Prevalence and Predictors of Intimate Partner Violence During Pregnancy in Northern Ghana: A Cross‐Sectional Study Authors: Fairuza Marian Salifu, Joseph Lasong, Yula Salifu Published: 2026 ABSTRACT ABSTRACTBackground and AimsThe persistence of intimate partner violence (IPV) in developing countries is attributable to patriarchal socio‐cultural norms that endorse male dominance, compounded by limited enforcement of legal protections. IPV during pregnancy poses significant risks to both maternal and fetal health. This study aimed to determine the prevalence and predictors of IPV during pregnancy in the Tamale Metropolis, Northern Ghana.MethodsA hospital‐based cross‐sectional study was conducted from April to May 2023. Systematic random sampling was used to recruit 260 pregnant women from three public health facilities in the Tamale Metropolis. Women with high‐risk pregnancies and non‐consenting participants were excluded. Data were collected using structured questionnaires and analyzed using IBM SPSS version 25. Binary logistic regression was performed to identify independent predictors of IPV. All statistical tests were two‐tailed, with statistical significance set at p < 0.05.ResultsThe overall prevalence of IPV during the current pregnancy was 60.0%. Compared with women aged ≤ 20 years, those aged 21–30 years (AOR: 0.3; 95% CI: 0.1–0.8; p = 0.02) and those aged ≥ 31 years (AOR: 0.2; 95% CI: 0.1–0.6; p = 0.01) had significantly lower odds of IPV. Women whose spouses had witnessed their fathers beating their mothers during childhood also had lower adjusted odds of IPV (AOR: 0.3; 95% CI: 0.1–0.9; p = 0.03); however, the bivariate analysis showed the opposite direction, and this adjusted finding warrants cautious interpretation and longitudinal investigation.ConclusionIPV during pregnancy was highly prevalent in the Tamale Metropolis. Older maternal age was associated with reduced odds of IPV. The inverse adjusted association between spousal childhood exposure to paternal IPV and current IPV risk was unexpected and requires further investigation. Targeted interventions for young pregnant women, integration of IPV screening into antenatal care, and community‐based education campaigns are recommended.Keywords: antenatal care, cross‐sectional study, Ghana, intimate partner violence, predictors, pregnancy Abbreviations antenatal care adjusted odds ratio confidence interval Demographic and Health Survey intimate partner violence Sustainable Development Goal Sub‐Saharan Africa variance inflation factor World Health Organization Introduction Intimate partner violence (IPV), encompassing physical, sexual, and emotional abuse by a current or former partner, remains one of the most pervasive public health challenges worldwide [1]. The World Health Organization (WHO) estimates that approximately one in three women experiences IPV during their lifetime [2]. A systematic review and meta‐analysis reported that the pooled prevalence of IPV during pregnancy ranged from approximately 25% to 40% in Sub‐Saharan Africa (SSA) [3]. In Ghana, despite legal frameworks such as the Domestic Violence Act [4], IPV remains a significant concern. National estimates from the Ghana Demographic and Health Survey (DHS) indicate a gradual decline in IPV rates from 34% in 2008 [5] to 30% in 2022 [6]; however, these rates remain disproportionately elevated relative to global targets. Among married or cohabiting women, the prevalence of sexual violence was 15%, while physical violence ranged from 8% in the Greater Accra region to 21% in the Upper East region [6]. IPV rates as high as 47% were recorded in the Savannah region [6]. In contrast, a study conducted in Accra reported that 11.2% of pregnant women experienced some form of IPV [7]. The discrepancy between reported and estimated IPV cases is partly attributable to underreporting, driven by cultural stigma, fear of reprisal, and inadequate awareness of legal rights [2, 8, 9]. Risk factors consistently associated with IPV during pregnancy include younger age, lower educational attainment, and economic dependency [8, 10]. Psychosocial factors—including a partner's substance use, prior exposure to parental violence, and unilateral household decision‐making—have also been associated with elevated IPV risk [9, 11]. IPV during pregnancy is associated with adverse feto‐maternal outcomes, including preterm birth, low birth weight, and pregnancy loss [2, 10]. Persistent IPV is linked to depression, anxiety, and impaired health‐seeking behavior [2]. These consequences impede progress toward Sustainable Development Goal 3 (reduction of maternal and neonatal mortality) and Sustainable Development Goal 5 (gender equality and elimination of violence against women) [12]. In Ghana, the Domestic Violence Act [4] provides legal protection; however, its effectiveness is constrained by enforcement challenges and limited