPV and other factors regarding pregnancy plan/care (unplanned pregnancy, inadequate ANC visits (i.e., less than eight), or birth outcomes (abortions, premature rupture of membranes, PTB, induction of labor, episiotomy, laceration, or postpartum hemorrhage) (46). Also, Afkhamzadeh and colleagues examined APOs based on IPV during pregnancy, identification on any lifetime IPV (i.e., not necessarily only the time of pregnancy) was lacking (47).

All these inconclusive and inconsistent findings on different APOs suggest a need to investigate the relationship between lifetime IPV and various types of APOs (21). Understanding the connection between a lifetime IPV exposure and APOs can inform prevention and intervention strategies aimed at reducing these risks. Despite the high prevalence of IPV in Jordan, the link between the lifetime IPV and different types of APOs (i.e., pregnancy loss, pregnancy complications including LBW, PTB) are poorly understood, which could vary depending on a country's culture and context or region. Our hypothesis was that those who experienced lifetime IPV are at increased risk of any type of APOs. As such, the purpose of this study was to examine the influence of lifetime IPV experiences and social determinants on APOs among Jordanian married women.

As per Harman's trauma theory when people experienced traumatic events such as IPV, they have subsequent negative health issues (48, 49). These health issues include PTSD (50, 51), cognitive problems (4), insomnia (51–53), worsening night vision during pregnancy (54), higher level of somatic problems, and confusion (55), which all can contribute to APOs. Trauma history has substantial effects on prenatal anxiety and stress (56), which could subsequently lead to microsomic babies (51, 57) or pregnancy loss (58). For example, prior scholars had identified that the major stress around conception had impact on infants including conotruncal heart defects or isolated cleft palate, or neural tube defects (59). Selye's theory of stress postulated that stress can be caused by diverse stimuli regardless of the events (60). And the response to the stress differed based on individuals and stages of stress (i.e., alarm reaction, resistant stage, and exhaustion stage) (60).

Women can face IPV at any time despite the frequency, types, and extent are varied (61). For example, a meta-analysis identified that the IPV prevalence was highest before pregnancy, lower during pregnancy, and higher again after childbirth (61). Additionally, Walker's cycle of violence theory posited that women can be revictimized repeatedly as women's motivation to respond reduced over time, however, these repeated victimizations subsequently contribute stress and anxiety throughout their life (62).

A study by Garabedian and colleagues identified that history of any IPV victimization (i.e., sexual, physical, or stalking) had significant association with increased risk of PTB, pre-eclampsia, or bleeding during third trimester (63). Specifically, their study found that sexual IPV victimization with or without other IPVs was associated with all these APOs, those with physical IPV victimization had bleeding and PTB, and those with stalking experience had pre-eclampsia (63). Additionally, Bonomi and colleagues identified that those with history of IPV victimization had higher risk of depression, anxiety, substance use, or other chronic diseases, reproductive system-related conditions such as menstrual disorders and sexually transmitted infections, all of which can negatively impact on their future pregnancies (64). IPV victimization can also be a barrier for perinatal health services used (17, 42). For example, a study in Ethiopia found that almost 23.6% of pregnant women were restricted by their partners for their ANC use during their pregnancy (65). Additionally, a recent study in Egypt identified that almost 58.8% of women reported psychological IPV victimization were found to be associated with delayed ANC, experienced hemorrhage, and premature rupture of membrane (66). Another scoping review identified an association between delayed ANC and any type of IPV (despite the study did not specify timing of IPV victimization event) (67).

Along with inadequate ANC, IPV victimization during pregnancy is associated with miscarriage or complications such as insufficient weight gain, premature rupture of membrane, placental abruption, placenta previa, PTB, hemorrhage, uterine rupture, and maternal death. IPV victimization during the perinatal period may increase the risk of PTB, LBW infants, and stillbirths (18). A systematic review and meta-analysis of seventeen studies found that IPV victimization during pregnancy were three times more likely to suffer perinatal death in comparison to those who did not (17). Considering the seriousness of the adverse impacts of IPV victimization, researchers have further explored the association between various types of IPV victimization and APOs. Physical and sexual IPV victimization was significantly associated with PTB, LBW, and increased hospitalization, while emotional IPV victimization was associated with PTB (18–20).

More broadly, IPV victimization has been linked to pregnancy health, various maternal complications, and reduced utilization of adequate ANC, all of which contribute to different types of APOs (68, 69), and therefore, understanding risk factors specifically among IPV victims are important and this study is needed.

