Intimate Partner Violence Against Women During the COVID-19 Lockdown in Italy: A Multicenter Survey Involving Anti-Violence Centers.

This study explores intimate partner violence (IPV) evolution during the lockdown with a sample of 238 women (44% cohabitating and 56% not cohabitating with the perpetrator), attending five antiviolence centers in Italy (June-September 2020). Questions included 12 items on IPV and, for each item, a question about whether violence increased/stayed the same/decreased during lockdown; an indicator of IPV modifications was constructed. Two distinct patterns, confirmed after adjustment for socio-demographic factors, emerged: IPV increased for 28% of cohabitating and decreased for 56% of non-cohabitating women. Such results suggest the efficacy of physical distancing-strictly controlled by the State-in the prevention of IPV.

Introduction

Violence against women is a human rights issue, represents an obstacle to development, and is considered a serious public health problem; it affects the women involved, their children, and the entire community. Wars, natural disasters, and pandemics are associated with increased violence (Peterman et al., 2020a). More particularly, pandemics are associated with several factors that may intensify intimate partner violence (IPV), such as forced cohabitation all day long with the violent partner because of quarantine, the inability to escape and to get help, social isolation, and stress linked to economic insecurity and the fear of losing a job (Peterman et al., 2020a). As soon as it was clear that COVID-19 was a serious epidemic, many voices, including governmental bodies, international organizations, and women's associations, alerted to the risk of increased IPV and a worsening of victims’ situations (Roesch et al., 2020; United Nations Population Fund, 2020). Dubravka Simonovic, the UN Special Rapporteur on violence against women, warned: “It is very likely that rates of widespread domestic violence will increase, as already suggested by initial police and hotline reports. For too many women and children, home can be a place of fear and abuse. That situation worsens considerably in cases of isolation such as the lockdowns imposed during the COVID-19 pandemic.” (Simonovic, 2020).

However, until now, data on the consequences of the lockdown during the COVID-19 pandemic on IPV concerned women's requests for help more than actual changes in violence during confinement. Data from administrative sources, such as calls or complaints to the police, visits to hospital emergency departments, or an analysis of requests for help to dedicated services or helplines from various countries were inconclusive; calls, complaints, and requests for help increased during the lockdown in some countries and for some services, and decreased in others (Peterman et al., 2020b; Peterman & O’Donnell, 2020). In Italy, for instance, calls to the national helpline 1522—a free number dedicated to violence against women and stalking—increased during the lockdown (Istat, 2020a), whereas visits to specialized hospital emergency departments and formal complaints of sexual or domestic violence or stalking decreased markedly (Barbara et al., 2020; Ministero dell’Interno, 2020). In a hospital in the United States, accesses to a specialized IPV service during the pandemic were compared to accesses in the same months in the previous 3 years: During the pandemic, there was an overall decrease in the total number of IPV victims, but the incidence of physical violence and the severity of injuries, including strangulation and stab wounds, was greater (Gosangi et al., 2021). The authors hypothesized that, during the pandemic, victims were reporting in later stages of violence; if the violence was only emotional, or less severe, they might have decided to avoid resorting to hospital services (Gosangi et al., 2021).

To date, only a few studies have analyzed the evolution of IPV during the lockdown. In a large online survey of Spanish women 18–60 years old who were cohabitating with a male partner, both the lockdown period and financial stress were associated with increased psychological and sexual violence but not physical violence (Arenas-Arroyo et al., 2020). The increase was greater when IPV was present before the lockdown, there were children involved, women's education was low, and the perpetrator's professional situation worsened in this period. In another online survey of a large sample of Argentinian women under lockdown and cohabitating with a male partner, with partners under or not under lockdown, when women's partners were placed under quarantine, the prevalence of emotional, physical, and sexual IPV increased (Gibbons et al., 2020). Analysis suggested that the increase in hours spent together mediated the effects of the quarantine on IPV. The authors found no effect of drug or alcohol consumption on the results. In yet another online survey in the United States, among male and female respondents screened for IPV, 54% stated that victimization remained the same, 17% stated it worsened, and 30% said it got better (Jetelina et al., 2020); however, the authors themselves expressed doubts about the validity of their sample. For these online surveys, the response rate was very low (5% in Gibbons et al., 2020; 6% in Jetelina et al., 2020) or not clearly specified (Arenas-Arroyo et al., 2020).

