Prevalence of Emotional and Physical Intimate Partner Violence Among Married Women in Pakistan

Abstract

Intimate partner violence (IPV) affects millions of women across the world, and Pakistan is no exception. However, there is wide variance in reported frequencies worldwide and even within the same population. No standardized representative national and subnational estimates were available for IPV in Pakistan. We analyzed 3,666 ever-married women (15-49 years), from the most recent Pakistan Demographic and Health Survey (PDHS; 2012-2013), to determine the prevalence of emotional and physical violence among women in Pakistan and its major subpopulations. The Conflicts Tactics Scale (CTS) was used to ascertain violence. Furthermore, we examined the socioeconomic and demographic characteristics, which have the potential to increase the risk of emotional and physical violence. Weighted multivariate logistic regression was carried out to determine the association of IPV with the potential risk factors using Stata version 13.0. The prevalence of emotional and physical violence was, respectively, 36.4% (95% confidence interval [CI]: [33.8, 39.1]) and 18.4% (95% CI: [16.4, 20.6]) in Pakistan. The frequency of emotional and physical violence was highest in Khyber Pakhtunkhwa (KPK; 54.9% and 36.4%) followed by Balochistan (50.0% and 25.5%), Punjab (35.9% and 15.8%), and Sindh (24.7% and 13.3%) provinces. The prevalence of any form of violence (emotional or physical or both) was higher in rural than in urban areas (45.2% vs. 30.6%). Higher age of the husband and lower socioeconomic status were associated with emotional violence in KPK. Furthermore, alcohol intake by husbands and lower educational status of women or their husbands were associated with greater risk of emotional and physical violence across the provinces. IPV is high in Pakistan and large variations exist in the prevalence of violence across the provinces. The predictors of emotional and physical violence included occupation and education of the husband, alcohol intake, woman’s occupation, and wealth index. Further studies are needed to explore the underlying factors of violence and reasons for variation across different regions in the country for contextual interventions.

Keywords

intimate partner violence Pakistan prevalence

Introduction

Intimate partner violence (IPV) is “any physical, sexual or psychological harm by a current or former partner or spouse” (Centers for Disease Control and Prevention, U.S. Department of Health and Health Services, 2017). IPV is the most common form of violence against women globally (UN Women, 2016). IPV could manifest as “the use of verbal and non-verbal communication with the intent to harm another person mentally or emotionally and/or exert control over another person,” considered as “emotional violence,” and as “the intentional use of physical force with the potential for causing death, disability, injury, or harm,” considered as “physical violence” (Breiding, Basile, Smith, Black, & Mahendra, 2015). Operationally, physical violence includes scratching, pushing, shoving, throwing, grabbing, slapping, biting, choking, shaking, hair pulling, punching, hitting, burning, use of a weapon (gun, knife, or other objects), and use of restraints or one’s body, size, or strength against another person (Breiding et al., 2015). Thus, IPV includes a broad sphere of acts and actions. Variable definition has been used for IPV across studies; many studies have used a more sensitive and broad definition and, therefore, have been criticized to be overestimating the burden of IPV (Nicolaidis & Paranjape, 2009).

One in every three (35%) women, globally, was reported to be a victim of IPV at least once during her lifetime (World Health Organization, 2017). About 38% of female homicides were committed by a male intimate partner (World Health Organization, 2017). The frequency of IPV varies, ranging from 23.2% in high-income countries and 24.6% in the Western Pacific region to 37% in the Eastern Mediterranean region and 37.7% in the South-East Asia region (World Health Organization, 2017). Besides that there is variability in the prevalence of IPV across different populations, there are several factors that influence its prevalence. Some of the risk factors of IPV among South Asian women include gender role expectations, need to protect family honor, cultural normalization of abuse, arranged marriage system, abusive partner characteristics, and women’s fear of losing children and being on her own (Sabri, Simonet, & Campbell, 2018). These factors also vary among different populations and have not been studied adequately.

Pakistan is a developing country where IPV is a common public health concern. Few qualitative and quantitative studies have been conducted in selected subpopulation of Pakistan, therefore are deficient in providing representative information about burden of the problem. Studies conducted before 2010 may be outdated and reported high variability and burden of IPV (Ali & Bustamante-Gavino, 2007; Farid, Saleem, Karim, & Hatcher, 2008; Fikree & Bhatti, 1999; Fikree, Jafarey, Korejo, Afshan, & Durocher, 2006; Shaikh, 2003; Shaikh, Shaikh, Kamal, & Masood, 2008). However, more recent studies in selected population of Karachi also indicated high variability and burden of the problem (35%-97.5%) of IPV. (Ali, Asad, Mogren, & Krantz, 2011; Ali, Ali, Khuwaja, & Nanji, 2014). These figures are not only staggering but also raised questions about its validity. A systematic review of IPV of available studies in Pakistan reported 48% to 84% for emotional violence and 16% to 80% for physical violence (P. A. Ali, Naylor, Croot, & O’Cathain, 2015). The large variation in estimates of IPV is due to differences in measurement tool, approach of the interviewers, and nonrepresentative hospital-based or selected subpopulation, questioning its validity and use. Thus, a standardized tool for assessment administered across different population strata and with representative sampling technique may provide more reliable estimates of the burden of the problem and also an opportunity for across- and within-country comparison over time. This may serve as a benchmark of IPV for intervention where follow-up studies could determine its progression.

