Violence against Pregnant Women: Prevalence and Associated Factors

OBJECTIVE: To identify the factors associated with domestic violence against pregnant women. METHODS: Interviews were conducted with 1,379 pregnant women undergoing antenatal care in basic health care units of the Brazilian Health System, within the municipality of Campinas (Southeastern Brazil). A structured questionnaire on domestic violence, validated in Brazil, was applied between July 2004 and July 2006. The fi rst and second interviews in a cohort study were analyzed. Descriptive and multiple logistic regression analysis of the data were conducted. RESULTS: Psychological violence was reported by 19.1% (n=263) of the total sample of pregnant women and physical/sexual violence was reported by 6.5% (n=89) of them. The factors associated to psychological violence were: adolescent intimate partner (p<0.019) and the pregnant woman had witnessed physical aggression before she was 15 years old (p<0.001). The factors associated to physical/sexual violence were: diffi culties encountered by the pregnant woman in attending her antenatal appointments (p<0.014), intimate partner uses drugs (p<0.015) and does not work (p<0.048). The factors associated to psychological and physical/ sexual violence were: low level of education of the interviewee (p<0.013 and p<0.020, respectively), the pregnant woman being responsible for the family (p<0.001 and p=0.017, respectively) pregnant woman had suffered physical aggression during childhood (p<0.029 and p<0.038, respectively), presence of common mental disorder (p<0.001) and intimate partner consumes alcoholic beverage twice or more weekly. (p<0.001). CONCLUSIONS: A high prevalence of different categories of domestic violence by an intimate partner during pregnancy was found as well as different factors associated with them. Appropriate mechanisms are necessary, particularly in primary health care, to identify and deal with domestic violence during pregnancy.


INTRODUCTION
Violence against women is widely acknowledged as a grave public health problem.A multicentric study on domestic violence coordinated by the World Health Organization (WHO) found that the prevalence of domestic violence perpetrated by an intimate partner against women throughout their lifetime varied from 15% in Japan to 71% in Ethiopia and the prevalence of physical/sexual violence in the past year ranged from 4% to 54%, respectively. 4regnant women are not free from domestic violence: in a review of the literature, prevalences varied from 0.9% to 20.1%.This variation in prevalences is attributed to the heterogeneity of the defi nitions of violence, to the different research methodologies utilized, as well as different sample sizes and sampling procedures. 5me women's life experiences have been described as factors associated to domestic violence: low socioeconomic status, low level of social support, black race/ ethnic group and being young. 3,13As to the reproductive history of women, the following events were found to be associated to domestic violence: under 19 years of age at fi rst sexual intercourse, unplanned pregnancy, partner's refusal to use condoms and the use of legal and illegal drugs. 3,13Pregnant women who had observed or suffered violence when they were young were also more susceptible to suffering violence during pregnancy. 3,5owever, there is no consensus concerning pregnancy as a risk factor for this type of violence. 5olence during pregnancy may have grave consequences for women's health, including hemorrhages and the interruption of pregnancy. 18As to the child's health, there is an increased risk for perinatal mortality and for newborns with low birth weight or prematurity. 2udies that can contribute to our understanding of this issue are of fundamental importance in order to help us confront it and, in this sense, to defi ne new approaches, particularly in the health services.
The objective of this study was to identify the factors associated to domestic violence, psychological, physical and/or sexual, perpetrated by an intimate partner during the gestational period.