public awareness of available legal recourse. Research indicates that only a minority of IPV cases are officially reported, creating a substantial gap between the actual burden and surveillance data [4, 6]. Most existing studies focus on IPV among women generally, with comparatively few targeting pregnant women specifically. Evidence on IPV during pregnancy in Northern Ghana—a setting characterized by documented socio‐economic disparities, cultural norms endorsing male authority, and limited research on this topic [13]—is particularly scarce. This study therefore sought to determine the prevalence and predictors of IPV among pregnant women attending selected antenatal care (ANC) clinics in the Tamale Metropolis of Northern Ghana. The findings are intended to inform the development of culturally sensitive, targeted interventions to reduce IPV, improve maternal and fetal health outcomes, and support Ghana's progress toward global health and gender equality targets. Methods Study Area This study was conducted in the Tamale Metropolis, one of the 16 administrative districts in Ghana's Northern Region. The Metropolis covers a total land area of 646.90 km² and has a population of 233,252, of whom 50.3% are female, and 81% reside in urban areas [14]. Three public health facilities within the Tamale Metropolis were selected: Tamale Teaching Hospital, Tamale Central Hospital, and Tamale West Hospital. The tertiary‐level facility (Tamale Teaching Hospital) was purposively selected on account of its referral role, staff capacity, and large client base, while two secondary‐level hospitals were randomly selected from a pool of four secondary facilities. This random selection of secondary facilities strengthens the representativeness of the study sample by increasing the likelihood of capturing the diversity of patient volumes, staffing levels, and service availability across secondary facilities in the Metropolis. Study Design A cross‐sectional study was conducted between April and May 2023 to determine the prevalence and factors associated with IPV among pregnant women in a current partnership. This design was appropriate as it enabled assessment of IPV prevalence and associated factors at a single point in time, consistent with the study's cross‐sectional objectives. Study Population Pregnant women attending ANC with ANC record books, in a current partnership of at least 12 months' duration, aged ≥ 18 years, residing in the Tamale Metropolis for at least 6 months, and providing written informed consent were eligible for inclusion. Women with high‐risk pregnancies (including those with placenta previa, severe pre‐eclampsia, or multiple gestations) were excluded because such conditions are associated with specialized clinical management protocols, more frequent hospital visits, and elevated psychological distress—all of which may confound IPV disclosure patterns and introduce information bias. Non‐consenting participants and those absent during the data collection period were also excluded. Sample Size and Sampling Technique Sample size was estimated using the Cochran formula: n = Z 2 α p(1 − p)/e 2, where n = sample size, Zα = 1.96 (corresponding to a 95% confidence level), p = 0.112 (IPV prevalence among pregnant women in Ghana) [7], and e = 0.05 (acceptable marginal error). Accounting for a 10% non‐response rate, the minimum required sample was 169 participants. A total of 260 participants were enrolled, as recruitment continued for the full planned data collection period across all three facilities, leveraging the complete pool of eligible participants. This exceeded the minimum requirement, thereby improving statistical power, enhancing the precision of prevalence estimates, and strengthening the robustness of subgroup comparisons. Probability proportionate‐to‐size allocation was applied to distribute the sample across the three health facilities (see Supporting Information S1). A sampling frame was constructed from the daily ANC attendance registers at each facility. Systematic sampling with an interval of three was applied, with the starting point randomly selected from the first three entries on each day's register, until the required sample allocation per facility was reached (kindly see Supporting Information S3: Figure 1). Data Collection and Procedure Five enumerators with Master of Public Health degrees, fluent in English and relevant local dialects (primarily Dagbani), were recruited and trained for 2 days prior to data collection. The Ghana 2022 DHS questionnaire was adapted for data collection among study participants [6]. Reviewed literature [15, 16] and expert consultation were used to adapt the questionnaire to the local socio‐cultural context. The questionnaire captured: (1) socio‐demographic characteristics of participants and their partners (age, educational level, occupation, religion, ethnicity); (2) IPV‐related