Pakistan is a country of diverse cultures, norms, and a male dominant society. It has four provinces, namely, Punjab, Sindh, Balochistan, and Khyber Pakhtunkhwa (KPK)—each with its unique customs and traditions. The culture of Punjab and Sindh promotes a more open atmosphere of respect and discussion among the couple, whereas the culture of Balochistan and KPK reflects tribal norms and customs where women have less say in decision making (Levesque, 2013; All Answers Ltd., 2017 ). The culture and norms also affect the relationship between men and women and frequency of IPV, as some cultures demonstrate acceptence of violence by women (Sabri et al., 2018).

During the third round of Pakistan Demographic and Health Survey (PDHS) in 2012-2013, a domestic violence module was included for the first time. PDHS is a nationally representative survey and used the Conflicts Tactics Scale (CTS) for measuring IPV. The frequency of IPV was reported to be 38.5% (National Institute of Population Studies, 2013). PDHS did not provide estimates of IPV across the provinces. As the ethnicities, cultures, and ways of life of the provinces are very distinct from one another, we aimed to explore IPV across them. We performed a secondary analysis of PDHS data to compare the prevalence and risk factors of emotional and physical IPV in the provinces of Pakistan, taking a more conservative approach for the IPV definition. This was done to have a more specific estimate for violence as the culture of Pakistan shows a trend of acceptance of IPV by women.

Method

Study Design and Population

Data for PDHS were collected from the four provinces of Pakistan, Gilgit-Baltistan (GB) region, and Islamabad capital territory (ICT). Azad Jammu Kashmir (AJK), Federally Administered Tribal Areas (FATA), and the restricted military and protected areas were not included (makes about 5% of population). The technique of two-stage stratified random sampling was adopted to draw a representative sample of the population of Pakistan. The details of the sampling technique are provided in the PDHS report 2012-2013 (National Institute of Population Studies, 2013).

A total of 14,000 households were selected and the domestic violence module was administered to one third of the household sample. Ever-married women between the ages of 15 and 49 years were included. Of the 3,743 eligible women, 3,687 were successfully interviewed. After excluding the incomplete records, final analysis was done on 3,666 women.

Measurement Tool

The CTS was adopted to measure IPV. The CTS has three domains: emotional, physical, and sexual violence (Straus, 1979). Although the questions of sexual violence were included, no data were available for analysis. Either this section was not administered or no data were collected due to the sensitivity of the issue. Thus, analysis was restricted to emotional and physical violence. The questions used for assessing emotional violence inquired about threatening, insulting, accusation, humiliation, limiting contact with the family, restriction to meet friends, insisting on knowing the whereabouts all the time, and being afraid of the husband. Physical violence included asking about slapping, pushing, shaking, twisting the arm, hair pulling, punching, kicking, dragging, burning, choking, and attacking with a knife, gun, or a weapon. The answers were recorded as yes (1) and no (0). All the answers in “yes” were added to calculate a composite score for individual women. Positive answers to at least two questions were considered for labeling emotional or physical violence.

Risk Factors for IPV

The relationships between socioeconomic and demographic characteristics with IPV were examined in the analysis. The variables included were the place of residence (rural/urban), age of the woman and husband (15-25, 25-35, 35-45 and >45 years), educational level of the woman and the husband (primary, secondary, and higher), occupational status of the woman and the husband, and the wealth index. The wealth index was constructed from a list of household assets (electricity, radio, television, landline telephone, refrigerator, cabinet, chair, room cooler, air-conditioner, washing machine, water pump, bed, clock, sofa, camera, sewing machine, computer, internet connection) by conducting principal component analysis. Education was recorded as years of schooling. Occupation was grouped as skilled (professional, technical, managerial, sales, services, skilled manual) and unskilled (agricultural, self-employed, household, domestic work) work. In addition, a behavioral variable of alcohol intake by the husband was also considered important and included.

Statistical Analysis

The analysis was done on STATA version 13.0 using the svy command. In the first stage, stratification was done by rural or urban setting of the primary sampling units (clusters). In the second stage households were from each strata and clusters. Weights were applied to get representative estimates of the provinces to adjust for unequal sampling. Weighted proportions along with 95% confidence intervals (CIs) were calculated for physical and emotional violence at the national level and for each province, separately. Multicollinearity among the variables was assessed using Cramer’s V.

To identify potential risk factors for IPV, univariate binary logistic regression, stratified by province, was done. The potentially significant factors having a significance of p <0.25 were carried forward to the multivariate model for further analysis. The multivariate model with the maximum value of Wald chi-square was considered for the final model. One model was developed for each province. The mutually adjusted odds ratios (aORs) for all the factors and 95% CIs were displayed. Multivariate logistic regression was not done separately for ICT and GB as their sample size was not sufficient.

The data of PDHS 2012-2013 are in public domain and freely available for use and analysis, and thus, no ethical approval was required for this analysis. However, PDHS undertook ethical clearance and permission as required (National Institute of Population Studies, 2013).

