Prevalence and correlates of physical violence perpetrated by an intimate partner against women who are pregnant in India

Abstract

Background

Intimate partner violence (IPV) against women is a public health concern and a human rights violation. Women in low- and middle-income countries such as India experience higher rates of violence than in high-income countries. Experiencing IPV during pregnancy is associated with worse maternal, foetal and child outcomes. There is relatively limited representative population-based data on the burden and related factors of IPV experienced by women during pregnancy in India. This study aimed to ascertain the prevalence of experiences of physical IPV during pregnancy and the factors associated with it in India.

Methods

This is a secondary analysis of data collected in the Fifth Indian National Family Health Survey (NFHS-5). The primary outcome was lifetime experience of physical violence during pregnancy perpetrated by a former and/or current partner. The relationship between this outcome and sociodemographic, relationship, societal and gender-based factors characteristics were examined. Data were analysed using multivariable logistic regression.

Results

Among 60,579 women who had ever been married or in a union and ever been pregnant, the lifetime prevalence of physical IPV during pregnancy was 2.4%. Experience of physical violence during pregnancy was significantly more common among women experiencing diverse social and economic disadvantages such as low household wealth, being a member of a scheduled caste, or a religious minority. Women whose partners had lower educational status or who drank alcohol had higher odds of experiencing IPV during pregnancy. It was less common among women who were more empowered including having more decision-making autonomy.

Conclusion

Despite low prevalence, the absolute numbers of women experiencing physical IPV during pregnancy in India are alarmingly high. There is an urgent need to screen and provide early interventions for IPV during pregnancy in India. Further research is needed to identify the prevalence of other forms of intimate partner violence including sexual and emotional abuse, financial abuse and coercive control experienced by women who are pregnant in India.

Background

Violence against women is defined as “any act of gender-based violence that results in, or is likely to result in, physical, sexual or psychological harm or suffering to women, including threats of such acts, coercion or arbitrary deprivations of liberty, whether occurring in public or private life” [1]. Violence against women is a public health concern and human rights violation with adverse impacts on social and economic participation and physical, mental, sexual, and reproductive health of women, and the health and development of children. The most common perpetrator of violence against women is an intimate partner [2–4].

Intimate partner violence (IPV) refers to behaviour by an intimate partner or ex-partner including physical aggression, sexual coercion, psychological abuse and controlling behaviours that causes physical, sexual, social, or psychological harm. Globally, it is estimated that about 35% of women experience IPV in their lifetimes with significant variations in prevalence within and between countries [5, 6]. However, pooled international estimates obscure the country-specific magnitude of IPV experienced by women and local factors associated with it. There is disproportionately higher lifetime prevalence (up to 42%) of IPV experienced by women living in low- and middle-income (LMICs) than in high-income countries (HIC) [2, 5, 7–9] and in different stages of life including the perinatal (pregnancy and postpartum) period.

Román-Gálvez et al. [10] pooled data from 155 cross-sectional and cohort studies in their systematic review to measure the global prevalence of IPV experienced by women who are pregnant. They found that one fourth of pregnant women experience any IPV and that there are significant variations in prevalence within and across the different regions and countries of the world. However, these figures are of limited value for individual nations, for national policies and programmes to address IPV in pregnancy. The WHO Multi Country Study on Women’s Health and Domestic Violence Against Women [2, 3] found marked variation in prevalence among countries with physical IPV in pregnancy ranging from 1% in Japan to 28% in Peru. Similar highly variable prevalence rates of IPV in pregnancy have been reported among countries such as China (7.7%) [11], Vietnam (5.9–32%) [12], Iran (51%) [13], Ethiopia (26%) [14], Nigeria (2.3–44.6%) [15] and Australia (2%) [16].

These variations in estimates of prevalence are partially attributable to methodological differences including in definitions (intimate partner versus family versus domestic violence (DV) and/or abuse) and measurement (varied assessment instruments, use of non-validated scales, study-specific semi structured questions) differences, reporting of lifetime prevalence versus IPV during the index or last pregnancy, study settings (clinic/hospital based, community sample or population surveys), design (small sample sizes, lack of comparison groups), and study quality [9, 10, 17]. However, these variations are predominantly attributable to differences in the social, economic, and cultural circumstances of women’s lives within and across countries. The highest risk of IPV is experienced by women in the most disadvantaged and marginalised positions. It has also been suggested that there is a complex interaction between socioeconomic disadvantage and harmful sociocultural norms including the status of women in a particular society, and individual and community attitudes towards IPV that influences likelihood of perpetration of IPV against women [14, 18–20].

In addition to measuring the extent of IPV, it is important to acknowledge that experiencing IPV during pregnancy is associated with worse maternal, foetal and child outcomes [21]. Women who have experienced IPV are less likely to attend antenatal care from early pregnancy and to attend for the recommended number of visits than those who have not [22]. They are at higher risk of experiencing intrauterine growth restriction, preterm birth and having a baby with low birth weight. Any type of IPV is associated with these negative outcomes but physical IPV has the most serious adverse consequences for maternal and child health [21]. Silva et al. [23] performed a meta-analysis of relevant longitudinal studies and found that children of mothers who experienced IPV in pregnancy are at increased risk of developing internalizing and externalizing problems in comparison to children with no maternal experience of IPV.

Systematic reviews have identified IPV as a clear risk factor for perinatal mental disorders (PMDs) [24–26]. However, Bacchus et al. [27] criticised these findings because they relied on cross-sectional studies in which causal direction could not be attributed, varied time periods of IPV experiences and restricted search criteria. Nevertheless, Bacchus et al.’s systematic appraisal of cohort studies found that IPV experiences in the previous year almost doubled the odds of postpartum depression. A similar systematic review of longitudinal studies [28] found strong association between IPV and perinatal depression and posttraumatic stress disorder with more consistent findings for IPV sustained close to or during the perinatal period than for lifetime IPV. Dai et al. [29] in their systematic review of association between IPV in the perinatal period and posttraumatic stress disorder found similar results. Importantly, IPV remains a modifiable risk factor, attributed to the unambiguous finding of higher prevalence of perinatal common mental disorders among women in LMICs than in HIC [24, 26].

