Health-related factors associated with intimate partner violence in women attending a primary care clinic in south-western Nigeria

Abstract

Background: Intimate partner violence (IPV) relates to poor physical and mental health in women. Women who have experienced it access healthcare facilities more frequently with symptoms that are often unrelated to violence. The objective of the study was to determine the prevalence of IPV and health-related factors in women in the study population.

Method: This clinic-based, cross-sectional study was carried out between November 2010 and January 2011 at the general outpatient clinic of the University College Hospital, Ibadan, Nigeria. Four hundred women aged 15 years and older with previous or current intimate partners were recruited by simple random sampling. Data were collected using an interviewer-administered questionnaire to obtain information on the respondents’ socio-demographic characteristics, lifetime experience of IPV and clinical data, i.e. presenting symptoms, body mass indices and blood pressure readings. The clinical data of women who had a lifetime experience of IPV were compared with those who did not.

Results: The overall lifetime prevalence of IPV was 89.2%. Reported IPV types were controlling behaviour (79.5%), psychological violence (35.7%), and physical (19.8%) and sexual violence (9.8%). Non-specific complaints (p = 0.028), female genital complaints (p = 0.037), poor sleep (p = 0.049), headaches (p = 0.011), abrasions or scars (p = 0.012), suicidal ideation (p = 0.001) and obesity (p = 0.021) were significantly more common in women who had experienced IPV than in those who had not. The logistic regression analysis showed suicidal ideation as the most significant factor relating to IPV (odds ratio 12.658, 95% confidence interval: 1.248–29.677).

Conclusion: The high prevalence of IPV in women who routinely present to this primary care clinic suggests that there is a need for IPV screening. The association of IPV with a spectrum of clinical factors signifies its ubiquitous occurrence and underscores the need for healthcare providers to have a high index of suspicion, especially in cases whereby women present with non-specific complaints. Suicide screening should also be performed on women found to be positive for IPV.

Keywords:

• domestic violence
• health effects
• intimate partner violence
• symptoms and signs

Introduction

Intimate partner violence (IPV) has been defined as a maladaptive behavioural pattern whereby an individual wilfully harms an intimate partner as a means of gaining or preserving power and control within the relationship.¹ It can be perpetrated by men against women or vice versa. The former is more common worldwide than the latter.² It can also be perpetrated among same-sex partners. The types of IPV perpetrated against women are usually different to those committed against men.³ In Nigeria, for example, women experience physical, psychological and sexual violence, including homicide, at the hands of men.⁴ Men also deny their wives leisure time and the ability to earn a living, accumulate wealth or access their children in cases of separation.^4,5

There is diversity with respect to IPV prevalence values across the world, mainly owing to the different ways in which IPV is measured in different cultures. The prevalence of IPV in a community-based study on Asian women in the USA by Raj and Silverman was 40%, while the prevalence reported by Bates et al. and Koenig et al. was 67% in Bangladesh and 52% in Uganda, respectively.⁶⁻⁸ Ameh and Abdul reported a prevalence of 28% in Zaria, northern Nigeria; Ilika et al. 46.3% in Nnewi, eastern Nigeria; and Okenwa et al. 47.7% in Lagos.⁹⁻¹¹ In Ibadan, Fawole et al. reported a prevalence of 31.3% in male and female civil servants, while Owoaje and Olaolorun reported prevalence to be 87% in women in a migrant community.^12,13

Violence perpetrated against women by their male partners, especially in Africa, is culturally sensitive and any attempt to expose it is often viewed as an intrusion into the “normal” family life of a couple.¹⁴ These socio-cultural perceptions, coupled with inadequate training, also affect the management of IPV cases by physicians who may not enquire about it.¹⁵

IPV has been reported to be associated with many physical, reproductive and psychological or mental health effects.^14,16−20 It was also found that women who experience IPV use medical services more frequently, and may even present with symptoms which are unrelated to the violence.²¹ Since these women do not usually initiate disclosure despite the presence of obvious signs, it is difficult for physicians to offer a diagnosis or help. Identification of certain symptoms or signs that are associated with IPV would help physicians to selectively screen for it. Prevention strategies could also be improved as increased insight into health problem could be gained through the experiences of identified victims.