Results

Table 1 shows the overall prevalence of emotional and physical violence of 36.4% (33.8-39.1) and 18.4% (16.4-20.6), respectively, for Pakistan. KPK reported the highest estimates of IPV: 54.9% and 36.4% of emotional and physical violence, respectively. It was followed by Balochistan (50.0% and 25.5%), Punjab (35.9% and 15.8%), ICT (30.4% and 15.7%), Sindh (24.7% and 13.3%), and GB (20.0% and 6.10%). The trend of difference in the prevalence of emotional and physical violence had similar variation across provinces, that is, those provinces with higher prevalence of emotional violence also had higher prevalence of physical violence and vice versa. The estimates of IPV were higher in rural than in urban areas (emotional violence, rural: 40.2%, urban: 20.8%; physical violence, rural: 28.8%, urban: 13.7%). Higher estimates of violence were reported if the husband was uneducated, did not work, consumed alcohol, was over the age of 45 years, the women was not educated, and the family belonged to poorer wealth quintile. The multivariate analysis confirmed the variability of IPV across the provinces with KPK having an aOR of 3.7 [2.6, 5.1] and 3.9 [2.7, 5.6] compared with the Sindh province. The odds of physical violence were twice if the age of the woman was more than 45 years compared with women aged 15 to 25 years. The odds of emotional and physical violence increased by 5.7 [3.4, 9.6] and 3.8 [2.5, 5.7] times if alcohol intake was reported by the husband.

Table 1.

Frequency and Mutually Adjusted Socioeconomic and Behavioral Factors of Emotional and Physical Intimate Partner Violence Among Married Women of Pakistan (n = 3,666).

Characteristics	Emotional Violence		Physical Violence
Characteristics	Weighted % (CI)	aOR (95% CI)	Weighted % (CI)	aOR (95% CI)
Pakistan	36.4 [33.8, 39.1]		18.4 [16.4, 20.6]
Place of residence
Rural	40.2 [37.3, 43.2]		20.8 [18.6, 23.1]
Urban	28.8 [25.4, 32.4]		13.7 [11.2, 16.6]
Province
KPK	54.9 [49.2, 60.5]	3.67 [2.62, 5.13]	36.4 [30.9, 42.2]	3.86 [2.65, 5.64]
Balochistan	50.0 [39.4, 60.6]	3.12 [1.93, 5.04]	25.5 [17.3, 36.0]	2.37 [1.39, 4.04]
Punjab	35.9 [32.2, 39.7]	2.03 [1.55, 2.66]	15.8 [13.3, 18.7]	1.37 [0.99, 1.89]
ICT	30.4 [23.5, 38.2]	2.71 [1.69, 4.34]	15.7 [10.9, 22.1]	2.75 [1.58, 4.79]
Sindh	24.7 [21.1, 28.6]	1	13.3 [10.6, 16.6]	1
GB	20.0 [13.4, 28.8]	0.63 [0.35, 1.12]	6.10 [2.5, 14.1]	0.35 [0.13, 0.97]
Current age of the woman (years)
15-25	33.4 [28.6, 38.5]	1	13.8 [10.5, 18.0]	1
26-35	35.0 [30.8, 39.4]	0.92 [0.61, 1.39]	19.0 [16.2, 22.2]	1.51 [1.01, 2.26]
36-45	38.1 [33.5, 42.8]	0.80 [0.53, 1.20]	19.1 [15.9, 22.8]	1.19 [0.74, 1.91]
>45	45.1 [37.4, 53.1]	1.14 [0.68, 1.89]	26.1 [19.5, 34.1]	2.00 [1.06, 3.76]
Current age of the husband (years)
15-25	26.8 [20.5, 34.2]	1	14.0 [9.4, 20.5]	1
26-35	32.9 [29.0, 37.0]	1.55 [0.89, 2.70]	15.0 [12.2, 18.3]	0.98 [0.55, 1.74]
36-45	38.1 [33.8, 42.5]	2.17 [1.21, 3.89]	19.9 [16.9, 23.3]	1.38 [0.74, 2.58]
>45	43.0 [38.2, 48.1]	2.24 [1.23, 4.08]	23.1 [18.9, 27.9]	1.31 [0.65, 2.63]
Woman’s education^a
No education	42.8 [39.4, 46.3]	1	23.2 [20.4, 26.2]	1
Primary education	34.8 [29.8, 40.2]	0.99 [0.73, 1.33]	15.3 [11.5, 20.2]	0.89 [0.61, 1.31]
Secondary education	28.7 [23.3, 34.7]	0.88 [0.60, 1.29]	12.5 [9.0, 17.1]	1.00 [0.62, 1.62]
Higher education	14.6 [10.3, 20.1]	0.45 [0.26, 0.78]	5.6 [3.3, 9.3]	0.62 [0.29, 1.29]
Husband’s education
No education	45.2 [41.1, 49.4]	1	25.2 [21.8, 28.8]	1
Primary education	36.0 [30.2, 42.2]	0.77 [0.56, 1.04]	19.7 [15.7, 24.5]	0.88 [0.64, 1.19]
Secondary education	34.1 [29.7, 38.8]	0.84 [0.63, 1.11]	16.5 [13.4, 20.0]	0.75 [0.54, 1.05]
Higher education	24.0 [19.9, 28.7]	0.77 [0.54, 1.10]	7.5 [5.3, 10.6]	0.37 [0.22, 0.61]
Woman’s occupation
Does not work	34.2 [31.2, 37.4]	1	16.9 [14.6, 19.5]	1
Unskilled work	42.6 [37.2, 48.1]	1.28 [0.97, 1.69]	22.2 [17.7, 27.4]	1.44 [1.00, 2.05]
Skilled work	35.2 [26.3, 45.3]	1.65 [1.04, 2.61]	20.2 [12.9, 30.1]	1.72 [0.97, 3.07]
Husband’s occupation
Does not work	49.1 [34.8, 63.4]	1	28.1 [16.8, 43.0]	1
Unskilled work^b	38.1 [35.0, 41.4]	0.79 [0.41, 1.53]	18.4 [16.2, 20.9]	0.71 [0.32, 1.57]
Skilled work^c	30.9 [27.1, 35.0]	0.76 [0.38, 1.50]	17.7 [14.4, 21.6]	1.00 [0.43, 2.30]
Wealth index
Richest	20.1 [15.4, 25.8]	1	7.1 [4.8, 10.3]	1
Richer	34.1 [28.7, 40.0]	1.60 [1.04, 2.47]	15.0 [11.4, 19.5]	1.84 [1.07, 3.16]
Middle	39.5 [34.5, 44.7]	1.66 [1.10, 2.50]	21.3 [16.9, 26.3]	2.29 [1.38, 3.82]
Poorer	45.4 [40.0, 50.9]	1.94 [1.24, 3.02]	27.5 [22.8, 32.7]	2.74 [1.61, 4.66]
Poorest	44.1 [39.1, 49.3]	2.27 [1.45, 3.53]	22.2 [17.5, 27.7]	2.27 [1.27, 4.06]
Alcohol intake by the husband
No	34.1 [31.6, 36.8]	1	16.6 [14.7, 18.7]	1
Yes	76.5 [66.2, 84.4]	5.70 [3.39, 9.57]	50.3 [40.5, 60.0]	3.75 [2.47, 5.70]

