Violence against pregnant women: prevalence and characteristics. A population-based study in Nicaragua

Authors


E. Valladares, Department of Epidemiology and Public Health, Umeå University, S-901 85 Umeå, Sweden.

Abstract

Objective  This study aims to estimate the prevalence and characteristics of partner abuse during pregnancy as well as to investigate associated social factors in León, Nicaragua.

Design  Cross-sectional community-based study.

Setting  All pregnant women from 50 randomly selected geographical clusters out of 208 in the municipality of León, Nicaragua.

Sample  A total of 478 pregnant women were included; only one woman refused to participate.

Method  The domestic violence questionnaire from the WHO-co-ordinated Multi-Country Study on Women's Health and Life Events was used with each participant being interviewed twice during pregnancy.

Main outcome measures  Prevalence and characteristics of partner violence during pregnancy.

Results  The prevalence of emotional, physical and sexual abuse during pregnancy was 32.4%, 13.4% and 6.7%, respectively. Seventeen percent reported experience of all three forms of violence. Two-thirds of the victims reported repeated abuse. Half of the abused women had experienced punches and kicks directed towards the abdomen and 93% had been injured. Most women had not sought health care in relation to the abuse, but those who did were usually hospitalised. Factors such as women's age below 20 years, poor access to social resources and high levels of emotional distress were independently associated with violence during pregnancy.

Conclusion  Violence against pregnant women in Nicaragua is common and often repeated. Although these women have poor access to social resources and high levels of emotional distress, they are rarely assisted by the health services. Innovative strategies are needed to provide support and counselling.

INTRODUCTION

Violence against women is a human rights violation and represents a serious medical, social and legal problem. Partner violence is increasingly reported during pregnancy. Victims have significantly greater risk of miscarriage, abruptio placenta, preterm delivery, perinatal mortality and low birthweight.1–4

Estimates of prevalence during pregnancy range from 0.9% to 29%5–10 depending on definitions and methods of assessment. The pregnancy itself has been suggested as a time of increased risk for abuse,11 although this is questioned by other authors.12

To date, the majority of studies carried out with pregnant women have been clinical or hospital based1,13,14 while the population-based studies have not focussed on pregnant women. In some studies of women of childbearing age, retrospective information on violence in past pregnancies has been reported.15,16 Therefore, although a considerable number of research reports have been published on this subject, the full extent of this problem is still not clear.

In Nicaragua, partner violence against women is reportedly common and supported by some cultural and religious norms in society.17 However, there is still little information about the frequency of abuse during pregnancy. This study aims to estimate the prevalence and characteristics of partner abuse during pregnancy as well as to explore associated social factors.

SUBJECTS AND METHODS

A cross sectional community base study was carried out in the municipality of León, the second largest city in Nicaragua. The survey was based on an ongoing demographic surveillance system, covering 50 randomly selected geographical clusters that included 11,000 households from urban and rural communities of the municipality of León (corresponding to 22% of the total population). From this, a sample of 483 pregnant women was identified by inserting a question regarding the presence of any women in the household who were thought to be pregnant. Once the pregnancy was confirmed (blood/urine test or ultrasound) the woman was invited into the study. Two households contained two pregnant women living in the same house. Due to the sensitivity of the issue only one woman per household was interviewed, the youngest in one case and the oldest in the other. One woman declined to participate and two women were found not to be pregnant. Thus, a total of 478 pregnant women were included in the study.