A study in Bangladesh (Hamadani et al., 2020) of a sample of mothers randomly selected from participants in a previous study, with a good response rate, found a sharp decrease in income and an increase in food insecurity and mothers’ psychological symptoms during the lockdown. Among women already experiencing emotional or physical IPV, more than half reported that it had increased since the lockdown.

These studies point to the increase in IPV during the lockdown and give valuable information concerning the role of children and financial and professional insecurity, but involve exclusively cohabitating couples or, as in Jetelina's survey (Jetelina et al., 2020), do not ask about cohabitation status. This is a serious limitation. Data indicate that violent partners often continue to exert violence against the ex-spouse even after the couple separates (Brownridge et al., 2008); this post-separation violence may involve children (Feresin et al., 2019; Radford et al., 1997). According to Italian national data (Istat, 2015), 5% of women are victims of physical or sexual violence from their current partner and 19% from their ex-partner; rates for psychological and economic violence are 26% and 46%, respectively. Considering IPV occurring in the past 5 years, 3% of women are victims of physical or sexual violence; the rate for recently separated women is 12%. In addition, violence after separation tends to be more severe than pre-separation violence (Istat, 2015). According to a smaller study carried out in northern Italy, among women who had presented at an anti-violence center (AVC), 3–5 years later, almost half were still experiencing partner violence (Pomicino et al., 2019). Notwithstanding its frequency and seriousness, post-separation violence tends to be poorly understood by professionals working with abused women and their children (Feresin et al., 2019). Investigating what happened to victims of post-separation violence during the lockdown is necessary to complete our picture of the consequences of COVID-19 on violence against women and their children.

Context of the Study

From February 2020, Italy was hit hard by the COVID-19 epidemic. When the situation was already dire, with about 600 deaths attributable to the virus, the Italian government imposed a strict lockdown at the national level. From March 9 to May 4, 2020, all social events were canceled, travel was forbidden, and most services and activities, such as schools, were suspended or closed down. If not cohabitating, people were told to stay home and not to visit even their closest relatives. Many were caught by the lockdown away from their family and could not join them until the end of the quarantine period. Going out was permitted only for buying groceries, because of emergencies, or to go to work for those employed in essential services. People could only go out alone, with a document certifying the reasons for leaving home. Police cars patrolled the cities, checking the reasons for being in the streets and issuing fines for non-compliance.

In Italy, AVCs are services dedicated to victims of violence, based on the principle of women's autonomy and empowerment. They offer counseling, legal advice, and advocacy; many are able to offer women a shelter in lodgings with a secret address (D.i.Re., 2018; Istat, 2020b). Most women presenting at AVCs or similar services experience high levels of violence and psychological distress (Bastiani et al., 2018). During the lockdown, most AVCs were closed; in-person encounters were suspended, but the advocates continued to counsel women via telephone or with web-based devices.

The aims of the present study were to analyze the impact of the COVID-19 lockdown on the evolution of IPV among cohabitating and non-cohabitating women and their children seeking the services of AVCs in an Italian region.

Methods

Study Design and Procedure

An observational study was carried out in five AVCs in Friuli Venezia Giulia, northeastern Italy. Between June 3 and September 30, 2020, 379 women sought services in person or by phone from an AVC in the region; 292 completed the standardized questionnaire (response rate 77%) and, among these, 238 were exposed to violence by a male partner or ex-partner and were included in the present analysis. Women were interviewed by AVC advocates in the context of their usual interactions; questionnaires were anonymous, and before the interviews, women signed an informed consent form. The study was approved by the ethics committee of the University of the first author.

Questionnaire and Measures

A standardized questionnaire was developed especially for this study. It is the result of the collaboration between the researchers and the advocates of the AVC, and is based both on a previous questionnaire used with a similar sample of women (Bastiani, 2016) and on the working experience of the advocates.