There is no standard protocol for management in Africa, and especially in Nigeria, owing to the high level of tolerance and trivialisation of IPV. Victims are often blamed for the violence.¹³ The true proportion of women who suffer physical and mental disabilities due to partner violence is not known in Africa as a result of gross under-reporting.^14,21 However, in order to address the needs of this group of “voiceless” women, an attempt needs to be made to draw attention to the burden and the health-related problems which they face.

This study determined the prevalence of IPV in women presenting at an ambulatory first-contact clinic in south- western Nigeria. It also assessed common symptoms and signs associated with IPV in women. It is hoped that the findings of this study will provide a better understanding of IPV and inform the development of public health strategies to address IPV in Nigeria.

Method

Study population

A total of 2 407 women presented to the general outpatient clinic of the University College Hospital, Ibadan, Nigeria, between November 2010 and January 2011.

Sample population

The minimum sample size, n, was calculated using the Leslie-Kish formula for descriptive studies, n = z²pq/d², where z, the standard normal deviate, was set at 1.96, corresponding to a 95% confidence level and where p = 0.5 was used to achieve the highest possible sample size in the absence of a similar local study at the time, and d, the desired degree of accuracy, was set at 5%. This gave a sample size of 385, which was approximated upwards to 400, given a 5% attrition rate.

Sampling technique

Given the study duration of 60 clinic days in which to recruit a sample size of 400 respondents, seven random numbers were computer-generated daily by means of Microsoft^® Excel^® (Windows^® 7). Women who selected any of these numbers daily through a balloting process, and who fulfilled the inclusion criteria, were interviewed until 400 of them were recruited. Inclusion criteria were consenting women aged 15 years and older, with a current or previous intimate partner. Women who required emergency care, or who were pregnant or psychotic, were excluded.

Procedure

Data were gathered by the researcher and trained research assistants using a structured interviewer-administered questionnaire. Interviewers were trained according to World Health Organization (WHO) guidelines.²² The questionnaire was used to collect data on the respondents’ biodemographic characteristics, their experience of abuse, as well as their clinical data. The portion of the questionnaire used to assess the experience of abuse was adapted from the WHO multi-country Study on Women’s Health and Life Experiences, a cross-sectional study carried out in 15 different regions across 10 countries around the world.¹⁸ The WHO study instrument was developed by specialists in the field, and validated with good internal consistency (Cronbach’s alpha of at least 0.7 in all domains).²³ The WHO instrument measured psychological, physical and sexual violence using the specific items shown in Table 1. Controlling behaviour was also measured by asking women if their previous or current partners prevented them from seeing friends, restricted their contact with their family of birth, insisted on always knowing where they were, ignored or treated them indifferently, became angry if they spoke to other men, frequently accused them of being unfaithful or controlled their access to health care.

Table 1: ^*Items used to measure types of intimate partner violence¹⁸

Type of IPV measured	Items used to measure IPV
Psychological violence	He insulted you, or made you feel bad about yourself.
	He humiliated or belittled you in front of others.
	He intimidated or scared you on purpose, e.g. by shouting and smashing things.
	He threatened you with harm directly or indirectly by threatening to hurt someone about whom you cared.
Physical violence	He slapped you.
	He threw something at you that could hurt you.
	He pushed or shoved you.
	He hit you with a fist or something else that would hurt.
	He kicked, dragged or beat you up, or choked or burnt you on purpose.
	He threatened you with or actually used a gun, knife or other weapon against you.
Sexual violence	He physically forced you to have sexual intercourse with him against your will.
	You had sexual intercourse with him because you were afraid of what he might do.
	He forced you to do something sexual which you found to be degrading or humiliating.