India is the most populous country in the world [30] with more than twenty-five million births annually [31]. Indian society is “collectivist” (promotion of interdependence and co-operation within the family) [32] and patriarchal with men and boys having privileged social positions. Male patriarchy and entrenched gender inequalities prevail in India [33, 34]. Despite the recent advances made in gender equality in India, most people still hold traditional gender norms in the family context [35]. Norms and practices such as dowry, viewing women as subordinates, preference for male children, sex selective abortions, female infanticide increase susceptibility of women to IPV [36–40]. Women in India, experience high lifetime burdens of IPV [41, 42].

Intimate partner violence experienced during pregnancy by women in India has received some research attention in the last two decades. Sarkar [43] reported varied prevalence rates (15 to 50%) of IPV in pregnancy in different parts of India, in a narrative review of studies published between 2000 and 2011. However, it was unclear whether the IPV reported was lifetime experiences or during the last pregnancy. A recent systematic review [44] involving sixteen studies of violence against women during pregnancy in India found wide prevalence rates of 7–53% with physical violence being the most common type experienced by women. Violence was assessed by different scales such as Violence Assessment Screening Box [45–48], questionnaires based on the WHO Multi Country Study on Women’s Health and Domestic Violence Against Women [49, 50] and the Violence Assessment Questionnaire, developed by the Indian Council of Medical Research [51]. In Jungari et al.’s review most of the included studies examined violence experienced during the most recent pregnancy. One [52] reported lifetime experiences during any pregnancy. In most, violence against women was perpetrated by their husbands/partners and/or, in-laws and at times by neighbours [53] or acquaintances [47]. Sample sizes varied from 165 [54] to 14,507 [40] with different sampling strategies. Most studies were clinic based and in urban settings across different parts of the country. These heterogeneous methods precluded pooling of prevalence rates and meta-analyses.

Jungari [55] recently examined the data collected about life experience of physical IPV during pregnancy in the Fourth National Family & Health Survey of India (2015-16) and reported 3.3% lifetime prevalence of physical IPV in pregnancy. Nath et al. [56] found a similar prevalence (3.7%) of any IPV in pregnancy among women in low or medium socioeconomic positions attending antenatal care in an urban setting. Neither of these studies were included in Jungari et al.’s [44] recent systematic review, but the reasons for their exclusion were not specified.

Relatively few studies have explored the risk and protective factors associated with IPV during pregnancy in India. Women occupying a lower socioeconomic position, who are members of a minority religion, of low education status, not earning an income, having a history of pregnancy loss and partner’s alcohol use have each been associated with higher odds of experiencing violence in pregnancy [44, 57].

To summarize, the true national picture of IPV in pregnancy experienced by women in India is unclear. There are relatively limited and non-nationally representative population-based recent data on the extent and correlates of IPV during pregnancy in India. In addition, the data on the correlates is predominantly from clinic-based populations and from different parts of India with limitations in generalizability of these findings.

To date,, there has been little consideration of gender-based factors influencing experiences of violence during pregnancy among women in India. Therefore, there is an identified need to describe the relationship between IPV and female gender disadvantage and disempowerment related factors, particularly in LMICs settings such as India. To meet these knowledge gaps, this study aimed to assess the extent and various correlates (including gender-based) associated with physical IPV during pregnancy among women in India using the national representative population data.

Methods

Study setting

India is a federal republic of states and union territories with a parliamentary system of government. There are twenty-eight states and 8 union territories in the country. Each state has its own administrative government. State governments share governing powers with the Union government. Union Territories are administered by the President through an appointed administrator. A district is an administrative division of an Indian state or union territory.

National family health survey (NFHS-5)

This study is a secondary analysis of data collected in the Fifth Indian National Family Health Survey (NFHS-5), also known as the Demographic Health Survey (DHS). National Family Household Surveys are conducted at five-year intervals under the direction of the Ministry of Health and Family Welfare, Government of India, in collaboration with the International Institute for Population Sciences, Mumbai as a nodal agency, with funding and technical assistance from the United States Agency for International Development. Their aim is to generate reliable self-reported and objective data about population and health indicators from nationally representative samples. This includes data related to maternal and child health, socio-economic welfare, and DV, to inform policy and programme development.

The NFHS-5 was conducted between 2019 and 2021 in India. It covers each state and union territory and provides district level information for all 707 districts, described as the district module. However, estimates of domestic violence, husband’s background and woman’s work, indicators of sexual behaviour, HIV/AIDS knowledge, attitudes, and behaviour are available only at the state/union territory and national level under the state module.

A stratified two-stage sampling design, which is representative at the national, state/union territory, and district level, was used. Each district was stratified into urban and rural areas. The primary sampling frame was based on the census of 2011. Villages in the rural and Census Enumeration Blocks in the urban areas formed the primary sampling units. A list of households was created by mapping each selected primary sampling unit and then selecting twenty-two households randomly.

In each selected household, four survey schedules/questionnaires for information pertaining to the Household, Woman, Man, and children were completed using a computer aided personal interview. All women aged 18–49 in the household were eligible to complete the Women’s Questionnaire in an interview. The DV module was part of the Woman’s Questionnaire.