The evolution of violence was assessed with a set of 12 items asking women if, during the lockdown, each type of violence increased, stayed the same, or decreased as compared with the period before (see items in the Appendix); when a type of violence was not present, the code “Not applicable” was attributed. Types of violence were: psychological; controlling behaviors; threats of hurting/killing the woman; physical; sexual; stalking; telephone, or cyber-violence, including revenge porn; economic violence; and threats of suicide. Three items assessed violence involving the children: threats to take the children away or hurt them; children witnessing violence against the mother; direct violence against children. From the types of violence reported by participants, we created a synthesis variable of the evolution of violence during the lockdown. Categories for violence were: (1) mostly increased; (2) stayed the same or similar; and (3) mostly decreased.

A question asked whether the perpetrator was a partner or an ex-partner. Another question assessed whether the woman cohabitated with the perpetrator during the lockdown (yes or no). This variable was used as the main explicative factor in the analysis of data. Financial stress was assessed by asking the woman to subjectively evaluate her financial situation. This factor was coded in one of 3 classes (good or very good; not good; very poor). The employment status of women at the time of the survey was assessed and included in the analysis a one of 3 classes (employed; precarious professional situation; housewife or not employed). The employment status of the partner was assessed by the women’s answers, and was coded as one of 4 classes (employed; precarious professional situation; not employed; unknown).

Questions were asked about common children, age, and nationality of the woman and perpetrator.

Strategy of Analysis

First, we described the characteristics of the women by cohabitation status and the frequency of the types of IPV experienced by cohabitating and non-cohabitating women before the lockdown. Then, we analyzed the evolution of IPV during the lockdown by cohabitation status with the perpetrator. The proportions were compared by Pearson's chi-square test; statistical significance was defined as p < .05. Finally, using polytomic logistic regression models, we computed odds ratios (ORs) and 95% confidence intervals (CIs) to quantify the associations between cohabitation status and IPV evolution, adjusting for perpetrator status (partner or ex-partner), women's employment, and financial situation, perpetrator's employment, and center. These factors were included in the multivariate models as possible confounders because their distribution differed between women who cohabitated and did not cohabitate with the aggressor, as observed in the first step of the analysis. In addition, the center (five classes) was included in the multivariate model to take into account the variability linked to the geographical, social, and cultural characteristics of the places where the AVCs were situated and the organization of each service. Data analysis involved using SAS v 9.4.

Results

The socio-demographic data for participants are given in Table 1. The characteristics of the participants corresponded closely to those of the women who presented at the same AVCs during 2019 (these data are routinely collected by the centers). For instance, in the study sample, 69% of women were ≥ 36 years old and 73% were Italians; in the sample of women presenting at the AVCs in 2019, these percentages were 67% and 76%, respectively.

Table 1.

Description of the Whole Sample of Women, and by Cohabitating Status.

Socio-demographic characteristics	Total womenn = 238		Cohabitatingwith perpetratorn = 104	Not cohabitating with perpetratorn = 134	p value*
	N	%	%	%
Anti-violence center (AVC)
1	114	47.9	45.2	50.0	.10
2	37	15.6	10.6	19.4
3	17	7.1	6.7	7.5
4	52	21.8	26.9	17.9
5	18	7.6	10.6	5.2
Women's age (years)
≤ 25	16	6.7	6.7	6.7	.88
26–35	51	21.4	19.2	23.1
36–45	89	37.4	37.5	37.3
> 45	82	34.5	36.5	32.8
Has children (whatever age) with perpetrator
No	72	30.2	31.7	29.1	.66
Yes	166	69.8	68.3	70.9
Women's citizenship
Italian	175	73.5	69.2	76.9	.18
Not Italian	63	26.5	30.8	23.1
Women's employment status
Employed	109	45.8	36.5	53.0	.04
Precarious professional situation	66	27.7	32.7	23.9
Housewife or otherwise not employed	63	26.5	30.8	23.1
Self-evaluated financial situation
Good or very good	96	40.3	32.7	46.3	.004
Not good	92	38.7	36.5	40.3
Very poor	50	21.0	30.8	13.4
Perpetrator's employment status
Employed	128	53.8	63.5	46.3	.001
Precarious professional situation	30	12.6	16.4	9.7
Not employed	42	17.6	20.2	15.7
Unknown	38	16.0	0.0	28.4
Perpetrator is
Current partner	132	55.5	84.6	32.8	.001
Ex-partner	106	44.5	15.4	67.2

*Comparing women cohabitating and not cohabitating by chi-square test.