This study used deidentified respondents’ data from women in the Jordanian Population and Family Health Survey (JPFHS), which was implemented from January to June 2023 (37). The JPFHS utilized a two-stage stratified random sampling design: First, households were randomly selected based on the 2016 Jordan's census framework, then, a participant from each selected household was randomly chosen and interviewed (37). The JPFHS included ever-married women aged 15–49 years (n = 12,595) with a response rate of 97% (37). For this study, the women who did not report their most recent pregnancy outcomes (i.e., recent live births, stillbirths, miscarriages, and induced abortions), those who reported don't know, and/or those who mentioned that their babies were not weighted (n = 69), and those who reported not knowing if their babies were too small or big (n = 3) were excluded from analyses, resulting in a final sample of 4,419 women (Figure 1).

Stata standard edition (SE) 19.0 was used for all the analyses (72). For descriptive statistics, univariate and bivariate [i.e., Pearson's Chi-square (χ²) test of independence] analyses were performed with the analytic sample (n = 4,419). The Pearson's χ² test is used to investigate the relationship between outcome variables and covariates (73). Only p-values of ≤0.05 were considered statistically significant.

Prior to analysis, we assessed multicollinearity among covariates by calculating variance inflation factors (VIF) (Supplementary Tables S1 and S2). Multicollinearity was assessed separately for the full sample and the subset of those who experienced IPV. Mean VIF (for the full sample) was within the acceptable range, indicating no serious multicollinearity (i.e., VIF greater than 5–10 is considered multicollinearity present) (74, 75). In the subset of IPV victims, higher VIF was observed. Nonetheless, based on the mean VIF, we included the covariates in our subset analyses (74, 75).

A series of binary logistic regression models were performed: (a) to predict the odds ratio of overall APOs, LBW, and pregnancy loss; and (b) to examine the relationship between the dependent variables and covariates. All socio-demographic variables and exposure variables were included guided by prior literature (4, 29–35), regardless of their statistically significant levels in bivariate analyses. Subgroup analyses were performed with samples limited to those who experienced IPV victimization (APOs, LBW, and pregnancy) [Note. For PTB, Micro/macro-somia, and pregnancy loss, subset analyses were not performed due to their limited sample sizes]. We used complete case analyses for our prediction models, i.e., missingness was not included. The detailed sample flow is shown in the sample flow-chart (Figure 1).

Additionally, we calculated marginal effects to quantify the effect of change on the outcome variables (Tables 2, 3) including overall APOs (Model 1), LBW (Model 1a) and pregnancy loss (Model 1b) among all sampled women. Moreover, we performed subgroup marginal effects analysis for APOs-Model 2, LBW (Model 2a, and pregnancy loss-Model 2b) among women who experienced IPV victimization.

All weighted analyses (i.e., descriptive, bivariate, logistic regression, and marginal analyses) were performed applying survey weights to account for complex sampling design (i.e., clustering, stratification, and sampling weights), as recommended by the Demographic and Health Survey Program (37–39). If we do not apply survey weights that account for complex sampling design, the calculation of point estimates and their standard errors will be incorrect. Also, lack of consideration on clustering can lead to underestimation of standard errors (76). Therefore, we applied “svy” commands using Stata software, which accounts for sampling weights, clustering, and stratification (77). For the sensitivity analysis, we evaluated differences in prediction estimates by comparing weighted and unweighted binary logistic regression models (Supplementary Tables S1 and S2) (78).

Supplementary Table S1 presents the weighted results of the binary logistic regression model identifying factors associated with APOs, LBW, and pregnancy loss. The women from the middle quintile to upper wealth quintile groups were less likely to experience APOs, compared to those from the poorest wealth quintile group, with the lowest odds ratio was observed among the richest wealth quintile group [adjusted odds ratio (aOR) = 0.50, p = 0.007]. Those living in the Northern region showed a higher odds ratio (aOR = 1.44, p = 0.001) of experiencing APOs compared to those living in the Central region. Additionally, women who gave birth to twins or multiples were more likely to experience APOs compared to those who delivered a singleton baby (aOR = 3.68, p = 0.001). Those who used SBAs were less likely to experience APOs compared to those who did not (aOR = 0.06, p = 0.020). Our model regarding among all sampled women showed that there is no significant association between the IPV victimization and APOs.