Note. CI = confidence interval; aOR = adjusted odds ratio; KPK = Khyber Pakhtunkhwa; ICT = Islamabad district; GB = Gilgit-Baltistan.

Primary education: Grades I to V; secondary education: Grades VI to VIII; higher education: greater than Grade VIII.

Agricultural, self-employed, household, domestic work, and unskilled manual work.

Professional, technical, managerial, sales, services, and skilled manual.

Table 2 shows the results of the multivariate logistic regression models for emotional violence, separately for each province. The risk of emotional violence in KPK was higher with higher age, low educational status, alcohol intake by the husband, and lower socioeconomic status of the family. In the province of Sindh, lower socioeconomic status and alcohol intake by the husband were important factors for violence. In Punjab, the higher education of women (odds ratio [OR]: 0.4, 95% CI: [0.2, 0.9]) and the richest quintile of wealth index (OR: 0.5; 95% CI: [0.3, 0.9]) had a protective effect on emotional violence, whereas alcohol intake by the husband increased the risk of violence.

Table 2.

Socioeconomic and Behavioral Factors for Emotional Violence Among Married Women Across the Provinces of Pakistan (n = 3,666).

Characteristics	KPK	Balochistan	Punjab	Sindh
	n = 681	n = 468	n = 1,090	n = 837
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Current age of the woman (years)
15-25	1	1	1
26-35	0.80 [0.45, 1.42]	2.20 [0.89, 5.41]	1.13 [0.60, 2.12]
36-45	0.61 [0.30, 1.24]	2.54 [0.65, 9.94]	0.97 [0.54, 1.75]
>45	1.25 [0.42, 3.66]	2.44 [0.56, 10.50]	1.41 [0.68, 2.91]
Current age of the husband (years)
15-25	1	1	1	1
26-35	3.36 [1.13, 9.92]	1.12 [0.33, 3.73]	1.14 [0.50, 2.62]	1.39 [0.76, 2.54]
36-45	3.84 [1.20, 12.28]	0.76 [0.18, 3.08]	1.53 [0.63, 3.67]	2.37 [1.29, 4.37]
>45	4.37 [1.42, 13.41]	0.54 [0.09, 3.07]	1.71 [0.70, 4.17]	1.82 [0.86, 3.85]
Woman’s education
No education	1		1	1
Primary education	0.94 [0.45, 1.93]		1.00 [0.68, 1.49]	0.78 [0.40, 1.54]
Secondary education	1.57 [0.78, 3.17]		0.70 [0.41, 1.19]	1.08 [0.50, 2.34]
Higher education	0.55 [0.20, 1.46]		0.38 [0.17, 0.88]	0.45 [0.17, 1.19]
Husband’s education
No education	1	1	1	1
Primary education	1.01 [0.50, 2.05]	0.55 [0.22, 1.39]	0.81 [0.54, 1.22]	0.69 [0.37, 1.31]
Secondary education	0.60 [0.34, 1.04]	0.49 [0.22, 1.09]	0.98 [0.66, 1.45]	0.78 [0.47, 1.29]
Higher education	0.41 [0.20, 0.84]	0.63 [0.24, 1.63]	1.03 [0.61, 1.74]	0.92 [0.45, 1.87]
Woman’s occupation
Does not work		1	1	1
Unskilled work		1.63 [0.82, 3.22]	1.31 [0.91, 1.88]	1.16 [0.69, 1.93]
Skilled work		3.02 [0.79, 11.54]	1.48 [0.72, 3.04]	1.76 [0.97, 3.22]
Husband’s occupation
Does not work			1
Unskilled work			0.67 [0.19, 2.38]
Skilled work			0.54 [0.14, 1.97]
Wealth index
Richest	1	1	1	1
Richer	2.73 [1.26, 5.88]	1.40 [0.33, 5.96]	1.51 [0.84, 2.70]	1.69 [0.87, 3.24]
Middle	1.94 [0.83, 4.52]	0.65 [0.17, 2.47]	1.52 [0.87, 2.65]	1.19 [0.56, 2.55]
Poorer	3.42 [1.48, 7.93]	0.90 [0.21, 3.92]	1.65 [0.86, 3.16]	2.03 [0.90, 4.54]
Poorest	2.72 [1.04, 7.10]	1.30 [0.35, 4.84]	2.04 [1.07, 3.91]	2.38 [1.07, 5.25]
Alcohol intake by husband	5.28 [2.61, 10.68]	3.30 [0.73, 14.85]	5.42 [2.07, 14.2]	8.89 [3.15, 25.04]

