Rape in Brazil and relationships with alcohol consumption : estimates based on confidential self-reports

This study presents the rape prevalence and its relationship with alcohol consumption based on the analysis of cross-sectional data from the 2012 Second National Alcohol and Drugs Survey, which used a cluster-stratified probabilistic sample of the Brazilian population. We included 1,918 men and 2,365 women, for a total sample of 4,283 individuals. Our results estimate the overlifetime rape prevalence at 2.6% for the entire population, 1.7% for men and 3.5% for women. For both sexes, the highest prevalences were concentrated among those aged between 26 and 59 years (3.3%), those with low educational levels (3.8%), and those who were single, divorced or widowed (3.1%). Among individuals who were diagnosed with alcohol use disorder, according to the DSM-5 (Diagnostic and Statistical Manual of Mental Disorders), 6% reported having been raped, as did 3.3% of those who reported binge drinking. Logistical and multivariate regression analyses showed that, for both women and men, age (those between 26 and 59 years), marital status (single), alcohol use disorder and binge drinking are factors associated with an increased probability of being raped, whereas a higher educational level (more than 9 years of schooling) was shown to be capable of reducing the odds of being raped. Awareness of prevalences and of victims’ sociodemographic profiles is necessary in order to identify the scope of this public health problem as well as to evaluate the urgency of implementing preventative and supportive measures. Rape; Sex Offenses; Alcoholism; Risk Factors ARTIGO ARTICLE This article is published in Open Access under the Creative Commons Attribution license, which allows use, distribution, and reproduction in any medium, without restrictions, as long as the original work is correctly cited. Massaro LTS et al. 2 Cad. Saúde Pública 2019; 35(2):e00022118 Introduction The Brazilian Penal Code, Law 12,015, 2009, defines rape as: “forcing someone, through violence or grave threat, to have sexual intercourse or to engage in, or permit someone else to engage in, another libidinous act” 1. It is a form of violence, considered to be a public health problem and a violation of universal human rights, which affects all social classes, ethnicities, religions, economies and cultures indistinctly 2. Data from 2005, regarding 100,000 individuals from 84 Brazilian cities, reveal that 67% of the population fears being sexually assaulted. Among men, 42% fear being raped, while among women that percentage reaches 90% 3. Victimization by rape may cause immediate trauma and long-term physical and psychological outcomes, such as injuries to genital organs, sexually transmissible infections, depression, alcohol use, anxiety disorders, post-traumatic stress disorder and attempted suicide 4. Another noteworthy severe consequence is pregnancy resulting from rape when there is a failure in the provision of prophylactic measures in public health services 5. Data published by the Brazilian Institute for Applied Economic Research (IPEA, in Portuguese) in 2014 shows a 66% increase in rape reports between 2011 and 2014, with one rape occurring every 11 minutes 6. According to the Ministry of Health, sexual abuse is the second most frequent type of violence in Brazil 7, with 527,000 individuals raped every year, 70% of which are children and teenagers under 17 years of age. It is believed that these data, though originating in official statistics, are underestimated, since only 10% of cases are reported to police 6. Before 2009, the law defined raped as “forcing a woman to have sexual intercourse, through violence or grave threat”, implicitly stating that only women could be victims of rape, while men were necessarily the aggressors. The change in the Brazilian Penal Code contributed to directing the focus to the crux of the problem: the lack of consent for a sexual act, regardless of the sex and gender of the victim or the aggressor. Rape and its impacts on the female population are relatively well-discussed in the national literature 8. The same cannot be said for male victims. Hohendorff’s 9 recent work has shown the invisibility of sexual violence against boys, whether due to the small number of reported cases or due to the disbelief and prejudice with which these cases are handled. It also highlights the network’s unpreparedness, which perceives the need for constant training. Parkill et al. 10 suggest that alcohol consumption and the context of rape may be connected. Considering this issue within the Brazilian population, this study seeks to present the prevalence of rape among men and women, as well as to investigate its association with problematic alcohol consumption, addressing binge drinking and alcohol use disorders (AUD). The study is particularly relevant because of its methodology, which, given the focus on drug use, prioritized participants’ confidentiality, thus avoiding the usual bias present in face-to-face interviews.


