Prevalence and Factors Associated with Gender-Based Violence among Pregnant women Attending Antenatal Clinics in Nigeria: A Systematic Review

Full text

Text extracted from the article PDF for reading and search. Tables, figures and equations are in the PDF, which remains the authoritative version.

Introduction

Gender based violence (GBV) during pregnancy involves physical, sexual, psychological, emotional, verbal and economic abuse as well as coercive and controlling behaviour affecting a pregnant woman, likely perpetrated by her husband or intimate partner. Women who are pregnant are not immune to violence and may be at a higher risk for violence due to financial dependence, changes in sexual relationships, disputed paternity, and stress in household relationships and decision-making. Non-respect and issues of bodily integrity are not the only implications, but also levels of psychological suffering, inadequate use of ANC and detrimental maternal and fetal outcomes. But there are indications in Nigeria of a high, though variable burden. Other studies in Nigeria found higher rates of domestic violence among attendees in Zaria (9.6%) and in Abuja (6.2%), as well as other Nigerian settings (Ameh & Abdul, 2004; Ezechi et al., 2004; Efetie & Salami, 2007) [4, 17, 15], with Fawole et al (2008) [19] recording a comparatively low prevalence of 2.3% in Abeokuta (Envuladu et al., 2012) [16], 44.6% in Abakaliki (Onoh et al., 2013) [34], 36.1% among tertiary-facility antenatal attendees (Ikeako et al., 2017) [22], 59.6% in selected Primary Healthcare centres in Abuja (Ayeni et al., 2021) [12], 43.0% in Enugu (Njoku et al., 2021) [28], 34.5% in Ilorin (Kofoworade et al., 2022) [26], 40.0% in Abakaliki (Egwu et al., 2023) and 27.3% in Chiejine et al. (2024) [13]. An extremely high IPV exposure of 72.0% has been reported in Oyo East by Ayodapo et al. (2017) [10], while Idowu and co-authors (2013) reported 7.4% for Northern Nigeria. There is a variety that could reflect actual differences in context. Still, it also clearly shows inconsistencies in the definition of violence, the instruments used, the data-collection timeframes, and the populations sampled. Research that includes psychological, emotional, verbal and/or controlling behaviours tends to show more impact than those that only include physical and/or sexual assault, indicating an underestimate of the amount of pregnancy-related violence in Nigeria when accounts of only physical and/or sexual violence are assessed.

The determinants of GBV in pregnancy occur at the individual, partner, household, clinical and community levels. It has been a continuous topic that low levels of education, unemployment and poverty, and with women's vulnerability to abuse (Abasiubong et al., 2010; Fawole et al., 2010; Iliyasu et al., 2013; Dada et al., 2022) [1, 20, 23, 14]. But the reasons and connections between economic independence and violence are not straightforward; Ansari et al. (2017) [7] discovered relations which became problematic if women's income and autonomy to use their income were contesting the existing gender dynamics. Low educational attainment, unemployment, alcohol use and multiple sexual partners are partner-level factors, while polygamous marriage, financial disputes, family dysfunction and past violence are relationship and familial-level factors linked to the pregnancies (Ameh et al., 2008; Envuladu et al., 2012; Isangadighi & Udeh, 2023; Ashimi & Amole, 2015; Ikeako et al., 2017; Kofoworade et al., 2022) [5, 16, 24, 9, 22, 26]. Consequences reported include injury, anaemia, depression and suicidal ideation, miscarriage/preterm delivery, and others (Onoh, 2013; Ayeni et al., 2021; Dada et al., 2022; and Kofoworade et al., 2022) [34, 12, 14, 26]. However, disclosure remains limited despite these consequences. As reported by Ayodapo et al. (2017) [10] and Tella et al. (2020) [35], women who experienced PV tended to report to family members and religious leaders, remain silent, avoid, or seek informal mediation. While the previous Nigerian systematic review included 17 studies up to June 2016, many studies have emerged from PHCs, communities, and tertiary facilities in Nigeria. The available evidence is thus larger, more geographically extensive, and methodologically sophisticated, and an update to the synthesis is needed. In view of this, this systematic review aimed at identifying the prevalence and pattern of GBV, individual, partnership, household and clinical factors associated with GBV, pattern of disclosure and health consequences, and differences by geographical zone, health care level, type of violence and methodological quality among pregnant women attending ante-natal clinics in Nigeria.