Note. KPK = Khyber Pakhtunkhwa; aOR = adjusted odds ratio; CI = confidence interval.

Table 3 shows the association of risk factors with physical violence across provinces. Alcohol intake significantly increased the risk of physical violence across all provinces, with highest risk in Punjab (OR: 6.9, 95% CI: [3.2, 14.9]). The richest quintile of wealth index was protected from physical violence in KPK and Sindh (OR: 0.2, 95% CI: [0.1, 0.6]; 0.2, 95% CI: [0.1, 0.7], respectively). The increasing age of the woman led to an increase in physical violence in Punjab. Women older than 45 years were 2.8 times more likely to be a victim of physical violence compared with women less than 25 years of age. The husband’s higher education and unemployed women were more protected against physical violence in Sindh.

Table 3.

Socioeconomic and Behavioral Risk Factors for Physical Violence Among Married Women Across the Provinces of Pakistan (n = 3,666).

Characteristics	KPK	Balochistan	Punjab	Sindh
	n = 681	n = 468	n = 1,090	n = 837
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Current age of the woman (years)
15-25 years	1	1	1	1
26-35 years	1.36 [0.68, 2.69]	1.30 [0.53, 3.18]	2.04 [1.14, 3.66]	0.80 [0.31, 2.05]
36-45 years	1.35 [0.56, 3.22]	1.26 [0.48, 3.29]	1.59 [0.88, 2.90]	0.99 [0.36, 2.72]
> 45 years	1.46 [0.55, 3.83]	2.53 [0.69, 9.22]	2.75 [1.28, 5.90]	1.95 [0.60, 6.36]
Current age of the husband (years)
15-25	1			1
26-35	2.40 [0.72, 8.02]			2.12 [0.79, 5.66]
36-45	2.90 [0.88, 9.59]			2.85 [0.92, 8.79]
>45	2.47 [0.69, 8.80]			2.73 [0.65, 11.47]
Woman’s education
No education	1		1	1
Primary education	0.71 [0.28, 1.82]		0.82 [0.49, 1.36]	1.13 [0.51, 2.48]
Secondary education	1.14 [0.44, 2.94]		0.77 [0.39, 1.52]	1.33 [0.48, 3.65]
Higher education	1.07 [0.34, 3.34]		0.47 [0.14, 1.55]	1.12 [0.30, 4.14]
Husband’s education
No education	1	1	1	1
Primary education	1.73 [0.96, 3.10]	0.79 [0.30, 2.09]	0.69 [0.42, 1.13]	0.98 [0.48, 1.98]
Secondary education	0.74 [0.42, 1.30]	0.60 [0.32, 1.12]	0.84 [0.51, 1.38]	0.89 [0.43, 1.87]
Higher education	0.58 [0.24, 1.36]	0.54 [0.20, 1.46]	0.43 [0.17, 1.06]	0.27 [0.09, 0.74]
Woman’s occupation
Does not work		1	1	1
Unskilled work		1.76 [0.86, 3.60]	1.26 [0.76, 2.10]	2.54 [1.46, 4.40]
Skilled work		Empty	0.81 [0.24, 2.77]	5.19 [3.00, 8.99]
Husband’s occupation
Does not work		1	1
Unskilled work		0.27 [0.06, 1.16]	0.54 [0.10, 2.90]
Skilled work		0.54 [0.09, 3.10]	0.58 [0.10, 3.40]
Wealth index
Richest	1	1	1	1
Richer	4.73 [1.64, 13.64]	1.12 [0.26, 4.74]	1.55 [0.77, 3.10]	3.02 [1.13, 8.05]
Middle	8.66 [3.31, 22.66]	0.83 [0.19, 3.60]	1.77 [0.91, 3.43]	1.44 [0.38, 5.39]
Poorer	13.31 [5.27, 33.61]	1.07 [0.27, 4.31]	1.83 [0.82, 4.07]	3.37 [0.97, 11.69]
Poorest	5.76 [1.77, 18.76]	1.95 [0.56, 6.78]	1.27 [0.51, 3.17]	4.59 [1.51, 13.93]
Alcohol intake by husband	3.76 [2.09, 6.76]	6.81 [2.36, 19.63]	6.92 [3.22, 14.85]

Note. KPK = Khyber Pakhtunkhwa; aOR = adjusted odds ratio; CI = confidence interval.