As reported in the Supplementary Table S1, among those who give livebirths (n = 3,919), higher odds of LBW babies were seen among women residents of the Northern region (aOR = 1.38, p = 0.009) and those who give birth to twins or multiple children (aOR = 5.48, p < 0.001). However, lower odds ratios of LBW babies were seen among those who were from the richer (aOR = 0.58, p = 0.011), and richest (aOR = 0.48, p = 0.23) wealth quintile groups, and those who used an SBA (aOR = 0.03, p = 0.002) compared to their reference categories.

As reported in the Supplementary Table S1, among the covariates, being in older age groups had higher odds of pregnancy loss, and the highest odds ratio was seen among those from aged 44–45 group (aOR = 66.47, p < 0.001), those who had a smoking history had higher odds of pregnancy loss compared to those who did not smoke (aOR = 1.63, p = 0.039). Contrarily, women with two or more children who were under 5 years old had lower odds of pregnancy loss compared to those who did not have children under 5 years old (aOR = 0.05, p < 0.001).

For the sensitivity analyses among all sampled women, we compared weighted and unweighted aORs. The detailed results of unweighted aORs of APOs, LBW, and pregnancy loss among all sampled women are described in Supplementary Table S1. The results of unweighted aORs were somewhat similar to the weighted aORs for APOs except for two variables, i.e., region and use of SBA. Regarding the region, we found higher odds ratio of APOs among women Southern residents (unweighted aOR = 1.32, p = 0.003). Additionally, we found no significant relationship between APOs and use of an SBA in unweighted model. Regarding LBW, the results were somewhat similar, and the weighted aORs were lower than unweighted aORs except the variable “region” (unweighted aOR = 1.33, p = 0.001). In the unweighted model with the pregnancy loss outcome, our study finding identified that the aORs were quite lower compared to the weighted models. For example, despite the highest odds were seen among women from the 45–49 age group (unweighted aOR = 16.01, p < 0.001), the odds ratio was somewhat lower than that of the weighted model.

Marginal analyses of APOs, LBWs, and pregnancy loss of all sampled women are described in Table 2. Our study identified that compared to the women from the reference categories, lower likelihood of APOs was seen among women from the middle wealth quintile group (β = −0.08, p = 0.017), those from the richest wealth quintile (β = −0.16, p = 0.004), and those who used an SBA (β = −0.52, p < 0.001). Contrarily, higher likelihood of APOs was seen among women from the Northern region (β = 0.09, p = 0.001) and those who have twins or more birth (β = 0.31, p < 0.001) compared to women from the reference categories. Among all sampled women, compared to the women from the reference categories, lower likelihood of LBW was observed among women from the richer (β = −0.09, p = 0.009) and those from the richest wealth quintile group (β = −0.12, p = 0.008), and those who used an SBA (β = −0.69, p < 0.001). Contrarily, higher likelihood of LBW was observed among those who had twins or more (β = 0.39, p < 0.001) and residents of the Northern region (β = 0.06, p = 0.010), compared to those from the reference categories. The women from all older age groups were more likely to have pregnancy loss compared to those from 15 to 19 age groups, and the greatest probability of pregnancy loss was seen among those from 45 to 49 age groups (i.e., 28 percentage point more likely to have pregnancy loss).

As presented in Supplementary Table S2, among IPV victims, those in the rural area had higher odds of APOs (aOR = 7.72, p = 0.001) compared to those living in the urban area, however, those in the Southern regions had lower odds of APOs (aOR = 0.14, p = 0.030) compared to those living in the Central region.

As presented in Supplementary Table S2, among IPV victims, higher odds of LBW were seen among those had primary education (aOR = 50.7, p = 0.010), secondary education (aOR = 23.92, p = 0.047), rural residents (aOR = 15.04, p < 0.001), those who had three or more children under age 5 years (aOR = 13.66, p = 0.055), and those whose health decisions were made by others (aOR = 7.21, p = 0.008) compared to those in the reference categories. Contrarily, lower odds of LBW were seen among those who were from middle (aOR = 0.20, p = 0.021) and richer wealth quintile groups (aOR = 0.15, p = 0.018), and those who had ANC ≥8 visits (aOR = 0.21, p = 0.001).

As presented in Supplementary Table S2, among IPV victims, the higher odds of pregnancy loss were seen those who were from older age groups compared to those 15–19 age groups, among the women residents of Southern region (aOR = 6.09, p = 0.039), who had twins or multiple births (aOR = 66.15, p = 0.001), those who were from the poorer wealth quintile, and those who had children under age 5 years compared to the reference categories.