The DV questions were included only in the state module. The state module consisted of a random subsample of 15% of households drawn from the district sample. Weights were applied to ensure that the DV subsample was accurately representative of the national population. One randomly selected woman per participating household completed the DV module, to meet the ethical requirements. Women who were victims of domestic violence were provided with a list of appropriate local organizations that they could contact if they wanted help.

The NFHS-5 collected information from 639,699 households, 72,4115 women (72,320 completed the domestic violence module) and 101,839 men across all the regions of India. The DV module was completed by a total of 72,056 women in NFHS-5. Only 4% of women who were eligible for the DV module were not interviewed due to privacy concerns. Details of the design and data collection of the survey can be found in the national report of NFHS-5 [58].

Study sample

The inclusion criteria were for any data contributed by women who had ever been married or in a union, been pregnant and completed the DV module.

Study variables

Experiences of violence

In the DV module, women were asked a single question (“Has anyone ever hit, slapped, kicked, or done anything else to hurt you physically while you were pregnant?”) with a possible response of yes or no. The follow up question enquired whether or not these experiences were perpetrated by a former and/or current partner during any pregnancy in their lifetime.

Factors associated with experiences of violence

We identified potential independent variables from the ecological framework of Heise [59], the ‘drivers’ of IPV [60], Survey-based Women’s emPowERment index [61] and factors identified previously from India and internationally and available in the NFHS-5 data. These variables were categorized into five sections: individual, relationship, family, societal and gender empowerment related factors. A detailed description, and the coding and scoring of these variables are described in Table 1.

Table 1.

List of variables, their categories, definitions, and scoring/measure

Variable	Category/definition/scoring/measure
Individual Factors
Age (in years)	Categorized as 18–29, 30–39 and 40+
Age at first marriage (in years)	Categorized as < 18 and 18 & more
Education	Categorized as no formal education, primary, secondary, and above
Religion	Categorized as Hindu, Muslim, and others
History of pregnancy loss/termination	Categorized as ‘yes’ or ‘no’ to the survey question
Ever witnessed father beat her mother	Categorized as ‘yes’ or ‘no’ to the survey question
Number of living children	Categorized as 0–1, 2–3 and 4 children or more
Relationship factors
Years of cohabitation (in years)	Categorized as < 5, 6–10,11–20 and > 20
Partner’s education	Categorized as no formal education, primary, secondary, and above
Partner’s alcohol drinking	Categorized ‘yes’ if partner drinks alcohol and ‘no’ otherwise
Partner’s occupation	Categorized into two Category one: professional/technical/managerial/clerical Category two: Sales’ Services/household and domestic; Agricultural; Skilled and unskilled manual and Others
Income difference	Categorized into 1. woman not earning, 2. husband not bringing any money, 3. less than or same as the partner and 4. more than the partner
Family Factors
Wealth index	Survey generated categories based on the household characteristics (number and kinds of household goods, drinking water, toilet facility, flooring and electricity and categorized into five quintiles from ‘very poor’ to ‘very rich’
Sex of household head	Categorized as either woman or man
Societal factors
Ethnicity/Caste	Categorized as belonging to schedule caste, schedule tribe, other backwards class or none of these
Residence	Categorised as rural or urban or not living at usual residence
Gender Empowerment factors
Acceptance of IPV	Categorized into ‘yes’ if women responded yes to any of the seven following questions related to their attitudes on IPV and ‘no’ otherwise. women were asked if hitting or beating a wife is justifiable in the following situations, if- 1. she goes out without telling him, 2. she neglects the house or children, 3. she argues with the partner, 4. she refuses sex, 5. she does not cook the food properly, 6. he suspects her of infidelity. 7. she shows disrespects for in-laws.
Decision making	Response to 3 survey questions (purchase of large household items, seeking health care for self, and visiting families, relatives, and friends) Categorized as yes (1)- if makes decisions herself or jointly and no (0) otherwise. Scoring: Total scores added of all individual responses with score range of 0–3: The higher the score, the greater the autonomy in decision-making.
Media access	Response to 4 questions (listening to radio, watching TV/cinema, reading newspaper/magazine and internet Categorised as yes (1), irrespective of the frequency (1) and no (0)- if not available Scoring Total scores added of all individual responses with score range of 0–4: The higher the score, the greater the media access
Asset ownership	Response to two questions related to ownership of property (house and land). Ownership scored into yes (1) if owned alone or jointly with a partner and no (0) if does not own. Scoring: Total scores added of both individual responses with score range of 0–2 The higher the score, the greater the asset ownership
Electronic empowerment	Response to ownership of mobile phone & using SMS Services for financial transactions. Ownership of a mobile phone scored into yes (1) if owned and no (0). Use of mobile phone for financial transactions categorized into yes (1) and no (0)- owned a mobile phone, but not using for financial transactions. Scoring: Total scores added of all three individual responses with score range of 0–2 The higher the score, the greater the electronic empowerment
Insurance	Categorized ‘yes’ if has health insurance or covered by any ‘health scheme’ and ‘no’ otherwise

Data management and analysis

Access to the data was granted following submission of a written proposal to the DHS program. Data were saved in a password protected folder on a secure Monash University server. Univariable comparisons were made between groups who had and had not experienced IPV during a pregnancy and each characteristic. Multivariable logistic regression was used to identify the constellation of factors that were significantly associated with physical IPV in pregnancy experienced by women in India. Analyses were adjusted for study design including stratification, primary sampling unit, and sampling weight pre-specified in the dataset. A p value below 0.05 was used to indicate statistical significance. Missing data were treated using a pairwise deletion method in which only the missing values were excluded from the specific analyses, rather than removing the entire case form the dataset. All the analyses were performed using STATA Version 16 [62].

