Profile of Gender Violence by Intimate Partners

Based on the master's dissertation presented at the Universidade Vale dos Sinos, in 2004. Abstract Objective To investigate the prevalence and profile of gender violence (physical, psychological, and sexual) perpetrated against women by current or former intimate partners. Methods This is a cross-sectional study carried out at a primary healthcare unit in the city of Porto Alegre, Southern Brazil. Our sample comprised 251 women aged 18-49 years who attended the healthcare unit between October and November 2003. Data were collected by means of a questionnaire and double-entered into a electronic spreadsheet. We carried out univariate and bivariate analyses and the Chi-square test. Conclusions The present study found high prevalences of gender violence perpetrated by intimate partners among the users of a primary healthcare unit. Such units play an important role in preventing violence against women.


INTRODUCTION
Gender violence against women, the subject of the present article, is understood as a public health issue by the World Health Organization. 13In order to address violence against women, it is necessary to construe gender as a constitutive element of social relationships, based on the differences between sexes and as a primary component of power relationships.'Gender' is a cultural concept related to the way in which a society constructs sexual differences, attributing different statuses to men and women.It refers to the social construction of sex, that is, the word 'sex' designates only the anatomic-physiological characterization of individuals, whereas gender refers to the social dimension of human sexuality. 18nder violence is conceived of as any act resulting, or having the potential to result, in physical, sexual, or psychological damage or suffering to the woman.This includes also the threat to carry out such acts, coercion, or arbitrary restriction from freedom in public or private life, as well as punishment, mistreatment, pornography, sexual aggression, and incest. 2,6,11,13ender violence by intimate partners Kronbauer JFD & Meneghel SN city of Porto Alegre, southern Brazil.The population attending this unit includes 7,000 users, of which 1,003 are women aged 18-49 years. 8Our sample comprised a group of women who attended this unit in October and November 2003.Sample size was calculated using Epi Info 6.04 software, considering a 2.0 risk ratio, a 95% confidence level, 80% statistical power, and an estimated prevalence of violence of 20%. 14This yielded a sample size of 207 women.An additional 15% were added to allow for possible nonresponders, totaling 238 women.The following inclusion criteria were established: age between 18 and 48 years, having used PHU services in October/ November 2003, and reporting current or past marital relationship (defined as the presence of a fixed spouse, husband, or partner).We used the standardized, pre-coded questionnaire employed in a study conducted by Schraiber et al. 17 This instrument, developed to be used in healthcare services, was translated, adapted to local culture, and pre-tested and comprises 36 questions divided into four groups of variables: demographic, socio-economic, sexual/reproductive, and violence-related.We used the classification proposed by the Associação Brasileira de Anunciantes [Brazilian Advertising Association] and Associação Nacional das Empresas de Pesquisa Mercado [Brazilian Market Research Association]* to categorize subject's social status.The reference used for calculating mean per capita income was the monthly minimum wage at the time of the survey -R$240.00(US$82.7 as of October 2003).The instrument used for the detection of minor psychiatric disorders was the Self-Report Questionnaire (SRQ-20), validated in Brazil and recommended by the World Health Organization. 10Seven or more positive answers were considered as indicative of minor psychiatric disorders.
All women fulfilling the inclusion criteria were invited to participate in the study.The sample was defined by means of consecutive interviews with the subjects until the desired sample size was reached.We studied a total of 251 women.Interviews were conducted by three university students previously trained to administer the questionnaire.Interviewers were oriented to refer women in situations of violence to specific care in the same institution.
Data was entered twice in an electronic spreadsheet to maximize consistency and prevent entry mistakes.Univariate and bivariate analyses were carried out using SPSS software.Some of the variables were dichotomized and associations were analyzed using the Chi-square test, with a 95% significance level.Studies 9,12,14 show a high frequency of gender violence in different societies, with prevalences ranging from 20% to 75%.This justifies carrying out surveys aimed at establishing the magnitude of this problem.However, issues still exist regarding the accuracy of the information on gender violence.More accurate estimates are hindered, among other things, by the use of different definitions for the phenomenon, by the variety of sources for information, and by the absence of populational surveys.
Surveys conducted in healthcare services 1,4,17 show annual prevalences of violence against women perpetrated by intimate partners ranging from four to 23%.Prevalences increase to 33-39% when considering the entire lifetime of these women.These studies have found that the highest prevalences are seen among poorer women, that abused women attend healthcare facilities three times more than other women, and that the number of appointments increases along with the severity of the aggression. 2,3,12However, healthcare professionals may fail to identify women that seek healthcare facilities in situations of violence, even in cases where the lesions presented are virtually pathognomonic of the phenomenon.It has been proposed that these professionals may create barriers that prevent them from caring for these women, and that this may be due to factors such as lack of time and resources, fear of offending the woman, lack of training, fear of opening 'Pandora's Box,' and frustration when dealing with the lack of response of the woman to the advice provided. 4nder violence has gained in visibility in Brazil in the last 20 years.However, as in other countries, this visibility is not yet apparent in healthcare services.Moreover, many of these women are referred to as 'polysymptomatic,' or as 'hypochondriacs,' among other pejorative denominations.We emphasize the importance of primary healthcare services in detecting the problem, and their ability to recognize and counsel women before the occurrence of sequelae or of more serious incidents. 17e aim of the present study was to measure the magnitude and characterize the profile of gender violence perpetrated by intimate partners, addressing the situations of violence expressed as physical, psychological, and sexual violence resulting, or with the potential to result, in damage to the woman's life or integrity.

