Prevalence of violence against women living in rural areas and associated factors: a cross-sectional study based on the 2019 National Health Survey

The aim of this cross-sectional study was to estimate the prevalence of violence against women living in rural areas, explore associated factors, and characterize cases of violence according to perpetrator, place of occurrence, and frequency. Based on data from the 2019 National Health Survey, using Poisson’s regression we calculated crude and adjusted prevalence ratios for violence committed during the last 12 months against women living in rural areas across Brazil, focusing on the following variables: sociodemo-graphic characteristics, income, social support, and self-reported health status. The prevalence of psychological, physical, and sexual violence was 18%, 4.4%, and 1.5%, respectively. Perpetrators were mainly people known to the victim and violence was mainly committed at home and repeated over time. Prevalence was highest among young women (24.2%), single and divorced women (20% each), women who had complete elementary school till not complete higher education (22% each), women with very poor (34%) and poor (30%) self-perceived health status; and women with a mental health problem (30%). After adjustment, the following variables were retained in the model: women aged 30-39 years and 40-49 years; married women; women with very poor, poor, and fair perceived health; and women diagnosed with a mental health problem.


Introduction
Violence against women is ingrained and normalized in our society.The UN Declaration on the Elimination of Violence against Women adopted by the United Nations General Assembly in 1993 states that violence against women "constitutes a violation of the rights and fundamental freedoms of women".The concept of violence against women adopted by the National Policy to Combat Violence against Women is based on the definition proposed by Article 1 of the Inter-American Convention on the Prevention, Punishment and Eradication of Violence against Women (Belém do Pará Convention): "any act or conduct, based on gender, which causes death or physical, sexual or psychological harm or suffering to women, whether in the public or the private sphere" 1 .Regarding the affirmation that violence against women is gender-based, Saffioti 2 states that gender violence is a broad concept that encompasses, besides women, children and adolescents of both sexes and is enrooted in power imbalances stemming from archetypes created in a patriarchal system.Viewed from this perspective, men hold the power to determine the conduct of social categories, receiving the approval, or at least the tolerance, of society to punish what appears to them to be a deviation.
According to Law 11,340, known as the "Maria da Penha Law", enacted in 2006, domestic and family violence against women is "any gender-based act or omission that causes death, injury, physical, sexual or psychological suffering and moral or property damage" 3 .This Law has played a fundamental role in curbing, punishing, and preventing violence against women and in establishing protection and assistance measures.
Women living in rural areas are subjected to all types of violence.The following factors in theses area can contribute to violence: geographic isolation, resulting in long distances to violence support and health services; poor road quality and conditions and transport; and lack of access to internet and other means of communication.In addition, lack of witnesses, distance from family and relatives, and cultural issues reinforce the invisibility and silence surrounding women who suffer from violence [4][5][6] .
Nationwide quantitative research on violence against women living in rural areas remain scarce in Brazil.Studies include a survey conducted by the Institute of Applied Economic Research (IPEA) with 611 women during the 2011 Marcha das Margaridas.The researchers found that 58%, 28%, and 23% of the women had suffered psychological/moral, physical, or sexual violence, respectively, during their lifetime 7 .It is important to highlight that the women who responded the survey were taking part in a union movement that addresses violence.They also broke the mold by "leaving home" and it is possible that they had a broader perspective of violence than other women 8 .Another study using data from the 2013 National Health Survey (PNS) showed that 3.0% and 1.41% of women living in rural areas reported having suffered violence committed by someone they know or a stranger, respectively, over the last 12 months.However, the 2013 PNS asked only about "violence or aggression" without specifying different types of violence, which may have limited the number of positive responses by respondents.
Given the relevance of violence against women living in rural areas to public health, its consequences for women's health and other areas of life, and the scarcity of national studies on the issue, further research in this area is fundamental to gain a better understanding of this problem.Moreover, nationwide studies of violence against women are crucial to help formulate and strengthen initiatives to tackle violence, guarantee women's rights, and ensure their safety and well-being.This study is unique in that it focuses on violence against women living in rural areas, using data from the reformulated version of the PNS, which distinguishes between psychological, physical, and sexual violence.To this end, the aim of this study was to estimate the prevalence of violence against women living in rural areas, explore associated factors, and characterize cases of violence according to perpetrator, place of occurrence, and frequency.