Discussion

Our results highlight the widespread IPV in Pakistan with the overall prevalence of emotional violence as 36.4% and physical violence as 18.4%. The principal risk factors identified for IPV were the socioeconomic status and alcohol intake by the husband. This is the first nationally representative study to compare IPV across the provinces of Pakistan. Emotional violence ranged from 24.7% to 54.9%, whereas physical violence varied from 13.3% to 36.4%. The province of KPK had the highest burden of IPV. The husband’s higher education had a protective effect, whereas the odds of emotional violence increased with increasing age of the husband in KPK only.

Our estimates are analogous to the available global estimates, depicting 35% women to be a victim of IPV at least once during their lifetime (World Health Organization, 2017). Our results second the national IPV estimates of Iraq, reporting 33.4% of emotional and 21.2% of physical violence, respectively (Ministry of Health, World Health Organization, Ministry of Planning and Development Cooperation, 2008). The overall estimates of IPV were comparable to one of the hospital-based studies conducted in Karachi in 2014, reporting the overall prevalence of IPV as 35% (Ali et al., 2014). However, variation was seen across other studies ( Ali & Bustamante-Gavino, 2007; Ali et al., 2011). Emotional and physical violence were reported to be 83.6% and 57.6% in 2011 in the married women of Karachi ( Ali et al., 2011). In a study conducted in 2007 in Karachi city of Pakistan, the figures were estimated to be 97.5% and 80.0% for emotional and physical violence, respectively ( Ali & Bustamante-Gavino, 2007). Both these studies were carried out in the poor marginalized, specific ethnic population of Karachi; hence, this may have overestimated the burden. The disparity with our results might also be explained by the difference in the measuring tools (Straus, 1979; García-Moreno, Jansen, Ellsberg, Heise, & Watts, 2005).

The variation in the proportion of IPV across the provinces draws attention to the important role of cultural norms and traditions. The anthropological studies in provinces of Punjab and Sindh reported that the culture of friendship, tolerance, and respect toward fellow human beings prevails (UK Essays, 2013). These provinces are also comparatively progressive in their cultures and social values. The culture of Balochistan is tribal in nature where women have less participation in decision making. KPK faces more sex segregation and women are hardly allowed in public life. Those who move outside their homes are usually covered from head to toe. Evidence also suggests that the tendency of accepting unequal distribution of power by less powerful members is high in KPK (Shah & Amjad, 2011). The customs and values of KPK and Balochistan are comparable with the neighboring country Afghanistan. The proportion of IPV in Afghanistan is 56% (37% emotional violence, 51% physical violence), giving analogous results (Ministry of Public Health, 2017). Similarly, the geographical boundaries of Punjab and Sindh touch with India. The traditions of these provinces also simulate the customs of neighboring provinces of India. In India, the proportion of emotional and physical violence was 12% and 34%, respectively, in 2007, relating closely to the proportions of Punjab and Sindh (International Institute for Population Sciences, Macro International, 2007). The overall proportion of IPV in India in 2015 was estimated to be 28.8%, signifying the ethnic and traditional similarities (International Institute for Population Sciences, 2015). However, in Afghanistan and India, the fraction for physical violence is more compared with emotional violence, whereas in Pakistan emotional violence is more compared with physical violence. We believe there is under-reporting of physical violence in our setting that may also partially explain the difference. Lack of a precise operational definition of violence can also be the cause of this difference. This geographical distribution of IPV needs to be further studied.

Alcohol intake by the husband and low socioeconomic status across all provinces were associated with IPV. Our finding was supported by data from 10 countries included in the WHO multicountry study on women’s health and domestic violence that reported alcohol abuse and low socioeconomic and educational status as potential risk factors for IPV (Abramsky et al., 2011). The link of alcohol intake and IPV has been extensively studied, although the temporality and mediator pathways of the observed association are not clear, yet it is established that there is a moderate association of alcohol abuse and male to female partner violence (Foran & O’Leary, 2008). The high proportion of emotional violence in KPK was associated with the lower husband’s education and his higher age. The low educational attainment has been identified as a risk factor in other small-scale studies conducted in Pakistan (Ali et al., 2011). Although the causative pathways between the risk factors and IPV remain to be fully explained, consistent findings have been reported in the literature (Foran & O’Leary, 2008; Gil-González, Vives-Cases, Alvarez-Dardet, & Latour-Pérez, 2006; Leonard, 2005). Some qualitative inquiry also highlights other risk factors for violence such as bringing insufficient dowry, inability to give birth to a male child, and lack of proper care of children as important factors of violence in this group (Nasrullah, Zakar, Zakar, Abbas, & Safdar, 2015). We could not capture these variables; we believe their assessment would have strengthened our results. The models of multivariate analysis cannot elucidate the wide variation across the provinces. Future studies need to further look in-depth into the reasons and risk factors leading to high IPV in KPK and Balochistan.