The prevalence and characteristics of violence were measured with the domestic violence questionnaire from the WHO Multi-Country Study on Women's Health and Life Events.18 It includes questions on all dimensions of partner violence (emotional, physical and sexual), injuries and gender roles in society as well as questions that provide context to the violent acts. Physical violence in the WHO questionnaire is approached by a modification of the Conflict Tactic Scale, which assesses specific behavioural acts of aggression. As the WHO instrument was not developed for the evaluation of events in pregnant women, we related every question to the time of the current pregnancy. Emotional abuse was defined as repeated yelling, humiliating or threatening acts by the male partner during the index pregnancy, while physical abuse as one or more intentional acts of physical aggression perpetrated by the male partner during the current pregnancy with the potential of causing harm, injury or death. Sexual abuse was defined as the use of force, coercion or psychological intimidation by the partner to force the woman to engage in a sex act against her will. Socio-economic status was measured through the Unsatisfied Basic Need Assessment, widely used and tested in the Nicaraguan setting.19 This assessment included four indicators: school enrolment among minors, housing conditions, access to sanitary services and economic dependency. Emotional distress was identified by means of the Self-Reported Questionnaire (SRQ-20), which includes 20 items representing symptoms associated with emotional distress. We defined distress as a score of ≥7 as recommended for Nicaragua by a validation study.20

Social resources, the resources that an individual has at his/her disposal to handle the demands of daily life, were measured using a questionnaire developed by Hanson et al.21,22 and adapted to a pregnant population.23 This assessment is built on two main concepts. First, the social network (the characteristics of social relationships surrounding the individual) was estimated through the measurement of social anchorage and social participation. Second, social support (defined as information leading the subject to believe that she/he is cared for, loved and belongs to a network of mutual obligations) was measured by means of emotional, instrumental and child's father support, maternal or parental support and control of daily life. Job support, included in the original scale, was not used in this study because more than 90% of participating women were housewives.

Questions on general and reproductive health were also included. The questionnaire was piloted in a group of pregnant women and minor corrections made.

Data were collected by three trained female interviewers and their assessments were standardised. All forms were reviewed by a field supervisor and inspected by the principal investigator (EV). Forms with missing data or inconsistencies were returned to the field for correction. Random control interviews were performed in order to assess the validity. Double data entry was performed a random sample of 10%.

Two interviews were performed with every respondent in order to enhance the disclosure of violence and as a follow up of pregnancy outcome and complications. The first visit was made immediately after identification and confirmation of pregnancy (on the average at 19 weeks). The study was presented to the households with the aim to explore the health of pregnant women. All interviews were done in complete privacy and the subject of abuse was revealed only to the interviewee. If privacy was not possible to achieve, women were offered to choose for another time or place, in four cases interviewers visited the woman twice or three times until man was not at home. Interviewers were provided with alternative topics if somebody arrived during the interview. The full questionnaire including scales to measure violence, availability of social resources, self-report questionnaire, unsatisfied basic need assessment and questions on general and reproductive health was applied to all 478 eligible women. All women received a second visit, except nine in the urban area and four in the rural area, who had migrated out from the study area. The mean gestational age at the second visit was the 36th week, and 49% of participating women had already delivered. This interview included repeated screening for violence as well as enquiring about pregnancy complications and recording maternal and perinatal outcomes. For those still pregnant the perinatal outcome was excerpted from medical registers of the hospital or primary health care centres. Counselling services (psychological, medical or legal) were offered to participating women with experience of violence, and 17% of them attended. A clinical psychologist was recruited to individually approach those who had attempted suicide during the current pregnancy. The first two visits were performed in the women's own houses, while any further appointments were made at the ‘Centre for Psychological Attention’. A session of peer debriefing and psychological support was organised for the field workers every week.

One trained data clerk entered and cleaned the data. Data were stored and logical validation checks were constructed in Access 2000 databases. Violence was considered to have occurred if reported in either visit. Data cleaning, descriptive as well as multivariate analyses were performed by using SPSS software version 11.0 (SPSS, Chicago, Illinois, USA). Prevalence of the different types of violence as well as characteristics of the abuse during the period of pregnancy was analysed, followed by the construction of a multivariate model, which included potential confounding or explanatory variables.

The study applied the WHO ethical guidelines for research on domestic violence.24 Informed and signed consent was obtained from each respondent at both visits. Ethical review and clearance was obtained from the Research Ethics Committee of León University, Nicaragua, and from the Research Review Committee at the Medical Faculty, Umeå University, Sweden.