Methodology

Review Design and Eligibility Criteria This systematic review was done based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines and looked at the prevalence and factors associated with gender-based violence among pregnant women attending antenatal clinics in Nigeria. The exposure was gender-based violence, including domestic violence, intimate partner violence, spousal violence, physical violence, sexual violence, psychological violence, emotional violence, verbal violence, controlling violence, and economic violence. The outcomes were the prevalence of gender-based violence, the pattern of gender-based violence, the perpetrators of gender-based violence, and the factors associated with gender-based violence. Eligible studies for inclusion were original studies that focused on quantitative or mixed methods and that had pregnant women from Nigeria as participants, and assessed the prevalence of violence during the pregnancy of interest (index pregnancy) or during a previous pregnancy. The studies were designed in Primary, Secondary, and Tertiary settings, and in Public, Private, or Community/Public settings, and the events took place during pregnancy and/or the antenatal period. If a postpartum experience did not have a pregnancy experience, this was not included in the postpartum study. Reviews, editorials, protocols, case reports, studies with only qualitative results, and studies of mistreatment of healthcare workers during childbirth were excluded. If several reports had the same participants, then they were regarded as 'companion' publications and were considered as one study if there was no indication of different results.

Information Sources, Search Strategy and Search History A systematic search was conducted (2000 - 26 August 23). A series of databases, which include: PubMed/MEDLINE, Scopus, Web of Science Core Collection, CINAHL, African Journals Online, Google Scholar, and Nigerian institutional repositories, were queried. Searches were conducted to examine the reference lists of eligible studies and prior systematic reviews, and to perform a forward search to identify newer publications. The key words used were: (“gender-based violence” OR “intimate partner violence” OR “domestic violence” OR “spousal violence” OR “partner abuse” OR “violence against women” OR “physical abuse” OR “sexual violence” OR “psychological abuse” OR “emotional abuse”) AND (“pregnant women” OR pregnancy OR antenatal OR prenatal OR “antenatal clinic” OR “maternal health services”) AND (prevalence OR magnitude OR pattern OR determinants OR predictors OR correlates OR “associated factors” OR “risk factors”) AND (Nigeria OR Nigerian). Controlled vocabulary (Medical Subject Headings (MeSH)), truncation, proximity operators, and database-specific field codes were utilised. An iterative search was done. The initial search was only for pregnancy-related searches conducted at Primary Health-care Centres, which yielded four potentially eligible reports – three of those were conducted at Primary Health-care Centres, and one used both Primary and Secondary health-care Centres. However, this excluded a significant amount of evidence from Nigeria, and the decision was made to expand the setting to include antenatal clinics (ANCs) across all levels of care. The search terms were subsequently varied to include a search for ‘gender-based violence', in addition to ‘domestic violence', ‘intimate partner violence', ‘spousal violence', ‘partner abuse' and ‘specific form of violence'. To identify studies that came before this systematic review, a backward citation search was used, while a forward citation search was used to identify studies that came after the previously conducted systematic review in Nigeria, up to June 2016 (17 articles). These were then checked for author, title, location, and DOI to certify the records and to determine whether there is a companion or duplicate record. 35 potentially eligible reports were found with the complete search. Based on preliminary screening of population, setting, outcomes, and sample overlap, 33 reports were initially retained for detailed full-text reviews.

Study Selection and Data Extraction All the identified references were placed in the review database and deleted if they were duplicates based on Title, Author, Year of publication, Setting, and Digital Object Identifier(s). Titles and abstracts were screened independently, and potentially relevant full texts were assessed by two reviewers. Reports that did not report pregnancy-specific outcomes, did not specify the type of report (qualitative only or quantitative/qualitative), the setting was other than Nigeria, the population description was incorrect, or there was no numerical data were excluded. Where disagreements arose, they were raised, with the assistance of a third reviewer where necessary. Categorisation of exclusion was performed, and the full-text review and selection process (study selection) were presented using the PRISMA 2020 flow diagram. The data extraction form included information on authorship, publication year, state, geopolitical zone, healthcare level, study design, sampling procedures, number of participants, participant characteristics, definition of violence type, measuring instrument, reference period, perpetrator, overview of prevalence, overview of the various violences, associated factors, effect estimates, confidence intervals, and limitations. Violence was identified as being physical, sexual, psychological/emotional, verbal or economic/controlling behaviour. The data was collected from one of the scores sheets and then cross-checked against the entire report using a second score sheet, which a different person reviewed.