In the study sample, cohabitating and non-cohabitating women were similar in age, having children with the perpetrator, and nationality. Non-cohabitating women were more frequently than cohabitating women to be employed (p = .04) and to describe their financial situation as good or very good (p = .004), and were less frequently informed of the employment status of their ex-partner (p = .001). Among cohabitating women, 85% defined the perpetrator as a current partner as compared with 33% of non-cohabitating women (p < .001).

Women reported high levels of violence before the lockdown: For instance, 98% reported psychological violence; 66% physical violence; 76% controlling behaviors; 62% threats; and in 31% of cases the perpetrator threatened to kill himself (data not shown). Table 2 presents the frequency of the types of IPV for cohabitating and non-cohabitating women. There were few differences concerning psychological or economic violence, controlling behaviors, and threats of suicide. Cohabitating women reported physical violence significantly more often (p = .03); noncohabitating women reported stalking in person and violence via phone or web significantly more often. Noncohabitating women also reported partners’ threats to take the children away or to hurt them significantly more often (p = .02), and partner's direct violence to children, although this difference was not significant (p = .08).

Table 2.

Description of Intimate Partner Violence (IPV) Types Present Before Lockdown by Cohabitation Status.

	Women cohabitating		Women not cohabitating
Type of IPV	N	%	N	%	P value*
Psychological violence	104	97.1	134	98.5	.46
Controlling behaviors	104	66.4	134	67.2	.89
Threats of hurting/killing the woman	104	65.4	134	59.7	.37
Physical violence	104	73.1	134	59.7	.03
Sexual violence	104	34.6	134	26.1	.15
Stalking (physical)	104	40.4	134	61.9	.001
Violence via phone or web (insults, threats, revenge porn, sexual harassment)	104	44.2	134	73.1	.001
Economic violence	104	72.1	134	69.4	.65
Threats of suicide	104	33.6	134	32.1	.80
Threats to take the children away, to hurt them**	71	39.4	95	57.9	.02
Children witnessing violence against the mother**	71	73.2	95	77.9	.49
Violence to children**	71	42.2	95	55.8	.08

*Comparing women cohabitating and not cohabitating by chi-square test.

**Among women who had children with their partner.

Table 3 presents the evolution of IPV during the lockdown for each type of violence. The synthesis indicator of IPV evolution showed significant differences by cohabitation status: Women cohabitating with the perpetrator more often reported increased violence, whereas noncohabitating women more often reported decreased violence (p = .001). The difference in IPV trends by cohabitation status was significant for 11 of the 12 types of violence considered; the only exception was violence via phone/web, despite a similar trend.

Table 3.

Evolution of IPV During the Lockdown by Cohabitation Status With the Perpetrator.

	Cohabitating with perpetrator					Not cohabitating with perpetrator
Type of violence	N	% Increased	% Stable	% Decreased	p value*	N	% Increased	% Stable	% Decreased
Psychological violence	101	54.5	29.7	15.8	***	132	18.9	29.6	51.5
Controlling behaviors	69	34.8	46.4	18.8	***	90	17.8	22.2	60.0
Threats of hurting/killing the woman	68	44.1	41.2	14.7	***	80	10.0	18.8	71.2
Physical violence	76	43.4	30.3	26.3	***	80	1.2	3.8	95.0
Sexual violence	36	38.9	25.0	36.1	***	35	2.9	5.7	91.4
Stalking (physical)	42	28.6	33.3	38.1	***	83	15.7	12.0	72.3
Violence via phone, web (insults, threats, revenge porn, sexual harassment)	46	32.6	34.8	32.6	.25	98	23.5	29.6	46.9
Economic violence	75	46.7	48.0	5.3	***	93	25.8	33.3	40.9
Threats of suicide	32	28.1	25.0	46.9	***	41	4.9	7.3	87.8
Threats to take the children away, to hurt them	31	38.7	41.9	19.4	***	59	10.2	28.8	61.0
Children “witnessing violence”	58	44.8	37.9	17.2	***	79	10.1	21.5	68.4
Violence to children	33	39.4	45.4	15.2	***	56	16.1	12.5	71.4
Average synthesis index** 12 types violence	104	27.9	59.6	12.5	***	134	8.2	35.8	56.0

*Comparing cohabitating and non-cohabitating women by chi-square test; ***p < .001.