METHODS
A cross-sectional analysis of the fi rst and second interview of a cohort study initiated in March 2004 and concluded in July 2006 is the object of this paper.The interviews were conducted with pregnant women who were receiving antenatal care at primary health care units of the Sistema Único de Saúde (SUS) [Brazil's Unifi ed Health System] within the municipality of Campinas, Southeastern Brazil.
The questionnaire validated in Brazil by Schraiber a (2003) was adopted.It includes questions that refer to psychological violence characterized by insults, humiliation, intimidation and threat; physical violence, when women are slapped, pushed or shoved, hit with a fi st, kicked, choked or threatened or actually hurt with a knife or gun; and sexual violence, when women report that they were forced to have some form of sexual intercourse they did not desire or had sexual intercourse because they felt afraid not to do so.After conducting a review of the literature, the importance of investigating the history of violence suffered or observed during adolescence became clear.Thus, interviewees were asked: "Before you were 15 years old, did you observe any type of physical violence within your family?When you were under 15 years old, did you suffer any type of physical aggression from a family member?When you were under 15 do you recall if anyone touched you sexually or forced you to engage in any form of sexual activity you did not desire?
The questionnaire included variables related to: demographic and socioeconomic characteristics of the pregnant women and their intimate partners, maternal reproductive history and situation of the current pregnancy, puerperal situation and infant feeding, smoking, consumption of alcohol and illicit drugs by the intimate partner.The intimate partner was considered the husband/companion, boyfriend with whom the woman was having sexual relationships or the father of the child that she was carrying in her womb.
The classifi cation elaborated by the Associação Nacional das Empresas de Pesquisa de Mercado (ANEP) [National Association of Market Research Enterprises] was utilized in the elaboration of the social stratifi cation of the sample.The Alcohol Use Disorders Identifi cation Test (AUDIT) 15 was applied during the fi rst interview in order to attain an understanding of the use of alcohol by pregnant women.In order to identify possible mental disorders, the SQR-20-Self-Report Questionnaire was utilized in both the fi rst and second interviews. 9e interviewers were over 20 years old and had completed high school.
A sample of 1,400 pregnant women was calculated for the cohort study.Among these, 1,379 women were selected in a non-random manner when they came to the service for their antenatal consultation.The level of statistical signifi cance adopted in this study was 5% and the sample error was 1%.
Pregnant women were included in this study between July 2004 and July 2006, when they came to the health care unit for antenatal consultations.On this occasion they were invited to participate, and informed of the objectives and procedures undertaken by the study.They were also asked to sign an Informed Consent form.When the pregnant woman was 18 years of age or under and legally dependent of her parents or guardians, this Consent form was signed by her and by her legal guardian.
The interviews were conducted with all pregnant women who agreed to participate in this study, regardless of their gestational age.
Only one interview was conducted with pregnant women whose gestational age was over 28 weeks as well as with those whose gestational age was up to 28 weeks and who reported in their fi rst interview that they had suffered physical and/or sexual violence during their current pregnancy.
More than one interview was conducted during pregnancy and particularly during the last trimester with the objective of minimizing the false negative responses and verifying, among the women interviewed during the fi rst trimester, violent events that occurred after the fi rst interview.
During the entire period of fi eldwork meetings were held every two weeks with the interviewers and the fi eld coordinator, under the supervision and orientation of a psychologist.
Initially, descriptive procedures used in calculating prevalences and bivariate tests for estimating non-adjusted risks were conducted, considering the types of violence (psychological and physical/sexual) as dependent variables and the sociodemographic characteristics and lifestyles as independent variables.After these analyses were completed, procedures involving a multiple model approach were undertaken by means of logistic regression, including all the independent variables that demonstrated an association with the two dependent variables -psychological and physical and/or sexual violence -in the model, with a 10% level of signifi cance.
The stepwise forward approach was utilized when running the multiple regression model.The variable remained in the model if p≤0.05.The strength of the association between the independent and dependent variables was expressed as crude and adjusted estimated odds ratios (OR), and their respective 95% confi dence intervals.Adjustments were made considering the independent variables described above and also including the variables race/ skin color and marital status, which differed among the pregnant women who were and were not followed-up in this study.Data was digitalized in a 6.04 version of EpiInfo.All the questionnaires included in the databank were entered and the consistency of the data was also checked.The 13.0 version of the SPSS program was utilized in data analysis.