Quality Assessment and Data Synthesis Each methodological quality assessment was conducted independently by two appraisers using the Joanna Briggs Institute Critical Appraisal Checklist for Studies Reporting Prevalence Data. The appropriateness of the sampling frame, sampling procedure, sample size, description of participants and settings, sample coverage, violence-measurement instrument, measurement consistency, statistical analysis, and response rate was evaluated in the appraisal. The analytical studies were also evaluated to control for and identify confounding factors. Studies were categorised as low-, moderate-, or high-risk of bias. Quality was not the only criterion for exclusion, and sensitivity analysis and interpretation of quality rating were performed. Each study was initially summarised narratively and presented in evidence tables that summarised its characteristics and findings. Distinctiveness of prevalence estimates was coded by geographic region (geopolitical zone), level of the healthcare (scope of healthcare), timeframe of the publication (date) and by measure of the violence (measurement period) and form of violence measured (type of violence).

Ethical Considerations

Ethical approval was not required because the review used published data and involved no direct contact with human participants.

Results and Discussion

Fig 1: PRISMA 2020 Flow Diagram Showing the Identification, Screening, Eligibility Assessment, and Inclusion of Reports in the Systematic Review The process of identifying and screening papers, as well as paper-number evaluation for eligibility and recursion in the present systematic review, is shown visually in Figure 1. A total of 1903 records were identified during the search across databases and other sources, indicating an increase over the preliminary search and signifying a substantial number of records. 33 reports were ultimately included after completing the sequential screening and eligibility assessment procedure. There were 28 of these reports, 7 that provided sufficiently detailed prevalence estimates in their statistics to be synthesised into the prevalence synthesis, and 5 that had a primary focus on outcomes, disclosure, or consequences, which could be synthesised into the thematic synthesis. This contrast can provide a template in which to integrate the critical contextual and experiential elements of violence towards women while using quantitative data. To ensure broader coverage of the review, an inclusive search strategy was used, encompassing primary and antenatal care, secondary and tertiary healthcare facilities, community settings, and mixed facilities. Including only studies conducted within the PHC limits the risk of missing applicable evidence originating at other levels of care. Other studies that were conducted in primary healthcare and population-based research studies include Envuladu et al. (2012) [16]; Iyanam et al. (2018) [25]; and Ayeni et al. (2021) [12]; whilst others were conducted at the tertiary level, such as Iliyasu et al. (2013) [23]; Ikeako et al. (2017) [22]; and Kofoworade et al. (2022) [26]; Nwafor et al. (2024) [30]; and Chiejine et al. (2024) [13]. In addition, the literature indicates continued interest in the research, with efforts by Ameh and Abdul (2004) [4], Ezechi et al. (2004) [17], Egwu et al. (2023), Nwafor et al. (2024) [30], and Olabisi et al. (2026) [32]. It should be noted that the 33 reports included in the study are not necessarily representative of 33 distinct study populations. Multiple publications can stem from the same cohort, institution, or geographical context, leading to double-counting. This is important to consider for publications from Zaria (Ameh et al., 2008) [5], where a careful comparison of recruitment periods, sample sizes, study settings, and reported outcomes is required. In addition, the results showed that the current evidence base is broader and more contemporary than the previous systematic review, which, after methodological appraisal, included only 17 of 19 potentially eligible articles. A large number of records (1,903) were found. In general, Figure 1 illustrates that the broader search scope resulted in full coverage of the evidence, and the PRISMA-based screening process was an appropriate way to narrow the broad literature sample to 33 included studies, which form the basis of the current review.