**A summary of trend: among the types of violence reported by the participants, this index is a synthesis from 1, mostly increased to 3, mostly decreased.

Among cohabitating women, the types of violence that frequently increased during the lockdown (>40% of respondents reported an increase) were psychological violence, threats, physical violence, economic violence, and children witnessing violence against the mother. Some types of violence decreased during this time; in particular, 47% of women said that their partner's suicide threats declined during lockdown. This was the type of IPV more likely to decrease in this group of women (Table 3).

Among women not cohabitating with the perpetrator, the types of violence that more frequently decreased (more than 70% of respondents reporting a decrease) were threats of hurting/killing the woman, physical and sexual violence, stalking, threats of suicide, and the father's direct violence to the children (Table 3).

Table 4 presents the results of the polytomic regression model showing the probability of an increase/stability/decrease in IPV during the lockdown by cohabitation status and adjusting for perpetrator status (partner or ex-partner), women's employment, financial situation, perpetrator's employment, and center. Even after adjustment, the association between cohabitation status and the evolution of IPV remained highly significant (p < .001). Cohabitating women were more likely to report an increase in IPV (adjusted OR 2.61 [95% CI 0.82–8.29]), and less likely to report a decrease (adjusted OR 0.07 [95% CI 0.03–0.18]) as compared with noncohabitating women.

Table 4.

Evolution of IPV During the Lockdown and Women's and Perpetrator Characteristics (n = 238 Women).

	Synthetic index of violence
Characteristics	Increased	Stable	Decreased
	aOR [95% CI]	Reference	aOR [95% CI]
Cohabitating
Yes	2.61 [0.82–8.29]	1	0.07 [0.03–0.18]
No	1	1	1
		<0.001
The perpetrator is
The current partner	1.47 [0.53–4.05]	1	0.52 [0.24–1.13]
An ex-partner	1	1	1
		0.13
Women's employment status
Employed	1	1	1
Precarious professional situation	1.22 (0.47–3.16]	1	1.63 [0.72–3.66]
Housewife or otherwise not employed	0.79 [0.24–2.56]	1	2.72 [1.16–6.42]
		0.15
Self-evaluated financial situation
Good or very good	1	1	1
Not good	1.28 [0.51–3.20]	1	0.96 [0.47–1.99]
Very poor	1.12 [0.34–3.67]	1	0.44 [0.16–1.19]
		0.49
Partner's employment status
Having a paid job	1	1	1
Precarious professional situation	3.92 [1.24–12.4]	1	3.89 [1.23–12.18]
Not employed	1.64 [0.56–4.77]	1	1.82 [0.74–4.50]
Unknown	1.96 [0.46–8.37]	1	1.15 [0.47–2.82]
		0.16

Adjusted odds ratios (ORs) (95% confidence intervals [CIs]) calculated by multivariate analyses with one polytomic logistic regression model including all the variables shown in this table + the center (five classes) that does not contribute significantly to the multivariate model (p = .49).

From this model, the status of the perpetrator (partner or ex-partner) and women's self-evaluated financial situation were not significantly related to the evolution of IPV during the lockdown. Nor was the employment status of the woman and that of her partner significantly related to IPV evolution. However, some associations are of interest. For women, a decrease in IPV was associated with being a housewife or otherwise not employed versus being employed (adjusted OR 2.72 [95% CI 1.16–6.42]). Concerning the perpetrators, the risk of an increase and decrease in IPV was associated with having a precarious occupational situation during the lockdown as compared with being employed (adjusted OR 3.92 [95% CI 1.24–12.4] and 3.89 [95% CI 1.23–12.18]).

