INTRODUCTION Body image can be conceptualized as a multidimensional construct that represents how individuals think, feel, and behave in relation to their physical attributes. Problems with body image can range from moderate dissatisfaction and concern with the body to profound dissatisfaction with physical appearance ¹ . Studies have revealed a high level of dissatisfaction with body image in adolescents, being more pronounced in females. Overall, research shows that dissatisfaction with body image may be linked to unhealthy weight control practices, eating disorders (anorexia, bulimia, and binge eating), and negative self-perception in relation to weight, health, and well-being ¹ . The literature suggests that body dissatisfaction has a multifactorial etiology and may be the result of physical characteristics: body size versus culture, social environment (emphasis on thinness and appearance), psychological factors (low self-esteem and depression), and a differentiated ideal of body image between the sexes (thin and sexy for women and strong and muscular for men) ¹ . Socioeconomic factors – education, race/ethnicity, income, and economic class – are also identified as influencing body dissatisfaction ² . With advances in knowledge of the social determinants of health, research in the area of ​​body image began to consider violence as another important aspect in its development. Kearney-Cooke and Striegel-Moore ³ state that in the presence of violence, a negative feeling or shame about the body – body dissatisfaction – may arise and favor unhealthy or inappropriate eating practices. Following this same line of argument, other studies suggest a relationship between sexual abuse and eating disorders in young women ⁴ . However, it is important to mention that other types of abuse, such as psychological and physical violence, are also identified as possible determinants of body image perception ⁵ . Therefore, uncovering the causes of body dissatisfaction is of utmost importance in order to minimize its consequences on the health of individuals, especially among adolescents. Exploring other types of violence will enable health professionals and the general public to better understand the relationship between family violence and possible changes in body image. This study aimed to investigate the relationship between physical and psychological family violence suffered by adolescents and dissatisfaction with body image. METHOD The study was conducted at a municipal health unit in Rio de Janeiro ( Policlínica Hélio Pellegrino ) and involved 201 adolescents aged 10 to 19 who received benefits from the Bolsa Família Program (PBF). These young people were selected through a convenience sample and presented similar characteristics regarding socioeconomic status . For example, the following prerequisites are required by the program: low family income, regular school attendance, and proof of periodic monitoring of their health status at a health unit. These adolescents were monitored by the nutrition service of this health unit from 2008 to 2009. Data were obtained through a pre-tested questionnaire, consisting of questions regarding the outcome (dissatisfaction with body image), central exposure (physical and psychological domestic violence), and covariates (demographic and personal information). Dissatisfaction with body image was measured using the Portuguese version of the Body Area Scale ⁶ . The scale consists of 15 items that cover the degree of dissatisfaction with the following body parts: face, hair, buttocks, hips, thighs, legs, stomach, waist, chest/thorax, back/shoulders, arms, muscle tone, weight, height, and all other areas. For each area, the adolescent self-assessed himself/herself according to a scale ranging from very dissatisfied (score 1) to very satisfied (score 5). Body image was assessed according to the average of the total scores for all body areas ⁷ , i.e., face + hair + buttocks + all areas/16 (total number of body areas), and classified as satisfactory (average >3) or unsatisfactory (average ≤3) ⁸ . Physical violence was measured by a Portuguese version of the Conflict Tactics Scales (CTS-1) instrument ⁹ . The CTS-1 is designed to measure the strategies used by family members to resolve possible disagreements and, indirectly, to capture a situation of family violence. The family relationships considered in this study were those of parents (father and mother) with adolescents. The scales of minor physical violence (items km) and serious physical violence (items ns) were classified as follows: did not occur in the relationship and occurred at least once in the relationship ¹⁰ . The time period investigated was 12 months before the interview. Psychological violence was measured by a Portuguese version of the