Introduction
The Brazilian Penal Code, Law 12,015, 2009, defines rape as: "forcing someone, through violence or grave threat, to have sexual intercourse or to engage in, or permit someone else to engage in, another libidinous act" 1 .It is a form of violence, considered to be a public health problem and a violation of universal human rights, which affects all social classes, ethnicities, religions, economies and cultures indistinctly 2 .Data from 2005, regarding 100,000 individuals from 84 Brazilian cities, reveal that 67% of the population fears being sexually assaulted.Among men, 42% fear being raped, while among women that percentage reaches 90% 3 .
Victimization by rape may cause immediate trauma and long-term physical and psychological outcomes, such as injuries to genital organs, sexually transmissible infections, depression, alcohol use, anxiety disorders, post-traumatic stress disorder and attempted suicide 4 .Another noteworthy severe consequence is pregnancy resulting from rape when there is a failure in the provision of prophylactic measures in public health services 5 .Data published by the Brazilian Institute for Applied Economic Research (IPEA, in Portuguese) in 2014 shows a 66% increase in rape reports between 2011 and 2014, with one rape occurring every 11 minutes 6 .According to the Ministry of Health, sexual abuse is the second most frequent type of violence in Brazil 7 , with 527,000 individuals raped every year, 70% of which are children and teenagers under 17 years of age.It is believed that these data, though originating in official statistics, are underestimated, since only 10% of cases are reported to police 6 .
Before 2009, the law defined raped as "forcing a woman to have sexual intercourse, through violence or grave threat", implicitly stating that only women could be victims of rape, while men were necessarily the aggressors.The change in the Brazilian Penal Code contributed to directing the focus to the crux of the problem: the lack of consent for a sexual act, regardless of the sex and gender of the victim or the aggressor.
Rape and its impacts on the female population are relatively well-discussed in the national literature 8 .The same cannot be said for male victims.Hohendorff's 9 recent work has shown the invisibility of sexual violence against boys, whether due to the small number of reported cases or due to the disbelief and prejudice with which these cases are handled.It also highlights the network's unpreparedness, which perceives the need for constant training.
Parkill et al. 10 suggest that alcohol consumption and the context of rape may be connected.Considering this issue within the Brazilian population, this study seeks to present the prevalence of rape among men and women, as well as to investigate its association with problematic alcohol consumption, addressing binge drinking and alcohol use disorders (AUD).The study is particularly relevant because of its methodology, which, given the focus on drug use, prioritized participants' confidentiality, thus avoiding the usual bias present in face-to-face interviews.

Method
The data we analyzed come from the Second National Alcohol and Drugs Survey (LENAD II, in Portuguese), a repeated cross-sectional study carried out in 2012 which investigated the patterns of alcohol, tobacco and illicit drug use among the Brazilian population, possible risk and protection factors for the development of abuse and/or addiction, as well as factors associated with use.Additional information regarding LENAD II can be found in: https://inpad.org.br/wp-content/uploads/2014/03/Lenad-II-Relat%C3%B3rio.pdf.

Sampling
LENAD II used a cluster-stratified probabilistic sampling technique, with allocations proportional to each state (stratum), in order to guarantee that the final sample would be representative of all Brazilian regions.Brazilian aged 14 years or older were considered eligible.We excluded indigenous populations living in indigenous territories, Brazilian residents who do not speak Portuguese and people with mental disabilities who were not able to answer the questionnaire.The survey had a 77% response rate with a total of 4,607 participants.

Instruments
In this analysis, we used sociodemographic, alcohol consumption and sexual violence indicators.LENAD II used a mixed technique for data collection, including a face-to-face interview (mean duration: 50 minutes) and a questionnaire filled out by participants themselves.In order to minimize measuring bias, during the interview phase, carried out by a team of trained researchers, we used scales that had been validated for the Brazilian population, aimed to investigate use of and addition to alcohol, tobacco, marijuana and cocaine.In an attempt to attain the greatest possible accuracy for answers related to delicate issues (such as child sexual abuse, rape, unprotected sex and sexual orientation), and to minimize information bias, at the end of the interview, participants answered the questionnaire regarding these issues by themselves, in a separate environment, returning it to the interviewer in a sealed envelope.