Discussion

The current study investigated the evolution of IPV during the COVID-19 lockdown period (March and April 2020) in Italy with a sample of women seeking services from AVCs in northern Italy. Regardless of cohabitating or not with the perpetrator, even before confinement, these women experienced high levels of violence. Almost all reported psychological violence: ≥60% were exposed to threats to be hurt or killed, one quarter to one-third experienced sexual abuse, and >70% of their children witnessed the violence against their mother. Physical violence was reported more frequently by cohabitating than noncohabitating women, but stalking, phone/web harassment, and threats concerning children were more often reported by noncohabitating women.

In this already dramatic situation, the lockdown hit abruptly, forcing cohabitating couples to spend more time together in a stressful situation and at the same time raising a solid barrier between those partners or ex-partners who did not share housing.

Concerning women cohabitating with the perpetrator, our results confirm the dire expectations: According to the synthesis index of violence, 28% of women reported a global increase in the intensity of IPV. Threats, psychological violence, physical violence, and economic violence escalated for 43% to 54% of women; 39% of women reported an increase in sexual violence and of partner violence against children.

These results agree with those found by the few studies investigating the modifications of IPV in this period. During the lockdown in Spain, the incidence of IPV among cohabitating couples increased by 24% (Arenas-Arroyo et al., 2020). In a sample of women in lockdown in Argentina, when their partners were also placed in quarantine, women reported a higher prevalence of emotional, sexual, and physical IPV (Gibbons et al., 2020). In Bangladesh, among women experiencing emotional or “moderate” partner physical violence, more than half reported that the violence had increased since the lockdown (Hamadani et al., 2020). More particularly, more than 70% of women reported an increase in various types of emotional violence; sexual violence was less common, but among those experiencing violence, half reported an intensification in this period.

However, for noncohabitating women in our study, the situation was the opposite: For 56% of them, IPV actually decreased during the lockdown. More specifically, physical violence, sexual violence, and threats of suicide decreased for 95%, 91%, and 88% of women, respectively; threats concerning children, assisted violence, and direct violence to children decreased in >60% of cases. The decrease was less marked for economic violence, with a reduction for only 40% of women.

Cohabitating and noncohabitating women differed concerning their and the perpetrator's employment, financial situation, and perpetrator status (partner or ex-partner). However, results of the multivariate analyses confirmed, after adjustments on these variables, the strong bivariate associations between cohabitation situation and IPV evolution during the lockdown. Perpetrator status and women's financial situation were not associated with changes in IPV, but we found some associations with employment variables. The probability of reduced IPV increased more than twice for housewives or others not employed versus employed women. It is possible that, as suggested by an Italian report (Brunori & Caterino, 2020), in the difficult time of the lockdown, women with less autonomy “walked on eggshells” and accepted any compromise to avoid the partner's abuse, which resulted in decreased violence.

Concerning perpetrators, the likelihood of being both more and less violent during the lockdown was increased for those with a precarious professional situation as compared with employed perpetrators. Estimations are imprecise owing to the low numbers of respondents, with only 30 individuals in these situations, so interpretations should be cautious; moreover, information on men's professional situation was reported by the women. Yet, both trends could be tentatively explained: for some men, being precarious professionally may represent an attack on their male identity. In this case, IPV is a way to reaffirm their role (Ruxton & Burrell, 2020). For others, the same situation may lead them to refrain from using violence for fear of losing the partner's support. In a Spanish sample, the increase in IPV was higher when the man's professional situation worsened because of the lockdown (Arenas-Arroyo et al., 2020). The prevalence of one of these patterns probably depends on the social or community context: For instance, male unemployment rates are associated with lower rates of IPV in the United Kingdom (Anderberg et al., 2016), but the trend is reversed in developing countries (Bhalotra et al., 2020). In developing countries, female unemployment is associated with reduced rates of physical IPV but only where divorce is difficult to obtain for women (Bhalotra et al., 2020). Even if our sample is not large enough to test these hypotheses, these results point to the important and contradictory role that employment and financial autonomy of both partners can have in IPV (Bettio & Ticci, 2017).