items (emotional, physical, and sexual violence); (3) women's empowerment indicators (household decision‐making, financial autonomy); and (4) partner‐level behavioral factors (substance use, sexual behavior) (see Supporting Information S2). Data were entered into a Google Form to facilitate data capture. Participants who self‐identified as literate in English completed the form independently via a shared link, while enumerators administered the questionnaire orally to non‐literate participants. To ensure data accuracy, enumerators read back each participant's responses for verification, and corrections were made where necessary. Interviews were conducted in English or Dagbani, depending on participant preference. The experience of IPV during the current pregnancy was the primary outcome variable. “Current IPV” was operationally defined as any experience of emotional, physical, or sexual abuse by an intimate partner occurring during the index pregnancy. Three categories of IPV were assessed—physical (seven items), sexual (eight items), and emotional (nine items)—all pertaining to the current pregnancy (see Supporting Information S2 and Table 4). Women who responded “yes” to at least 1 of the 24 IPV items were coded as having experienced IPV (coded 1), while those who responded “no” to all items were coded as not having experienced IPV (coded 0). Table 4Forms of intimate partner violence reported by participants during the current pregnancy (n = 156*).ItemYes, N (%)No, N (%)Physical violence—“During this pregnancy, has your partner ever…”Slapped you86 (55.1)70 (44.9)Twisted your arm or pulled your hair6 (3.8)150 (96.2)Pushed, shook, or thrown something at you5 (3.2)151 (96.8)Punched you with his fist or with something that could hurt you7 (4.5)149 (95.5)Kicked, dragged, or beaten you up5 (3.2)151 (96.8)Tried to choke or burn you on purpose5 (3.2)151 (96.8)Threatened or attacked you with a knife, gun, or any other weapon8 (5.1)148 (94.9)Emotional violence—“During this pregnancy, has your partner ever…”Humiliated or embarrassed you in front of others141 (90.4)15 (9.6)Threatened you with harm24 (15.4)132 (84.6)Insulted you or made you feel bad about yourself103 (66.0)53 (34.0)Kept you from seeing or talking to your family or friends16 (10.3)140 (89.7)Kept you from having access to a job, money, or financial resources13 (8.3)143 (91.7)Mistrusted your emotions or accused you unfairly54 (34.6)102 (65.4)Accused you of flirting or cheating and blamed you for their problems33 (21.2)123 (78.8)Made you feel constantly confused and trapped, such that you could not express your emotions84 (53.8)72 (46.2)Made you feel like you were “walking on eggshells” around him25 (16.0)131 (84.0)Sexual violence—“During this pregnancy, has your partner ever…”Physically forced you into unwanted sexual intercourse61 (39.1)95 (60.9)Forced you into other unwanted sexual acts13 (8.3)143 (91.7)Made you agree to have sex only because you were afraid19 (12.2)137 (87.8)Refused to use or agree to safe sex practices25 (16.0)131 (84.0)Forced his decisions regarding birth control, pregnancy, or abortion on you55 (35.3)101 (64.7)Withheld sex as a form of punishment26 (16.7)130 (83.3)Persisted with sexual advances after you said “no”131 (83.97)25 (16.0)Insulted or criticized your body or private parts133 (85.3)23 (14.7) Note: Percentages are column proportions within this IPV‐positive subsample.*Denominator = 156 women who reported experiencing any form of IPV during the current pregnancy. Forms of intimate partner violence reported by participants during the current pregnancy (n = 156*). Note: Percentages are column proportions within this IPV‐positive subsample. Denominator = 156 women who reported experiencing any form of IPV during the current pregnancy. Independent variables included women's and spousal socio‐demographic characteristics, women's empowerment indicators, and spousal behavioral factors, selected based on prior evidence [11, 15, 16]. The questionnaire was piloted among 30 eligible pregnant women at the Tamale Reproductive and Child Health unit and revised for clarity before use in the main study. Interviews lasted approximately 20–45 min. Enumerators were trained to offer breaks as needed to mitigate respondent fatigue, and participants were informed of the estimated duration during the informed consent process. Ethical approval was obtained from the University for Development Studies Ethics Review Board (UDS/RB/date: 27/02/2023). Administrative permission was obtained from the Northern Regional Health Directorate, the Tamale Metropolitan Health Directorate, and all participating hospital administrations prior to data collection. Written informed consent was obtained from all participants before enrollment. Participation was entirely voluntary; participants were free to decline to answer any question or withdraw at any point without