psychological violence against adolescents scale ¹¹. This scale assessed the experiences of young people who had a significant person in their relationships who denigrated their qualities, abilities, desires and emotions or made excessive demands on them. It consists of 18 closed items, with five response options: never, rarely, sometimes, almost always and always, with scores of 1, 2, 3, 4 and 5, respectively. The score is calculated by adding the scores for each item, dividing this sum by the total points that each adolescent could score on the questionnaire and multiplying by 100. The presence or absence of psychological violence is guided by the median of this percentage of points, that is, presence of psychological violence (percentage of points ≥ median) and absence of psychological violence (percentage < median) ¹² . Demographic information and characteristics of the adolescents were collected, such as age, sex, race/color (black or non-black), health status (good or bad), racial discrimination (present or absent), housing conditions, social network and support, sexual maturation and nutritional status. Housing conditions were assessed by an instrument composed of items corresponding to the agglomeration of people, materials used to build the house, type of flooring, electricity, water supply, sewage system and the way in which garbage is collected. The data were scored and the total score was considered as satisfactory housing conditions when ≥ 9 and unsatisfactory housing conditions when < 9 ¹³ . The variables social network and support were assessed by a Portuguese version of the social support questionnaire used in the Medical Outcomes Study ¹⁴ . This instrument measures the perception of the availability of support. It is composed of five dimensions: 1) material; 2) affective; 3) positive social interaction; 4) emotional and 5) information. Each dimension comprises questions that are scored by scores ranging from 1 to 5, corresponding to the answers never, rarely, sometimes, almost always and always. By adding the scores of the questions of each of the dimensions, the score was calculated by dividing by the maximum total of scores for each dimension and multiplying by 100. The measure of social support was obtained by the average score of the five dimensions ¹⁵ . Social support was considered satisfactory when the average was equal to or greater than 75%. To measure the social network, the instrument presents items related to the number of friends and relatives, and the adolescent’s participation in social activities ¹⁵ . The information regarding the number of people the adolescent can count on was distributed as 0, 1, 2 and 3 or more. Sexual maturation was self-assessed by viewing the Tanner chart with photos representing the five stages of pubertal development ¹⁶ and the adolescent was classified as prepubertal or pubertal ¹⁷ . Nutritional status was assessed according to the body mass index (BMI) calculated by dividing body weight (kg) by height (m) squared. BMI was classified according to the criteria of the World Health Organization ¹⁸ and then dichotomized for logistic regression analysis into adequate weight/underweight and overweight. All analyses were conducted separately for male and female adolescents, considering that there are differences between the sexes regarding the perception of body image. Statistical analysis of the data proceeded as follows:

Univariate analysis, determining absolute and relative frequencies (prevalence) for categorical variables, and summary measure (mean, median, standard deviation) for numerical variables; Bivariate analysis, considering the body image variable as the outcome. For categorical variables, according to the contingency tables determined for each variable versus outcome, the Chi-square test was selected to test the association between them. For numerical variables, the unpaired Student’s t-test was used; Univariate logistic regression analysis (crude) between the exposure variables (minor physical violence, severe physical violence and psychological violence) and outcome variable (dissatisfaction with body image) and covariates of interest; Multivariate logistic models were created for each exposure variable (minor physical violence, severe physical violence, and psychological violence) and outcome variable (dissatisfaction with body image), adjusted for covariates that presented p < 0.20 in the crude analyses. The criterion of statistical significance for the multivariate logistic model was p < 0.05; The models were diagnosed using the Hosmer-Lemeshow test. The model was considered well adjusted when p > 0.05; and Odds ratios (OR) for each exposure variable in the multivariate logistic models, in addition to the respective 95% confidence intervals and p value.