• Binge drinking
We adopted the concept of binge drinking proposed by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) 11 , which defines it as ingesting four doses (for women) and five doses (for men) of alcoholic beverages in a two-hour interval.
A card specifying the correspondence between milliliters and dose for different types of alcoholic beverages was presented to participants in order to avoid measurement bias for this variable.
Binge drinking was evaluated through the question (yes/no): "Have you ever drink four (women) or five (men) or more doses of any alcoholic beverage in a 2 hour period?".

• Alcohol use disorders (AUD)
The alcohol consumption section was based on the HABLAS questionnaire, developed by Caetano et al. 12 , and has a subsection for evaluating abuse and addiction based on DSM-IV.Both questionnaires (2006/2012) also present the criteria for diagnosing AUD, according to the DSM-5.The 2012 questionnaire also uses the Alcohol Use Disorder Identification Test (AUDIT), translated from the English household survey 13 .

• Rape
Information regarding sexual violence was collected in sealed envelopes.The question we used to assess this subject was: "Have you ever been forced to have sexual relations with someone?".

Ethical aspects
This study was approved by the Federal University of São Paulo's (UNIFESP) Ethics Research Committee, under the report 1.833.235.Participants filled out an informed consent form which guaranteed anonymity when results were published.A copy of the form was provided to participants, with contact information for the principal investigator and the UNIFESP Ethics Research Committee, so they could obtain answers to any doubts regarding the study.For participants under the age of 16 years, we requested formal authorization from parents or guardians, following the requirements established by the National Research Companies Association's Ethical Guidelines.

Statistical analyses
We used the statistical software Stata 13SE (https://www.stata.com)for all statistical analyses.Given the probabilistic population design, all analyses were weighted in order to take into account the different selection probabilities at each data collection stage.We used the svy command for complex samples.The variable association models were calculated using logistical regression, using the specific commands in order to generate robust standard error and odds ratio values with a 95% confidence interval (95%CI).

Results
We analyzed a sample of 4,283 participants, 1,918 of which were male and 2,365 of which were female.According to Table 1, the prevalence of rape among Brazilians over 14 years of age was estimated at 2.6%, with prevalences of 1.7% among men and 3.5% among women.The age group between 26 and 59 years had the highest prevalence, 3.3%.This group's odds ratio of being raped was significantly increased by nearly four times, both for men (OR = 3.9; 95%CI: 1.0-15.9)and for women (OR = 3.6; 95%CI: 1.1-11.4),when compared with the other age group.Individuals who were single had double the odds of being raped when compared with those who were married (OR = 2.0; 95%CI: 1.1-3.7).
A higher educational level (more than nine years of schooling) significantly reduced the odds of being raped in the population as a whole (OR = 0.4; 95%CI: 0.2-0.8)and among women (OR = 0.3; 95%CI: 0.1-0.6).In the population as a whole, 6% of individuals who had AUD reported having been raped.The odds of being raped were 3.6 times bigger in this group (OR = 3.6; 95%CI: 1.9-7.1)when compared with individuals who do not have AUD.When looking only at alcoholic men, the odds were five times higher (OR = 5.2; 95%CI: 1.5-17.2).The rape prevalence among those who reported binge drinking reached 3.3%.These individuals had double the odds (OR = 2.3; 95%CI: 1.5-3.6) of being raped, both among men and among women.

Table 1
Rape prevalences over lifetime according to sociodemographic characteristics and associated factors.