consequence. Confidentiality and anonymity of all data were strictly maintained throughout. The study was conducted in accordance with the principles of the Declaration of Helsinki (2013 revision). The STROBE guidelines for cross‐sectional studies were followed in reporting this study. Data Analysis Data exported from Google Forms were cleaned in Microsoft Excel and analyzed using IBM SPSS version 25 (IBM Corp., Armonk, NY, USA). Descriptive statistics were summarized as frequencies and percentages. Chi‐square (χ²) tests of independence were performed to assess bivariate associations between IPV and each independent variable; results are reported with the test statistic, degrees of freedom, and p‐value. Variables significant at p < 0.1 in bivariate analysis were entered as candidates into the multivariable model—this threshold is an exploratory criterion for variable selection and is distinguished from the pre‐specified significance level of p < 0.05 for final inference. Binary logistic regression was performed to identify independent predictors of IPV during the current pregnancy, and adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were reported. Multicollinearity was assessed using the variance inflation factor (VIF); all VIF values were below 5, indicating no problematic multicollinearity. Missing data were handled using listwise deletion. A goodness‐of‐fit assessment was performed using the Hosmer–Lemeshow test, which was non‐significant (p > 0.05), confirming adequate model fit. All statistical tests were two‐tailed, with a pre‐specified significance level of p < 0.05. Results Socio‐Demographic Characteristics of Women Table 1 presents the socio‐demographic characteristics of participants and their bivariate associations with IPV during the current pregnancy. The mean age was 26 ± 5.3 years. Among participants, 42.3% had no formal education, 22.3% were unemployed, and 86.9% practised Islam. Table 1Socio‐demographic characteristics of participants and bivariate association with IPV during the current pregnancy (N = 260).VariableIPV: Yes, N (%)IPV: No, N (%)Total χ²(df); p‐valueAge of participant (years) χ²(2) = 8.80; p = 0.01* Mean age: 26 ± 5.3 years≤ 20 years13 (39.4)20 (60.6)3321–30 years106 (59.9)71 (40.1)177≥ 31 years36 (72.0)14 (28.0)50Highest educational level χ²(4) = 4.30; p = 0.37No formal education73 (66.4)37 (33.6)110Primary12 (63.2)7 (36.8)19Junior high school (JHS)13 (52.0)12 (48.0)25Senior high school (SHS)18 (54.5)15 (45.5)33Tertiary39 (53.4)34 (46.6)73Ethnicity χ²(2) = 3.75; p = 0.15Dagomba123 (58.9)86 (41.1)209Akan7 (43.8)9 (56.3)16Other25 (71.4)10 (28.6)35Occupational status χ²(3) = 7.12; p = 0.07Farmer21 (56.8)16 (43.2)37Trader72 (69.2)32 (30.8)104Unemployed29 (50.0)29 (50.0)58Public servant33 (54.1)28 (45.9)61Area of residence χ²(1) < 0.001; p = 1.00Rural25 (59.5)17 (40.5)42Urban130 (59.6)88 (40.4)218Religion χ²(1) = 3.90; p = 0.06Christianity15 (44.1)19 (55.9)34Islam140 (61.9)86 (38.1)226 Note: Percentages are row proportions within each variable category. IPV = intimate partner violence during the current pregnancy.* p < 0.05. Socio‐demographic characteristics of participants and bivariate association with IPV during the current pregnancy (N = 260). Note: Percentages are row proportions within each variable category. IPV = intimate partner violence during the current pregnancy. p < 0.05. In the bivariate analysis, women's age was the only socio‐demographic variable significantly associated with IPV (χ²(2) = 8.80; p = 0.01). Among women aged ≥ 31 years, 72.0% experienced IPV during the current pregnancy; the corresponding figures were 59.9% among women aged 21–30 years and 39.4% among those aged ≤ 20 years. No significant bivariate associations were observed for educational status (χ²(4) = 4.30; p = 0.37), ethnicity (χ²(2) = 3.75; p = 0.15), occupational status (χ²(3) = 7.12; p = 0.07), area of residence (χ²(1) < 0.001; p = 1.00), or religion (χ²(1) = 3.90; p = 0.06) (Table 1). Spousal Characteristics Table 2 presents spousal characteristics and their bivariate associations with IPV. In the bivariate analysis, spousal ethnicity (χ²(2) = 7.87; p = 0.02), having children with another woman (χ²(1) = 6.03; p = 0.02), spousal smoking (χ²(1) = 7.46; p = 0.01), a prior marriage or divorce (χ²(1) = 6.61; p = 0.01), multiple sexual partners (χ²(1) = 6.32; p = 0.02), and spousal childhood exposure to paternal IPV (χ²(1) = 8.26; p < 0.01) were each significantly associated with IPV during the current pregnancy. Among women whose spouses reported witnessing their father beating their mother during childhood, 81.1% experienced IPV during the current pregnancy, compared with 56.1% of women whose spouses reported no such exposure—a bivariate direction consistent with social learning theory [17, 18]. The direction of this association reversed in the adjusted analysis, discussed further in Section