Characteristics of the Included Studies As shown in Table 1, the included studies exhibited wide variation in geography, healthcare and methodological aspects, and sample size. Evidence was generated from each of Nigeria's six geopolitical zones, though not evenly distributed across them. The areas included Zaria, Kano, Sokoto, Jos, Abuja, Ilorin, Ekiti, Ibadan, Abeokuta, Benin City, Oleh, Uyo, Itu, Abia, Enugu, and Abakaliki. This coverage is relevant because factors such as socioeconomic inequality, marriage, gender norms, and family organisation relate to gender-based violence and vary across regions. For example, Ansari et al. (2017) [7] reported prevalence rates of 22% in Cross River State and 9% in Bauchi State, indicating subnational differences. On average, women recruited as samples in various studies conducted in PHCs or rural communities, such as Ashimi and Amole (2015) [9], Envuladu et al. (2012) [16], and Iyanam et al. (2018) [25], may not have the same socioeconomic or healthcare experience as women sampled in tertiary healthcare, as in Iliyasu et al. (2013) [23], Ikeako et al. (2017) [22], Kofoworade et al. (2022) [26], and Nwafor et al. (2024) [30]. Sample sizes also varied, from a minimum of 137 (Nwafor et al., 2024) [30] to a maximum of 500 prenatal attendees (Umeora et al., 2008). Estimates should not be given equal weight for two reasons: first, the statistical precision of the estimates varies, and second, the estimates reflect the sample sizes of the corresponding studies. Finally, because most of the research was cross-sectional, it is difficult to interpret the results and implications for related factors and violence, as these are measured at the same time. Additionally, there was variation in the time period measured (current pregnancy, last 12 months, lifetime experiences of PRVs). The screening devices, definitions of disclosure, and interview contexts may also impact disclosure. Table 2, therefore, confirms that evidence is available for Nigeria and indicates methodological differences.

Table 1: Characteristics of the included reports Study Location Setting/design Sample Principal outcome or reported prevalence Ameh and Abdul (2004)

Zaria, Kaduna State Antenatal facility; cross-sectional NR Prevalence and forms of domestic violence during pregnancy Ezechi et al. (2004) South-East Nigeria Antenatal clinic; cross-sectional 418 Current-pregnancy violence: 28.7% Efetie and Salami (2007) Abuja FCT National Hospital ANC; cross-sectional NR Domestic violence: approximately 37.4%; psychological abuse predominated Fawole, Hunyinbo and Fawole (2008)

Abeokuta, Ogun State Antenatal clinic; cross-sectional NR Current-pregnancy violence: 2.3% Ameh, Shittu and Abdul (2008)

Zaria, Kaduna State Antenatal clinic; risk-assessment study NR Developed risk indicators for domestic violence Umeora et al. (2008) South-East Nigeria Referral-centre prenatal clinic; cross-sectional 500 Violence prevalence: 13.6%; assessed patterns and determinants Awusi, Okeleke and Anyanwu (2009) Oleh, Delta State Suburban antenatal population; cross-sectional NR Domestic violence during pregnancy: prevalence and correlates examined Gyuse, Ushie and Etukidem (2009) Jos, Plateau State Hospital ANC; cross-sectional NR Pregnancy-related violence and its different forms Abasiubong et al.

Uyo, Akwa Ibom State Antenatal population; cross-sectional NR Demographic risk factors and domestic violence Fawole, Abass and Ibadan, Oyo State Antenatal clinics; cross-sectional NR Violence during pregnancy and associated Fawole (2010) factors Olagbuji et al. (2010) Benin City, Edo State Referral centre; repeated cross-sectional/trend analysis NR Pregnancy-related violence trends and determinants Adesina, Oyugbo and Olubukola (2011) Ibadan, Oyo State Antenatal clinic; cross-sectional NR Prevalence and pattern of violence in pregnancy Arulogun and Jidda (2011) Abuja FCT Selected hospitals; cross-sectional NR Experiences and forms of violence among pregnant women Envuladu et al. (2012) Jos North, Plateau State PHC antenatal clinic; cross-sectional 201 Current-pregnancy violence: 31.8%; previous-pregnancy violence: 28.9% Iliyasu et al. (2013) Kano State Tertiary antenatal facility; cross-sectional 392 analysed Current-pregnancy violence: 7.4% Onoh et al. (2013) Abakaliki, Ebonyi State Federal Medical Centre ANC; cross-sectional NR Index-pregnancy IPV: 44.6% Ashimi and Amole (2015) Northwest Nigeria Rural/community antenatal population; cross-sectional NR Domestic violence in pregnancy and independent predictors Ansari et al.

