Prevalence of physical intimate partner violence in the first six months after childbirth in the city of Rio de Janeiro , Brazil

O objetivo do estudo foi estimar a prevalência de violência física entre parceiros íntimos nos primeiros seis meses após o parto entre mulheres que frequentam unidades básicas de saúde (UBS) do Rio de Janeiro, Brasil, para o acompanhamento da criança. Este é o primeiro estudo sobre o tema realizado com uma amostra representativa de UBS do Município do Rio de Janeiro. Tratase de um estudo transversal, realizado entre junho e setembro de 2007, que incluiu 927 mães/bebês atendidos em 27 UBS, selecionadas por amostragem complexa e geograficamente representativas do município. As informações foram coletadas por meio de entrevistas face a face, por equipe previamente treinada, utilizando questionário estruturado. A história de violência física entre parceiros íntimos desde o nascimento da criança até a data da entrevista foi obtida por meio da versão nacional do instrumento Revised Conflict Tatics Scales (CTS2). Trinta por cento (IC95%: 26,2-33,8) das mães relataram ter vivenciado alguma forma de violência física entre parceiros íntimos no pós-parto e 14% (IC95%: 11,0-17,0) referiram violência física grave. Os abusos físicos ocorreram especialmente entre mães adolescentes, negras, que não viviam com o companheiro, em situação socioeconômica desfavorável, e que apresentavam falhas no cuidado pré-natal, na amamentação e na utilização do serviço de saúde. A ampla magnitude da violência física entre parceiros íntimos reforça a necessidade de enfrentamento imediato do problema. Para isso, é fundamental que os serviços da atenção primária estejam articulados a outras redes de apoio e os profissionais de saúde, preparados para lidar com o problema.


Introduction
Physical intimate partner violence is a global problem, occurring in all social, economic, religious, and cultural groups and in all phases of life 1 .High prevalence rates have been shown in international studies 1 and in Brazil 2,3,4 .Prior studies conducted elsewhere indicate high rates of physical intimate partner violence in the first months after childbirth 5,6,7,8 .In Brazil, as far as the authors know, only two studies have explored the occurrence of physical intimate partner violence in this period of the woman's life.The first study on the theme focused on women enrolled in the Family Health Strategy (FHS) in an area of Recife, Pernambuco State, in 2005 and 2006.According to the authors, 12.1% of the women reported having experienced physical intimate partner violence in the postpartum 9 .The second published study, in mothers of infants up to 5 months of age seen at 5 primary care clinics (UBS) in the city of Rio de Janeiro in 2007, estimated 16.2% prevalence of physical intimate partner violence in the postpartum 10 .
The problem's relevance in early infancy increases considerably due to the consequences, affecting not only the couple, but also the children and other family members.Research has documented the serious consequences of intimate partner violence for the woman's and child's health, including postpartum depression, interference in breastfeeding, infant malnutrition, and gaps in the child's immunization schedule and follow-up in health services 5,11,12 .The hostile and unsafe environment and violence witnessed by the child at an early age appear to have lasting emotional effects that can manifest throughout life, threatening the individual's growth and development 13,14 .In addition, such violence may not be limited to the couple, and may directly affect the child, whether in the physical, psychological, and/or sexual form, or in situations involving neglect 13,15 .
As mentioned above, despite so many negative repercussions on the health of mothers and infants and the growing number of international publications on the theme, in Brazil the information on the magnitude of violence in the first months after childbirth is based on only two previous studies that involved women enrolled in specific health services.In order to contribute to knowledge on this issue in Brazil, the current study estimates the prevalence of physical intimate partner violence in the first six months after childbirth in a wide, representative sample of users of primary care in the city of Rio de Janeiro, identifying the subgroups most vulnerable to the problem.The hope is that publication of the findings will expand the debate on the issue and include it on the maternal and child health agenda.Wider circulation of the study's findings should encourage measures for early detection and immediate response to this important public health problem in the city.

Study design and strategy for selection of participants
The current study was nested in a larger research project entitled Quality Assessment of Care for Children under Six Months in the Unified Health System in the City of Rio de Janeiro, coordinated by the Sergio Arouca National School of Public Health, Oswaldo Cruz Foundation (ENSP/Fiocruz).This was a cross-sectional study of mothers of infants up to six months of age who brought their children for pediatric or childcare appointments at 27 UBS in the Brazilian Unified National Health System (SUS) in the city of Rio de Janeiro from June to September 2007.
Participants were selected by two-stage cluster sampling.The primary sampling units were UBS and the secondary units were mothers of children seen in the selected clinics.To obtain a geographically representative sample of the city, UBS were ordered according to Euclidian distance, calculated on the basis of the clinics' geographic coordinates in relation to the administrative headquarters of the Rio de Janeiro Municipal Government (Figure 1).Next, the clinics were selected systematically in spiral fashion, with the selection probability proportional to the monthly mean number of appointments for children under six months of age in the previous year.The secondary sampling units (mothers) were selected systematically according to the order in which they left the appointments.