For the sensitivity analyses among the IPV victims, we compared weighted and unweighted effect sizes (i.e., aORs). The detailed results of unweighted aORs of APOs, LBW, and pregnancy loss among women IPV victims are described in Supplementary Table S2. In the model with APOs outcome, those living in the rural areas had higher odds (aOR = 7.72, p < 0.001) in the weighted model, compared to unweighted model (aOR = 2.57, p = 0.026). Despite the significant relationship between APOs and Southern region being identified in the weighted model (aORs = 0.14, p = 0.026), we saw significant relationship between APOs and Northern region in the unweighted model (aOR = 2.11, p = 0.024). Additionally, although we did not see any significant association between APOs and decision making in the weighted model, a significant association between decision making by partner only and APOs (aOR = 2.66, p = 0.048) was found in the unweighted model.

In the weighted model with LBW, despite seeing significant association between LBW and primary or secondary education, we did not find any significant relationship between LBW and education in the unweighted model. Also, there were differences in significant association between LBW and wealth quintiles between weighted and unweighted model. Additionally, although we found significant association between LBW and place of residence in both models, the odds ratio was higher in weighted model (aOR = 15.04, p < 0.001) compared to that of unweighted model (aOR = 3.69, p = 0.003). The results of pregnancy loss were slightly different. All age groups, wealth quintile, region, the number of children under age 5 years old, showed lower odds in the weighted model compared to the unweighted model. Despite we found significant association between twins or multiple births with pregnancy loss (aOR = 66.15, p = 0.010) in the weighted model, we did not see any significant association in the unweighted model.

The results from marginal analyses among IPV victims are presented in Table 3. Our findings identified that higher likelihood of APOs was observed among rural residents compared to the urban residents (β = 0.47, p < 0.001). Contrarily, women from the Southern region had lower likelihood of APOs (β = −0.46, p = 0.034) compared to the residents from the Central region. The higher likelihood of delivering LBW babies was found among women who had primary education (β = 0.43, p = 0.011) or who had secondary education (β = 0.35, p = 0.032) compared to those who had no formal education. Similarly, higher likelihood of delivering LBW infants was observed among women who had 3 or more children under age 5 years (β = 0.29, p = 0.018), those living in the rural (β = 0.30, p = 0.001), and those with other people who made the pregnant women health care decision (β = 0.22, p = 0.009) compared to the reference categories. Contrarily, women from middle (β = −0.18, p = 0.028) and richer (β = −0.21, p = 0.029) wealth quintiles, and those who had ANC≥8 visits (β = −0.17, p < 0.001) were less likely to have LBW babies. Additionally, our findings also showed that higher likelihood of pregnancy loss was seen among those who had twins or more compared to women with singleton delivery (β = 0.62, p = 0.013). Contrarily, lower likelihood of pregnancy loss was observed among the poorer (β = −0.09, p = 0.014) or richer (β = −0.08, p = 0.041) wealth quintile groups, and those who had three or more children under age 5 years (β = −0.38, p = 0.037) compared to the respective reference categories.

Overall, the prevalence of lifetime IPV victimization was 6.7%. IPV victimization increases during pregnancy due to the stress and instability that can be brought on by pregnancy and the prospect of raising a child (83). Worldwide prevalence of IPV during pregnancy ranges between 2% and 67%, with one study finding that when several forms of IPV combined were reported, the prevalence of any kind of IPV was 25%, and another finding that some countries have rates as high as 67% (48–50, 99–101). Experiencing IPV across the life span, including the pregnancy period, negatively impacts the mother's health across her life span (102–107) as well as the birth outcomes (11, 108). Therefore, it is critical to detect IPV early and prevent further health consequences.

Among those IPV victims, women in the rural area had higher odds of APOs compared to those living in the urban areas. This finding is partially aligned with a prior study in which the urban residents had lower odds of pregnancy loss (109). The societal acceptance of IPV in the rural areas may create undisclosed or repeated violence, that could lead to APOs. For example, a prior study in Jordan had revealed that men living in the villages had higher acceptance of attitude toward emotional and physical violence (110) which might influence their perpetration as normal, leading to higher perceived societal acceptance on violence, creating stressful situation for women (83), which could further lead to higher APOs. Or women's lack of decision making may create delayed health care or limited ANC visits (111). It is also possible that health care workers missed screening the IPV during the ANC. For example, a study in the US (despite geographic differences) showed that there was a higher chance of not screening for IPV in the rural areas during the ANC visits (111). All the prior research and our research findings suggest an urgent need to ensure early and routine ANC, and IPV screening for every health care visit including before, during, and after pregnancy for every woman (111), specifically to the women in the rural area. Interestingly, our findings revealed that those in the Southern region had lower odds of APOs compared to those in the Central region of Jordan. Although not exactly the same, the prior study in Jordan identified that the Central region had a higher risk for LBW, one of the pregnancy complications (112). Perhaps the women in the Central region may have less social support, especially among those who had IPV, putting them stressful situation which could lead them to experience APOs. All these mixed findings suggest a need for further investigation.