IPV: intimate partner violence

^* : Adapted from the World Health Organization multi-country Study on Women’s Health and Life Experiences

The clinical data of women who had a lifetime experience of IPV (IPV Yes)²⁴ were compared with those who did not (IPV No). For the purposes of analysing the health-related effects of IPV, women who had experienced controlling behaviour alone were classified as “IPV No”.²⁵

The presenting complaints of the respondents were grouped using the World Organization of Family Doctors’ (WONCA’s) International Classification of Primary Care-Second edition (ICPC-2) to promote easy data manipulation and comparison with other studies.²⁶ This questionnaire was translated into Yoruba, the local language, field tested and translated back into the original language before being pre-tested in the general outpatient clinic of Our Lady of Apostle, Oluyoro, a similar general outpatient clinic, to validate it within the study environment. Each interview lasted approximately 30 minutes.

Ethical considerations

The study was approved by the University of Ibadan/University College Hospital Ethics Committee (Assigned Number UI/EC/10/0062). The approval of the Head of Family Medicine Department, University College Hospital, was also obtained before the study was carried out. Informed consent was obtained from respondents before the interview. The guidelines on ethical recommendations for violence studies were followed.²⁷

Follow-up

Respondents were managed for their health complaints, and women who had a positive history of any form of violence were referred to a clinical psychologist in the outpatient clinic for psychotherapy. The clinical psychologist had been previously sensitised about the study. Referral to other specialists was carried out, where necessary, using the standard protocol of referral in place in the general outpatient clinic. A non-governmental organisation involved with the care and help of women who are exposed to violence was also contacted and informed about the study, and women who consented to receive further help in this regard were given the phone number of the contact person in the organisation.

Data analysis

At the end of each study day, the administered questionnaires were checked, sorted and coded serially. SSPS^® (version 17) software was used for data entering, cleaning and analysis. Frequency tables were generated for the socio-demographic characteristics, IPV experience and health-related factors of the respondents. Chi-square statistics was used to assess an association between the categorical variables. Significance was set at p ≤ 0.05. Logistic regression analysis was carried out on the variables that were significant with IPV.

Results

Table 2 shows the bio-data of the respondents. Only consenting patients were recruited until the sample size of 400 patients was attained.

Table 2: Respondents’ biodata

Biodata	n = 400
	n (%)
Age group (years)
20–29	49 (12.2)
30–39	139 (34.8^*)
40–49	94 (23.5)
50 and older	118 (29.5)
Ethnicity
Hausa	17 (4.3)
Igbo	32 (8.0)
Yoruba	340 (85.2^*)
Others	10 (2.5)
Relationship type
Monogamous	236 (59.0^*)
Polygamous	145 (36.3)
Engaged or cohabiting	19 (4.7)
Religion
Christian	222 (55.5^*)
Muslim	174 (43.5)
Traditional	4 (1.0)
Educational level
None	71 (17.8)
Primary	95 (23.8)
Junior secondary	19 (4.8)
Senior secondary	101 (25.2^*)
Post secondary	114 (28.4^*)
Income level
Below N 6 000	275 (68.8^*)
N 6 000 and above	125 (31.2)
Occupational class
Class I	37 (9.2)
Class II	35 (8.8)
Class III	66 (6.5)
Class IV	12 (3.0)
Class V	220 (55.0^*)
Class VI	30 (7.5)

N: Nigerian naira

^* : Bold items signify statistical significance (p < 0.05)

Twenty-eight respondents could not complete the interview either due to personal time constraints or the interviewer omitted important parts of the questionnaire. Response rate was thus 93%.

The prevalence of intimate partner violence with respect to the respondents, by type, is highlighted in Figure 1.

Figure 1: Intimate partner violence by type and proportion

IPV: Intimate partner violence *: includes multiple responses

One hundred and seventy women (42.5%) had a lifetime experience of one or more forms of psychological, physical or sexual IPV, i.e. “IPV Yes”, while 230 women (57.8%) had not experienced any psychological, physical or sexual IPV, i.e. “IPV No”. For the purposes of analysing the health-related effects of IPV, women who had experienced controlling behaviour alone were classified as “IPV No”.

Table 3 shows the association of respondents’ symptoms with their experience of IPV.