Discussion
Though rape is undeniably a priority in both security and health, obtaining accurate rape estimates is clearly difficult.For the most part, data are based on police reports and medical charts, both sources that are known to be under-reported.Investigation through population surveys is also a challenge, given the lack of appropriate interview strategies that preserve participant confidentiality when reporting such a delicate event 6,7 .The sealed envelope technique used in this study is a methodological innovation in investigations of rape, guaranteeing, for the first time, participant confidentiality with regard to interviewers and obtaining accurate data about the prevalence of rape in a representative sample of the Brazilian population.Our results indicate that the prevalence of rape is 2.6% among the population as a whole, 1.7% among men and 3.5% among women.For both sexes, the highest prevalences are concentrated among individuals aged between 26 and 59 years (3.3%), those with low educational levels (3.8%) and those who were single, divorced or widowed (3.1%).Among individuals who had AUD, 6% reported having been raped, as did 3.3% of those who reported binge drinking.
The regression analyses indicated that age between 26 and 59 years, being single, having an AUD and binge drinking were associated with an increased probability of being raped, whereas a higher educational level (over 9 years of schooling) reduced the odds of being raped.
There are no studies in Brazil with a compatible methodology that offer comparable estimates.In 2013, based on data on victimization, IPEA estimated that every year, in Brazil, 0.26% of the population is sexually assaulted, a percentage that is ten times smaller than our results show.The study's author explains that, due to methodological issues, this estimate should perhaps be considered the lower limit of the country's prevalence 6 .
Epidemiological data regarding sexual violence in the United Kingdom, also collected through the sealed envelope methodology, estimated the rape prevalence for the population over 16 years of age at 1.6%, reaching 0.3% among men and 3% among women 14 .In the data comparison, our results point to considerably higher prevalences, especially among men.American data collected over the phone estimated the national, over-the-lifetime rape prevalence at 19.3% among women and 1.7% among men 15 .Although the prevalence among men corroborates our findings, the prevalence among women is nearly six times higher than our results show.Even though the literature shows an increase in the number of cases in these countries, constant epidemiological surveillance coupled with the organization of an efficient service network, including legal, medical and psychological care, has led to a gradual reduction in subnotification rates 15,16 .In Brazil, unfortunately, this is not yet a reality.In addition to a lack of epidemiological studies with adequate methodology, capable of producing systematically updated rates, the services offered to rape victims are known to often be inefficient in terms of health care and social assistance 17 , as well as in terms of the police reportinh process, being far from able to meet this population's demands 18 .
Regarding the proportion among the sexes, there is a consensus in the national and international literature that women are the most common victims 6,14,15 , but we must consider that 1.7% of Brazilian men have been victims of rape, being subjected to the harms that this type of violence causes.Though male rape was acknowledged by the Brazilian Penal Code in 2009 1 , opening up this discussion, this has not contributed to greater identification of these cases or greater support for male victims.The scarce Brazilian literature that has been produced regarding male rape is restricted to specific populations, such as inmates 19 , and to regional studies 20 , but it does raise the lack of basic resources needed to provide care to these individuals, whose heavy stigma and gender biases lead to silence and unnotified cases.
Data from health services provided to lesbians, gays, bisexuals and trans individuals (LGBT) victims of violence in the Brazilian Northeast show the presence of homophobic discourses among health professionals, who are influenced by social stereotypes regarding sexual orientation.Thus, the stigmatization of sexual minorities in health services stands in opposition to what is recommended by the Brazilian National Humanization Policy (PNH, in Portuguese) and the National Integral Health Policy for the LGBT Population, both of which are aligned with the Brazilian Unified National Health System (SUS, in Portuguese) principles by emphasizing the need for guaranteeing integral care and equity, seeking strategies to expand individuals' citizenship and rights within the system 21 .
Cad. Saúde Pública 2019; 35(2):e00022118 The mean age described in national surveys that exclude individuals under 12 years of age is 23.7 years.Our results indicate greater risk and prevalence among individuals aged between 26 and 59 years.The age difference may be explained by methodological differences in the categorization of the age variable.As to marital status, our results corroborate previously presented information which reveals greater rates among those who are single, divorced or widowed 6 .
Previously published studies are consistent with the results presented in this article regarding education, justifying the higher prevalences among individuals with lower educational levels.The data presented here in the regression analysis indicate that higher educational levels are significantly associated with a reduction in the odds of victimization by rape in the general population and among women.The role of education as a crucial tool for perceiving and reacting to signs of abuse and for preventing violence in general, and sexual violence in particular, is largely ignored, particularly in public safety policies and in the development of universal and basic prevention programs.A study that included 10,000 primary school children in 2008 carried out child abuse prevention actions directed at children aged 6 to 10 years and had positive results in terms of promoting self-protection skills against sexual abuse, in addition to early identification of cases, avoiding even higher costs to child and adolescent victims of sexual abuse 22 .Estimates released by recent national studies suggest that the probability of an individual with up to seven years of schooling being murdered in Brazil is 15.9 times higher when compared with individuals with some university education and that, for each 1% increase of young people aged 15 to 17 years enrolled in schools, there is a 2% decrease in the murder rate 23 .It is speculated that, due to the known association between education and income, individuals with higher educational levels reside in regions with greater protection and lower incidence of crimes, including sexual crimes.
Exposure to violence, whether physical, sexual or psychological, associated with the use of alcohol and drugs has been documented in many countries with regard both to aggressors and to victims 24 .Women undergoing treatment for alcohol or other drug use report high rates of victimization 25 .This cycle is established when the excessive consumption of alcohol reduces decision-making skills, thus increasing the odds of being involved in sexual violence episodes 26 .
This epidemiological study described and analyzed the relationship between alcohol consumption and victimization by rape.The results indicate that individuals with AUD and those who binge drinks have their odds of being raped increased by 3.6 and 2.4 times, respectively.These data reinforce the urgency of focusing on this more vulnerable population and the lack of effective strategies for preventing abusive alcohol consumption.From a legal standpoint, victims' vulnerability due to alcohol is addressed by the Brazilian Penal Code as rape of a vulnerable person, punishable by 10 to 30 years of incarceration 1 .
An important limitation of this study was our inability to use the income variable, due to nonresponse, in order to estimate rape prevalences in different income categories.However, the high correlation between education and income 27 enables us to infer the rape/income relationship as similar to what we observed for the education variable.Therefore, the higher rape prevalences would be found among individuals with lower educational levels and lower income.Another limitation worth mentioning is the lack of sample segmentation into vulnerable groups, such as LGBT.We chose to use the variable sex for statistical analysis, since the previously-mentioned stratification did not yield satisfactory sub-samples for other analyses.