The major strength of this study was nationally representative data. Various small studies have been done in Pakistan depicting violence, but no study has compared violence across the provinces of Pakistan. This is the first study to bring into light the variation in IPV across provinces. Our data include aspects of differences such as socioeconomic status and age. We also applied robust techniques of analysis, making the results more compelling. However, our results do not come without limitations. The cultural adaptability of CTS for Pakistan is questionable, as it has not been validated in this setting before. There is also evidence available that CTS underestimates the proportions of IPV (Archer, 1999). We only included the variables available in PDHS data for examining the factors associated with violence. Therefore, we missed out on some of the important risk factors for IPV, such as maltreatment of the husband as a child, exposure of husband to violence in the family, husband’s attitude accepting of violence and gender inequality, exposure of wife to abuse during childhood, her exposure to violence between the parents, and having an attitude accepting of violence and gender equality (Fulu, Jewkes, Roselli, & Garcia-Moreno, 2013; World Health Organization, 2017). Furthermore, there were no data available regarding sexual violence. Either this portion of the module was not administered or the women did not respond to this section. Finally, we cannot rule out the possibility of under-reporting of violence in our country, as women tend to hide their personal information.

Conclusion

The prevalence of IPV was very high in Pakistan. The variation in its prevalence across the provinces stresses on the pivotal role of traditions, norms, and culture. We found an association of IPV with an increasing age of the husband, blue-collar occupation, low educational status, alcohol intake, and low wealth index of the household. However, we need more in-depth assessment in this direction to understand the differences across cultures, which leads to amplified violence. As policies are implemented at a provincial level in Pakistan, the risk factors of IPV across provinces and the perceptions of men and women need to be clearly studied and understood to have policy implications.

Footnotes

Acknowledgements

The authors acknowledge the assistance of Dr. M. Arslan Khan in the initial analysis of the study.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Meesha Iqbal

Author Biographies

Meesha Iqbal is a post graduate trainee in the Department of Community Health Sciences at the Aga Khan University, Karachi. Her area of prime interest is researching on vulnerable groups. She has also worked on a WHO-EMRO funded project to assess the health and social issues of the child laborers of Pakistan. She has also been associated in voluntary capacity with several nongovernmental organizations (NGOs) of Pakistan that deal with vulnerable groups such as Sahil, Society for the Protection of the Rights of the Child (SPARC), Aahung, and PILER. She was the president of the Pakistan’s largest charitable society run by undergraduate students, Students’ Patients Welfare Society (SPWS), during her college days at King Edward Medical University. She also worked as a volunteer in CMH Muzaffarabad when the city was struck by the devastating earthquake.

Zafar Fatmi is a public health academician, researcher and practitioner with a terminal public health degrees (Fellowship in Community Medicine from Pakisan and PhD from MRC Lifecourse Epidemiology Unit, Faculty of Medicine, University of Southampton, UK). He currently works as professor and Unit Lead, Environmental, Occupational Health & Injuries at the Department of Community Health Sciences of Aga Khan University, Karachi, Pakistan. He was director of Community Medicine Residency Program from 2007 to 2016 and served as coordinator since 2003 at Aga Khan University. He is also supervisor in Community Medicine at the College of Physicians & Surgeons, Pakistan, since 2008. He has supervised thesis/dissertation work for more than 25 students of FCPS (Community Medicine) and MSc graduates (Epidemiology & Biostatistics and Health Policy & Management) at Aga Khan University. He directs research and academic activities of a group of public health faculty pertaining to environmental health and injuries. He has over 100 peer-reviewed publications and authored a chapter on a book by APHA (2011) titled “Megacities and Global Health.” He has collaborated for research with several international institutes and organizations, for example, Institute for Health and the Environment; University at Albany, New York; Wadsworth Center; New York State Department of Health; University of Alabama at Birmingham; Emory University, USA; University of Southampton, UK; Jichi Medical University, Japan; and King Abdul Aziz University, Saudi Arabia. He is also a member of “WHO Global Platform on Air Quality and Health” to represent the region.

References

Abramsky

Watts

C. H.

Garcia-Moreno

Devries

Kiss

Ellsberg

. . . Heise

(2011). What factors are associated with recent intimate partner violence? Findings from the WHO multi-country study on women’s health and domestic violence. BMC Public Health, 11(1), Article 109.

Ali

N. S.

Ali

F. N.

Khuwaja

A. K.

Nanji

(2014). Factors associated with intimate partner violence against women in a mega city of South-Asia: Multi-centre cross-sectional study. Hong Kong Medical Journal, 20, 297-303.

Ali

P. A.

Naylor

P. B.

Croot

O’Cathain

(2015). Intimate partner violence in Pakistan: A systematic review. Trauma, Violence, & Abuse, 16, 299-315.

Ali

T. S.

Asad

Mogren

Krantz

(2011). Intimate partner violence in urban Pakistan: Prevalence, frequency, and risk factors. International Journal of Women’s Health, 3, 105-115.

Ali

T. S.

Bustamante-Gavino

(2007). Prevalence of and reasons for domestic violence among women from low socioeconomic communities of Karachi. Eastern Mediterranean Health Journal, 13, 1417-1426.

All Answers Ltd. (2017, November). Culture diversity in Pakistan cultural studies essay. Retrieved from https://www.ukessays.com/essays/cultural-studies/culture-diversity-in-pakistan-cultural-studies-essay.php?vref=1

Archer

(1999). Assessment of the reliability of the conflict tactics scales: A meta-analytic review. Journal of Interpersonal Violence, 14, 1263-1289.