To perform the analysis, the study used Epi Info 2000 to create the database, the software R Studio version 0.96.0331, and the R version 2.15 with MKmisc and epicalc packages and extensions for a more in-depth analysis. Only adolescents and their legal guardians who agreed and signed the Free and Informed Consent Form were included in the research. The research was approved and authorized by the Research Ethics Committee of the Municipal Health Department of Rio de Janeiro (opinion no. 122A/2007). RESULTS Table 1 shows the demographic and personal information of the adolescents studied. Regarding body image, approximately 20% of the adolescents were dissatisfied. Female adolescents had higher values ​​of dissatisfaction with body image than male adolescents, but this difference was not statistically significant (p = 0.454). Table 2 shows the frequency of each type of family violence perpetrated by parents against adolescents. Almost half of the sample experienced psychological violence, but female adolescents were the main victims. With regard to other types of violence, 30% of both sexes suffered serious physical violence and approximately 40% suffered minor physical violence. Based on the results obtained from the bivariate analysis (Table 3), it was observed that, for male adolescents, their nutritional status was significantly associated with body image (p = 0.018). Those who were overweight were 4 times more likely to be dissatisfied with their body image. For female adolescents, in addition to nutritional status (p = 0.019), psychological violence (p = 0.015) and housing conditions (p = 0.041) were also significantly associated with dissatisfaction with body image, that is, overweight, the presence of psychological violence and inadequate housing conditions generate approximately 3 times more chances of dissatisfaction with body image. The information from the multivariate models in relation to the OR and their respective confidence intervals are shown in Table 4. The models used the explanatory variables (types of violence), nutritional status and age for male adolescents and for female adolescents the types of violence, nutritional status and housing conditions. Para os adolescentes do sexo masculino (Tabela 4), as análises mostraram que nenhum tipo de violência foi significativamente associado à insatisfaçao com a imagem corporal (violência física menor – p = 0,668; violência física grave – p = 0,670 e violência psicológica – p = 0,854). Para as adolescentes do sexo feminino (Tabela 4), apenas a violência psicológica esteve significativamente associada à insatisfaçao com a imagem corporal (p = 0,012). Assim, as adolescentes do sexo feminino que sofreram violência psicológica tiveram 4,2 vezes mais chances de estarem insatisfeitas com a imagem corporal do que aqueles que nao sofreram violência psicológica. Apesar de nao serem estatisticamente significativas, as associaçoes foram positivas entre a violência física e a insatisfaçao com a imagem corporal das adolescentes (violência física menor – p = 0,242 e violência física grave – p = 0,076). Quanto mais grave a violência física, maior é a insatisfaçao com a imagem corporal – violência física menor (OR=1,8; IC=0,67-4,80) e violência física grave (OR=2,5; IC=0,91-7,01). O diagnóstico do modelo logístico multivariado mostrou, em ambos os sexos, valor de p superior a 0,05 para todos os tipos de violência familiar (dados nao mostrados). DISCUSSAO Este estudo investigou a relaçao entre a violência familiar realizada pelos pais e a insatisfaçao com a imagem corporal dos adolescentes monitorados pelo Programa de Assistência à Família. Era esperado que o adolescente vítima de violência apresentasse maior insatisfaçao com sua imagem corporal. Essa hipótese foi suportada somente para as adolescentes do sexo feminino vítimas de violência psicológica, tanto em análises bivariadas quanto multivariadas. A exposiçao a fatores estressantes nos estágios iniciais da vida pode afetar a imagem corporal. Murray, Byrne e Rieger¹⁹, em seu estudo envolvendo 533 adolescentes australianos, examinaram a relaçao entre estresse e imagem corporal e descobriram que as adolescentes do sexo feminino com maiores níveis de estresse apresentaram mais sintomas de depressao, maior insatisfaçao com a imagem corporal e baixa autoestima do que os adolescentes do sexo masculino. Quando se considera que a existência de violência familiar pode ser uma situaçao estressante para os adolescentes²⁰, a imagem corporal é consequentemente afetada de forma negativa, o que leva aos mesmos resultados do presente estudo. Desta forma, as adolescentes do sexo feminino expostos à violência mostram maior insatisfaçao com a imagem corporal. Em relaçao aos tipos de violência, os estudos que investigaram a relaçao entre violência e imagem corporal em adolescentes avaliaram a violência sexual^21-23, a violência física^{21, 23} e a negligência²². Schaaf e McCanne²³, researching female students, found no relationship between physical and sexual violence and body image distortions, despite the fact that young females who suffered physical abuse had higher personal expectations and greater difficulty in identifying emotions and sensations of hunger and satiety. However, Logio ²¹ , when evaluating 1571 adolescents in New Castle (USA), observed that those who suffered physical and sexual violence showed changes in the perception of their body image towards excess weight, with greater evidence for female