RESULTS

A total of 478 pregnant women were included in the study. More than 40% of the respondents and their partners were either illiterate or had not completed primary school. Eighty percent of the pregnant women were living with the father of the child at the time of the second visit and 39% were primiparas (Table 1). One hundred and fifty-five women (32%) reported partner abuse during the current pregnancy (Table 2). All of these had been emotionally abused, while 64/478 (13%) had been physically abused, and 32/478 (7%) sexually abused. Seventeen percent of the abused women (26/155) had been victims to all three types of abuse during the current pregnancy. The overlap between the different types of abuse, presented in Fig. 1, indicates the complexity and severity of the problem.

Table 1.  Description of the total sample of pregnant women and women experiencing violence during the index pregnancy.
CharacteristicsLevelTotal pregnant women (N= 478)Type of abuse during the index pregnancy
n%Emotional (N= 154), n (P)Physical (N= 64), n (P)Sexual (N= 32), n (P)
Mother's age (years)≤19 years13728.6632616
≥20 years34171.491 (0.000)38 (0.02)16 (0.008)
 
Father's age (years)<204910.323125
>1942889.7131 (0.02)52 (0.02)27 (0.3)
 
Mother's education≤3rd. grade of Primary22446.9803518
>3rd. grade of Primary25453.174 (0.1)29 (0.1)14 (0.2)
 
Father's education≤3rd. grade of Primary19641763218
>3rd. grade of Primary2825978 (0.13)32 (0.13)14 (0.09)
 
Marital statusAlone8116.92476
With partner39780.1130 (0.69)57 (0.21)26 (0.8)
 
ParityPrimiparous18538.7592112
Multiparous29361.395 (0.9)43 (0.3)20 (1.0)
 
Place of residencyRural14730.8582011
Urban33169.296 (0.02)44 (1.0)21 (0.69)
 
Economic conditionsLow27357.1943920
High20542.960 (0.27)25 (0.50)12 (0.58)
 
Unwanted pregnancyYes22246813925
No2565473 (0.01)25 (0.00)7 (0.00)
 
Antenatal controlsYes15532.4603116
(Delayed booking or no AC)No32367.694 (0.03)33 (0.04)16 (0.03)
 
Social resourcesLow17236713818
High3066483 (0.002)26 (0.00)14 (0.02)
 
Perceived emotional distress
(SRQ ≥ 7)Yes19140874321
(SRQ ≤ 6)No2876067 (0.000)21 (0.000)11 (0.003)
Table 2.  Prevalence and description of abuse against pregnant women.
 Current pregnancyLifetime
  • *

    Because of multiple responses the addition of the percentages is not 100.

Any act of violence(n= 478)(n= 478)
 155 (32%)259 (54%)
 
Types of Violence(n= 478)(n= 478)
Any act of emotional violence155 (32%)252 (53%)
Any act of physical violence64 (13%)148 (31%)
Any act of sexual violence32 (7%)72 (15%)
 
Reported acts of physical violence*(n= 64)(n= 148)
Pushed55 (86%)131 (89%)
Slapped35 (55%)89 (60%)
Beat with fist32 (50%)90 (61%)
Punched or kicked in the abdomen24 (37%)46 (31%)
Choked or burned17 (27%)46 (31%)
Kicked14 (22%)44 (30%)
Use of weapon9 (14%)24 (16%)
 
Reported acts of sexual violence by a partner*(n= 32)(n= 72)
Physically forced to have sex29 (91%)69 (96%)
Forced to have sex by fear27 (84%)62 (86%)
Forced to engage in acts of sex she found degrading17 (53%)37 (51%)
Figure 1.

Overlap of types of partner abuse during pregnancy. Community-based study of violence against pregnant women. León, Nicaragua, 2004.