Figure 1
Distribution of primary care clinics (UBS) selected for the study (black dots) and other UBS (gray dots).Rio de Janeiro, Brazil.
The initially estimated sample size was 1,080 interviews, or 40 interviews in each of the 27 UBS.Among the 1,082 women interviewed, 927 reported living with their partners at the time of the interview and were included in the current study.

Fieldwork and research instruments
Data were collected in interviews with the mothers, using a structured questionnaire applied by previously trained interviewers.Interviews were conducted after the pediatric appointment in a reserved space and without the partner's presence, thus guaranteeing the woman's privacy.
Information on the occurrence of physical intimate partner violence from the date of the child's birth to the date of the interview was obtained with the use of the Portuguese-language version of the Revised Conflict Tactics Scales (CTS2), adequately adapted for use in Brazil 16,17,18 .Women were defined as having experienced physical intimate partner violence if they responded affirmatively to at least one of the 12 items on the physical violence scale referring to violence perpetrated by the woman and/or her partner.
The household's socioeconomic status was assessed by the household assets index (IB), calculated as: where i varies from 1 to 10 assets; b i is equal to 1 or 0, respectively, in the presence or absence of the following: radio, refrigerator or freezer, DVD or videocassette player, clothes washing machine, microwave oven, landline telephone, computer, TV, private car, and air conditioner.The weight assigned to each item's presence was the complement of the relative frequency (f i ) of each item in the total sample; the rarer the item's presence, the higher the weight assigned to it.Classification of the quality of prenatal care used the Adequacy of Prenatal Care Utilization Index (APNCU), also known as the Kotelchuck index, previously adapted for use in Brazil 19 .The scale combines information on the time of initiation of prenatal care and the number of prenatal visits 20 .The other variables are self-explanatory and are presented in the Table 1.

Data analysis
Stata 14.0 (StataCorp LP, College Station, USA) was used for the analysis.Fisher's exact tests were performed for heterogeneity of physical intimate partner violence prevalence rates in the population subgroups.Differences in prevalence between the subgroups were considered statistically significant when the p-value was less than 0.05.The analysis was performed with the svy routine in Stata 14.0 to deal with the fact that the data came from a complex cluster sample.

Ethical issues
The study was approved by the Institutional Review Boards of ENSP/Fiocruz (case review n. 132/06, March 7, 2007) and the Rio de Janeiro Municipal Secretariat of Health and Civil Defense (case review n. 74A/2007, June 18, 2007).Data were collected after the participants had signed a free and informed consent form, which guaranteed confidentiality of the information.All the study's procedures complied with the Declaration of Helsinki.

Results
As shown in Table 1, the study sample consisted predominantly of adult mothers of children between 2 and 6 months of age, living with a partner, and with white or brown skin color.Slightly more than half of the mothers (54.1%) had 2 or more children, and most of the households had only one child less than 5 years of age.Nearly half of the families lived in precarious economic conditions.Most of the women had at least a complete primary education and were not formally employed at the time of the interview.Most of the mothers had begun their prenatal care in the first trimester, with at least 4 visits during the pregnancy.This resulted in only a small proportion of inadequate prenatal care.Approximately 7% of the children had a history of low birth weight (< 2,500g) and 8% had been hospitalized at some time.At the time of the interview, 60% of the infants were exclusively breastfeeding.Nearly onefourth of the mothers reported difficulties taking the child for pediatric follow-up at the primary care clinic (Table 2).
As shown in Table 3, 30% of the women reported having experienced at least one act of physical intimate partner violence from the date of the child's birth to the interview, whether as victim or perpetrator of the violence.Some 28% had been involved in acts of minor physical violence in terms of potential severity, while 14% reported acts of severe physical violence.Both minor and severe acts of physical violence were frequently committed by the woman.
Table 4 shows the prevalence rates for minor and severe physical violence in different subgroups.Both forms of physical violence occurred mainly among black adolescent mothers with less schooling and who were unemployed or underemployed at the time of the interview.Minor physical violence was the only form that increased in prevalence with the child's age.Women not living with a partner and who lived in households with more than one child less than 5 years of age were more frequently involved in acts of severe physical violence.This form of violence occurred predominantly in women with a history of inadequate or no prenatal care and fewer than 4 prenatal visits.Prevalence of severe physical violence was also higher in mothers not practicing exclusive breastfeeding and among those reporting difficulties in attending the UBS.