Ethics and consent

The NFHS- 5 survey protocol was approved by the ethics review board of the International Institute for Population Sciences, Mumbai, India. Participants gave written consent to contribute data to the survey. This study used only de-identified data. Separate ethics approval for this secondary analysis was not required. This study was conducted in accordance with the Declaration of Helsinki.

Results

Sample

In total, data contributed by 60,579 women who had ever been married or in a union and ever been pregnant were included in analyses. Most (70%) of the women were aged at least 30 years. Only 30% were employed in income-generating work. Two fifths had first been married before the age of eighteen. Two thirds of the sample did not have access to a radio, newspapers or magazines or the internet. One third of the women had health insurance and 80% had an account in a bank or a financial institution. The sociodemographic profile of the sample is described in detail in Table 2.

Table 2.

Characteristics of the study participants

	N (%)
Total	60,579
Individual factors
Age (in years)
18–29	18,668 (30.8)
30–39	24,657 (40.7)
40+	17,254 (28.5)
Age at marriage (in years)
< 18	24,099 (39.8)
≥ 18	36,480 (60.4)
Education
No education	18,182 (30.0)
Primary	8,957 (14.8)
Secondary	27,348 (45.1)
Higher	6,092 (10.1)
Religion
Hindu	45,995 (75.9)
Muslim	7,234 (11.9)
Others/no religion	7,350 (12.1)
History of pregnancy loss/termination
No	50,500 (83.4)
Yes	10,079 (16.6)
Ever witnessed father beat her mother
No	47,835 (79.0)
Yes	10,924 (18.0)
Do not know	1,820 (3.0)
Number of Living Children
0–1	14,111 (23.3)
2–3	37,064 (61.2)
4+	9,404 (15.5)
Relationship factors
Not currently living with a partner	3034 (5.0)
Year of cohabitation (in years)
<5	7372 (12.2)
5–10	13,870 (22.9)
11–20	22,603 (37.3)
20+	16,734 (27.6)
Partner’s education
No education	11,157 (18.4)
Primary	8,922 (14.7)
Secondary	32,233 (53.2)
Higher	8,048 (13.3)
Do not know	219 (0.4)
Partner’s alcohol drinking
No	43,541 (71.9)
Yes	17,038 (28.1)
Husband/partner’s occupation
Professional/technical/managerial AND Clerical	5803 (9.6)
Sales’ Services/household and domestic; Agricultural; Skilled and unskilled manual; and Other	54,776 (90.4)
Income difference
More than him	3,260 (20.1)
Less than him	8,721 (53.8)
About the same	3,185 (19.7)
Husband/partner does not bring in money	825 (5.1)
Do not know	205 (1.3)
Family factors
Wealth index
Poorest	13,485 (22.3)
Poorer	13,789 (22.8)
Middle	12,583 (20.8)
Richer	11,218 (18.5)
Richest	9,504 (15.7)
Sex of household head
Male	50,933 (84.1)
Female	9,646 (15.9)
Societal factors
Ethnicity
Schedule Caste	11,546 (19.1)
Scheduled Tribe	11,575 (19.1)
Other Backward Class	23,274 (38.6)
None of the above	14,084 (23.2)
Usual residence
Urban	14,431 (23.8)
Rural	45,037 (74.3)
Not living at usual residence	1,111 (1.8)
Gender Empowerment factors
Acceptance of IPV
No	38,479 (63.5)
Yes	22,100 (36.5)
Decision Making
Who usually decides on large household purchases
Woman alone	4,238 (7.4)
Woman and partner	42,249 (73.4)
Partner alone	9,581 (16.6)
Someone else	1,093 (1.9)
Other	384 (0.7)
Who usually decides on woman’s health care
Woman alone	5,624 (9.8)
Woman and partner	41,757 (72.6)
Partner alone	9,466 (16.4)
Someone else	476 (0.8)
Other	222 9 (0.4)
Who usually decides on visits to family or relatives
Woman alone	4,545 (7.9)
Woman and partner	43,009 (74.7)
Partner alone	8,949 (15.6)
Someone else	770 (1.3)
Other	272 (0.5)
Media Access
Listening to radio
Not at all	53,053 (87.6)
Less than once a week	5,291 (8.7)
At least once a week	2,235 (3.7)
Watching television
Not at all	18,065 (29.8)
Less than once a week	12,921 (21.3)
At least once a week	29,593 (48.9)
Reading newspaper or magazine
Not at all	45,000 (74.3)
Less than once a week	9,412 (15.5)
At least once a week	6,167 (10.2)
Internet use
Never	44,738 (73.9)
Yes	15,841 (26.1)
Asset Ownership
Owns a house alone or jointly
Does not own	29,884 (49.3)
Alone only	9,190 (15.2)
Jointly only	10,412 (17.2)
Both alone and jointly	11,093 (18.3)
Owns land alone or jointly
Does not own	36,805 (60.8)
Alone only	6,139 (10.1)
Jointly only	8,437 (13.9)
Both alone and jointly	9,198 (15.2)
Electronic empowerment
Owns a mobile telephone
No	25,779 (42.6)
Yes	34,800 (57.4)
Use of mobile telephone for financial transactions
No	28,306 (81.3)
Yes	6,494 (18.7)
Insurance
No	39,740 (65.6)
Yes	20,839 (3.4)

Prevalence and subgroup comparisons of physical IPV experienced by women during pregnancy

The lifetime prevalence of physical IPV experienced by ever married/partnered women during any pregnancy was 2.4%.

We made comparisons between groups who had or had not experienced physical IPV in pregnancy and variables classified as related to individual, relationship, family, societal, and gender empowerment related factors (Table 3).

Table 3.