Table 3: Association of respondents’ symptoms with intimate partner violence

ICPC-2 class of symptoms	“IPV Yes”	“IPV No”	χ^2	p
	n = 170	n = 230
General or unspecified	41 (24.1)	50 (21.8)	0.315	0.028^*
Female genital	21 (12.4)	23 (10.0)	0.553	0.037^*
Digestive	19 (11.2)	39 (17.0)	2.634	0.048^*
Poor sleep (psychological)	10 (5.9)	6 (2.6)	3.724	0.049^*
Musculoskeletal	22 (12.7)	38 (16.6)	0.983	0.049^*
Headaches (neurological)	12 (7.1)	15 (6.6)	0.045	0.011^*
Urological	03 (1.8)	2 (0.9)	0.635	0.361
Cardiovascular	03 (1.8)	1 (0.4)	1.746	0.208
Scars or abrasions	07 (4.1)	1 (0.4)	6.764	0.012^*
Eye complaints	19 (11.2)	31 (13.5)	0.474	0.049^*
Ear complaints	3 (1.8)	4 (1.7)	0.000	0.635
Respiratory	6 (3.5)	9 (3.9)	0.400	0.841
Pregnancy related	3 (1.8)	10 (4.4)	2.074	0.136
Endocrine	1 (0.6)	1 (0.4)	0.046	0.831

ICPC-2: International Classification of Primary Care-2, IPV: intimate partner violence, χ²: chi-square value

^* : Bold items signify statistical significance (p < 0.05)

A significantly higher proportion of women who had experienced one or more forms of psychological, physical or sexual IPV reported suicidal ideation, as portrayed in Table 4.

Table 4: Suicidal ideation in the respondents

Suicidal ideation	“IPV Yes”	“IPV No”	χ^2	p
	n = 170	n = 230
	n (%)	n (%)
Present	21 (12.4)	08 (3.5)	11.449	0.001^*
Absent	149 (87.6)	222 (96.5)

IPV: intimate partner violence, χ²: chi-square

^* : Bold items signify statistical significance (p < 0.05)

Furthermore, a significantly higher proportion of women who had experienced one or more forms of psychological, physical and sexual IPV were obese (Figure 2).

Figure 2: Association of intimate partner violence with respondents’ body mass indices

BMI: body mass index, IPV: intimate partner violence,*: p = 0.039, chi-square = 4.253

Although a higher proportion of women who had ever had IPV also had elevated blood pressure, this difference was not statistically significant, as indicated in Figure 3.

Figure 3: Association of intimate partner violence with respondents’ blood pressure

BP: blood pressure, *: p = 0.66, chi-square = 0.194

Further analysis by logistic regression of the significant variables with IPV was performed (Table 5). This revealed that suicidal ideation was the most significant factor to result from IPV (odds ratio 12.658, 95% confidence interval: 1.248–29.677).

Table 5: Logistic regression of analysis of the significant variables with respect to intimate partner violence

Significant variables	β	p	OR	95% CI for OR
Non-specific complaints	1.023	0.300	2.782	0.402–19.246
Female genital complaints	−0.600	0.958	0.942	0.100–8.857
Poor sleep (psychological)	0.749	0.505	2.115	0.234–19.144
Headaches (neurological)	2.545	0.102	12.749	0.602–27.018
Abrasions or scars	0.502	0.745	1.652	0.081–33.915
Suicidal ideation	2.543	0.032^*	12.658	1.248–29.677
Obesity	0.909	0.088	2.482	0.874–7.051
Constant	−0.997	0.464	0.369

CI: confidence interval, OR: odds ratio

* : Significant association was set at the 5% level of significance

Discussion

The finding of a high prevalence of IPV (89.2%) in the respondents mirrors the findings of Owoaje and Olaolorun, who also studied women in south-western Nigeria, and who found a high prevalence thereof (87%). However, Owoaje and Olaolorun studied immigrant women who were mostly Hausas or Fulanis and Muslims.¹³ Our study population was predominantly Yorubas and Christians. Therefore, it may be implied that ethnicity and religion do not affect a woman’s experience of IPV, probably because many African cultures are inherently patriarchal, and often a woman’s position in most African countries is to be subservient to men. It is also noteworthy that Owoaje and Olaolorun’s study was community based, while this study was clinic based, and yet the prevalence of IPV was similarly high in both. Since community-based studies are a better indication of the true situation in a general population, this study showed that the random selection of respondents from a general outpatient clinic, where a fairly diverse number and type of women were present, added credence to the results of this clinic-based study. Hence, the results can be generalised in that population.