Cross River and Bauchi states Household survey with qualitative component Large two-state sample Physical IPV in the last pregnancy: 22% in Cross River and 9% in Bauchi Ayodapo et al. (2017) Oyo East LGA, Oyo State Three PHCs and one secondary facility; cross-sectional 350 IPV exposure: 72.0%; disclosure among exposed women: 28.6% Ikeako et al. (2017) Enugu State Tertiary antenatal clinic; cross-sectional NR IPV during pregnancy: 36.1% Iyanam, Udoh and Morgan (2018)

Itu, Akwa Ibom State PHC antenatal clinic; cross-sectional 200 Domestic violence prevalence and risk factors Iwuoha, Ekeleme and Ezirim (2018) Abia State Tertiary institutions; cross-sectional NR Determinants and consequences of IPV Ezeudu et al. (2019) Enugu State Maternal and child health facilities: mixed-method study NR IPV before and during pregnancy; prevalence during pregnancy is approximately 37.2% Oche et al. (2020) Sokoto State Antenatal clinic; cross-sectional NR IPV knowledge and experience; psychological and controlling behaviours assessed Tella et al. (2020) Niger Delta region Rural and urban antenatal settings; comparative cross-sectional NR Domestic violence, risk factors, coping and help-seeking Ayeni, Yusuf and Musa (2021)

Karu and Jikwoyi, Abuja FCT Two PHCs; cross-sectional 384 GBV prevalence: 59.6% Njoku et al. (2021) Enugu State Pregnant antenatal attendees; cross-sectional 400 Ever abused during pregnancy: 43.0% (2022)

Ilorin, Kwara State University teaching hospital ANC; cross-sectional 370 Domestic violence: 34.5% Dada et al. (2022) Ado-Ekiti, Ekiti State Tertiary antenatal clinic; cross-sectional NR IPV prevalence: 7.9% Egwu et al. (2023) Abakaliki, Ebonyi State Federal teaching hospital ANC; cross-sectional 165 Any domestic violence: 40.0% Nwafor et al. (2024) Nigeria Booked antenatal population; cross-sectional 137 IPV prevalence: 52.6% (72/137) Chiejine et al. (2024) Asaba, Delta State Antenatal clinic; cross-sectional NR Overall IPV prevalence: 27.3% Olabisi et al. (2026)

Nigerian healthcare facilities Antenatal attendees; cross-sectional NR IPV experience: approximately 28.4% Prevalence and Patterns of Gender-Based Violence From the above analysis in Table 2, the expectant women are exposed to different levels of gender-based violence (GBV), with the range starting from 2.3 % as per Fawole et al (2008) [19] to 72.0 % per Ayodapo et al (2017) [10] in OYO East. It was a descriptive median of ~31.8%, which did not allow interpretation as a pooled national prevalence because it did not account for sample size, variance, or methodological quality, and it presented a high level of violence among attendees at the antenatal clinic. The variations may be attributed in part to variations in the definition of violence and the screening devices used. Some of the studies were only for physical violence, and mostly, there was a lower level of prevalence of violence. In contrast, studies of violence of any type, including psychological, verbal, sexual and controlling, also had high prevalence, including a study by Ansari et al. (2017) [7]. For instance, Ayeni et al (2021) [12] indicated the prevalence at 59.6%, and nearly all the cases were emotional abuse, while Chiejine et al (2024) [13] indicated a higher prevalence for psychological violence and controlling behaviour than physical violence. A regional difference was also observed due to the difference between the reports by Iliyasu et al. (2013) [23] (Northern Nigeria), which found 7.4, Onoh et al. (2013) [34] at 44.6 in Abakaliki, Njoku et al. (2021) [28] at 43.0 in Enugu and Kofoworade et al. (2022) [26] at 34.5 in Ilorin. The higher prevalence estimates of 40.0% by Egwu et al. (2023) and 52.6% by Nwafor et al. (2024) [30] may reflect better sensitisation and reporting of nonphysical violence rather than an increase in prevalence over time. Psychological and/or verbal abuse was the most prevalent form of abuse by Onoh et al (2013) [34], Ikeako et al (2017) [22], Ayeni et al (2021) [12], Kofoworade et al (2022) [26] and Egwu et al (2023), which implies that any screening tool that does not identify psychological and/or verbal abuse will severely underestimate the problem. For this reason, in statistical pooling, it is important to distinguish overall violence, physical, sexual and psychological violence and also to distinguish between current pregnancies (where there is violence) and violence in life.

Table 2: Numerical synthesis of reported prevalence estimates Study Overall prevalence Physical violence Sexual violence Psychological/emotional/verbal violence Ezechi et al.

(2004) 28.7% Reported Reported Reported Fawole et al.

(2008), Abeokuta 2.3% Principal measured form NR NR Umeora et al.

(2008) 13.6% Reported Reported Verbal/psychological abuse is prominent Envuladu et al.

