Abstract
This paper investigates the impact of the COVID-19 pandemic in the time series of female homicides in the state of Pernambuco, Northeast of Brazil. Microdata on lethal violent crimes from the Department of Social Defense from 2015 to 2020 was used. The homicide rates were analyzed using joinpoint regression models. The results showed an upward trend in homicide rates in the Sertão mesoregion and within the 20 to 39 age group in 2020. It is concluded that confinement and social distancing aggravated the violence against women, showing the need for intersectoral planning and measures to prevent and reduce female homicides.
Introduction
Violence against women can occur in different ways, with different degrees of severity and is not an isolated occurrence; it is part of a series of increasing episodes (Leite et al., 2019). The homicide is the ultimate expression of violence in the patriarchal regime, in which women are subject to the control of men who are part of their daily lives (Gomes, 2018; Iratzoqui & McCutcheon, 2018). Most causes of female homicides are not related to the pathological conditions of the perpetrators, rather to the women’s desire for possession (Bejarano Celaya, 2014; Meneghel & Portella, 2017).
In 2017, there were 87,000 female homicides, with an estimated global rate of 2.3/100,000 women (United Nations, 2019). Approximately 58% of these women were killed by a family member, in other words, 137 women are killed every day by their intimate partners or by family members (United Nations, 2019). In Brazil, the female homicide rate in 2018 was 4.3/100,000 women, showing a 9.3% decrease when compared to the previous year (Instituto de Pesquisa Econômica Aplicada [IPEA], 2020). In Pernambuco, the rate was 4.9/100,000 women, higher than what was found in Brazil and among the states with the highest rates (IPEA, 2020). Female homicides are more frequent in places with greater social vulnerability, where there is a high occurrence of interpersonal violence, involvement with drug trafficking, alcohol use, poverty, among others (Justino et al., 2021; Meira et al., 2019).
Studies acknowledge that major disasters, whether natural (earthquakes, hurricanes), environmental (dam failures, oil spills), epidemics (Ebola, Zyka), or those caused by man (shootings), exacerbate factors related to behaviors that victimize women and consequently increase the numbers of violence (Nigam, 2020; Peterman et al., 2020).
The Coronavirus (COVID-19) outbreak caused by the new Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) emerged between late 2019 and early 2020 in Wuhan, Hubei Province, China (Lai et al., 2020; Singhal, 2020). With great potential for the dissemination of the disease and migration during the Chinese New Year, the epidemic grew, bringing several impacts in various regions of the world (Freitas et al., 2020).
In Brazil, the first case was confirmed on February 26, 2020. The patient had a recent travel history to Italy. In Pernambuco, responses to the pandemic were given quickly after the confirmation of the first two imported cases of COVID-19 reported on March 12 (W. V. Souza et al., 2020). Several decrees were published during the month of March and in subsequent months, with social distancing measures of various natures (W. V. Souza et al., 2020). Many countries adopted strategies aiming to slow down the spread of the disease, among them, social isolation and quarantine (Aquino et al., 2020; Kraemer et al., 2020; Person et al., 2021).
The COVID-19 pandemic caused by the new coronavirus, SARS–CoV-2, brought with it several consequences to the lives of the women who were already suffering from domestic violence (Campbell, 2020; Usher et al., 2020). Considering that most cases of violence occur in the domestic environment, social distancing is configured as a threat to women, increasing the exposure to those who were already living in abusive relationships (Anurudran et al., 2020; Mazza et al., 2020). Restricted access to victim support services during this period is also an aspect that should be noted (de Paz et al., 2020). In addition to the decrease in police interventions, the limitation of health services, and the courts being closed resulted in a difficulty to report violence cases and contributed to impunity (John et al., 2020; Machado-Rios et al., 2020).
Adding to the difficulties encountered, the stress caused by the confinement during the pandemic, the economic impact and the use of legal and illegal drugs in the home environment, are factors that simultaneously increased the levels of disturbance in families (Henson, 2020; Usher et al., 2020). In Brazil, in 2020, the number of telephone complaints regarding domestic and family violence through Ligue 180 had a 12.3% increase when compared to the previous year (from 67,438 to 75,753; Vilela, 2021).
