Controlling behavior manifests as part of patriarchal dominance which can be present along with physical and sexual violence within a society [1]. Controlling behavior is often manifested as monitoring, coercing or threatening the partner. It is experienced by 5.2% to 56.6% women globally and often results in their physical and mental ill health. Controlling behavior has a higher prevalence and lower severity as compared to exposure to physical and sexual violence [2]. Partner violence is defined as a behavior within an intimate relationship that causes physical and/or psychological harm, including acts of physical aggression, sexual coercion and psychological abuse [3]. The focus of this study is physical and sexual violence experienced by women and perpetrated by her domestic partner [4].

Presence of controlling behaviors are found to be simultaneously associated with the presence of physical and sexual violence, which can increase the stress of the woman who was experiencing it [2,5]. Controlling behavior could be manifested in different ways i.e. anger, restricting movement, restricting communication with family, limiting decision-making ability about health and promoting economic dependence. In developed countries, the prevalence of physical Intimate Partner Violence (IPV) ranges from 7% to 57% while the prevalence of sexual IPV ranges from 2.5% to 24.7% [4,6]. The prevalence of controlling behavior ranges from 5.2% to 67% depending on the population studied [6].

The risk of IPV is greater for women compared to men [7]. In low and middle income countries lifetime prevalence of physical abuse is between 13% to 61% while sexual violence by a partner ranges from 6% to 58% [2]. Pakistan has lifetime prevalence for physical and sexual IPV of 57.6%, and 54.5% respectively [8]. IPV if not controlled in the initial stages can result in extreme consequences like intimate partner homicide and suicide [9].

Researchers have given significant importance to psychological factors to explain the causal framework of IPV [10]. Positive and negative representation of self and the partner, low self-esteem and anger are factors that affect the type of relationship and thus contribute to the occurrence of IPV [11]. Associations between anti-social personality disorder and IPV have also been reported [12]. Recently, two theories (i.e. crime spin theory and victim-perpetrator interactive spin theory) have been proposed to explain the psychopathology of those who commit IPV [13,14]. IPV is associated with a number of psychiatric morbidities including depression, postpartum depression, post-traumatic stress disorder, eating disorders, and suicidality [15-20]. IPV is associated with physical health problems including neurological, gynecological, and functional gastrointestinal disorders in addition to the risk of injury from physical abuse [20,21].

Gender bias is commonly seen in Pakistan and other south Asian countries due to cultural Taboos [22]. Pakistan is a patriarchal society where male partner wants an ideal wife. This concept is embedded in society; women who are not considered ideal are vulnerable to controlling behaviors and IPV [23]. No study has been done to find out the prevalence of controlling behavior in Pakistan. Therefore, this study was conducted to find out the prevalence of physical or sexual abuse and different types of controlling behaviors in the context of Pakistani society. Moreover, the study also explored possible associations between these controlling behaviors and pain or discomfort, sadness or depression and suicidal thoughts. The study further explored whether three outcome variables i.e. pain or discomfort, sadness or depression and suicidal thoughts are associated with both controlling behavior and IPV together.

A cross-sectional study was conducted in Karachi city of Sindh Province, Pakistan. Karachi has a population of 16 million, divided into 18 towns. In these study 759 married women, aged 25-60 years, 27% living in two of the towns with approximately 720,000 inhabitants were included. This data was collected by community-based midwives and the response rate of 93.70%.

Women who live in lower and middle income areas of Karachi have limited movement due to cultural taboos (24, 25) and they mostly stay at home. In order to collect data it was necessary to get linked to the local NGOs whose workers are known to the community. We got support of Health and Nutrition Development Society (HANDS) for getting access to the community. HANDS is a Non-Governmental Organization (NGO) that works closely with the government health services. HANDS provided basic healthcare education and income generating opportunity to empower people in low and middle income areas of Karachi and have a workforce of skilled people who take responsibility of local health care services at primary level (maternal, child health and immunization). These services include Oral rehydration therapy, control of diarrheal disease, growth monitoring, nutrition counseling and treatment of minor illnesses. HANDS only provides services to lower and middle income families, therefore the data was collected only from women in these socioeconomic classes.

Another reason for involving NGO in data collection was that public sector is reluctant to work in areas of violence against women due to fragile law and order conditions in some areas of Karachi. Since this NGO was only catering the services to the lower and middle socioeconomic groups, that’s why in this study the upper SES women could not be enrolled.

Data collection instrument

WHO has developed a data collection instrument called multicountry study on women’s health and life experience. It comprises of questionnaires for studies on interpersonal violence and is named Violence Against Women (VAW) [26] and is used in more than 15 countries. It has a variety of abuse questions and other abuse assessment scales Index of Spouse Abuse (ISA) and Conflict Tactic Scale (CTS) with confirmed reliability and validity [27,28]. The final questionnaire of this study contained items addressing socio-demographic and psychosocial factors, different forms of violence, and the health effects of the violence inflicted. Information on controlling behavior was elicited by asking questions about the husband’s role in the woman’s decisions making related to her general and reproductive health, her contact with friends, husband’s faith in his wife and money which wife gets from her husband.