The high level of tolerance for IPV in Nigerian, as well as African, culture, may account for the high prevalence of IPV in our study. Tolerance of IPV is a reflection of the cultural norm. For example, the Tiv people of Nigeria regard wife beating as a sign of love, and it is socially accepted and sometimes even encouraged by the women.²⁸

This study demonstrates a wide variation in the prevalence of IPV with inclusion or exclusion of controlling behaviour in the classification of IPV. It has been suggested that a wide variation is indicative of the fact that such behaviour may be normative in some cultures, including the patriarchal community in which this study was conducted.¹⁸ In fact, when some of the respondents were asked if their partner insisted on knowing where they were at all times (a measure of controlling behaviour), they promptly responded that they would not even go out without first informing their partner. Therefore, this makes it difficult to classify “controlling behaviour” and the items used to measure it as a reliable indicator of IPV in patriarchal communities. This is because “controlling behaviour” was accepted or even welcomed by the women. Thus, it may not produce the negative and uncomfortable effects associated with “violence”. However, it is difficult to differentiate between this and a situation in which a woman has accepted her fate as “normal” although she neither likes nor enjoys it. A similar difficulty was encountered in a study conducted in China, which is a patriarchal country like Nigeria, where controlling behaviour was excluded from the analysis because it interfered with the definition of physical and sexual IPV such that it was unclear if it was an independent predictor of violence or a component of IPV.²⁵ If controlling behaviour was excluded as a component of IPV in this study, the overall lifetime point prevalence of IPV would reduce to 42.5%. This underscores the importance of appropriately defining controlling behaviour in IPV prevalence studies. Further research should also be carried out to assess cultural variations that may exist with respect to the definition of controlling behaviour in academic research.

Slightly over a third of the women in this study had experienced psychological violence, which is nearly two-and-a-half times the 14% among women who attended family practice clinics in the USA, as reported by Coker et al. The higher value in our study may be owing to the differences in the way in which the psychological violence was measured in the two studies. Coker et al. assessed psychological violence as a function of a feeling of loss of power and control, as well as entrapment in the relationship.¹⁷ Power and control over women in a predominantly patriarchal society, like Nigeria, may not necessarily denote much as the women willingly relinquish it to their men, and the loss of it may not translate to a feeling of violation. Therefore, low values, similar to those reported in the study by Coker et al, may have been described if the criteria used by the latter were used in our study.

The value of 35.7% reported for psychological violence in our study is also higher than the 23% reported by Okenwa and Lawoko with respect to women attending obstetrics and gynaecology clinics in Lagos, south west of Nigeria.¹¹ The percentage might have been affected by a high proportion of pregnant women in their study, whereas pregnant women were excluded from our study. It is possible that pregnancy may be protective against physical and psychological violence because a woman may be favoured and treated warmly once pregnant, especially in cultures in which a woman is defined by her reproductive capabilities.²⁹ The high prevalence of psychological IPV in our study underscores the need for physicians to have a low threshold for screening, as psychological violence may not be as visible as physical or sexual violence against women.

One in every five women in our study experienced physical IPV. This is approximately half the 36% reported by Ameh and Abdul in Zaria, northern Nigeria; and may be explained by cultural differences in different regions of the same country.⁹ For example, polygamy and purdah are customary for Muslims in northern Nigeria. These practices often render the women economically dependent on their husbands, a situation that promotes violence. However, in a study conducted in Eastern Nigeria, Ilika et al. reported the prevalence of physical violence in women of childbearing age to be 15.8%. This lower value may be explained by the fact that beating pregnant women is culturally unacceptable in the Eastern part of the country where the study by Ilika et al. was conducted, although the proportion of pregnant women in that study was not reported.¹⁰ On the other hand, our study specifically excluded pregnant women. It should be noted that in the study by Ameh and Abdul, the respondents were all pregnant, and yet physical violence was still high at 36%.⁹ Therefore, culture and the mentality of people play a strong role in the perpetration and acceptance of violence against women.