Conclusion, 4 (Table 2). Table 2Spousal characteristics and bivariate association with IPV during the current pregnancy (N = 260).VariableIPV: Yes, N (%)IPV: No, N (%)Total χ²(df); p valueAge of spouse (years) χ²(2) = 2.57; p = 0.28≤ 257 (41.2)10 (58.8)1726–46140 (60.9)90 (39.1)230> 468 (61.5)5 (38.5)13Highest educational level χ²(2) = 4.49; p = 0.11No formal education68 (66.0)35 (34.0)103Middle school9 (42.9)12 (57.1)21Tertiary78 (57.4)58 (42.6)136Ethnicity χ²(2) = 7.87; p = 0.02* Dagomba141 (62.7)84 (37.3)225Gonja6 (54.5)5 (45.5)11Other8 (33.3)16 (66.7)24Occupation χ²(3) = 2.00; p = 0.41Farmer63 (65.6)33 (34.4)96Trader12 (54.5)10 (45.5)22Unemployed5 (45.5)6 (54.5)11Public servant75 (57.3)56 (42.7)131Religion χ²(1) = 0.94; p = 0.34Christianity16 (51.6)15 (48.4)31Islam139 (60.7)90 (39.3)229Spouse has children with another woman χ²(1) = 6.03; p = 0.02* Yes50 (72.5)19 (27.5)69No105 (55.6)84 (44.4)189Spouse smokes χ²(1) = 7.46; p = 0.01* Yes35 (77.8)10 (22.2)45No120 (55.8)95 (44.2)215Spouse drinks alcohol χ²(1) = 1.19; p = 0.32Yes20 (69.0)9 (31.0)29No135 (58.4)96 (41.6)231Spouse has a prior marriage or divorce χ²(1) = 6.61; p = 0.01* Yes23 (82.1)5 (17.9)28No132 (56.9)100 (43.1)232Spouse has multiple sexual partners χ²(1) = 6.32; p = 0.02* Yes60 (70.6)25 (29.4)84No95 (54.3)80 (45.7)175How a woman responds to disagreements with her spouse χ²(2) = 5.10; p = 0.08Keeps calm74 (55.2)60 (44.8)134Talks back40 (72.7)15 (27.3)55Apologizes41 (57.7)30 (42.3)71Spouse witnessed his father beating his mother during childhood χ²(1) = 8.26; p < 0.01* Yes30 (81.1)7 (18.9)37No125 (56.1)98 (43.9)223 Note: Percentages are row proportions within each variable category.* p < 0.05. Spousal characteristics and bivariate association with IPV during the current pregnancy (N = 260). Note: Percentages are row proportions within each variable category. p < 0.05. Women's Empowerment Characteristics Table 3 presents women's empowerment indicators and their bivariate associations with IPV. Decision‐making on large household purchases by the spouse alone (χ²(2) = 7.83; p = 0.02), decision‐making on visiting family by the participant alone (χ²(2) = 8.25; p = 0.02), and spousal control of contraceptive decisions (χ²(2) = 13.29; p = 0.001) were significantly associated with IPV in the bivariate analysis. Decision‐making on daily household purchases also reached significance (χ²(2) = 6.40; p = 0.04) (Table 3). Table 3Women's empowerment characteristics and bivariate association with IPV during the current pregnancy (N = 260).VariableIPV: Yes, N (%)IPV: No, N (%)Total χ²(df); p valueWho decides on the participant's healthcare? χ²(2) = 5.38; p = 0.07Participant alone41 (66.1)21 (33.9)62Spouse alone97 (61.0)62 (39.0)159Both jointly17 (43.6)22 (56.4)39Who decides on large household purchases? χ²(2) = 7.83; p = 0.02* Participant alone35 (71.4)14 (28.6)49Spouse alone101 (60.5)66 (39.5)167Both jointly19 (43.2)25 (56.8)44Who decides on daily household purchases? χ²(2) = 6.40; p = 0.04* Participant alone40 (71.4)16 (28.6)56Spouse alone94 (59.1)65 (40.9)159Both jointly21 (46.7)24 (53.3)45Who decides on visiting family or relatives? χ²(2) = 8.25; p = 0.02* Participant alone62 (68.1)29 (31.9)91Spouse alone77 (59.2)53 (40.8)130Both jointly16 (41.0)23 (59.0)39Does the participant have control over her own earnings? χ²(1) = 0.05; p = 0.87Yes130 (59.9)87 (40.1)217No25 (58.1)18 (41.9)43Does the participant have control over the spouse's earnings? χ²(1) = 1.11; p = 0.42Yes7 (46.7)8 (53.3)15No148 (60.4)97 (39.6)245Does the participant know the amount the spouse earns? χ²(1) = 3.00; p = 0.09Yes54 (52.9)48 (47.1)102No100 (63.7)57 (36.3)157Who decides on contraceptive use? χ²(2) = 13.29; p = 0.001* Participant alone57 (69.5)25 (30.5)82Spouse alone21 (38.9)33 (61.1)54Both jointly77 (62.1)47 (37.9)124Who decides on the number of children? χ²(2) = 5.43; p = 0.07Participant alone6 (54.5)5 (45.5)11Spouse alone48 (71.6)19 (28.4)67Both jointly101 (55.5)81 (44.5)182 Note: Percentages are row proportions within each variable category.* p < 0.05. Women's empowerment characteristics and bivariate association with IPV during the current pregnancy (N = 260). Note: Percentages are row proportions within each variable category. p < 0.05. Prevalence of IPV During Pregnancy Overall, 60.0% (95% CI: 54.1%–65.9%) of women experienced at least one form of IPV during the current pregnancy (n = 156/260). Emotional violence was the most prevalent form, followed by sexual violence, with physical violence recorded at the lowest prevalence of the three types (kindly see Supporting Information S4: Figure 2). Types of IPV During Pregnancy Physical Violence Among participants who experienced IPV during the current pregnancy (n = 156), the most frequently reported physical act was being slapped (55.1%), followed by being threatened or attacked with a weapon (5.1%) and being punched with a fist or object (4.5%). Other physical acts were each reported by approximately 3%–4% of affected women (Table 4). Emotional Violence Humiliation was the most prevalent form of emotional abuse, reported by 90.4% of women who experienced IPV during the current pregnancy. Other commonly endorsed items included being insulted or made to feel bad about oneself (66.0%), feeling constantly confused and trapped such that emotions could not be expressed (53.8%), and being mistrusted or unfairly accused (34.6%). Threats of harm were reported by 15.4%, and social isolation (being kept from family or friends) by 10.3% (Table 4). Sexual Violence Insulting or criticizing the participant's body was the most reported sexual IPV act (85.3%), followed by persisting with sexual advances after the participant refused (84.0%). Physical force into unwanted intercourse was reported by 39.1% of affected women, while forced reproductive decision‐making occurred in 35.3%. Withholding sex as punishment (16.7%), refusing safe sex (16.0%), and agreeing to sex out of fear (12.2%) were also endorsed (Table 4). Predictors of IPV Among Pregnant Women Table 5 presents the results of the multivariable binary logistic regression analysis. After adjusting for all variables simultaneously, two factors were independently and statistically significantly associated with IPV during the current pregnancy: maternal age and spousal childhood exposure to paternal IPV. Table 5Multivariable predictors of intimate partner violence during the current pregnancy—Binary logistic regression (N = 260).VariableIPV: Yes, N (%)IPV: No, N (%)TotalAOR (95% CI) p valueMaternal age (years)≤ 20 years (Reference)13 (39.4)20 (60.6)331.0 (Reference)—21–30 years106 (59.9)71 (40.1)177 0.3 (0.1–0.8) 0.02 ≥ 31 years36 (72.0)14 (28.0)50 0.2 (0.1–0.6) 0.01 Spousal ethnicityDagomba141 (62.7)84 (37.3)2250.4 (0.1–1.1)0.07Gonja6 (54.5)5 (45.5)110.4 (0.1–2.1)0.27Other (Reference)8 (33.3)16 (66.7)241.0 (Reference)—Spouse has children with another womanYes50 (72.5)19 (27.5)690.9 (0.4–2.1)0.89No (Reference)105 (55.6)84 (44.4)1891.0 (Reference)—Spouse smokesYes35 (77.8)10 (22.2)450.8 (0.3–2.0)0.63No (Reference)120 (55.8)95 (44.2)2151.0 (Reference)—Spouse has a prior marriage or divorceYes23 (82.1)5 (17.9)280.5 (0.1–1.5)0.20No (Reference)132 (56.9)100 (43.1)2321.0 (Reference)—Spouse has multiple sexual partnersYes60 (70.6)25 (29.4)840.8 (0.1–0.9)0.64No (Reference)95 (54.3)80 (45.7)1751.0 (Reference)—Spouse witnessed his father beating his mother during childhoodYes30 (81.1)7 (18.9)37 0.3 (0.1–0.9) 0.03 No (Reference)125 (56.1)98 (43.9)2231.0 (Reference)—Decision‐maker for large household purchasesParticipant alone35 (71.4)14 (28.6)490.4 (0.03–5.1)0.47Spouse alone101 (60.5)66 (39.5)1670.5 (0.01–3.3)0.45Both jointly (Reference)19 (43.2)25 (56.8)441.0 (Reference)—Decision‐maker for daily household purchasesParticipant alone40 (71.4)16 (28.6)561.3 (0.0–15.1)0.81Spouse alone94 (59.1)65 (40.9)1591.8 (0.3–12.0)0.57Both jointly (Reference)21 (46.7)24 (53.3)451.0 (Reference)—Decision‐maker for visiting family or relativesParticipant alone62 (68.1)29 (31.9)910.8 (0.3–2.2)0.63Spouse alone77 (59.2)53 (40.8)1301.0 (0.3–3.0)0.97Both jointly (Reference)16 (41.0)23 (59.0)391.0 (Reference)—Decision‐maker for contraceptive useParticipant alone57 (69.5)25 (30.5)820.7 (0.3–1.3)0.23Spouse alone21 (38.9)33 (61.1)542.0 (0.9–4.2)0.06Both jointly (Reference)77 (62.1)47 (37.9)1241.0 (Reference)— Note: Bold values indicate statistical significance at p < 0.05. Hosmer–Lemeshow test: p > 0.05 (adequate model fit). VIF values all < 5 (no multicollinearity). All tests two‐tailed. Reference categories in parentheses.Abbreviations: AOR, adjusted odds ratio; CI, confidence interval. Multivariable predictors of intimate partner violence during the current pregnancy—Binary logistic regression (N = 260). Note: Bold values indicate statistical significance at p < 0.05. Hosmer–Lemeshow test: p > 0.05 (adequate model fit). VIF values all < 5 (no multicollinearity). All tests two‐tailed. Reference categories in parentheses. Abbreviations: AOR, adjusted odds ratio; CI, confidence interval. Compared with women aged ≤ 20 years, those aged 21–30 years had approximately 70% lower adjusted odds of experiencing IPV (AOR: 0.3; 95% CI: 0.1–0.8; p = 0.02), and those aged ≥ 31 years had approximately 80% lower adjusted odds (AOR: 0.2; 95% CI: 0.1–0.6; p = 0.01). Women whose spouses reported having witnessed their father beating their mother during childhood had approximately 70% lower adjusted odds of experiencing IPV compared with those whose spouses reported no such exposure (AOR: 0.3; 95% CI: 0.1–0.9; p = 0.03). This adjusted inverse association contrasts with the bivariate pattern, in which spousal childhood exposure was associated with substantially higher IPV prevalence (81.1% vs. 56.1%). The remaining variables entered into the model were not independently