RESULTS
Descriptive data as well as the bivariate analysis of the sociodemographic factors associated to psychological and physical/sexual violence perpetrated against the pregnant woman by her intimate partner are presented in Table 1.
In the fi rst interview, 16.3% (n=225) of the pregnant women reported that they had been victims of psychological violence, 5.7% (n=79) had been victims of physical violence and 1.3% (n=18) had been victims of sexual violence.In the second interview with participants whose gestational age was under 28 weeks and who had not reported violence in the fi rst interview (n=806), psychological violence was reported by 8.3% (n=114) of the pregnant women, physical violence was reported by 1.7% of them (n=24) and sexual violence by 0.8% (n=11).If occurrences of violence reported in both interviews are summed up, the prevalence of psychological violence is 19.1% (n=263) and of physical/sexual violence combined is 6.5% (n=89).
The average age of the pregnant women was 23.8 years (sd=5.50),with 23.6% being adolescents.Those who declared their skin color to be white or yellow represented 56.4% of the sample.The pregnant women had a low level of schooling, 47.1% had concluded junior high school and 1.0% had college education.The majority of the women (81.2%) reported they were either married or in a stable union.Half of them declared they were catholic and a third that they were evangelical.Approximately one fourth of the pregnant women were working at the time they were interviewed; almost half belonged to the D/E economic strata and the remaining women belonged to the C strata.
Among the women interviewed, 5.5% declared they were responsible economically for their families.The prevalence of the use of alcoholic beverages during pregnancy was 1.4%, 13.6% were cigarette smokers.The prevalence of pregnant women who reported some experience of violence during their childhood was 55.8%.Among the latter, 31.3%witnessed physical violence in their families, 17.8% were victims and 6.7% suffered some type of sexual abuse.There was no signifi cant statistical association (p>0.05) between any type of violence and current age, declared race/skin color, current work status and cigarette smoking among the women interviewed.On the other hand, the following were considered risk factors for domestic violence ( psychological and physical/sexual) against pregnant women: low educational level, non stable union, the pregnant woman alone and/or she and her partner being responsible for sustaining the family economically, and having witnessed or been a victim of some kind of violence during her childhood.Belonging to the catholic religion was a protective factor for psychological violence, but did not remain so in the logistic regression model.The use of alcoholic beverages by pregnant women increased by four times the chances of them suffering physical/sexual violence and if they were in unstable unions, this increased their probability of having such experiences more than twice (Table 1).Having had the fi rst sexual intercourse and the fi rst pregnancy before the age of 16 years, diffi culty in arriving at the antenatal appointments, being a primipara and presenting common mental disorders constituted situations that increased the chances of the occurrence of physical/sexual and/or psychological violence.
The average age of the intimate partner was 27.6 years, with 8.5% being adolescents.Like their partners, the men had a low educational level: 53.8% had completed junior high school and 2.1% college education.The majority of the women worked (87.5%) and 81.9% lived with their spouses.Among the latter, 26.3% had lived together less than a year and 21.2% had children with another man.The prevalence of cigarette smoking among pregnant women was high (33.4%)and so was the use of alcoholic beverages by intimate partners (30% drinked twice or more times per week).Six per cent of the intimate partners used illicit drugs.All variables except duration of cohabitation with the pregnant woman were associated to violence perpetrated against her.The results of logistic regression analysis are presented on Table 4. Low educational level -up to eight years of schooling -increased the probability of psychological violence 1.5 times and almost doubled the chances of physical and sexual violence taking place.The presence of common mental disorders increased the probability of the occurrence of psychological violence more than one and a half times and almost tripled the chances of physical and sexual violence taking place.Having witnessed within the family or experienced violence before the age of 15 years also increased the probability of psychological violence occurring to almost twice as much as among those who had not witnessed or experienced such violence before that age.It also increased one and a half times the chances of physical and sexual violence occurring during the current pregnancy.When the pregnant woman was fi nancially responsible for her family the chances that she would suffer psychological and physical/sexual violence was almost twice as great.
Having reported diffi culties in getting to her antenatal appointments more than doubled the chances of the pregnant woman becoming a victim of physical and sexual violence.If the frequency with which alcoholic beverages were consumed by the intimate partner was twice or more times per week, this more than doubled the chances that psychological violence and physical and sexual violence would occur.Likewise, the consumption of illicit drugs more than doubled the chances of perpetration of physical and sexual violence against the pregnant woman.
The intimate partner's age presented a positive association with psychological violence.The fact that the intimate partner was unemployed increased in 77% the probability that physical and sexual violence would occur.