Prevalence and subgroup comparisons of physical IPV experiences during pregnancy

	No (%)	Yes (%)
Total	59,133 (97.6)	1446 (2.4)
Individual factors
Age (in years)
18–29	18,252 (97.8)	416 (2.2)
30–39	24,089 (97.7)	568 (2.3)
40+	16,792 (97.3)	462 (2.7)
Age at marriage (in years)
< 18	23,392 (97.1)	707 (2.9)
≥ 18	35,741 (98)	739 (2)
Education
No education	17,635 (97)	547 (3)
Primary	8,719 (97.3)	238 (2.8)
Secondary	26,763 (97.9)	585 (2.1)
Higher	6,016 (99.8)	76 (1.2)
Religion
Hindu	44,854 (97.5)	1,141(2.5)
Muslim	7,058(97.6)	176(2.4)
Others/no religion	7221 (98.2)	129 (1.8)
History of pregnancy loss/termination
No	49,415(97.9)	1,085(2.1)
Yes	9,718(96.4)	361(3.6)
Ever witnessed father beat her mother
No	47,104(98.5)	731(1.5)
Yes	10,245(93.8)	679(6.2)
Don’t know	1,784(98)	36(2)
Number of Living Children
0–1	13,822(97.9)	289(2.1)
2–3	36,182(97.6)	882(2.4)
4+	9,129(97.1)	275(2.9)
Relationship factors
Duration of cohabitation (In years)
< 5	7239(98.2)	133(1.8)
5–10	11,084(97.8)	246(2.2)
10–20	22,394(97.8)	515(2.2)
20 +	18,416(97.1)	552(2.9)
Partner’s education
No education	10,780(96.6)	377(3.4)
Primary	8,660(97.1)	262(2.9)
Secondary	31,522(97.8)	711(2.2)
Higher	7,957(98.9)	91(1.1)
Do not know	214(97.7)	5(2.3)
Partner’s alcohol drinking
No	43,025(98.8)	516(1.2)
Yes	16,108(94.5)	930(5.5)
Partner’s occupation
Professional/technical/managerial AND Clerical	5712(98.4)	91(1.6)
Sales’ Services/household and domestic; Agricultural; Skilled and unskilled manual; and Other	53,421(97.5)	1355(2.5)
Income difference
More than him	3,102(95.2)	158(4.8)
Less than him	8,429(96.7)	292(3.3)
About the same	3,115(97.8)	70(2.2)
Husband/partner does not bring in money	801(97.1)	24(2.9)
Do not know	199(97.1)	6(2.9)
Family factors
Wealth index
Poorest	13,092(97.1)	393(2.9)
Poorer	13,392(97.1)	397(2.9)
Middle	12,266(97.5)	317(2.5)
Richer	11,004(98.1)	214(1.9)
Richest	9,379(98.7)	125(1.3)
Sex of household head
Male	49,784(97.7)	1,149(2.3)
Female	9,349(96.9)	297(3.1)
Societal factors
Ethnicity/Caste
Schedule Caste	11,199 (97.0)	347(3.0)
Schedule Tribe	11,334 (97.9)	241(2.1)
Other Backward Classes	22,785 (97.5)	589(2.5)
None of the above	13,815(98.1)	269(1.9)
Usual residence
Urban	14,117(97.8)	314(2.2)
Rural	43,941(97.6)	1,096(2.4)
Not living at usual residence	1,075(96.8)	36(3.2)
Gender empowerment factors
Acceptance of IPV
No	37,842(98.3)	637(1.7)
Yes	21,291(96.3)	809(3.7)
Decision making
Person who usually decides on large household purchases
Woman alone	4,070(96)	168(4)
Woman and partner	41,559(98.4)	690(1.6)
Partner alone	9,201(96)	380(4)
Someone else	1,071(98)	22(2)
Other	372(96.9)	12(3.1)
Person who usually decides on woman’s health care
Woman alone	5,401(96)	223(4)
Woman and partner	41,094(98.4)	663(1.6)
Partner alone	9,098(96.1)	368(3.9)
Someone else	466(97.9)	10(2.1)
Other	214(96.4)	8(3.6)
Person who usually decides on visits to family or relatives
Woman alone	4,364(96)	181(4)
Woman and partner	42,309(98.4)	700(1.6)
Partner alone	8,581(95.9)	368(4.1)
Someone else	756(98.2)	14(1.8)
Other	263(96.7)	9(3.3)
Media Access
Frequency of listening to radio
Not at all	51,799(97.6)	1,254(2.4)
Less than once a week	5,157(97.5)	134(2.5)
At least once a week	2,177(97.4)	58(2.6)
Frequency of watching television
Not at all	17,614(97.5)	451(2.5)
Less than once a week	12,5809(97.4)	341(2.6)
At least once a week	28,939(97.8)	654(2.2)
Frequency of reading newspaper or magazine
Not at all	43,902(97.6)	1,098(2.4)
Less than once a week	9,184(97.6)	228(2.4)
At least once a week	6,047(98.1)	120(1.9)
Use of internet
Never	43,556(97.4)	1,182(2.6)
Yes	15,577(98.3)	264(1.7)
Asset Ownership
Owns a house alone or jointly
Does not own	29,178(97.6)	706(2.4)
Alone only	8,906(96.9)	284(3.1)
Jointly only	10,166(97.6)	246(2.4)
Both alone and jointly	10,883(98.1)	210(1.9)
Owns land alone or jointly
Does not own	35,935(97.6)	870(2.4)
Alone only	5,955(97)	184(3)
Jointly only	8,236(97.6)	201(2.4)
Both alone and jointly	9,007(97.9)	191(2.1)
Electronic Empowerment
Owns a mobile telephone
No	25,075(97.3)	704(2.7)
Yes	34,058(97.9)	742(2.1)
Use mobile telephone for financial transactions
No	27,686(97.8)	620(2.2)
Yes	6,372(98.1)	122(1.9)
Insurance
No	38,780(97.6)	960(2.4)
Yes	20,353(97.7)	486(2.3)

Among the individual factors, prevalence did not vary greatly across the different age groups, but women with younger age at marriage and with low literacy experienced higher prevalence. Higher prevalence was also observed in women who had experienced prior termination or loss of pregnancy. Similarly, women, who witnessed their fathers beating their mothers when they were children experienced higher rates than women who had not.