Study design and sampling plan
We conducted a cross-sectional study using data from the 2019 PNS.The PNS is a representative national household survey conducted in urban and rural areas across all regions of the county, encompassing all states, capital cities and metropolitan regions.The survey was conducted between August 2019 and March 2020 by the Brazilian Institute of Geography and Statistics (IBGE) in partnership with the Ministry of Health and Oswaldo Cruz Foundation (FIO-CRUZ).
The data were collected using a three-stage cluster sampling design including the following levels: census tracts or cluster of tracts (primary unit); households (secondary unit); and adult residents (tertiary unit).A detailed description of the method and further details can be found in publications produced by the IBGE 9,10 .
The study population was women aged 18-59 years living in rural areas.The total sample comprised 6,677 women, giving a weighted sample size of 7,958,079.

Study variables
The study outcome was prevalence of violence among women living in rural areas.We used data from women who responded the 2019 PNS violence module, which consists of questions on psychological, physical, and sexual violence.A yes answer to one of the following questions was taken to indicate that the woman had experienced violence: Psychological The module also contains information on the perpetrator of violence (current or ex-boyfriend/ partner/husband, relative, friend, stranger, other); place of occurrence of the act of violence (at home, at work/education establishment, public thoroughfare/public space/bar, restaurant, on the internet, social media or cell phone -only psychological violence, others); and frequency of violence (often; sometimes; rarely).
The sociodemographic variables were as follows: region (Midwest, Northeast, North, Southeast, and South); age (18-29 years, 30-39 years, 40-49 years, and 50-59 years); race/color (white, black/brown, yellow/Asian/indigenous); marital status (married, divorced, widow or single); education level (no schooling/not completed elementary school, completed elementary school/ not completed high school, completed high school/not completed higher education, and completed higher education); per capita income (up to ½ minimum wage, between ½ and 1 minimum wage, between 1 and 2 minimum wages, between 2 and 5 minimum wages, and more than 5 minimum wages); support (whether the respondent had family/relatives/friends on who they can rely during good and bad times), self-reported health status (very good, good, fair, poor, very poor); mental health problems (whether the respondent had been diagnosed with depression; whether the respondent had been diagnosed another mental illness, such as anxiety disorder, panic disorder, schizophrenia, bipolar disorder, psychosis, obsessive-compulsive disorder).

Data analysis
Descriptive statistics were used to analyze the data and the prevalence of violence was calculated according to sociodemographic characteristics adopting a 95% confidence interval.Pearson's chi-squared test was used to examine statistically significant differences adopting a 5% significance level (p-value < 0.05).Descriptive statistics were also used to analyze cases of psychological, physical, and sexual violence according to perpetrator, place of occurrence, and frequency of violence.All analyses were performed using the sample parameters of the 2019 PNS and individual weighting.
To explore factors associated with violence against women living in rural areas, we analyzed explanatory the variables mentioned above related to sociodemographic characteristics, social support, and mental and physical health status.The outcome variable was having suffered either psychological, physical, or sexual violence during the last 12 months.Crude and adjusted prevalence ratios and their respective 95% confidence intervals were calculated using Poisson regression.All variables with p-value < 0.20 were included in the initial model.Variables with p-value < 0.05 were retained in the final model.The analyses were performed using R version 4.2.2 and the survey and tableone packages for complex samples.

Ethical aspects
The 2019 PNS protocol was approved by the National Health Council's National Research Ethics Committee (CONEP) on 23 August 2019 (approval code 3.529.376).All respondents signed an informed consent form.