Ligue 180 is a free telephone service, available 24 hours a day, to assist women who are victims of violence (Brasil, 2022). A study carried out in Brazil, comparing mortality from aggression in women with and without prior notification of violence, revealed that 15.9% of homicide victims had a notification, which demonstrates the chronic nature of violence against women (Barufaldi et al., 2017). In addition to the escalation in the number of complaints, there was an increase in cases of femicide in the country between January and June 2020, corresponding to an increase of 1.9% (648 deaths; Brazilian Forum for Public Security, 2020). The data used in the Brazilian Public Security Forum are based on information provided by the state public security secretariats and the civil, military, and federal police, enabling the identification of femicides (Brazilian Forum for Public Security, 2020).
In an attempt to circumvent the difficulties encountered by victims in making complaints in person in Brazil, the Ministry of Women, Family and Human Rights (MWFHR) incorporated, in early April 2020, consultations through digital platforms to the National Human Rights Institution (NHRI), via BR Human Rights app and ouvidoria.mdh.gov.br website (Brasil, 2020). The expansion of the digital platforms was created to receive complaints by receiving photos, audios, and videos, as a silent way to report abuse (Brasil, 2020). A strategy created with the aim of facilitating reporting was the “Red Light for Domestic Violence” campaign, offering women victims of violence a simple and silent channel as a means of denouncing (Association of Brazilian Magistrates, 2020).
In addition to improving access to reports in order to restrain aggression, two bills were drafted in an attempt to protect women during the COVID-19 pandemic. Bill 1796/2020 provides for the urgency of cases related to domestic and family violence, prohibits the suspension of deadlines, and procedural acts, allowing the inclusion of remote assistance to the parties (Federal Senate, 2020a). Bill 1798/2020 deals with the permission to record the occurrence of domestic and family violence against women, which can be carried out through emergency telephone numbers or the internet, as well as listening to the victim at her residence (Federal Senate, 2020b).
Although social isolation measures to stop the rapid spread of COVID-19 have direct effects on disease transmission, its impact on violent deaths still needs further studies. It is understood that social isolation may have become a trigger for the increase in violence against women. Research on violence against women during the quarantine period due to the COVID-19 pandemic can contribute to planning measures developed in the women’s support network. This study aimed to analyze the impact of COVID-19 on the time series of female homicides in the state of Pernambuco, Northeast region of Brazil.
Methods
This is an ecological time series study of the number and rates of female homicides in Pernambuco. The state has a territorial extension of 98,312 km², it is composed of 184 municipalities and the state district of Fernando de Noronha, spread over five mesoregions: Agreste (71 municipalities), São Francisco (15 municipalities), Sertão (41 municipalities), Metropolitan Region of Recife (15 municipalities), and Zona da Mata (43 municipalities; Figure 1). The metropolitan mesoregion of Recife has the highest population concentration, the highest rates of urbanization, the best socioeconomic indicators, and most of the gross domestic product. Although the geopolitical division of the state of Pernambuco is composed of five mesoregions, for the actions of the public security plan, data from the Sertão and Sertão do São Francisco mesoregions are jointly released by the Pernambuco Social Defense Department. These mesoregions have geographic, climatological, social, economic, and cultural patterns similarities that justify their aggregate analysis.

Regionalization of the state of Pernambuco by mesoregions and the distribution of the network for combating violence against women.
Currently, the network to combat violence against women has 654 units distributed in the state of Pernambuco, among them: Specialized Police Stations for Attendance to Women, Courts for Domestic and Family Violence against Women, Specialized Centers for Attendance to Women, and Comprehensive Health Care Services for Women and Adolescents in Situations of Domestic and Sexual Violence.
The data source consisted of the microdata of intentional lethal violent crimes—ILVC (Pernambuco, 2021). This group of violent crimes has in common the fact that it causes the death of its victims, whether it is intentionally sought by the aggressor agent or an indirect consequence of malicious criminal action. Intentional homicides, robbery murder, grievous bodily harm followed by death are all characterized by of this type of crime. Femicide was included in the year 2017, in the ILVC and is characterized by the deaths of women committed by their partners or former partners. The Police Information System of the Department of Social Defense of the State of Pernambuco aggregates the cases of autopsy reports from forensic medicine institutes, civil police stations, military police reports, and expert reports from criminalistic institutes (Pernambuco, 2020).
The study population consisted of women aged 10 years or older, victims of homicides, residing in the state. The semestral time series of the number and rate of homicides by age group (10–19, 20–39, 40–59, and 60 years or more) were analyzed in each of the mesoregions of the state of Pernambuco starting from 2015 until 2020.