In relationship related factors, higher proportions of women whose partners had poor educational status or who drank alcohol experienced violence during pregnancy. Women earning more than their husband/partner experienced more violence than those who did not. In relation to family and societal factors, women belonging to a family with a higher wealth index were less likely to experience violence and women from a scheduled caste background experienced more violence than others.

Among the gender empowerment factors, women who held beliefs accepting of IPV reported higher rates of IPV than their counterparts. Women with higher decision making,.,, higher electronic empowerment and access to insurance reported less extent of violence. Better access to media was associated with an increased likelihood of experiencing violence.

Factors associated with intimate partner physical violence among women in pregnancy

Results of the multiple logistic regression analysis are presented in Table 4. The odds of experiencing physical violence in pregnancy were higher if women were younger (< 18 years) at their first marriage (aOR 1.17, 95% CI 1.03–1.32) belonged to the Muslim rather than the Hindu religion (aOR 1.58, 95% CI 1.32–1.90), or had experienced a previous pregnancy termination (aOR 1.52, 95% CI 1.34–1.73). Women who had witnessed as a child, their father perpetrate violence against their mother were more likely to experience violence than those who had not (aOR 2.84, 95% CI 2.54–3.17).

Table 4.

Multivariate analysis of factors influencing IPV disclosure among pregnant women

	Odds ratio	p-value	95% CI
Individual factors
Age (in years)
18–29	Ref.
30–39	1.03	0.789	0.85	1.24
40+	1.12	0.411	0.86	1.46
Age at marriage (in years)
< 18	Ref.
18+	1.17	0.014	1.03	1.32
Education
No education	1.06	0.705	0.77	1.46
Primary	1.06	0.705	0.77	1.46
Secondary	1.05	0.748	0.79	1.38
Higher	Ref.
Religion
Hindu
Muslim	1.58	< 0.001	1.32	1.90
No religion/Others	0.88	0.219	0.72	1.08
History of pregnancy loss/termination	1.52	< 0.001	1.34	1.73
Ever witnessed father beat her mother	2.84	< 0.001	2.54	3.17
Number of Living Children
0–1	Ref.
2–3	0.98	0.817	0.84	1.15
4+	1.11	0.313	0.91	1.36
Relationship factors
Years of cohabitation
< 5 years	Ref.
5–10 years	1.04	0.730	0.83	1.31
11–20 years	1.01	0.931	0.77	1.34
20 years +	1.15	0.438	0.81	1.63
Partner’s education
No education	1.53	0.004	1.15	2.04
Primary	1.46	0.009	1.10	1.94
Secondary	1.39	0.010	1.08	1.79
Higher	Ref.
Partner’s alcohol drinking	4.07	< 0.001	3.61	4.57
Partner’s occupation
Professional/technical/managerial AND Clerical	Ref.
Sales’ Services/household and domestic; Agricultural; Skilled and unskilled manual; and Other	0.95	0.690	0.75	1.21
Income difference
Women not working	Ref.
Husband/partner does not bring in money	1.33	0.147	0.91	1.94
Less than or same him	1.17	0.022	1.02	1.33
More than him	1.92	< 0.001	1.60	2.30
Family factors
Wealth index
Poorest	1.74	< 0.001	1.34	2.27
Poorer	1.70	< 0.001	1.33	2.17
Middle	1.39	0.006	1.10	1.76
Richer	1.12	0.335	0.89	1.42
Richest	Ref.
Sex of household head
Male	Ref.
Female	1.35	< 0.001	1.18	1.55
Societal factors
Ethnicity/Caste
Schedule caste	1.04	0.694	0.87	1.24
Schedule Tribe	0.74	0.003	0.60	0.90
OBC	1.01	0.915	0.86	1.18
No caste/tribe	Ref.
Residence
Rural
Urban	1.13	0.104	0.98	1.31
Not a resident	1.44	0.044	1.01	2.05
Gender Empowerment factors
Acceptance of IPV	1.54	< 0.001	1.38	1.72
Decision making	0.78	< 0.001	0.74	0.81
Media Access	1.13	0.001	1.05	1.21
Asset ownership	1.00	0.968	0.94	1.06
Electronic empowerment	0.96	0.373	0.87	1.05
Insurance	0.83	0.002	0.74	0.93

Women with partners who had no or little formal education (aOR 1.53, 95% CI 1.15–2.04) and who used alcohol (aOR 4.07, 95% CI 3.61–4.57) were at increased the odds of experiencing violence. Women in a low socioeconomic position (aOR 1.74, 95% CI 1.34–2.27) were more likely to be subjected to violence compared to women in the richest wealth quintiles. There was higher likelihood of violence among women from the female led household (aOR 1.35, 95% CI 1.18–1.55) versus male led households.

Women who held attitudes accepting of violence (aOR 1.54, 95% CI 1.38–1.72) were significantly more likely to experience violence compared to those who did not accept violence. Gender empowerment related factors such as having higher total scores on the decision-making autonomy (aOR 0.78, 95% CI 0.74–0.81) and having insurance (aOR 0.83, 95% CI 0.74–0.93) were less likely to experience physical violence than their counterparts. In contrast, women with higher levels of media access (aOR 1.13, 95% CI 1.05–1.21) were more likely to report experiencing violence than women with lower media access.