This is a cross-sectional study based on a sample of users of the Primary Healthcare Units (PHU) in the
The project was approved by the Unisinos Research Ethics Committee, according to statute 196/96 of the Brazilian National Health Council.Participation was voluntary and anonymous, and all subjects were previously informed of the aims of the study and that they could interrupt or terminate the interview at any time.

RESULTS
Mean age was 30 years (sd=9.4);most women reported being literate (94%); 51% were black, and 67% reported practicing Catholic religion; 55% lived in common spaces, slums, or invaded areas.Most women (48%) belonged to class D, and the income of 75% of the women was below one minimum-wage.
Regarding marital life, most women reported having a male partner; 37% reported first sexual intercourse before age 15 years; 86% of women had at least one prior pregnancy, and 16% had delivered low-birthweight babies; 77% of women reported not using condoms.
Violence during pregnancy was reported by 17% of women, 69% of which, even in situations of violence, did not consider such situations as abuse.
Regarding gender violence, we found a prevalence of psychological violence of 55% (95% CI: 49-61) that is, 139 women reported having suffered at least one episode of verbal abuse, humiliation, intimidation, or threats by their partners.There was 38% prevalence of physical violence (95% CI: 32-44), with 96 women having been subjected to slapping, shoving, punches, kicks, beatings, strangling, and fire arm use.We found 9% prevalence of sexual violence, (95% CI: 5-11), with 22 women having been forced to perform humiliating sexual practices or subjected to rape (Table 1).In the present article we will not report on the associations between the variables studied and sexual violence.
Figure shows that most women suffered more than one type of violence.More than half of the 145 women reporting episodes of violence in life reported two or more types of violence, with 76 women (52%) reporting two, and 18 women (12%) reporting three types of violence.
Table 2 presents the results of the bivariate analysis of all demographic variables in relation to outcomes physical and psychological violence.Older women (p=0.004) reported greater prevalence of psychological, but not of physical, violence.We found 86% prevalence of psychological violence (p=0.019) and 64% of physical violence (p=0.039)among women without formal education.Schooling was associated with both types of violence (p=0.012;0.023), women with lesser schooling showing greater prevalences.
The population attending the PHU comes predominantly from lower social strata.Women who lived in slums showed greater prevalence of psychological (61%, p=0.04) and physical (48%, p=0.000) violence than those living in regular housing conditions.As to social class, the levels of violence among women from less favored classes (D and E) were greater than those found among more favored classes (B and C) (p=0.006;0.000).Despite the lack of an association between income and violence, women with per capita income below one minimum wage showed greater prevalence of gender violence.Partner variables significantly associated with psychological an physical violence were age (p=0.001;0.008), years of schooling (p=0.004;0.000) and occupation (p=0.015;0.001).Older men, with lesser schooling, and who were unemployed or retired perpetrated violence(s) more frequently.Demographic variables not associated with any of the two types of violence were skin color, religious practice, place of origin, migration, and woman's occupation.Nevertheless, we found that black women reported psychological (59%) and physical (40%) violence more frequently than white women (51% and 36%, respectively).
We investigated the relationship of the group of variables related to the marital, sexual, and reproductive life of women and the occurrence of psychological and physical violence (Table 3).Characteristics associated to the two types of violence included living for more than 10 years with the same partner (p=0.006;0.005); greater number of pregnancies (p=0.018;0.037); greater number of living children (p=0.028;0.019); and partners who refused to use condoms.Women with history of abortions reported more episodes of physical (p=0.011),but not of psychological, violence.Women whose first sexual intercourse happened before age 15 years showed 46% prevalence of physical violence, versus 34% among those whose first intercourse took place at a later age (p=0.05).
The prevalence of minor psychological disorders among the studied women was statistically associated with psychological and physical violence.Women with scores of seven or more in the Self Re-port Test (68%) reported psychological violence more frequently (p=0.000)than those scoring less than seven points (38%); this was also observed in relation to physical violence (p=0.000).Women who reported having had suicidal feelings showed 86% prevalence of psychological violence (p=0.000) and 70% prevalence of physical violence (p=0.000).