The series were made available by month, however, to facilitate the analyses, the series were grouped by semester, that is, the series that were evaluated on a monthly basis are now being evaluated every 6 months. As notation, the first 6 months of the year (first semester) will be denoted by the suffix .1, for example, for the first semester of 2015 we will have 2015.1. As for the last 6 months of the year (second semester), the suffix .2 will be used, for example, for the second semester of 2015 we will use 2015.2. Therefore, the total number of observations will be 12. This strategy was adopted to guarantee a total number of cases so that the application of statistical methods could be tangible.
The analyses were performed using joinpoint models based on the methodology proposed by Muggeo (2003). Joinpoints are models that automatically find trend change points in time series, these points are called breakpoints (Muggeo, 2003). The aforementioned models make it possible to empirically identify changes in the trends of the series. Therefore, if breakpoints are found for the 2020.1 semester, we will have evidence that will guarantee the verification of the research hypothesis of this study.
To calculate the homicide rates, the women population of each mesoregion was used, using data from population estimates from the years 2015 to 2020, provided by the Brazilian Institute of Geography and Statistics (IBGE). All calculations and graphs were generated using the R programming language version 4.0.3, where the segmented package version 1.3-3 was used to estimate the joinpoint models (Muggeo, 2008).
The data used are accessible in the public domain, and personal information about the victims is not available, therefore, approval by a Research Ethics Committee was not required.
Results
Between 2015 and 2020 there were 1,222 female homicides in the state of Pernambuco, with an average homicide rate of 2.07 (±0.64) per semester. The metropolitan region of Recife concentrated 42% of the total number of homicides, during which the average homicide rate in the Zona da Mata was higher 2.42 (±0.71) per 100,000 women. The average age of the victims was of 30 years (±12), with a minimum age of 10 and a maximum age of 99 years.
The estimates of the joinpoint models for the indicators in the study according to the mesoregions and the state of Pernambuco are shown in Table 1. All mesoregions had breakpoints for the number of homicides, with the exception of the Zona da Mata, which had a downward trend (Figure 2a). The Sertão was the only one that showed trend changepoints in 2020, where the period prior to 2020 (year 2019) had an upward trend of 0.49 homicides/semester, while in the semesters during the COVID-19 pandemic, the trend was 4.6 homicides per semester, a 9.4-fold increase.
Estimates for Joinpoint Models for the Number and Rate of Female Homicides (per 100,000 women), by Mesoregion, Pernambuco, Brazil.

Semestral time series for the number (a) and rate of female homicides (b) and their trend lines for each period found by the joinpoint model.
Regarding the homicide rate per 100,000 women, breakpoints were found for the Metropolitan Region of Recife, Sertão, and for the state of Pernambuco. The Sertão and Pernambuco stand out, with trend changepoints during the first semester of the COVID-19 pandemic (Figure 2b). Before 2020, the Sertão had an increase of 0.09 homicides per 100,000 women/semester, while in 2020, it continued with an upward trend of 0.84/semester, with a 9.3 increase in average growth (p < .001). The state of Pernambuco, on the other hand, had dropped −0.07 homicides per 100,000 women/semester before 2020, and then started to increase by 0.14/semester, returning to rates before COVID-19 (p < .001).
By analyzing the number of female homicides by age group, breakpoints were found in the 20 to 39 years old age group (Table 2 and Figure 3a). Homicides had fallen by −10 deaths/semester before 2020, and increased again by approximately 27 deaths/semester in 2020 (p = .01), which made the total number of homicides equal to the amount which had been occurring in the previous three semesters.
Estimates for Joinpoint Models for the Number of Homicides and the Female Homicide Rate (per 100,000 Women), by Age Group, Pernambuco, Brazil.

Semestral time series for the number (a) and rate of female homicides (b) by age group and their trend lines for each period found by the joinpoint model.
As for the female homicide rate by age group, breakpoints were observed in the stratum of the 20 to 39 years old age group (Table 2 and Figure 3b). With a decrease of −0.91 homicides per 100,000 women/semester before 2020 and an increase of 2.35 homicides per 100,000 women/semester during the COVID-19 period (p = .01), the indicator showed a regression of three semesters prior to COVID-19.