significant at p < 0.05 in the adjusted analysis (Table 5). Discussion This study determined the prevalence and predictors of IPV during pregnancy among women attending ANC facilities in the Tamale Metropolis, Northern Ghana. The overall prevalence was 60.0%, substantially higher than both national (30%) and regional SSA estimates. Two independent predictors were identified in the adjusted analysis: maternal age—inversely associated with IPV—and spousal childhood exposure to paternal IPV, whose adjusted inverse association contrasted with its positive bivariate direction, indicating a complex relationship requiring further investigation. The prevalence observed in this study exceeds IPV rates reported in prior studies from Ghana (30.5%) [15] and Nigeria (31.5%) [19] and is also higher than pooled estimates from SSA (52%–54%) [20, 21]. Substantially lower rates of 2.2%–8.9% have been reported in high‐income settings [22, 23]. These differences likely reflect variability in socio‐cultural contexts, study populations, measurement tools, operational definitions of IPV, and reporting behaviors. The high prevalence in this setting is consistent with evidence of elevated IPV burden in Northern Ghana, where patriarchal norms and limited awareness of legal protections persist [13]. These findings underscore a critical gap in IPV prevention and response—particularly for pregnant women—and emphasize the need to integrate systematic IPV screening into ANC. Older maternal age was independently associated with reduced odds of IPV during pregnancy. This is consistent with prior research indicating that younger women may face heightened vulnerability, potentially due to greater economic dependency, more limited relationship negotiating capacity, and less developed social networks [24, 25]. Although some studies present conflicting evidence [26], inconsistencies likely reflect contextual differences in study populations and analytical approaches. It is plausible that older age is associated with greater economic independence and stronger relationship‐negotiating capacity, both of which are empirically associated with lower IPV risk in some contexts. However, these represent potential explanatory pathways warranting empirical testing in longitudinal research and should not be interpreted as established causal mechanisms. These findings highlight young pregnant women as a priority group for targeted intervention. The multivariable model identified that women whose spouses had witnessed their fathers beating their mothers during childhood had approximately 70% lower adjusted odds of experiencing IPV (AOR: 0.3; 95% CI: 0.1–0.9; p = 0.03). This finding is counterintuitive relative to social learning theory, which posits that childhood exposure to domestic violence increases the risk of perpetration and victimization in adulthood [10, 17, 18]. Critically, the bivariate analysis showed the opposite direction: women whose spouses had witnessed paternal IPV had substantially higher IPV prevalence (81.1% vs. 56.1%; χ²(1) = 8.26; p < 0.01), consistent with social learning theory. The reversal in direction after multivariable adjustment may reflect a suppressor variable effect, whereby the association between spousal childhood exposure and current IPV is modified by co‐variables simultaneously controlled in the model—such as spousal ethnicity or relationship characteristics—that were not independently significant but may collectively alter the direction of this association. Measurement heterogeneity in respondents' interpretation of “witnessing paternal IPV” may also have contributed. Protective factors such as psychosocial resilience or higher educational attainment, not captured in this study, may further complicate the relationship. This finding should be interpreted with considerable caution. It does not establish a protective mechanism. Rigorous longitudinal and qualitative research is warranted to investigate this complex association. Overall, these findings emphasize the multidimensional nature of IPV risk during pregnancy and the need for context‐sensitive, multi‐level interventions targeting individual vulnerabilities and broader socio‐cultural determinants. Strengths and Limitations This study addresses a documented evidence gap on IPV during pregnancy in Northern Ghana, a setting with limited prior research. Key methodological strengths include: use of a validated, contextually adapted instrument (Ghana DHS questionnaire); multi‐facility sampling across one tertiary and two randomly selected secondary facilities, improving sample representativeness; a clearly defined and operationalised outcome measure; and adherence to STROBE reporting guidelines. The study provides empirical prevalence estimates and identifies two independent predictors—maternal age and spousal childhood IPV exposure—to guide intervention targeting. Several limitations