adolescents. At the same time, there was a strong correlation between adolescents with a history of abuse and the presence of eating disorders and dieting behavior. Ramalhete and Santos ²² , in a study with female adolescents who were victims of neglect and sexual abuse, observed that these adolescents had a more negative view of the body, regardless of the type of violence. These findings identify past experiences of abuse as a risk factor for unhealthy eating practices and an unhealthy self-assessment of body image. It is important to emphasize that there are few studies on the subject ^5,21,22 and these employ different methodologies to assess body image and measure family violence. Regarding body image, instruments/scales can assess body image disturbances according to perceptual and attitudinal aspects. Studies that express the perceptual component measure body image by the degree of accuracy with which body size is perceived, while the attitudinal component expresses the affective and cognitive components ²⁴ . Most of the studies mentioned above measured body image according to the attitudinal aspect, despite the use of different scales/instruments. It is worth mentioning that this study also assessed body image by the attitudinal aspect. In family violence, there is disagreement among researchers regarding the definition of which types of behaviors or punishments should be considered abuse and also regarding temporality – past versus current ²⁵ . Consequently, it is not always possible to compare the results found in the few existing investigations. Yet, evidence suggests that a stressful family environment creates serious problems in the development of children and adolescents ²⁶ . Despite the relative lack of studies examining the relationship between psychological violence and body image dissatisfaction in adolescents, existing research shows a link between violence in childhood and body image problems in adulthood ²⁶. Violence suffered in childhood has negative consequences and can lead to the emergence of psychopathologies in adulthood. Psychological violence has been associated with self-criticism, dissatisfaction with body image, depression and low self-esteem among adults ⁵ . Therefore, it is important to have a safe and supportive family environment to develop a healthy body image in individuals, particularly in the case of adolescents. The body is not simply a biological machine, but a body that comprises a history of relational experiences that will assist in the development of the individual and in the ability to deal with stressful situations. It is also a means of regulating affections or resolving problems of a relational/personal nature ²⁷ . Eubanks et al. ⁴ , in their study with 38 university students, observed that victims of physical and psychological abuse felt that their parents loved them less, gave them less support and acted as a negative role model in their lives. This shows once again how the existence of physical or psychological violence within the family can negatively affect the lives of the individuals involved. From this perspective, healthy relationships that adolescents enjoy with other family members are potentially decisive in leading to adequate development and good social relationships ²⁷ . This study has certain limitations: the first concerns the sample design, which was based on a cross-sectional study, which is not capable of inferring the temporality of the associations described. The second limitation refers to the type of sample. The present study used a convenience sample of individuals registered in the Family Assistance Program, with similar socioeconomic characteristics. Another limitation to be considered refers to the variables that may be involved in the relationship between violence and dissatisfaction with body image, which were not addressed here and which can generate body image problems, such as: shame, self-criticism, mood swings, perfectionism, puberty, self-objectification, identification of gender roles, sexuality, anxiety and other experiences that are potential mediators in the relationship between childhood maltreatment and body image problems. ⁵. Furthermore, given that some children who suffer abuse do not develop serious psychological consequences or other health problems, it is an important task for the future to identify the factors that protect these children from the negative consequences of abuse. From this point of view, it is important that further research investigate the relationship between family violence and body image, developing longitudinal studies and considering other aspects, such as compensatory mechanisms, psychological factors and the perception of family relationships that, when favorable, help in overcoming trauma. CONCLUSION Despite the limitations, this was the first study on the relationship between physical and psychological violence and body image dissatisfaction in Brazilian adolescents. Based on the results of this study, it can be assumed that psychological violence is linked to body image dissatisfaction in female adolescents. This knowledge can help to support strategies that allow adolescents and their parents to understand the detrimental effects that an unfavorable family environment can have on an adolescent’s growth and development. It should also be clear to society that violence, culturally accepted as part of the educational process, can have negative consequences for the formation of body image.