Discussion

This study reports data from the most recent survey of a representative sample of women recruited from every state and union territory in India on the extent of physical IPV experienced during pregnancy, and with a broad set of correlates encompassing individual, family, societal, relationship related variables and noteworthy inclusion of gender-based factors. These data reveal a relatively low prevalence, but because of the large population in India the absolute numbers of women experiencing physical violence in pregnancy are alarmingly high. Physical violence during pregnancy was significantly more common among women experiencing diverse social and economic disadvantages including low household wealth, being a member of a scheduled caste, or a religious minority, having a partner who was not educated or employed, or abused alcohol. It was less common among women who were more empowered including having more decision-making autonomy.

Our findings are consistent with the NFHS-4 data collected in 2015-16, analysed using a similar method and reported by Jungari et al. [55]. In that survey there were slightly higher lifetime rates (3.3%) of physical IPV during pregnancy, which might indicate that there has been a reduction in the perpetration of IPV during pregnancy in India over the last 5 years. Indian DHS surveys before NFHS-4 did not collect information on IPV limiting potential for establishing longer term trends. Nevertheless, it is also important to emphasize that reporting of national prevalence obscures the regional, state or district level variations of IPV experienced by women during pregnancy in India. This is particularly relevant for countries like India with marked geographical, cultural, and socioeconomic variations between different states and regions of the country [63, 64].

These estimated prevalence rates are much lower than the reported prevalence (9–31%) of physical IPV during pregnancy in site-specific studies in India [46, 47, 52, 53, 57, 65]. We hypothesize that enquiry by or with the assistance of known treating health professionals, often encountered in the care of clinic-based venues, leads to relatively more supportive environment to explore sensitive subject such as IPV. This may assist higher disclosure of IPV by women. In addition, relatively small sample size in these settings may results in overestimation of prevalence estimates in such situations. It is also important to note that almost all previous studies except Peedicayil et al. [52] assessed violence during the most recent pregnancy. However, a recent randomly selected sample in an obstetric clinic-based study Nath et al. [56] found that only 3.7% reported experiencing any type of violence during a pregnancy, but they did not report whether this was in their entire lifetime or just during the index pregnancy.

It is possible that despite the confidentiality and ethics protocols, the NFHS household interview method does not create a sufficient sense of safety for informants [66, 67]. Under reporting may also be a result of the fear of disclosure related potential repercussions along with a sense of individual and family shame, and absence of therapeutic or care relationship with a data collector or enumerator in these surveys. Low reported rates could also be due to a lack of differentiation of certain behaviours as transgressive or acceptance of violence among women in a society with traditional gender roles, particularly in LMICs including India [68]. Together, these suggest that the prevalence of physical IPV might vary between study settings or by methods of data collection.

The findings of this study confirm the significant associations between social and economic disadvantages at the individual (belonging to a religious minority) and family (low household economic status) and societal domain (scheduled caste) with higher risks of IPV during pregnancy are consistent with the results of previous studies [49, 50, 52, 53, 55, 69].

Higher likelihood of violence in women with younger age of marriage (< 18) further supports the association of higher risk of IPV in socially and economically disadvantaged settings. Underage marriages are more prevalent in rural and poor socioeconomic regions of India. Underage marriages also restrict the abilities of women to attain education and employment opportunities, increase power imbalances between marriage partners, further increasing women’s vulnerabilities and the risk of victimization. There are also important roles of not only cultural and traditional factors but also the safety and security concerns for the female child. Child marriage has also been previously associated with higher risk of physical IPV in India [70]. We also consider that child marriage itself, is a form of violence towards women and that it reflects undervaluing of women in a male dominated society like India.

The finding of markedly increased risk in women with husband’s alcohol use is in consonance with a consistent finding from the previous studies from India [49, 50, 52, 53, 55, 65, 69] and other countries such as Nigeria, Ethiopia, China [11, 14, 18]. Various suggested mechanisms are the disinhibitory effect of alcohol, reduced ability for conflict resolution and increased financial burden on the family which facilitate violence towards women.

We found that higher education of the husband/partner is protective, in accordance with the previously reported findings [45, 50, 57, 69] who found increased odds of violence in women with husband/partner of low education attainment. This could be due to better knowledge of husband about wrongfulness in committing violence, more respectful relationship, higher psychological maturity and better emotion regulation skills It speaks to the theories about male sense of shame and inadequacy as a factor leading to perpetration of violence and that perhaps more mature educated men do not feel these or act on them.

Women earning more than their husbands were at highest risk of exposure to physical violence in pregnancy compared to women with no income. Similarly, we wonder if female led household led to increase vulnerability. We speculate if increased female empowerment in a male dominated society is perceived as a threat by males with retribution by males towards gender equality by perpetrating violence. Use of physical violence to control and maintain their dominant position may leads to justification of wife beating with increased perpetration of IPV against women. It is not possible to draw causal conclusions from cross-sectional data. On one hand it is possible that in female led households, women felt safer to disclose experiences of violence. On the other hand, It is possible that women who have experienced IPV have separated and so are leading their household.

Higher odds of violence in women who are accepting of IPV in comparison to women who are non-accepting might reflect the prevailing gender inequality and subordinate position of women in Indian society. It is also possible that higher risk in women who witnessed parental violence in childhood, may have contributed to normalization of physical violence leading to increasing their own risk in adulthood, apart from influencing development of their attitude towards IPV. This is consistent with the concept of intergenerational transmission of violence with development of more tolerant views towards IPV with higher risk of experiencing or perpetrating IPV [71–73].