should be considered. First, reliance on self‐reported data raises concerns about social desirability bias and potential underreporting of IPV; future studies should consider multi‐informant methods to improve disclosure. Second, the cross‐sectional design precludes causal inference regarding identified associations; longitudinal designs are needed to establish temporal relationships. Third, the facility‐based, single‐metropolis sampling restricts generalizability to the broader Ghanaian pregnant population; future research should employ population‐based, multi‐regional sampling. Fourth, the absence of knowledge, attitude, and perception measures limits understanding of the social and cognitive drivers of IPV reporting; incorporating these domains would enrich future investigations. Fifth, the study excluded women with high‐risk pregnancies, which may limit applicability to this subpopulation. Conclusions This study documented a high prevalence of IPV during pregnancy (60.0%) in the Tamale Metropolis, Northern Ghana. Older maternal age was independently associated with substantially lower odds of IPV, underscoring young pregnant women as a priority target for prevention. The unexpected inverse adjusted association between spousal childhood exposure to paternal IPV and current IPV contrasts with the bivariate direction and requires cautious interpretation pending longitudinal research. Based on these findings, the following responses are recommended. First, routine IPV screening and brief counseling should be integrated into ANC across all public health facilities, with prioritized attention to young women; this approach is supported by WHO guidance on IPV identification in health settings. Second, community‐based education programmes engaging men and boys should challenge norms that normalize IPV and promote gender‐equitable relationship models. Third, psychosocial support employing trauma‐informed approaches—defined here as care that acknowledges and responds to the impact of prior trauma on current health behavior and relationships—should be available for affected women and high‐risk couples. Fourth, district health directorates should strengthen institutional frameworks for routine IPV data collection and monitoring. Finally, longitudinal and mixed‐methods research is recommended to generate robust evidence on the mechanisms and trajectories of IPV during pregnancy in this and comparable settings. Author Contributions F.M.S., Y.S., and J.L. conceptualized the research plan. F.M.S. led data collection. Y.S. and J.L. conducted statistical analyses, interpreted the results, and drafted the manuscript. F.M.S., Y.S., and J.L. reviewed and edited the manuscript. All authors approved the final version. Funding The authors have nothing to report. Ethics Statement Ethical approval was obtained from the University for Development Studies Ethics Review Board (UDS/RB/date: 27/02/2023). Administrative permission was obtained from the Northern Regional Health Directorate, the Tamale Metropolitan Health Directorate, and all participating hospital administrations. Written informed consent was obtained from all participants before enrollment. Participation was voluntary; participants could withdraw or decline to answer any question without consequence. The study was conducted in accordance with the principles of the Declaration of Helsinki (2013 revision). Consent The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest. Transparency Statement The lead author, Joseph Lasong, affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained. Supporting information Supporting File 1 Supporting File 2 Supporting File 3 Supporting File 4 Supporting File 5 Acknowledgments We thank all participating hospitals for granting permission to conduct this study. We are grateful to all enumerators and study participants. We thank Mr. Torjim Salifu and Mr. Gordon Dandeebo for proofreading the manuscript. Written permission has been obtained from all named individuals for inclusion in this section. Salifu F. M., Lasong J., and Salifu Y., “Prevalence and Predictors of Intimate Partner Violence During Pregnancy in Northern Ghana: A Cross‐Sectional Study,” Health Science Reports 9 (2026): e72351. 10.1002/hsr2.72351. Fairuza Marian Salifu and Yula Salifu contributed equally to this study. Data Availability Statement All data generated or analyzed during this study are available upon reasonable request to the corresponding author (jlasong@uds.edu.gh). No publicly archived dataset was created or used. References Associated Data Supplementary Materials Supporting File 1 Supporting File 2 Supporting File 3 Supporting File 4 Supporting File 5 Data Availability Statement All data generated or analyzed during this study are available upon reasonable request to the corresponding author (jlasong@uds.edu.gh). No publicly archived dataset was created or used.