31.8% current; 28.9% previous pregnancy Reported Reported Reported Iliyasu et al.

(2013) 7.4% 58.6% of abused women 13.8% of abused women Reported Onoh et al. (2013) 44.6% Approximately 21% of reported forms combined with emotional abuse Reported Verbal abuse accounted for approximately 60.1% of reported forms Ansari et al.

22% Cross River; 9% Bauchi Physical IPV measured Not the principal outcome Not the principal outcome Ayodapo et al.

(2017) 72.0% Reported Reported Reported Ikeako et al.

(2017) 36.1% Reported No sexual violence reported Verbal abuse: 14.4% Ayeni et al. (2021) 59.6% 28.5% of survivors/cases 13.6% 99.4% of survivors/cases Njoku et al. (2021) 43.0% ever abused in pregnancy Reported Reported Reported (2022) 34.5% Physical assault with injury: 3.5% Reported Psychological aggression: 74.8% of the violence pattern Dada et al. (2022) 7.9% Included in WAST assessment Included Included Egwu et al. (2023) 40.0% The second most common form Reported Psychological violence is the most common Nwafor et al.

(2026) Approximately 28.4% Reported Reported Reported Associated Factors Table 3 suggests that there is interactive individual, partner, household, clinical and community predictors of pregnancy-related gender‐based violence and not any one predictor. A shared theme was that of socioeconomic disadvantage. Regardless, low maternal education was associated with adverse outcomes in Fawole et al. (2008) [19], Abasiubong et al. (2010) [1], and Iliyasu et al. (2013) [23], and unemployment and low income were significant factors in Dada et al. (2022) [14]. These factors may reduce women's knowledge and awareness of access to effective protection options, reduce agency within the household to make decisions and place a greater financial strain on women who endorse a need to exit an abusive relationship. Some research has, however, shown that women earners could be more vulnerable to violence if there is an idea of danger to the established gender dynamics in the home, as this is perceived as endangering their access to income (Ansari et al., 2017) [7]. Partner attributes were of great importance, too. Unemployment of partners (Ikeako et al., 2017) [22], alcohol consumption (Envuladu et al., 2012) [16], low education status of sexual partners, low level of education and HIV status of the clients were all factors found to be associated with violence. According to Ameh et al. (2008) [5], a combination of indicators could be a better way to identify a child for early help than relying on a single indicator. Polygamous marriage was a risk factor in Envuladu et al. (2012) [16] and Ashimi and Amole (2015) [9], which could be attributed to the jealousy, competition for household resources and unequal decision making. Family dysfunction was seen to be a significant correlate of violence in adolescent pregnancy. It was also blamed on the same violence-adolescent pregnancy association by Kofoworade et al. (2022) [26], while Dada et al. (2022) [14] attributed anaemia to it. Community norms were also considered, as the extent of violence reported in the community positively predicted an individual's risk, consistent with Ansari et al. (2017) [7]. All of these indicate the need for a multidimensional assessment during pregnancy, including assessment of socioeconomic status, partner behaviour, family functioning, and maternal status. The majority of these studies were cross-sectional in design and therefore identified only associations; prospective studies are needed to understand the direction of these associations better.

Table 3: Thematic synthesis of factors associated with gender-based violence Domain Associated factor Direction of association Supporting studies Woman’s characteristics Low educational attainment Generally increased risk Fawole et al. (2008); Iliyasu et al. (2013); Abasiubong et al. (2010)

Woman’s characteristics Unemployment Increased vulnerability Dada et al. (2022); Iliyasu et al. (2013) Woman’s characteristics Low income/economic dependence Increased risk Dada et al. (2022); Ansari et al. (2016/2017) Woman’s characteristics Personal income with control over spending Context-specific increased risk Ansari et al. (2016/2017) Woman’s characteristics Younger maternal age Frequently associated with increased vulnerability Partner characteristics Low partner education Increased risk Envuladu et al. (2012); Fawole et al. (2008) Partner characteristics Partner unemployment or unstable occupation Increased risk Ikeako et al. (2017); Envuladu et al. (2012) Partner characteristics Alcohol consumption Strongly increased risk Envuladu et al. (2012); Partner characteristics Multiple sexual partners Increased risk Envuladu et al. (2012) Relationship factors Polygamous marriage Increased risk Envuladu et al. (2012); Ashimi and Amole (2015)