The protective role of decision making, and health insurance status underlines the potential importance of gender empowerment related factors in addressing violence against women. It emphasizes the contribution of gender disadvantage to the heightened risk of violence experienced by women. Electronic empowerment and asset ownership, other indirect measures of gender enfranchisement show similar but non-significant trend. Increased risk of violence with better media access may be reflective of improved reporting with better knowledge and awareness of gender-based violence, their rights, and resources. Alternatively, increased demands of gender equality with better understanding of women with higher media access results in heightened risk of violence in a patriarchal society [74].

These findings confirm the heightened vulnerability and dependence experienced by women in deprived circumstances with increased risk of exploitation and reduced opportunities for help seeking. Further, they highlight the importance of identifying and addressing family and societal disadvantages, which cannot be resolved by women individually. They highlight the need to identify and address partner’s alcohol use, as a meaningful strategy to reduce the risk of physical violence towards women in pregnancy. Reducing access and affordability of alcohol may be helpful but it eventually calls for change in men’s behaviour. This finding also confirms the presence of factors beyond women’s control to reduce the risk of victimization. Similarly, enhancing women’s empowerment and their social position in the Indian society is an important strategy to diminish the hazard of gender-based violence. Overall, our findings support the complex interplay between individual and family factors, shaped by the cultural and social norms in influencing the risk of physical violence experienced by women during pregnancy in India.

Implications

This work contributes national estimates with the examination of diverse determinants including gender related factors at different contextual levels to formulate a framework for the identification of women at risk for and prevention of physical violence in the heightened vulnerable state of pregnancy. The findings support calls for screening and early interventions to address IPV during the perinatal period including during pregnancy in India, consistent with the recommendation of WHO perinatal care and existing protocols in high income countries [75–77]. This evidence will inform the policy and programs with targeted interventions incorporating the principles of proportionate universalism. Effective structural socioeconomic interventions [78], development and scaling of locally developed programs, like SASA! Activist Kit to Prevent Violence Against Women [79] in India is urgently needed. Increased contact with health professionals during the perinatal period is an opportunity to intervene for reducing the risk of negative maternal and child outcomes, in addition to preventing IPV.

Besides screening and interventions, it is important to target training of community health professionals operating at ground level to equip them with required skills to identify and address IPV. Moreover, cross-sectoral coordination and approaches beyond the health sector are warranted. Policies and programs to reduce poverty and improve the educational and economic status of women, in addition to promotion of gender equity, will help to reduce and address the important public health concerns of IPV in general and during the pregnancy.

It is important to recognize that life prevalence rates of physical violence during pregnancy may not accurately reflect the frequency and severity of experiences of DV that can impact negatively on women’s health and well-being. However, given the strong association of IPV during pregnancy with perinatal mental health and high disease burden of PMDs and suicide among women in India [80–83], these findings, coupled with high prevalence of lifetime experience of DV by women in India [41], makes IPV in pregnancy, a matter of national significance for public health and maternity care systems. It further reinforces the need for an integrated maternity care system with inclusion of maternal mental health and violent victimization in the national maternity and mental health policy and programmes [84, 85].

There is a need for further research on IPV during pregnancy in India using sound methodology and incorporation of domains other than physical IPV, such as sexual and emotional violence, financial abuse and coercive control experienced by women with investigation in longitudinal cohort studies of their impact on maternal and child health outcomes. We call for inclusion of these realms in pregnancy specific IPV data in future NFHS surveys, similar to the current exploration of IPV in the non-perinatal periods. There is a need for regular data collection and monitoring, especially from other known or trusted sources, to monitor prevalence and trends of IPV during pregnancy among women in India.

Strengths and limitations

The current analysis has various strengths. It used a large data set from a nationally representative sample collected using a standardized methodology across the country, with considerable confidence about the generalizability of the findings. A wide range of factors derived from well recognized frameworks of IPV at different contextual levels were assessed and explored. Additionally, this study addressed the identified knowledge gaps on the potential role of gender related correlates about physical violence during pregnancy in India.

We nevertheless acknowledge some limitations. Self-reporting of physical IPV with proneness for recall bias along with fear and stigma about disclosure may have led to underreporting with concerns about the true extent of the population prevalence estimates. Causative conclusion about the determinants is not possible with the cross-sectional design only enabling associations to be identified. Similarly, we could only use pre-determined variables for analysis, which may not be the full list of factors related to physical IPV during pregnancy. For example, we were unable to assess the role of maternal/family preference for a male child. However, we believe that we analysed a thorough list of possible correlates with in the restrictions of available data.

Conclusion

This study documents physical IPV during pregnancy experienced by women across India. Our study highlights the potential contribution of social and gender-related correlates, including many beyond individual control, with higher prevalence of physical IPV in pregnancy among women from socially and economically underprivileged backgrounds and those experiencing higher gender disadvantage. Nevertheless, there is a priority need for further research to capture the total picture of not only physical but any type of IPV during pregnancy experienced by women in India.

Authors’ contributions

HK, TT and JF conceptualized this study. HK and TT managed the data. HK, TT, HN conducted the analysis. HK wrote the first draft of this paper. TT, JF, PC and HS provided critical inputs. All authors reviewed and agreed to the final draft.

Funding

No funding was received for this study. TT is supported by an Early Career Fellowship from the National Mental Health Research Council, Australia. JF is supported by the Finkel Professorial Fellowship which receives funding from The Finkel Family Foundation.

Data availability

This is a secondary analysis of data collected in the NFHS-5. The dataset supporting the conclusions of this article is included within the article. Details of the design and data collection of the NFHS-5 are available on https://www.dhsprogram.com/pubs/pdf/FR375/FR375.pdf.

Declarations

Ethics approval and consent to participate

The NFHS- 5 survey protocol was approved by the ethics review board of the International Institute for Population Sciences, Mumbai, India. Participants gave written consent to contribute data to the survey. This study used only de-identified data. Separate ethics approval for this secondary analysis was not required. This study was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.