Neither educational level, socio-economic situation, marital status nor parity were related to abuse during pregnancy, although violence was more commonly reported by young women. Abused pregnant women reported that the pregnancy was unwanted more frequently. They also reported late booking at antenatal care (after 20 gestational weeks) or no visits at all (Table 1). They often mentioned that their partner had prohibited the visit.

Fifty-four percent (259/478) of the pregnant women reported a lifetime history of partner abuse, and a significant association with abuse during the current pregnancy (P < 0.001). The prevalence of abuse and characteristics are presented in Table 2.

Half of the physically abused pregnant women (32/64) reported at least one severe act of violence. The most frequent types were punches and kicks: in 37% (24/64) these were directed at the abdomen (Table 2).

Acts of abuse perpetrated many times (>4) during the current pregnancy were reported by 65% of the emotionally abused pregnant women, 66% of the physically abused and 69% of the sexually abused. Despite the fact that 93% of the victims reported at least one type of injury, only 14% sought health care (Table 3). Most of those who contacted the health services were admitted to hospital. Two women in their second interview reported kicks in the stomach with subsequent hospitalisation due to bleeding and pain. One of the women, who was in the second trimester, spontaneously aborted while the other, who was at the end of pregnancy, had an abruptio placenta. The reports were corroborated by hospital records.

Table 3.  Injuries and health needs among pregnant women.
 Current pregnancy (n= 64)
Injuries among physically abused
Abrasion, bruises60 (94%)
Cuts, punctures, bites7 (11%)
Sprains, dislocations2 (3%)
Deep cuts2 (3%)
Fractures1 (2%)
Others10 (16%)
 
Self reported unconsciousness30 (47%)
Health care
 Sought health care9 (14%)
 Admitted to hospital6 (9%)
 Told health worker real cause of injury2 (3%)

Two hundred and nineteen of the total respondents had no history of violence against them. Of these, 26% reported that the partner abuse started during the current pregnancy. Of those with a history of previous abuse, one-third reported that the violence increased in frequency and severity during the current pregnancy.

Jealousy, disobedience and refusal to have sex were the most commonly mentioned conditions reportedly leading to violence, with one of these features mentioned by more than 80% of the abused women. Other common stimulants were drunken husbands, failure to have meals ready at home, family problems and economic difficulties. The pregnancy itself was mentioned by almost half of the women victims as a precipitating factor.

Different coping strategies had been employed. Forty-seven percent (73/155) had fought back as an act of self-defence and 21% (32/155) had left home to escape from the severe physical abuse on one or more occasions, even if most of them (98%) had returned back home after a few hours. The reason to return was usually an effort to keep the family together because of the new baby. Forty-six percent of the pregnant victims had never talked to anybody about the abuse, but one-third had disclosed it to their parents. Very few had made any contact with authorities.

No differences were found between abused and non-abused women regarding the attitudes towards partner violence. More than 80% of both groups expressed the view that family problems should not be discussed outside the family, and almost 45% percent agreed that outsiders should not intervene and that a woman must obey her husband even if she disagrees with him. When asked about specific reasons for a man to hit his wife, most women stated that there were none, although 18% felt that it was a justified response to unfaithfulness and 9% to disobedience.

High levels of emotional distress as well as low social resources were more frequently found among abused women in the age-adjusted multivariate model (Table 4).

Table 4.  Women's age, social resources and emotional distress in relation to reported abuse during pregnancy, adjusted in a multivariate regression model.
 Abuse during pregnancy
Crude OR/95% CIAdjusted OR/95% CI
Mother age (<20 years)2.41 (1.56–3.72)2.40 (1.56–3.69)
Emotional distress (high)2.81 (1.86–4.25)2.61 (1.73–3.94)
Social resources (low)1.93 (1.28–2.92)1.55 (1.02–2.35)

DISCUSSION

We have demonstrated a high prevalence of emotional, physical and sexual violence during pregnancy in Nicaraguan women. Violence was more common among young pregnant women. Abused women frequently did not want the current pregnancy, were late booking for antenatal services, had low access to social resources and a high level of emotional distress.