Relationship factors Family dysfunction Increased risk Kofoworade et al. (2022) Relationship factors Financial disagreement Common precipitating factor Iliyasu et al. (2013); Efetie and Salami (2007) Relationship factors Previous history of violence Increased likelihood of recurrence Envuladu et al. (2012); Ezechi et al. (2004) Clinical factors HIV-positive status Increased risk Envuladu et al. (2012) Clinical factors Anaemia Associated with violence Kofoworade et al. (2022) Mental health Depression Positively associated with IPV Dada et al. (2022) Mental health Suicidal ideation Associated with IPV Dada et al. (2022) Community factors Community-level normalisation of violence Increased individual risk Ansari et al. (2016/2017) Protective factors Greater household wealth Reduced risk Ansari et al. (2016/2017) Protective factors Higher partner education Reduced risk Ansari et al. (2016/2017); Envuladu et al. (2012)

Perpetrators, Disclosure and Help-Seeking Table 4 shows that husbands and current intimate partners are the most prevalent actors of violence against pregnant women, which agrees with the results of Efetie and Salami (2007) [15], Onoh et al. (2013) [34], Ikeako et al. (2017) [22] and Ayeni et al. (2021) [12]. This only corroborates an earlier report that intimate partner violence (IPV) is the most prevalent type of pregnancy-related gender-based violence (GBV) in Nigeria. But, numerous cases of abuse by boyfriends, relatives, co-wives and in-laws have been reported, suggesting that the notion of screening based on the assumption of abuse by a husband is incorrect. This is particularly pertinent in the context of extended-family/having more than one partner, where conflictual situations regarding financial resources/property, reproductive aspirations/choices, and access can involve more than one family member. The low disclosure rate recorded by Ayodapo et al. (2017) [10], for instance, is alarming. 350 respondents (350 women) reported exposure, and 72 women (28.6% of survivors) reported exposure. This suggests that the burden is significantly underestimated in both formal case reporting and police statistics. In addition to disclosure to trusted family members, there are also informal coping and help-seeking interventions that were identified by Tella et al. (2020) [35] as silence, avoidance, family intervention and religious help. Informal networks can provide emotional support and help facilitate continued living and/or reconciliation without addressing further risk issues. 46.2% of the respondents indicated that they would seek legal action in the case of physical violence, which may be attributed to fear of repercussion, economic dependence and stigma (as reported by Ikeako et al., 2017) [22], as well as their fear of the breakup of the family, and limited trust in formal institutions. It is relevant to fieldwork in antenatal care. Violence screening should be conducted in the absence of partners/relatives by trained personnel using culturally sensitive, nonjudgmental questions. Ongoing mental health care/social welfare, emergency housing, legal counsel, and/or law enforcement protection must be available, and there should be a confidential referral system within health care facilities that connects survivors with these services. Community and religious leaders, who are usually the first to be affected during times of disclosure, should also receive appropriate training. If no confidential, well-resourced and survivor-centred system is in place to respond, then screening may be costly.

Table 4: Perpetrators, disclosure and responses to violence Theme Main result Study evidence Primary perpetrator Husband or current intimate partner was the dominant perpetrator Efetie and Salami (2007); Onoh et al. (2013); Ikeako et al. (2017); Ayeni et al. (2021)