Discussion and Conclusions
The results of the study showed that there was an upward trend of female homicides in two mesoregions and for the state of Pernambuco in the first year of the COVID-19 pandemic. In Pernambuco, responses to the pandemic were fast in the first months of 2020, involving social distancing with the total closure of educational establishments, suspension of cultural equipment activities, gyms, suspension of commerce and non-essential services, and lockdown in five municipalities in the Metropolitan Region (Pernambuco 2020; W. V. Souza et al., 2020).
During the containment measures and epidemiological emergency management of the COVID-19 pandemic, some studies have shown that the circumstances of social distancing increased social problems and brought new stressors, including risks to physical and psychological health, isolation and loneliness, many school and business closures, economic vulnerability, and job losses (Henson, 2020; Usher et al., 2020). Due to these reasons, women were more vulnerable to the risk of domestic violence, as the victims remained confined in the same place with their aggressors, making it difficult to seek help (John et al., 2020; Machado-Rios et al., 2020; Pattojoshi et al., 2021). The quarantine also reduced the options for women to seek immediate assistance, as did the limited health infrastructure, and social and legal services (Fawole et al., 2021; John et al., 2020).
To assess trends in female homicides, Joinpoint Regression was used to identify inflection points where the respective trends significantly changed during the observation period. Some results must be emphasized. First, the Sertão mesoregion showed an increase in the number of homicides in 2020, marked by the COVID-19 pandemic. It increased approximately nine times compared to the previous one. Cultural factors may explain this result. It is geographically the farthest mesoregion from the state capital, is the least densely inhabited and is still characterized by the strong presence of gender hierarchy cultural patterns (Barros et al., 2021).
In regions that are farther away from urban centers, it is still common for women not to report domestic violence. For reasons such as the fear of being judged, which can be practiced by the local society, fear of reprisals, in addition to socioeconomic inequalities, financial dependence, and low education levels, that reflect the vulnerability of these women (Barbosa et al, 2019). In addition, women’s police stations and shelters are unevenly distributed, being more present in capitals and metropolitan areas, something that hinders the effectiveness of such services to combat violence in a uniform manner among municipalities (Cerqueira et al., 2015). Difficulty in transportation, communication, emergency shelters, and access to other means constitute additional barriers for women living in regions that are farther away from urban centers (Northcott, 2011).
The joinpoint model estimates for female homicide rates showed changes in trends in Pernambuco, which showed a decrease from the second half of 2017 to the first half of 2019, and in 2020 evidenced an increasing trend. For the mesoregions, the trend in the Sertão is predictable, corroborating with the result found in the number of homicides. The Zona da Mata presented a downward trend during the analyzed period. A recent spatial analysis of intentional women homicides in the state of Pernambuco pointed to this region as the one with the highest rates and critical clusters for homicides (Barros et al., 2021).
Although the Metropolitan Region of Recife did not show a breakpoint trend in the year 2020, it maintained a downward trend from the second half of 2017 to the first half of 2019. After this period, there was an increase in the homicide rate that was sustained during the first year of the COVID-19 pandemic. The Metropolitan Region of Recife concentrated 42% of all homicides, this being the mesoregion that brings together most of the protective equipment and services (Vieira et al., 2019).
Homicides can be considered the “tip of the iceberg” of violence against women (Bejarano Celaya, 2014). When the result of this study showed an upward trend in this indicator, it is possible that other types of violence against women may have also increased. In this regard, female homicide rates are an important indicator of gender violence. Very frequently, femicides are a result of an aggression cycle, which many women are exposed to and do not seek help even after repeated episodes of violence (Ali et al., 2020).
In Brazil, there was an average increase of 14.1% in the number of complaints made to Ligue 180 in the first 4 months of 2020 compared to the previous year. It should be noted that notifications of violence represent a small fraction of real incidents. In turn, the number of official notifications of violence against women decreased in police stations in every state in the country. The female homicide rate dropped 2.1%, from 3.7 women killed per 100,000 women in 2019 to 3.6 deaths per 100,000 in 2020 (Bueno et al., 2021). These data need to be carefully evaluated, as they may represent an underreporting and not a diminution in the levels of violence against women.