The prevalence of sexual violence in our study was roughly 10%. According to the WHO multi-country Study on Women’s Health and Life Experiences, a wide range of values were reported in the different regions, i.e. 4%, 46% and 56%, in Japan, provincial Bangladesh and Ethiopia, respectively.¹⁸ This wide range could be explained by the subtlety in the definition of sexual violence within an intimate relationship. For example, a woman may not be willing to have sex with her partner, but may consent because of fear of rejection. In other instances, consenting to a relationship is taken to mean agreeing to fulfil her partner’s sexual desires at all times, irrespective of her feelings. This was often the case in this study. Therefore, sexual violence in an intimate relationship is scarcely reported, hence its “low prevalence.”

Analysis of the presenting complaints by respondents showed that general and non-specific symptoms were significantly more common in IPV victims than in non-victims in this study. This is similar to reports by Guth and Patcher that victims may present with symptoms that do not relate to IPV.¹⁹ It is also similar to findings by other researchers that somatic symptoms, which are usually vague and non-specific, relate to IPV.^30,31 However, it is important to note that very commonly, vague symptoms are presented to the family physician as an undifferentiated disease. Therefore, general and non-specific symptoms may not directly relate to IPV as the logistic regression analysis did not show a significant association of non-specific symptoms with IPV in this study.

As classified by WONCA’s ICPC-2,²⁶ female genital complaints such as breast pain, vaginal discharge, vulva pruritus and vaginal pain, as well as symptoms relating to menstrual periods, were significantly more common in women who had experienced IPV than in those who had not. This is similar to the findings by Campbell who concluded that gynaecological complaints were the most consistent and long-lasting finding in women who had experienced IPV.¹⁴ Coker et al. also found that pelvic pain was a significant physical symptom in these women.¹⁷ This may be explained by the direct mechanisms of sexual violence which may predispose them to pelvic inflammatory disease, vaginal discharge, pruritus and pain. However, further research needs to be carried out in this regard to assess other psychosocial mechanisms through which gynaecological symptoms may be associated with sexual IPV, especially in cultures in which the women are defined by their reproductive capabilities. This is because such women may psychosomatically express their violation through genital complaints.

However, digestive system complaints, such as generalised abdominal pain and epigastric pain, were significantly more common in women who had not experienced IPV than in those who had experienced it. This is in contrast to reports by Heise et al. that abdominal pain and other digestive system problems, like irritable bowel syndrome, were associated with IPV.³² It is expected that complaints such as epigastric pain, which may be suggestive of acid reflux disease, may be more common in women who have experienced IPV, owing to the stress relating to exposure to violence. However, this contrasting association was not confirmed by multivariate analysis in this study.

In this study, we found that musculoskeletal pain, such as limb and back pain, were significantly more common in women who had not experienced IPV. This is in contrast to the findings by Obi and Ozumba, who reported that 91.4% of women who were physically abused by their male respondents complained of muscle aches and pains.³³ The significant finding in our study, of musculoskeletal complaints being reported more by women who had not experienced IPV than by those who had, contradicts expectations that abuse in any form relates to musculoskeletal pain. This may be due to the fact that pain is a very common reason for presentation to a primary care physician.³⁴

This study identified a strong significant association between IPV and suicidal ideation. This is similar to reports by other authors.^16,35 Logistic regression analyses have demonstrated that women who experience any one of the forms of psychological, physical or sexual violence are 13 times more likely to experience suicidal ideation. Given the seriousness of suicidal ideation as a risk for actual suicide, a standard protocol for the suicidal ideation screening of women who have a lifetime experience of IPV is required.