Discussion
This was the first study to estimate the prevalence of intimate partner violence in the first six months after childbirth in a representative, comprehensive sample of families attending UBS in the city of Rio de Janeiro, Brazil.Publication of the results should thus expand the debate on the theme, reinforcing its importance on the maternal and child health agenda in Rio de Janeiro.
As discussed above, nearly one-third of the women were involved (as victim and/or perpetrator) in situations of physical intimate partner violence in the first 6 months after childbirth, indicating the problem's relevance in this period of the woman's life.The estimate is higher than one obtained in a previous study in the city of Rio de Janeiro (16.2%), including only 5 UBS 10 .The estimated rates reporting only physical violence against the woman also exceeded the rates in two previous studies in    Brazil.While 18.3% of the women interviewed in this study reported having been victims of physical intimate partner violence from the child's birth to the day of the interview (up to six months postpartum), a study in 2005 and 2006 with women enrolled in the FHS in a region of Recife, estimated a 12.1% prevalence rate for postpartum victimization 9 .
Comparing the current study's estimates to the international literature, the problem's relevance in Rio de Janeiro becomes even more evident.Studies that assessed the magnitude of physical violence against women in the first year after childbirth in the United States 7,8,21,22,23 , Canada 24 , England 25 , Sweden 26 , Australia 27 , China 28 , India 29 , and South Africa 30 showed prevalence rates varying from 1.8% (England) to 13.5% (South Africa), far lower than the estimated rates in UBS in Rio de Janeiro.Minor violence was not the only form that proved to be commonplace among the women interviewed here.Acts of serious physical violence such as punching, throwing the other against a wall, beating, suffocating, strangling, burning, or even brandishing a knife or firearm also showed high rates.These are obviously threatening situations for the mothers and their children, given the total dependence on maternal care during infancy.
The differences between the prevalence rates estimated in this study in Brazil and those from studies elsewhere in the world may have resulted from socioeconomic differences between the studies' samples, since most women that use SUS for their infants' follow-up have lower socioeconomic status than the women included in international studies.This is corroborated by the higher likelihood of physical intimate partner violence among unemployed black adolescent or mothers with low schooling, both in this study and in the literature 31,32,33 .
Importantly, the estimates could have been even higher if this had been a population-based survey rather than a sample of women attending UBS.As shown in a previous study, mothers involved in violent relationships were less likely to attend health services for routine pediatric follow-up of their infants, thus hindering their uptake by studies conducted only in health services 34 .A more comprehensive picture of the situation would require population-based studies to include women that rarely or never attend UBS.Some methodological issues in the strategy used to detect situations of violence may also have contributed to the higher physical intimate partner violence rates found here as compared to studies conducted in other contexts.The use of an instrument with good psychometric properties, widely used in the literature and adapted for use in Brazil 16,17,18 , probably added value to the measurement process and decreased the odds of underestimating the violence.Of the non-Brazilian studies cited here, only Gartland et al. 27 and Hellmuth et al. 8 used a structured instrument to assess the occurrence of violence in the couple.The others used single questions to detect physical intimate partner violence, which may have led to the underestimation of prevalence rates.In addition, application of the questionnaire by a properly trained team, without the intimate partner's presence, and guaranteeing the mother's privacy in relation to the other mothers in the waiting room, may also have contributed to more fine-tuned measurement of physical intimate partner violence.
Another question deserving debate is the fact that the women were more perpetrators than victims of physical abuse, consistent with the literature 33,35,36,37,38 .A previous study addressing prevalence of violence during pregnancy in the city of Rio de Janeiro had also called attention to the reciprocity of violence within the couple 39 .This profile of violence may result from the study sample, including only women that attend health services.In general, studies with samples of health services users mainly detect situations of mild to moderate violence, practiced more frequently by women.Meanwhile, studies in shelters and other services for women's protection capture more serious cases, usually practiced by the male partners 40 .
Some women may also have avoided disclosing acts of violence they suffered for fear of retaliation by their partners, or even to "protect" them, realizing the illegality of violence against women 40 .Future studies should thus seek to obtain relevant information from both members of the couple.At any rate, although the evidence indicated higher prevalence of physical intimate partner violence practiced by women against their partners, several previous studies indicate that the negative consequences of intimate partner violence are much more drastic for the woman 35,40 .As mentioned in the introduction, such repercussions not only substantially affect the victim's health and well-being, but also end up affecting the child as well, either directly -violence against the child per se -or indirectly, since the situation particularly affects the mother's capacity to care for her children 13,14 .
Cad. Saúde Pública 2017; 33(8):e00141116 Consistent with the literature 6,31,32,33,41 , although physical intimate partner violence occurred in all the population subgroups, unemployed or underemployed black teenage mothers not living with the partner, with less schooling, and living in households with more than one child under 5 years of age showed higher prevalence of this form of violence.The results indicate that characteristics related to maternal and child health and care may also be associated with higher rates of physical abuse.Mothers who had received unsatisfactory prenatal care or reported difficulty in attending the primary health clinics were involved more frequently in violent relationships, making them even more vulnerable.Likewise, physical intimate partner violence was more common among mothers that were not practicing exclusive breastfeeding, so important in the infant's first six months of life.
The identification of this profile of women involved in situations of physical intimate partner violence can be quite useful for primary healthcare staff.Gaps in prenatal care, breastfeeding, and use of health services by women with these socio-demographic characteristics may signal the occurrence of intimate partner violence, facilitating the detection of suspected cases, and such gaps are routinely seen by primary healthcare professionals.The suspicion should serve as the basis for more detailed approaches with the use of specific instruments and other screening methods that can produce elements for the identification of actual situations and subsequent intervention.
The study's results should be analyzed in light of its limitations.As discussed, the study was only conducted in standard primary healthcare clinics and thus did not include families enrolled in and regularly using the more comprehensive FHS.The estimates might have been even higher if the latter clientele had been part of the sample, since the implementation of the FHS in the city of Rio de Janeiro has prioritized areas that are known to be more vulnerable to various forms of interpersonal violence.The decision to only interview the child's mother during attendance at the clinics may also have underestimated the findings, since only one member of the couple was heard.However, as suggested in previous studies 42,43 , the decision to focus on violence in the couple, defining positive cases as situations in which the woman or partner committed acts of physical intimate partner violence, increases the strategy's sensitivity to identify abuse, thus mitigating the limitation.Another potential limitation is the fact that the data were collected several years ago.With the more recent economic downturn in Brazil as a whole and the state of Rio de Janeiro in particular, with repercussions are also felt in the city of Rio, the current situation tends to be just as bad, or worse, than reflected in the study's data.Due to the scarcity of studies on the issue, and since this was the first such study with a representative sample of women users of primary care services in the city of Rio de Janeiro, the results can serve as the point of departure for future studies and actions.
The great magnitude of intimate partner violence in the study reinforces existing claims by researchers and health professionals concerning the need for immediate measures to deal with the problem.Primary health services are strategic places to detect risks and identify cases of intimate partner violence, since victims' use of more specialized services is still quite stigmatized, and with limited access.Early infancy is a time of many contacts between the mother and health services, and the opportunities to screen for situations of violence should not be wasted.Health professionals that work with families during this phase of the life cycle should be alert to the issue of violence and prepared to deal with it, to promote and a safe and welcoming environment, favorable to disclosing the situation.
It is also crucial for healthcare services to be linked to a protective network involving different sectors of society, such as child protection services, the courts, law enforcement, and social services.Given the magnitude, serious consequences, and complexity of intimate partner violence, public policies and action strategies integrated in networks are essential to reduce this serious public health problem.Hopefully discussion of the problem's magnitude, especially in certain population subgroups, will increase its visibility, raising the awareness of policymakers and health professionals concerning the importance of actions aimed at prompt identification and management in the city of Rio de Janeiro.