Results
Table 1 shows that that most of the women in the sample were from the Northeast (50.3%), with similar proportions across all age groups.Most of the sample were black/brown (67.5%) and single (51.8%) or married (41.5%).Most of the women had no schooling or had not completed elementary school (48.7%) and 30.0% of the sample had completed high school but had not completed higher education.Over half of the women (55.5%) had a per capita income of up to half a minimum wage.Most of the women (75.2%) said that they had family/relatives/friends they could rely on during good and bad times.Almost 60% reported that they had good health status, while 26.0% said they had fair health.Sixteen per cent reported depression or other mental health problems.
Table 2 shows the prevalence of violence against women living in rural areas according to sociodemographic and economic characteristics, social support, and health.The findings show that prevalence was highest in the youngest age group (18-29 years), where 24.2% of women had experienced violence, and among single and divorced women, with prevalence being around 20% in each group.Regarding education level, prevalence was highest among women who had completed elementary school but not completed high school and who had completed high school but not completed higher education (21.2% and 21.7%, respectively).With regard to health, 33.6% of women who reported very poor health status, 29.1% of those who reported poor health status, and 24.0% of those with fair health reported experiencing violence, while 30% women who reported having been diagnosed with depression or another mental health problem reported having suffered violence.The other variables were not statistically significant.
Table 3 shows cases of psychological, physical and sexual violence during the last 12 months according to perpetrator, place of occurrence, and frequency.The findings show that the most prevalent type of violence was psychological violence (18.0%), followed by physical violence (4.37%) and sexual violence (1.42%).The most common type of perpetrator of acts of psychological and physical violence were people known to the victim: in 32.5% of cases of psychological violence and 48.1% of cases of physical violence the perpetrator was a current or ex-boyfriend, partner, or husband, while in 33.2% of cases of psychological violence and 28.1% of cases of physical violence the perpetrator was a relative; the main perpetrator in cases of sexual violence was a current or ex-boyfriend, partner, or husband (61.5%).
The most common place of occurrence for all types of violence was at home: psychological violence, 61.2%; physical violence, 70.9%; and sexual violence,75.4%.The most common category of frequency of violence during the last 12 months was "sometimes" for both psychological and sexual violence (44.6% and 49.1%, respectively).The most common category for physical violence was "rarely" (41.5%), followed by "sometimes" (39.9%).The percentage of women who reported that the frequency of violence was "often" was almost 20.0% for both psychological and physical violence and 13.4% for sexual violence.
Table 4 shows the crude and adjusted prevalence ratios for the variables included in the logistic regression model for the outcome occurrence of either psychological, physical or sexual violence during the last 12 months among women living in rural areas.In the multivariate model, prevalence of violence was highest in women in the 30-39 year (PR:1.81;95%CI 1.37-2.39)and 40-49 year (PR: 1.37; 95%CI 1.07-1,76) age groups.Prevalence was lower in married women (PR:0.65;95%CI 0.54-0.77)than in the other categories of marital status.The chance of having suffered some type of violence increased with deteriorating self-reported health, with the prevalence of violence in women who reported very poor health status being two times higher (PR:2.43;95%CI 1.31-4.49)than in those who reported very good health.The prevalence of violence was higher in women who reported having been diagnosed with a mental health problem (PR:1.65;95%CI 1,36-2,01) than in those who did not.