Breiding

M. J.

Basile

K. C.

Smith

S. G.

Black

M. C.

Mahendra

R. R.

(2015). Intimate partner violence surveillance: Uniform definitions and recommended data elements (Version 2.0). Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention. Retrieved from https://stacks.cdc.gov/view/cdc/31292

Centers for Disease Control and Prevention, U.S. Department of Health and Health Services. (2017, August 22). Intimate partner violence. Retrieved from https://www.cdc.gov/violenceprevention/intimatepartnerviolence/index.html

10.

Farid

Saleem

Karim

M. S.

Hatcher

(2008). Spousal abuse during pregnancy in Karachi, Pakistan. International Journal of Gynecology & Obstetrics, 101, 141-145.

11.

Fikree

F. F.

Bhatti

L. I.

(1999). Domestic violence and health of Pakistani women. International Journal of Gynecology & Obstetrics, 65, 195-201.

12.

Fikree

F. F.

Jafarey

S. N.

Korejo

Afshan

Durocher

J. M.

(2006). Intimate partner violence before and during pregnancy: Experiences of postpartum women in Karachi, Pakistan. Journal of the Pakistan Medical Association, 56(6), 252-257.

13.

Foran

H. M.

O’Leary

K. D.

(2008). Alcohol and intimate partner violence: A meta-analytic review. Clinical Psychology Review, 28, 1222-1234.

14.

Fulu

Jewkes

Roselli

Garcia-Moreno

(2013). Prevalence of and factors associated with male perpetration of intimate partner violence: Findings from the UN Multi-country Cross-Sectional Study on Men and Violence in Asia and the Pacific. The Lancet Global Health, 1(4), E187-E207.

15.

García-Moreno

Jansen

H. A.

Ellsberg

Heise

Watts

(2005). WHO multi-country study on women’s health and domestic violence against women: initial results on prevalence, health outcomes and women’s responses. World Health Organization.

16.

Gil-González

Vives-Cases

Alvarez-Dardet

Latour-Pérez

(2006). Alcohol and intimate partner violence: Do we have enough information to act? The European Journal of Public Health, 16, 278-284.

17.

International Institute for Population Sciences. (2015). National family health survey (NFHS-4) 2015-16 (India fact sheet). Mumbai: Ministry of Health and Family Welfare, Government of India.

18.

International Institute for Population Sciences, Macro International. (2007, September). National family health survey (NFHS-3): 2005-06. Mumbai: Ministry of Health and Family Welfare, Government of India.

19.

Leonard

K. E.

(2005). Alcohol and intimate partner violence: When can we say that heavy drinking is a contributing cause of violence? Addiction, 100, 422-425.

20.

Levesque

(2013). Managing diversity in Pakistan: Nationalism, ethnic politics and cultural resistance [Review essay],South Asia. Multidisciplinary Academic Journal. Retrieved from http://journals.openedition.org/samaj/3551

21.

Ministry of Health, World Health Organization, Ministry of Planning and Development Cooperation. (2008). Republic of Iraq: Iraq family health survey report 2006-07. Retrieved from http://www.who.int/mediacentre/news/releases/2008/pr02/2008_iraq_family_health_survey_report.pdf

22.

Ministry of Public Health. (2017). Afghanistan demographic and health survey, 2015. Kabul, Afghanistan: Central Statistics Organization.

23.

Nasrullah

Zakar

M. Z.

Abbas

Safdar

(2015). Circumstances leading to intimate partner violence against women married as children: A qualitative study in Urban Slums of Lahore, Pakistan. BMC International Health and Human Rights, 15(1), Article 23.

24.

National Institute of Population Studies. (2013, December). Pakistan demographic and health survey 2012-13. Islamabad, Pakistan: National Institute of Population Studies, ICF International.

25.

Nicolaidis

Paranjape

(2009). Defining intimate partner violence: Controversies and implications. In Mitchell

(Ed.), Intimate partner violence: A health-based perspective (pp. 19-30). New York: Oxford University Press .

26.

Sabri

Simonet

Campbell

J. C.

(2018). Risk and protective factors of intimate partner violence among South Asian immigrant women and perceived need for services. Cultural Diversity & Ethnic Minority Psychology, 24, 442-452.

27.

Shah

S. A. M.

Amjad

(2011). Cultural diversity in Pakistan: National vs provincial. Mediterranean Journal of Social Sciences, 2, 331-344.

28.

Shaikh

M. A.

(2003). Is domestic violence endemic in Pakistan: Perspective from Pakistani wives. Pakistan Journal of Medical Sciences, 19(1), 23-28.

29.

Shaikh

M. A.

Shaikh

I. A.

Kamal

Masood

(2008). Domestic violence and pregnancy—Perspective from Islamabad and Rawalpindi. Journal of the College of Physicians and Surgeons Pakistan, 18, 662-663.

30.

Straus

M. A.

(1979). Measuring intrafamily conflict and violence: The Conflict Tactics (CT) Scales. Journal of Marriage and the Family, 41, 75-88.

31.

UK Essays. (2013, November). Culture diversity in Pakistan cultural studies essay. Retrieved from http://www.ukessays.com/essays/cultural-studies/culture-diversity-in-pakistan-cultural-studies-essay.php?vref=1

32.

UN Women. (2016, February). Facts and figures: Ending violence against women. Retrieved from http://www.unwomen.org/en/what-we-do/ending-violence-against-women/facts-and-figures

33.

World Health Organization. (2017). Violence against women. Author. Retrieved from http://www.who.int/mediacentre/factsheets/fs239/en/