It was also demonstrated in this study that psychological symptoms (poor sleep mostly) were significantly more common in women who had experienced IPV. However, this relationship was not confirmed by logistic regression analysis. The paradoxical nature of the simultaneous presence of love and violence/hate may explain the psychological response of women exposed to IPV.^16,35

This study showed that obesity was significantly more common in women who had experienced one or more of the forms of psychological, physical and sexual IPV. This is similar to the findings in a study performed in Ethiopia where although obesity was endemic, it was found to be significantly associated with partner violence.³⁶ The identification of a relationship between obesity and IPV in this study could link to an association of IPV with psychological symptoms that affect a woman’s eating habits, activity level and general healthcare practices. IPV has been found to be associated with higher levels of depression, anxiety, sleep and appetite problems.¹⁶ Could this lead to obesity in some women? Or could it be that because women become unattractive to their husbands when they become obese, extramarital affairs result and violence ensues? There is the remote possibility that obesity may be both a cause and a consequence of intimate partner violence. However, this association was not confirmed by logistic regression analysis of the significant variables in this study. Hence, future research should be carried out to examine the possibilities suggested by this study.

A significant association between IPV and hypertension was not demonstrated in this study, contrary to reports by Campbell that cardiac problems such as hypertension and chest pain are associated with IPV.¹⁴ The contrary finding in this study may be owing to the fact that many other factors, such as genetic predisposition, family history or salt intake might predispose to hypertension, and were not taken into account.

Conclusion

The high prevalence of IPV in women who did not specifically present on that account highlights the importance of screening for this often-hidden phenomenon. The fact that IPV may be associated with specific symptom categories suggests that women who present accordingly should also be screened for IPV. Given that women who are exposed to IPV are 13 times more likely to have suicidal ideation as seen in this study, standard protocols of routine screening for suicidal thoughts or attempts should be done once a woman is found to have IPV exposure.

Study strengths

Considering the paucity of local literature on this subject, this study adds to the body of knowledge by describing some of the presenting symptoms and health-related factors that may be associated with intimate partner violence in south-west Nigeria. Although this was a cross-sectional study, its comparative nature adds credence to the results. In this study, a standardised questionnaire was used and administered in private. The interviewers were also carefully trained according to the WHO guidelines. It has been demonstrated that these factors help to improve disclosure of a sensitive issue, like IPV.^37,38 Thus, this was a further strength of our study. Lastly, translation of the questionnaire into the local language, as well as field testing and its translation back into the original language, improved its validity.

Study limitations

This was a cross-sectional study. Therefore, conclusions cannot be drawn on the causal relationships of the observed associations. It should not be overlooked that as it was a clinic-based study, the women who experienced IPV were more likely to seek care. The timing of the violence was not taken into consideration, yet the health effects of IPV may be further influenced by its duration as its effects have been found to be “dose dependent”.¹⁴ Furthermore, recall bias may have existed as a time limit was not imposed on how far back in the past the IPV occurred. Also, information bias may have existed, given the sensitive and intimate nature of IPV, and especially sexual IPV.

The details of the intimate relationships and the characteristics of the intimate partners were not analysed in our study. This may have affected the way in which IPV was perceived, and had an influence on the responses of the women experiencing it.

Conflict of interest

The authors declare that there was no conflict of interest that might have inappropriately influenced this study.

Acknowledgements

This study used the WHO Violence Against Women Instrument as developed for use in the WHO Multi-Country Study on Women’s Health and Domestic Violence and adhered to the WHO ethical guidelines for the conduct of violence against women research. Additionally it incorporates questions from the WHO study questionnaire used to measure Intimate partner violence. The authors acknowledge Dr. Garcia Moreno Claudia and Dr. Christine Pallitto for sending these online materials for our use in this study. We further wish to thank the Head of Department of the Family Medicine Department (formerly General outpatient department, University College Hospital, Ibadan, Nigeria, the body of consultants and residents for helping to make this possible. I wish to acknowledge Dr. Garcia Moreno Claudia and Dr. Christina Pallito PhD, of the WHO for sending me on line materials, copies of those used in the WHO Multi-country study on women’s health and life experiences without which the work would have been impossible. I also wish to thank the Head of Department of the Family Medicine Department (formerly General Outpatient Dept) of the University College Hospital, Ibadan, Nigeria, the body of Consultants and residents for helping to make this possible.