Discussion
Our findings show that 18.6% of the women living in rural areas reported having suffered some type of violence (psychological, physical and/ or sexual) during the last 12 months.The most common type of violence was psychological violence, followed by physical and sexual violence.In a study using data from the 2013 PNS to examine the effect of violence on self-perceived health among women aged between 20 and 49 years, Cruz and Irffi (8) found that almost 3.0% of women reported having suffered violence committed by someone known to them (such as a husband, ex-husband, father/mother, brother, employer, or relative) and 1.41% had suffered at the hands of a stranger (thief, police officer, mugger etc.) during the last 12 months.However, as mentioned above, the questionnaire used for the 2013 PNS only contained questions about "violence or aggression", meaning that other types of violence may have been underreported.The questionnaire used for the 2019 PNS distinguishes between three types of violence, broadening the perception of violence.Psychological violence is silent and ongoing and reduces the victim's self-esteem and adversely affects women's health.This is generally the first type of violence experienced, intensifying over time and co-occurring with physical and sexual violence 11,12 .According to Saffioti 13 , different types of violence (physical, sexual, psychological, property, or moral violence) do not occur in isolation and psychological violence is always present regardless of the form of aggression, frequently preceding other types of abuse.
Violence also adversely affects women's health.The findings of the present study show that prevalence of violence increased with deteriorating self-reported health.Cruz and Irffi 8 observed that women living in rural areas who suffered violence committed by someone they know were less likely to report having very good or good health and more likely to report fair health.It is important to stress that the health consequences of violence can be short-, mediumand long-term.Resulting health problems can be physical -including injuries, headaches, gastrointestinal disorders, limited mobility -mental -such as depression, post-traumatic stress, anxiety, eating problems, and suicide attemptsor sexual and reproductive -such as unwanted pregnancies, abortions, and sexually transmitted infections 14 .Prevalence of violence was higher among women who reported having a mental health problem than in those who did not.In a systematic literature review, Silva et al. (2019) 15 highlight that high prevalence of common mental disorders is associated with factors such as poverty, marriage, gender violence, invisibility, overburden of work and caring for children, and lack of social support, which can cause physical exhaustion and psychic suffering symptoms.According to Frazão et al. 16 , violence and depression can have serious consequences for the lives of women, such quitting work, lack of motivation to perform daily activities, social isolation, development or deterioration of health problems, and feelings of hopelessness.
Our findings showed that the prevalence of violence was higher in younger women (30-49 years).Other studies have reported that women aged between 20 and 49 years are the main victims of violence [17][18][19] .Some authors suggest that this may be associated with the fact that this is the period of life in which women have most of their relationships, are more sexually active, and have greater reproductive capacity 20,21 .The fact that prevalence of violence was higher among single women may also be related to the high prevalence of violence in younger women.The high prevalence of violence among divorced women may be associated with the non-acceptance of the end of a relationship by the perpetrator, who consequently stalks, threatens, and commits different types of violence, or the possibility that the separation occurred during the 12-month period.Being married appears to be a protective factor against violence.However, this finding may be masked by the fact that married women may be less likely to report violence than other women, due to fear or coercion 20 .In this respect, the 2019 PNS was conducted in households and some women may have felt awkward in responding such a sensitive question or insecure because they were in the presence of the perpetrator.According to the survey data, 76.3% of the questionnaires answered by women living in rural areas were filled in by the interviewer and privacy was ensured during the administration of the violence module of the questionnaire in around only 10% of cases.
With regard to the higher prevalence of violence among women who had completed elementary school but had not completed high school and those who had completed high school but had not completed higher education, educa-   tion level may linked to factors such as income and work.In general, women with a lower level of education may experience violence because they do not have sufficient income to leave home or break up with an abusive partner, while women with a slightly higher level of education (completed high school but have not completed higher education) may be better informed about violence and therefore be more likely to answer yes to the questions on violence 20 .
Our data show that the main perpetrators of violence were current or ex-boyfriends, partners, or husbands or relatives, showing that violence was mainly perpetrated by someone known to the victim, and that the most common place of occurrence was at home.Cruz and Irffi 8 found that more than half of the women living in rural areas who suffered violence were abused by someone they knew.This finding is corroborated by a study conducted by Soares et al. of violence against women aged 18-59 years living in rural areas using data from the Notifiable Medical Conditions System (SINAN) for the period 2010-2012.The results show that 67.1% of the cases of violence took place at home and that the main perpetrator was the husband (36.2%).Furthermore, violence was repeated in 42.3% of cases 22 .
Another point worth highlighting from our findings is that 61.5% of respondents who reported having suffered sexual violence were abused by a current or ex-boyfriend, partner, or husband and 50% reported that the frequency of abuse during the last 12 months was "sometimes".The control of female sexuality, consent, and a wife's "conjugal duties" towards her husband are anchored in patriarchal culture, which means that men often turn a blind eye to sexual violence and women are conditioned to feel guilt and moral shame, refraining from talking about the problem or seeking help.Violence has serious consequences for women's physical and emotional health and leaves a profound mark on their lives 23,24 .
But why is it important to provide a snapshot of violence against women living in rural areas?Besides the lack of research on this topic in Brazil, which contributes to the invisibility of violence, rural areas act as an obstacle to tackling violence.Problems such as geographic isolation, lack of witnesses, and cultural and religious factors reinforce the permanence of women in situations of violence 8 .As this study shows, younger women, especially those who are single or divorced, who have completed elementary school or high school but have not completed higher education, with poor self-perceived health status, and diagnosed with a mental health problem were more likely to report having suffered some type of violence during the last 12 months.Most cases of violence are committed at home by people known to the victim and were repeated acts of abuse.Tackling violence against women requires the implementation of policies and joint actions across different sectors, including health, public security, the justice system, education system, social services, among others.However, policies and actions need to take into account both the complexity of this problem and the area where these women live to develop solutions that address patriarchal culture, strengthen women's financial autonomy, and provide easily-accessible places that offer humanized support and assistance.Victims also need to have access to effective means of communication so that they can seek help, report abuse, keep themselves informed, and maintain close and supportive relationships with family and friends 25,26 .
An example is support services for women in situations of violence, which are generally concentrated in major large urban centers.Raquel Braz of the Federation of Rural Farm Workers and Family Farmers of the State of Alagoas (FE-TAG-AL) highlights that "It is necessary to "interiorize" the delegacias da mulher [police stations that deal specifically with crimes involving female victims] to serve rural areas.Women living in rural areas, the forests, and waters face difficulties reporting cases of violence due to distance, lack of access to internet, fear, and lack of witnesses to cases in more isolated areas" 27 .The question is how rural areas are thought of and approached: they cannot be seen simply as a space outside urban areas, opposite, inferior, a place at odds with modernity and technology, neither as an idyllic place, where calm reigns and people are always happy and carefree.These simplistic views of rural settings fuel stereotypes that skew our way of looking at these places and the people who reside there.The "rural world" is not isolated, but rather interspersed with the specificities of rural life and work, which hold economic, social, cultural, and patrimonial potential 28 .
This study has both strengths and limitations.Limitations include those inherent in cross-sectional studies and possible underreporting of the prevalence of violence as some interviewees may feel awkward responding questions on violence or pressured not to answer when responding in the presence of the perpetrator.It is also important to broaden the scope of the violence questionnaire to include property and moral violence, which are also encompassed by the Maria da Penha Law.Study strengths include the fact that the questions in the violence module are broader than in the previous PNS, insofar as the new questionnaire distinguishes between psychological, physical, and sexual violence.Furthermore, due to the sensitive nature of this issue, the interview is provided with the option of a self-administered questionnaire 10 , enabling the respondent to provide more detailed information about violence.The PNS is a major population-based nationally representative survey that provides data broken down into type of census tract (rural and urban areas), thus allowing us to understand the magnitude of the problem of violence against women living in rural areas at national level.Finally, it is important to stress that studies focusing on women living in rural areas are particularly important to draw attention to the invisibility of violence and incorporate this issue into discussions and public policy making.