Conclusion
The scientific literature corroborates that violence, though a universal phenomenon, can be avoided 23 .The data presented in this article reveal the vulnerability of single young women and men with low educational levels to victimization by rape and reiterate the connection between alcohol consumption and sexual violence, which affects 2.6% of Brazilians over their lifetimes.
Considering the protective effect of education, setting up an educational system the provides, in addition to academic knowledge, opportunities for promoting human capital and citizenship is an action with great protective potential.Universal and focused prevention programs, as well as health services, could also benefit from the protective effect of education, addressing alcohol consumption Cad.Saúde Pública 2019; 35(2):e00022118 as a risk factor for exposure to violence, following the example of what has been done with regard to the combination of alcohol and dangerous driving.
For those whom prevention actions could not reach, there needs to be services capable of addressing the needs and costs generated by rape.From the clinical standpoint, according with the literature on best practices for rape cases 28 , humanized care by trained professionals must follow protocols that guarantee the clinical treatment of possible injuries, prevention of sexually transmitted infections, evaluation of pregnancy risk and quick access to psychological care.
The development of efficient prevention strategies and treatment protocols that meets the different needs and characteristics of rape victims in our population requires knowledge of prevalences and sociodemographic profiles.This article seeks to make this contribution and suggestion of epidemiological studies that explore rape among the LGBT population.

Contributors
L. T. S. Massaro participated in the analyses, interpretation, discussion of statistical data according to the updated bibliographical research and in writing the final text.L. Adesse participated in the analyses, interpretation, discussion of statistical data according to the updated bibliographical research and revised the text for publication.R. Laranjeira contributed substantially to the study development in its different phases, from project conception to data collection and tabulation, literature review and writing the final text.R.