It should be noted that the registration of homicides and non-lethal violence is carried out differently in Brazil. In violent deaths, or in suspected cases, the forensic medical necropsy is mandatory under the country’s legislation. In these circumstances, the bodies must be sent to the nearest Forensic Medical Institute for inspection and expert analysis by a forensics examiner, and the primary cause that led to the outcome must also be investigated. The medical examiner is responsible for completing the Death Certificate (Brasil, 1941; Bordoni et al., 2017; Federal Council of Medicine, 2005).
Therefore, the records of violent deaths are more reliable because they are accounted for in public safety and health systems. Domestic violence, on the other hand, depends on women’s complaints, and during the COVID-19 pandemic, public services were partially closed, making complaints more difficult.
The estimated trends in the mesoregions that showed stationary behavior are unsettling and may be underestimated, due to events in which the intention is undetermined. A study that analyzed the trend of female homicides over the last 35 years in Brazil, showed a progressive increase in rates (E. R. Souza et al., 2017). However, the comparison of trends in female homicide rates according to different regions of the country, showed stationarity in some, which may be influenced by events with undetermined intent, underestimating female homicide rates. This fact may indicate problems in recording external causes (E. R. Souza et al, 2017).
The analysis by age group showed an upward trend for the number and rate of female homicides in the group of 20 to 39 years of age. Homicides at this stage of the life cycle result in many years of potential life lost, and have a socioeconomic impact, causing psychological suffering for the victims’ families (Meira et al., 2019). Globally, this age group represents the majority of the women victims of homicide and may be at greater risk of exposure to intimate partner or family violence than other age groups (United Nations, 2014). Likewise, a study conducted in all states of the Northeast region of Brazil, revealed an upward trend in homicides in this age group (Meira et al., 2019).
To achieve gender equality around the world, combating violence against women is a priority. The global crisis caused by the COVID-19 pandemic, in addition to being a threat to public health, it also compromises the women’s dignity, security, and human rights (El-Serag & Thurston, 2020; Standish & Weil, 2021). During the COVID-19 pandemic, ensuring protection programs for women victims of domestic violence, with services via telephone or internet is necessary as well as emergency services and shelters with multidisciplinary teams (including social workers, psychiatrists, psychologists, and legal assistance; Mazza et al., 2020). Furthermore, the ongoing trauma and the need for psychological and psychiatric support for the women should be considered, in order to recover from the abuse suffered during the period of social isolation and after the COVID-19 pandemic.
There are limitations in this study that should be noted. First, the use of secondary data, with possible undercounting of female homicides, either due to the absence of registration or classification as unspecified external causes, or of undetermined intent. Admittedly, homicide data are limited by the accuracy and precision of the death investigation. This fact may underestimate the homicide indicators presented, particularly in the mesoregions where there is no Institute of Forensic Medicine. In addition, the Department of Social Defense’s public domain database does not have the reference variable and type of femicide (intimate partner, child/parent, state violence, stranger, etc.), which is a relevant categorization for female homicides analysis, considering that the intimate partner is the main aggressor and most aggressions occur in the home. However, it is the official source of records of deaths from violent causes in the state and is important for the monitoring of female homicides, as a strategy for evaluating its temporal evolution and the impacts of policies to prevent and reduce violence against women.
In addition, the fact that the regions further away from the capital present cultural patterns of gender hierarchy and a large part of the efforts are concentrated in the metropolitan region (due to 42% of the cases being in this region). Such characteristics may be causing underreporting to adjacent indicators. Furthermore, they may also be an indirect cause of the high growth trend for the homicide rate found in the COVID-19 period for the Sertão region.
Given these limitations, the study found that confinement and social distancing aggravated crimes of violence against women, as the victim spent more time in the company of her aggressor. Combating domestic violence requires effective public policies from the authorities. However, these policies were fragilized in all areas, making it difficult to face these challenges during the COVID-19 pandemic. Although the bills that were developed in 2020 in an attempt to protect women in the period of the COVID-19 pandemic are yet to be concluded in the legislature, judges weave jurisprudential constructions in this regard. In this sense, urgent protective measures to combat domestic violence should be strengthened, not only during the period of social isolation, but contributing to the prevention and coping of aggression, as well as the reduction of violence. As future research, the follow-up of these series, assessing the persistence or changes in trends, and the individualized analysis of children and adult women, may serve as a basis for future research and for the formulation of strategies to combat violence against women.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Coordenação de Aperfeiçoamento de Pessoal de Nível Superior—Brazil (CAPES), funding code 001.