Final considerations
This study reveals the worrying prevalence of violence against women living in rural areas during the last 12 months, especially against younger single or divorced women and those who have completed elementary school or high school but have not completed higher education, with poor self-perceived health status, and diagnosed with a mental health problem.In addition, women who had suffered psychological, physical, or sexual violence during the last 12 months also reported that the most common place of occurrence was at home and that violence was mainly perpetrated by someone known to them and repeated over time.These findings therefore show that abuse is an everyday occurrence in the lives of women living in rural areas and that greater attention needs to be paid to this problem.Violence profoundly limits and damages the lives and health of women, stripping them of their rights, including the right to live a full life.There is therefore an urgent need for violence prevention policies aimed at reducing the incidence of violence and the provision of protection and support for women in this situation who need to break the cycle of violence in their lives and the lives of future generations, guaranteeing their rights, safety, and well-being.

Collaborations
L Stochero was responsible for the conception, analysis and writing of the manuscript; LW Pinto was responsible for reviewing the data analysis, writing and critical analysis of the manuscript.

Table 1 .
Sociodemographic, economic, health, and social support characteristics of the study sample, Brazil 2019.

Table 2 .
Prevalence of violence against women living in rural areas during the last 12 months according to sample characteristics, Brazil 2019.

Table 3 .
Cases of psychological, physical and sexual violence against women living in rural areas during the last 12 months according to perpetrator, place of occurrence and frequency, Brazil 2019.

Table 4 .
Crude and adjusted prevalence ratios of the variables included in the Poisson regression model: occurrence of either psychological, physical or sexual violence against women living in rural areas during the last 12 months, Brazil, 2019.