DISCUSSION
Violence against women in whatever phase of their lives represents a grave social and public health issue that must be confronted in Brazil.Violence during pregnancy demands special attention from the health services because it affects women in a moment of great physical and emotional vulnerability.This is the fi rst cohort study conducted in Brazil that followed-up pregnant women that were undergoing antenatal care in public primary health care units.These cross-sectional analyses were based on two interviews undertaken with pregnant women during their pregnancy.Approximately 2% of the women interviewed in this study reported they had suffered physical violence during the second interview.In the studies in which domestic violence questions were asked more than once during pregnancy or during the third trimester, higher prevalences of domestic violence were found, varying from 7.4% to 20.1%. 5 The high prevalences of domestic violence perpetrated against pregnant women in this study corroborate the results of the multicentric study undertaken by the WHO.Among the countries included in that study, wide variations of prevalences of physical and sexual violence were registered.The smallest rate was observed in Japan (8%), followed by Siberia and Montenegro (13%), Thailand (11%) and the highest rates were registered in Brazil, in the cities of the Zona da Mata [Forrest Region] in Pernambuco (32%), and in a province in Peru (44%). 4rge variations in the prevalences were also reported by other authors in literature reviews on violence against women during pregnancy. 5,8Caution is necessary in interpreting these results, for these differences may be related to the diverse characteristics of the populations studied, to the defi nitions of violence and to the diversity of instruments and methodologies utilized in collecting data.
Among the factors associated to domestic violence during pregnancy are low educational level, frequent use of alcohol, unemployment and low income 5,7,8,14 of the pregnant women and their intimate partners.However, no other studies were found that dealt with some of the variables investigated in the present study and that constituted factors associated with psychological or physical/sexual violence, such as: the pregnant woman being the person responsible for the family economically or the pregnant woman reporting she faced diffi culties in coming regularly to her antenatal appointments.
One may suppose that the latter is a woman who suffers various kinds of constraints such as jealousy, threats and lack of economic resources that may result in restrictions with respect to her liberty, being her irregular attendance at antenatal appointments indicative of this state of things.The fact that the greater probability of violence was observed when the pregnant woman was the person responsible for her family's income may be related to her partner's unemployment status, as observed in this study.A revision of the literature 8 indicates that unemployment of the intimate partner is a risk factor for violence.There is also a report according to which violence is greater when women begin to assume non-traditional roles or when they start to work. 14other frequently observed indicator of risk for violence is the frequent use of alcoholic beverages by the intimate partner.Some authors believe that the use of alcohol facilitates acts of violence since it modifi es behavior patterns, creating conditions for discussions, offenses, curses, insults and threats that may culminate in sexual and physical aggressions.In the present study, consumption of alcohol and illicit drugs by the intimate partner represents a greater probability of the occurrence of violence against pregnant women.Such a situation may lead to delays in searching for help and, consequently, in interventions that could minimize the effects of or interrupt these acts. 1,11cording to Schraiber 17 (2003), studies of men and women in a situation of domestic violence indicate a multifactorial condition that acts as a precursor of this type of violence.Although alcohol and poverty favor violence, they cannot be considered its direct causes.Another factor that is consistently related to an increase in the risk for violence is the pregnant woman having witnessed domestic violence during her childhood. 17his condition, observed in the present study, may indicate that violence initiated in the pregnant woman's adolescence may currently be experienced as a "natural" part of life, contributing to her low self-esteem and to the lack of autonomy to create mechanisms that may help to modify this situation. 3 was observed in this study that common mental disorder presented a strong association with the characteristics of violence during the pregnancy currently investigated.However, since this is a cross-sectional analysis, it was not possible to verify if this was a causal association or not.This relationship was also observed in a study conducted with women who reported they were victims of violence during pregnancy, attended in public health services. 3olence against women is a complex and multidimensional problem that has gradually been approached as an issue of public health. 10The health sector has an important role in combating this type of violence by means of research, case notifi cations, the organization of reference services for victims as well as other proposals involving intervention.However, none of these strategies for combating violence can ignore the issue of the cultural roots of these abuses besides, evidently, attending to the immediate needs of the victims.According to Heise 6 (1994), this signifi es challenging attitudes and social beliefs that sustain violence perpetrated by men against women, and creating ways of negotiating power between genders at all levels of society. 16gh prevalences of different categories of domestic violence perpetrated by the intimate partner against pregnant women were found in this study.These are associated to diverse factors related to the socio-economic, demographic and health conditions of the woman and her intimate partner.Appropriate methods for identifying and approaching domestic violence during pregnancy are necessary, particularly in primary health care.
The study was approved by the Research Ethics Committee of the Faculdade de Ciências Médicas da Universidade Estadual de Campinas [School of Medicine of the State University of Campinas] (Processo nº 116/2004).

Table 1 .
Bivariate analysis of the sociodemographic characteristics, unhealthy habits and violent situations experienced by the pregnant woman.Campinas, SoutheasternBrazil, 2004-2006.N=1,379 ** Both: Pregnant woman and her intimate partner are reported by the pregnant woman to be responsible for sustaining the family

Table 2 ,
presents the variables representing the reproductive and mental health histories of these pregnant women and the bivariate analyses with each respective OR, confi dence interval and p value may be observed.More than half of the pregnant women had their fi rst sexual intercourse at the age of 16 years or less and 23.4% became pregnant for the fi rst time at that age.

Table 2 .
Bivariate analysis of the characteristics of reproductive health and mental health conditions of the pregnant woman.
* Diffi culty in keeping or going to an antenatal appointment (fi nancial problem, cannot schedule appointments, doesn't have a person with which to leave her children, does not like the way she is being attended by professionals, does not believe that antenatal care is important, cannot leave work to go to her antenatal appointment)

Table 3 .
Bivariate analysis of the sociodemographic characteristics, unhealthy habits of the pregnant woman's intimate partner.

Table 4 .
Factors associated to psychological violence and physical/sexual violence, perpetrated by the intimate partner during pregnancy.Campinas, SoutheasternBrazil, 2004-2006.