DISCUSSION
The present study provides evidence of the magnitude of gender violence in the studied region.The limitations of the present study include the use of a sample restricted to the clientele of a healthcare facility, the cross-sectional design (which does not allow for inferences regarding causality), and the potential problems concerning the accuracy of the information provided.Regarding the latter aspect, however, there was an increase in the number of women who sought the healthcare center in order to participate in the survey.
One of the important findings of the present study was the evidence of a high prevalence of violence against women among the users of this PHU.Moreover, we emphasize the fact that these situations were unknown to the health professionals that cared for these women.In a similar survey, conducted in the Municipality of Sao Paulo, Southeastern Brazil, Schraiber et al 17 (2002) found a prevalence similar to that of the present study.Furthermore, we observed the presence of multiple types of violence, a finding also reported in other studies, showing that physical www.fsp.usp.br/rsp

Gender violence by intimate partners Kronbauer JFD & Meneghel SN
gender violence is often accompanied by psychological coercion and sexual abuse. 1,17further finding was the 'masking' of violent events by the women themselves, who naturalize, banalize, and relativize the violence they suffer and, even worse, fail to perceive these events as violence.
Early studies of violence against women showed that the phenomenon could be observed across all social classes. 7However, there is evidence for a superposition of the domination and exploitation systems constituted by gender, racial/ethnic, and social relationships, laying upon poor black women a heavier burden and greater exposure to the different types of violence. 16ecent studies confirm that family poverty and low schooling of the male partner are predictors of physical mistreatment of women.Factors associated with higher risk of violence against women include low salaries and economic pressures. 9In addition, unemployed men are more violent with their wives and children. 5he finding that violence is enhanced in women from less favored social strata was corroborated in the present study, where a number of markers of social vulnerability were associated with greater prevalence of psychological and physical violence.
Regarding marital and reproductive life, we wish to highlight, in first place, the high numbers observed with respect to violence during pregnancy (17%), given that the prevalence of this event is estimated at around 10%. 15 In addition, women suffering violence had been with the same partner for longer periods and had had more children and more abortions.Gen-der violence can have an effect on the woman's reproductive health: the risk of spontaneous abortion is two times greater and the risk of having a child with low birthweight is four times greater. 7The prevalence of physical violence was greater among women who had had their first sexual intercourse during adolescence, a stage of development in which they are exposed to other hazards or risk situations such as STD and undesired pregnancy. 12e partner's refusal to use condoms in order to prevent STD and AIDS was associated with the two forms of violence studied.This provides evidence of the difficulty women have in carrying through this negotiation, which in turn is related to low self-esteem and to the male-dominated context of domestic relationships, thus exposing women to other types of violence.In a survey of women's health among sex workers, it was found that women report not requiring the use of condoms during sexual intercourse with their boyfriends or husbands. 3Studies indicate that "male power is manifest in the lack of control from the woman's part regarding when, with whom, and in what conditions she will have intercourse, and this clearly illustrates the difficulty or impossibility of negotiating condom use."*Women more prone to minor psychological disorders were more likely to report emotional and physical violence.Furthermore, there was an association between suicidal feelings and episodes of psychological and physical violence.We found that violated women may suffer permanent effects in terms of self-esteem and self-image and become less sure of www.fsp.usp.br/rspGender violence by intimate partners Kronbauer JFD & Meneghel SN their own worth and more prone to depression.Gender violence may be related to suicide, homicide, and maternal mortality. 7The administration of SRQ-20 in routine healthcare may constitute a indicator of situations of violence among users; we thus suggest the use of this instrument in healthcare facilities.
We observed a series of events while conducting the present study that are worthy of note.These included the mobilization of the local healthcare team regarding the need for a reference center for victims of gender violence; the favorable of the local community association regarding the survey; and the increase in the number of users in the healthcare unit, showing a desire to participate in the survey, and indicating that women in situations of violence wish to be heard.
Few studies have investigated the presence of gender violence among healthcare service users.With respect to healthcare professionals, we emphasize that it is necessary to promote the discussion of this subject whenever possible.The present survey worked as a means of promoting discussion within the unit, motivating the team to deal with the problem.It also met a repressed local need, as suggested by the women's interest in and demand for answering the questionnaire.
In conclusion, women in situations of gender violence seek healthcare services more often and talk more about the violence imposed upon them as long as they are given the necessary support and attention.Therefore, we recommend that primary healthcare facilities open specific spaces aimed at listening to, understanding, and facing gender violence.

Table 1 -
Distribution of women according to type of violence.Porto Alegre, Brazil, 2003.
Figure -Number of women victims of gender violence seen at a healthcare facility.Porto Alegre, Brazil, 2003.