Validity and reliability of the VAW instrument

The reliability of this instrument VAW was tested by WHO teams in ten different countries. Cronbach’s alphas for physical abuse was 0.81 and sexual was 0.66 [29]. Cronbach’s alpha in a study in Brazil was 0.78 for psychological, 0.83 for physical and 0.78 sexual violence construct. Face validity study was done by pre-testing the questionnaire at similar field level and content validity by a psychologist, sociologist, and public health specialist. Cronbach’s alpha for physical, sexual, and psychological violence sub-scale was 0.87, 0.79, and 0.93 respectively [30].

Study setting

Pakistan is a low income Muslim country with a population of 176 million; out of which Karachi has about 16 million inhabitants. Karachi is a city with multiethnic societies which includes Mohajirs, Sindhis, Pushtoons, Baluch and Punjabis.

Karachi is divided into 4 districts; each district is sub-divided into towns and again divided into union councils UCs (87-88). There are 18 towns and 178 UCs in Karachi. The data collected from two towns of lower and middle socio-economic strata.

Sample size

To detect a 1.6 fold risk increase of physical/sexual psychological violence and abuse with 80% probability and estimated exposure ranging from 20-40% in the study sample, a sample size of about 664 individuals was required. It was decided to aim for 800 respondents eventually 759 women were included in the study and response rate was 93.7%.

Data collection procedure

HANDS employed Community Midwives (CMs) who collected data from March to August 2008 using multistage random sampling technique in selected area. A list of registered families was developed and six field sites were selected. Each site was monitored by surveillance system set up by (CMs). Required number of households were randomly selected/ (using computer generated numbers in EPI info) where women of required ages were available. Ten women refused participation at the initial stage and were replaced by women of the same age. 41 women decided to discontinue in the middle of study but they were not replaced. Either youngest or oldest woman was selected alternatively from house where more than one eligible woman was present. Information related to current husbands was collected. Six CMs were trained for a week in the national language. Training was conducted by local researchers and members of the Pakistan Women’s Lawyers Association (PAWLA; an NGO supporting women’s rights) and HANDS [31]. The training included information on the causes of IPV, management techniques, referral resources, study rationale, and ethical considerations. The study was presented as a women health study to household members until conversation was safe from being overheard, as the topic of violence was sensitive. The interviews were done in respondent’s home in the national language (Urdu) and privacy was ensured. To ensure quality of data about 5% of participants were re-interviewed by random selection and only minor differences were detected in response.

Socio-demographic factors and prevalence of different types of IPV

Out of 759 women included in this study 536 (70.6%) had no education. Husbands of 528 (69%) women were unskilled and unemployed. Majority of the women i.e. 517 (68.1%) in our sample were from medium and high socio-economic status within the study subjects. Shoulder and neck pain was experienced by 314 (41.4%) women. Suicidal thoughts were experienced by 313 (41.2%) women while the feeling of worthlessness was found among 438 (57.7%) women.

The prevalence of past-year physical and sexual violence was 68% (n = 516) while controlling behavior was experienced by women was 51.6% (n = 392) (Table 1)

Table 1: Socio – demographic and psychosocial factors and health effects N= 759.

Among the different types of controlling behavior refusal of the husband to give money for the household chores had the highest prevalence (Table 2).

Table 2: The controlling behavior scale N= 759.

Association between violence and its physical and psychological consequences

Bi-variate analysis showed that there was a significant association between women who had a past-year physical or sexual violence experience with shoulder and neck pain (OR = 3.04; 95% confidence interval: 2.21 -4.16), feeling worthlessness (OR=1.99; 1.44 – 2.74) and suicidal thoughts (OR = 5.73; 4.11 – 7.99). Similarly controlling behavior experience from current husband was also significantly associated with all the three adverse consequences and association was strongest for suicidal thoughts (OR = 5.77; 4.20 – 7.93) (Table 3).

Table 3: Associations between socio eco demographic and psychological factors and health conditions N= 759.

Bi-variate analysis to check the association between the three health consequences and other factors showed that suicidal thoughts and shoulder and neck pain had significant association with respondents’ age, women education, husband’s age, husband’s education, number of children, number of family members and socio-economic status. Whether the women chose her husband or not, were the factors associated with shoulder and neck pain but not with suicidal thoughts. Factors associated with feeling worthlessness at Bi-variate level included respondents’ education, husband’s occupation and socioeconomic status.

An important objective of this study was to find out whether women who had exposure both variables i.e. past-year physical or sexual violence and experience and controlling behavior experience from current husband had significant association with the three mentioned health outcomes. For this purpose a crude and adjusted odds ratio were computed for the four levels of this variable as shown in Table 4. The association was strongest between controlling behaviors and subjected to past year physical/sexual violence (OR=15.08; 9.70 – 23.43) and it remained significant when adjusted for important covariates which were significant at the Bi-variate level.

Table 4: Interaction of controlling behavior on the association between past-year physical/ sexual violence and health conditions n= 759.