A high prevalence of violence against pregnant women has been reported from other low-income countries,8 but most previous studies have been retrospective. This study is, to the best of our knowledge, the first population-based study prospectively conducted in pregnant women.

One-third of pregnant women in this study reported to be victims of emotional abuse during the current pregnancy and almost one out of six were physically abused. This is a high figure in comparison with reports from community studies in the United States,16,25 but consistent with results from other developing countries.15 A broad range of prevalence, from 0.9% to 29% of pregnant women, has been reported.5–10 The variation could partly be explained by differences in the methodology. In most cases the studies assessing violence in pregnancy have been conducted in hospitals or clinics, and these have shown a higher prevalence than population-based studies.9 The population-based design increases the external validity of the results. Efforts were made to identify all pregnant women in the study area during the study period. As has been pointed out in previous studies, overreporting of violence is highly unlikely due to stigma attached to partner violence, the victimisation and the fear of reprisals.5,17 Further, pregnant women are likely to try to protect the father of her child. In order to enable the disclosure of violent experiences and minimise underreporting, the two interviews were conducted in privacy and the field workers trained to create an atmosphere of confidence within a secure environment. The re-visit at the end of pregnancy increased the reported prevalence of violence by three times. This is partly explained by violent acts occurring during the period between the two visits, but also by the confidence created and the repeated opportunity to disclose these experiences. Recall bias was unlikely in this study because of the prospective design.

A history of previous abuse was strongly associated with abuse during the current pregnancy, a finding consistent with international reports.26 Our design did not allow for an analysis of whether pregnancy increases the risk of violence. Nevertheless, it should be noted that in one-third of the pregnant women without a past history of abuse the violence started during the current pregnancy, and that half of the abused pregnant women considered the pregnancy itself a major risk factor for the abuse. This issue has been controversial, and some studies report a decrease in previous abuse during pregnancy10,13 while others claim the contrary.26,27

The type of physical acts as well as the reported frequency and injuries reflect the severity of the partner abuse against pregnant women in Nicaragua. As reported by previous studies26 any physical aggression against pregnant women was likely to be directed towards the abdomen, sometimes with fatal consequences as was the case with the late abortion and the abruptio placenta mentioned above. Consistent with results reported in previous studies28 women rarely sought health care, while those contacting health services usually were hospitalised due to more severe injury.

Gender inequality, docility and an acceptance of male dominance were demonstrated in the comments by the pregnant women. Most expressed the opinion that partner violence was a family or personal matter. These opinions were less prominent in previous studies of domestic violence conducted in the general population of women in the same area.29 The reasons could be a stronger socialisation of gender rolls during the pregnancy period, or the desire of a pregnant woman to keep the family together, thus providing the best conditions for the coming child.

The lack of association between education and social factors and experiences of violence are in agreement with previous findings from Nicaragua and internationally, demonstrating how partner abuse crosses social frontiers.17,30 Contrary to findings from previous research, economic conditions were not associated with abuse during pregnancy. A possible explanation is the relative socio-economic homogeneity with widespread poverty in the study area. If so, the economic factors may contribute to the stressors but remain undetected as an associated factor behind violence in pregnancy.

The poor access to social resources among women experiencing violence and the higher levels of emotional distress are consistent with previous research reports.4,19,29,31

From the findings we conclude that violence against pregnant women is not only common in Nicaragua, but often severe and repeated. The victims are characterised by poor access to social resources and high levels of emotional distress, but are rarely assisted by the health services. Innovative strategies are needed to provide support and counselling.

Acknowledgements

This study was financially supported by SAREC (Swedish Agency for Research Co-operation with Developing Countries). The authors would like to thank WHO for the use of the Women's Health and Life Event Questionnaire and Per Olof Östergren, University of Lund in Sweden, for the permission to use the Social Support Scale. We want to thank all women who shared their stories with us.

Accepted 22 December 2004

Ancillary