Table 1
Socio-demographic characteristics of the study population.Rio de Janeiro, Brazil, 2007.

Table 2
Characteristics of the study population: mother's and child's health.Rio de Janeiro, Brazil, 2007.

Table 3
Estimated prevalence of intimate partner physical violence according to perpetrator and in the couple.Rio de Janeiro, Brazil, 2007 (n = 923) *.

Table 4
Estimated prevalence of intimate partner physical violence according to subgroups.Rio de Janeiro, Brazil, 2007.
95%CI: 95% confidence interval; UBS: primary care clinic.* Number of individuals with complete data for the variables analyzed.Cad.Saúde Pública 2017; 33(8):e00141116 Cad. Saúde Pública 2017; 33(8):e00141116 Contributors C. L. Moraes participated in the study planning and execution, data analysis and interpretation, and writing of the article.A. G. S. Oliveira collaborated in the literature search, data analysis and interpretation, and writing of the article.M. E. Reichenheim contributed in the critical revision of the article.S. G. N. Gama e M. C. Leal participated in the study planning and execution and collaborated in the critical revision of the article.Os abusos físicos ocorreram especialmente entre mães adolescentes, negras, que não viviam com o companheiro, em situação socioeconômica desfavorável, e que apresentavam falhas no cuidado pré-natal, na amamentação e na utilização do serviço de saúde.A ampla magnitude da violência física entre parceiros íntimos reforça a necessidade de enfrentamento imediato do problema.Para isso, é fundamental que os serviços da atenção primária estejam articulados a outras redes de apoio e os profissionais de saúde, preparados para lidar com o problema.