Discussion identified

This study found high prevalence of controlling behaviors, and physical and sexual violence, which is in accordance with many studies conducted at the national and international level [2,32]. Furthermore, majority of women in this study were found to be uneducated out of which most of them were housewives dependent on their husband’s income. Since most of the husbands were unskilled workers and were earning limited amount of money. It has been observed that if husband’s income is limited, the wife is given very limited amount to fulfill the household’s needs. When housewives ask for more money from their husbands, they become victim of violence.

In this study SES index has been used for analysis purpose. Participants of study were rated as high and middle SES, but when we compared them to the general Pakistani population, they would fall under the low or middle classes. This study has identified that women with lifetime exposure to IPV and controlling behavior had higher reported rates of neck and shoulder pain, worthlessness, and suicidal thoughts in the past year, which may leads to the development of ill physical and emotional health effects, similar to findings from other studies [33,34]. A Pakistani study reported 6168 suicidal cases over a period of three years (2007-2009), however, it is likely that the number of attempts and the number of individuals with suicidal thoughts is much higher [35]. This is the indicator of the presence of depression within the population. According to a study the prevalence of depression is 40.6%, with the higher proportion in women than in men [36]. In our study the high proportion of having feeling worthlessness and suicidal thoughts are also signifies the presence of depression, which is a serious concern for the health professionals. This study has also identified that the persistent presence of both IPV and controlling behaviors may result in severe depression. Moreover, persistent presence of feeling of worthlessness and suicidal thoughts reflects state of stress [37]. Since stress response is a result of chronic exposure to IPV, it might result in an imbalance of cortisol level and other inflammatory mediators which might result in chronic pain [38].

High percentage of worthlessness and suicidal thoughts in the IPV women show that these women are at high risk of severe depression, which may lead to suicidal tendency and ultimately serious suicidal attempts. This finding is supported by a meta-analysis done to explore this relationship [15]. Impaired autonomy, rejection, self-sacrifice and economic subjugation as a result of IPV may lead to depression [39]. IPV and controlling can initiate different types of negative cognitions related to self and world, therefore lead to depression and suicidal tendency. The concept of ideal wife when combined with self-centered approaches of “I can’’ and “I must’’ leads to IPV. Controlling behavior when present along with IPV has a stronger association with shoulder pain, feeling of worthlessness and suicidal thoughts.

In Pakistan controlling behavior is considered as part of the culture and females are instructed to follow men’s order [40]. Women are less empowered for taking the major decisions of their lives like decision of marriage moreover they are not allowed to break the relation with their husbands even when there is proof of IPV [41]. Though there are laws in Pakistan to protect women rights but there is a lack of seriousness to accept and adopt these laws in different ethnic group [42]. In this situation a married woman who tolerates IPV for her life time with little or no resistance as she has no hope to improve her condition [43]. Marital bonding being a religious concept binds the male partner to fulfill his duties towards the female partner [44]. However, religion is presented as an argument to defend IPV though none of the formal religious decree has allowed any of such activity. It will be adequate to say that IPV is a cultural phenomenon which has been opposed by important religious stakeholders. There are explicitly defined laws and boundaries quite well set in different religions which safeguard rights of the women [45]. In Pakistani society where cultural phenomena like IPV is defended by presenting it as a part of religion, religious argument against IPV can be considered as an intervention along with all the established approaches. Further studies are needed to explore the religious misconceptions responsible for this IPV.

The strength of this study is that the data was collected by community midwives quite well verse with the local culture, are from same society and aware of different types of IPVs against women. The questionnaire of this study is widely used in different studies in different parts of the world. It was piloted and used after necessary modification. Expert opinion was used to do the content analysis of this questionnaire. This study had some limitations. As source information to record violence was women themselves which might have resulted in over-reporting of violence. Similarly, we did not explore what type and aspect of a violent act is associated with psychiatric morbidity. Another limitation of this study was that people from higher socio-economic class were not included in this study. This might have resulted in over estimation of prevalence since IPV is known to be directly associated with income.

This study identified those women in assessed (middle and lower) socio-economical groups are exposed to physical and sexual violence. Furthermore, we found that other than exposure to violence women are facing controlling behaviors from their partner which restrict their decisions particularly related to their health. The stress induced as a result of controlling behavior results in psychiatric illness. In clinical practice psychiatrists needs to be vigilant about these morbidities among women experiencing IPV. There is a need to confirm the findings of this study in sub-groups of different population. Qualitative inquiries are needed to identify thought constructs of the perpetrators of partner’s violence to design an intervention accordingly. Religious arguments should be made part of the interventions against IPV.

We proposed more studies and more variables to be explored for the future studies in the area of the outcome such as injuries and its related outcome (e.g., loss of a child, loss of ability to become pregnant, etc.). Furthermore, there is a need to measure about suicidal ideation, level of ideation and attempts. Additionally, there are a number of psychological consequences beyond worthlessness such as PTSD and other anxiety disorders, somatization disorders, perhaps personality pathology, etc. which could be measure in a separate study.