Among the IPV victims, higher odds of LBW were seen among those who had primary or secondary education compared to those who did not have any education. Despite a prior study in the US that found that mothers' illiteracy as the risk factor for LBW (113), our study found that those who have primary or secondary education had higher risk of LBW compared to those who did not have formal education. Perhaps these contradictory findings may be due to geographic differences or the way variables were defined, requiring further investigation. Additionally, a prior study in Jordan found that education is negatively associated with LBW babies. Perhaps, education facilitates better health service use including use of ANC or an SBA (98, 114). All these inconsistent findings suggest further investigation. Also, among the IPV victims living in rural areas had higher odds of LBW support the fact that the societal perception of acceptance of violence (83) or lack of social support (115) may create undisclosed stressful conditions for the women, leading them experiencing pregnancy complications.

Those who had three or more children under age 5 years compared to those who did not have children under 5 had a higher odd of LBW. This may also have been related to inadequate use of ANC due to potential restrictions of health services used by partners (65) or lack of social support (115). Not surprisingly, the individuals whose health care service use was decided by others compared to those who made their own decisions had higher odds of APOs. For example, as stated elsewhere, ANC use could be delayed due to the restriction from their spouses (65) or if they were traumatized due to their IPV victimization experiences (116). Adequate ANC used are important as health care providers can identify potential risks, IPV screening, and make appropriate action to minimize the risk. As expected, and widely known, our study identified that those who were from the higher wealth quintile groups (especially from the middle and richer wealth quintile groups) had lower odds of delivering LBW babies. Our findings regarding the association between the wealth quintile and LBW among IPV victims contradicts with a prior study conducted in South Africa, in which IPV victims from the rich wealth quintiles were more likely to have LBW babies, suggesting further investigation between wealth and LBW association, considering other stressful and traumatic factors. Additionally, our study confirmed that those who had ANC ≥8 visits have lower odds of LBW infant delivery.

Our findings regarding the higher odds of pregnancy loss among IPV victims, especially among women from the Southern region, those who had twin or multiple births, and those who were older those who were from the poorer wealth quintile, and those who had children under 5 somewhat aligned with the prior studies. Our findings somewhat aligned with the findings mentioned in the National Strategic plan (2015–2019) of Jordan, in which the Southern region had higher child mortality rate than the national average (117). The regional variation could be due to many different reasons including quality of health (98) or other sociodemographic factors, underscoring further studies.

In the sample consisting only IPV victims, our findings regarding showing discrepancies of aORs between weighted and unweighted logistic regression models may have been many reasons. Although we found aORs of some variables in the unweighted models had higher odds, which may be overestimated and may have been biased (96, 97). However, we saw higher odds among some variables in the weighted model, underscoring further investigation. Although there was no known exact mechanism for these discrepancies, small sample size could be one of the reasons. The other reason could be inclusion of covariates. Despite our models being based on literature when selecting the covariates, there could be other covariates that could be missed, leading to discrepancies in aORs. All these findings require further investigations.

Our findings on marginal analyses of IPV survivors aligned with the findings from logistic regression. Our study found that women IPV survivors living in the rural areas is associated with increased likelihood of APOs compared to the urban residents (90). This finding was not surprising as those in the rural areas may have different challenges including limited access to health information, services, and ability to decide for their health care use (90). Those who live in the Southern region had 46-percentage point decrease in the likelihood of APOs compared to those living in the Central region. Regarding LBW, the IPV victims with primary education or who had secondary education, or those who had 3 or more children under age 5 years, those living in the rural, and those with other people made the decision had an increased likelihood of LBW. Contrarily, those who were from middle, richer, and the rural residents, and those who had ANC ≥8 visits had a decreased likelihood of delivering LBW babies. Regarding pregnancy loss, the IPV victims from the poorer or richer wealth quintiles, and those who had three or more children under age 5 years were less likely to experience pregnancy loss compared to their reference categories. Those living in the Southern region had a decreased likelihood of pregnancy loss compared to those in the Central region. Contrarily, those who had twins or more had higher likelihood of pregnancy loss compared to their refence category.