Other perpetrators Boyfriends, relatives, co-wives and in-laws were occasionally identified Benin City and South-East studies Disclosure Only a minority of exposed women disclosed violence Ayodapo et al. (2017): 72/252 or 28.6% First disclosure point Relatives, friends and religious leaders were preferred Ayodapo et al. (2017); Tella et al. (2020) Formal reporting Reporting to health professionals, police, or legal authorities was uncommon. Ayodapo et al. (2017); Ikeako et al. (2017) Legal response Limited willingness to seek legal redress Ikeako et al. (2017): 23.8% indicated willingness in cases of physical violence Coping strategy Silence, endurance, avoidance, family mediation and religious support were reported. Tella et al. (2020) Medical consequences Injury, anaemia, miscarriage and preterm delivery were reported Kofoworade et al. (2022); Ayeni et al. (2021) Pregnancy and Health Consequences Table 6 reveals that there is an association between GBV during pregnancy and the mother experiencing physical injuries, psychological distress, clinical complications and adverse pregnancy outcomes. The findings in the study by Kofoworade et al. (2022) [26] revealed a significant relationship between domestic violence and anaemia, indicating that the survivors might receive poor nutrition, have restricted access to food, and might have problems taking medicine that enhances the absorption of iron, or frequent doctor checkups. Even the less common physical assault in that study does not suggest the assault was small in any regard; however, as there is no physical manifestation with some of the other forms of abuse, namely psychological and controlling abuse, certain health effects may be significant. Dada et al. (2022) [14] revealed that there is a positive correlation between IPV & depression suicidal thoughts, indicating a great mental health problem among the abused pregnant mothers. The results also confirm the literature, indicating that the level of disclosures is low (Ayodapo et al., 2017) [10] and the results of Tella et al. (2020) [35], which indicate that coping is a problem because silence or lack of support from others may worsen fear, isolation, and psychological distress. There were also other undesirable pregnancy complications. Of the 27.6% of the relevant physically-abused group who suffered from miscarriage, most of whom (6.9%) had suffered from previous preterm delivery. Unfavourable obstetric outcomes are also associated with domestic violence, as done by Ameh et al. (2008) [5] and Onoh et al. (2013) [34]. Those mechanisms can be a direct accident to the abdomen, over-stimulation of maternal stress response, inadequate food intake, greater risk of infection, postponing care and less adherence to carer guidelines. Most, however, were retrospective self-reports/cross-sectional, making it impossible to determine cause and effect, and there could have been recall bias. Prospective cohort studies should adjust for medical conditions, substance use, prior pregnancy and birth, poverty and age of mother. The clinical implications of the results indicate that the screening for violence should be a routine part of the assessment of injury, anaemia, depression, suicidal risk and obstetric complications. The social needs are strongly interwoven and linked with psychological and obstetric needs, and identified survivors cannot be referred but instead should be cared for in a coordinated manner.

Table 5: Maternal and pregnancy consequences Consequence Evidence Numerical finding Physical injury Ezechi et al.; Kofoworade et al. Physical assault with injury constituted 3.5% of the pattern in Anaemia Kofoworade et al. (2022) Significant association Depression Dada et al. (2022) r=0.151, p=0.012 Suicidal ideation Dada et al. (2022) p=0.04 Miscarriage Ayeni et al. (2021) 27.6% among the relevant previously abused group Preterm delivery Ayeni et al. (2021) 6.9% Ameh et al. (2008) and related studies - Reduced disclosure and care-seeking Ayodapo et al. (2017); Tella et al.

(2020) Disclosure among exposed women: 28.6% in Oyo East It is safe to suggest that from this systematic review, gender-based violence is a significant and embedded problem of maternal health for pregnant women attending antenatal clinics in Nigeria. Descriptive prevalence rates ranged from 2.3% to 72.0%, with a median of approximately 31.8% across the studies included. The overall span is due to the wide range of studies on their location, health care contexts, baseline characteristics of the persons studied, the definition of violence, the time period covered, and the instruments of measurement. It should not be interpreted as an overall national-level prevalence. The most common were psychological, emotional and verbal violence; followed by physical and sexual violence; thus, screening that is limited only to physical injuries will seriously underestimate the burden. Husbands or husbands-to-be and current boyfriends were the main perpetrators, but boyfriends, relatives, co-wives and in-laws were also reported as perpetrators of violence. Factors repeatedly mentioned were low levels of education, unemployment, poverty, polygamous marriage, partner drinking, frequent sexual partners, family dysfunction and prior victimisation/experience of violence and power inequality. Important consequences were also found for maternal and fetal well-being, with violence also associated with anaemia, depression, suicidal ideation, physical injury, miscarriage and preterm delivery. Disclosure and formal help-seeking were low, with survivors more likely to talk to relatives, friends, and religious leaders than to seek help from healthcare providers, law-enforcement agencies, or legal services. Antenatal clinics should thus provide for private and confidential screening for all forms of violence and referral pathways to link survivors with medical, psychological, social, legal and protection services should be made effective. Healthcare workers need to be trained to be able to identify nonphysical abuse, ascertain the level of safety and respond accordingly, without any bias or discrimination. Unhelpful gender norms, economic vulnerability, alcohol abuse and the imbalance in relationships must be addressed in the broader spectrum of prevention. Validated instruments, the same reference period, and the same samples/designs, with a national approach in future investigations, are recommended. All of these are critical to protecting pregnant women, improving pregnancy outcomes, and advancing progress towards SDG 3 and 5 in Nigeria.