A different set of hypotheses can be formulated to explain the increase in IPV for cohabitating women and its decrease for those not cohabitating. Lockdowns imply several risk factors for IPV (Peterman et al., 2020a): Cohabitating couples are forced, in a period dominated by anxieties, limitations, and irritation, to spend more time together, thus augmenting the time of exposure to IPV in case of a violent partner. In the study in Argentina, when both partners were placed in quarantine, they spent together, on average, 3.8 h more than couples where only the women had to stay at home (Gibbons et al., 2020). Isolation limits the occasional monitoring of abusive behavior by other people. Also, victims’ ability to get help and escape violence is restricted. In Milan, during March and April 2020, women's access to an emergency service dedicated to gender-based violence was halved as compared with the same period in 2019 (Barbara et al., 2020); complaints of sexual or domestic violence or stalking also decreased markedly (Ministero dell’Interno, 2020). In addition, social isolation coupled with stress and anxiety may lead to an increase in behaviors that in turn can be linked to violence, such as pornography or alcohol use (Mestre-Bach et al., 2020; Ruxton & Burrell, 2020).

In contrast, for noncohabitating women, the lockdown and the forced distancing from a violent man meant a welcome respite to the continued abuse because it was almost impossible for him to put into practice physical and sexual violence and stalking. It is also likely that, in the case of minor children, the visits with the father were suspended, thus suppressing another occasion for encounter and violence against the woman (Elizabeth, 2017; Radford et al., 1997; Stanley et al., 2019). Apparently, not having direct access to the women led to a decline also in other forms of violence, such as phone/web harassment and economic violence, which can be done from a distance. In these circumstances and with the woman not accessible, the stress caused by the epidemic and by financial worries may have left less room to try to control, insult, or otherwise abuse a noncohabitating partner.

Noncohabitating women manifested their will to be distanced from a violent man by living by themselves or with their children but, as studies on post-separation violence indicate (Pomicino et al., 2019; Radford et al., 1997), this does not prevent these men from continuing to consider the ex-partner their property and to harass or assault her. Nor are social services, the police, or courts unambiguously protective of victims of violence (Romito, 2008); women often exhibit a “credibility deficit” when they denounce the abuse (Manne, 2018; Silberg & Dallam, 2019). The Italian legislation contains various measures to separate the perpetrator from the victims, such as protection orders or precautionary measures, including the removal of the perpetrator from the family home, prohibiting them from approaching places frequented by the victim, and pre-trial detention. However, these measures are rarely applied in a timely and adequate manner, owing to the underestimation of the risks to which women are exposed and to insufficient training of the professionals involved (ECHR, 2017; GREVIO, 2020; Senato della Repubblica, 2018). Moreover, Italian authorities do not have a structured system to collect data on protection orders, which is an obstacle to effectively monitoring their application (ECHR, 2020). The physical distancing consequent to the lockdown was effective because it was imposed by the State with much rigor for reasons different from women's requests, reasons that do not imply accepting women's aspirations to be free from male violence.

Strengths and Limitations

Our study is the first to analyze the impact of the COVID-19 lockdown on IPV, considering also women who do not cohabitate with their perpetrators. Results are unique and show unambiguously that confinement was associated with an increase in IPV for cohabitating women but with a strong decrease for non-cohabitating women. We used a detailed questionnaire to collect information on 12 types of IPV, including violence involving children. A limit is that we did not test the qualities of the violence questions. However, we compared our data on violence before confinement with those collected routinely by the advocates of AVCs in the same region where the present study was carried out (Regione Autonoma Friuli Venezia Giulia, 2015): The regional report shows that women suffered mainly from psychological violence, 94%, followed by physical violence, 64%; in our study, the percentages were 98% and 66%, respectively. For some specific questions, we compared our data with those collected in a study by Bastiani (2016) also involving IPV victims presenting at an AVC in the same region. In Bastiani's study, 65% of women said the perpetrator “Threatened to hurt you physically,” and in the present study, 62% answered they experienced “Threats of hurting/killing you”; in the former, 38% said the man “Threatened to kill himself,” and in the latter, 31% of women said the man made “Threats of suicide.” The study has other limitations. Our sample consisted of women who sought the services of an AVC, which represents a minority of victims of IPV (Istat, 2015). As compared with all women experiencing some form of IPV, those seeking AVC services probably experienced more serious violence and have the possibility and the willingness to ask for help. Of note, the study was based in a region with a high presence of AVCs (Istat, 2020b), thus presumably reducing the selection effects on our sample. However, we obtained a high response rate, and the characteristics of the women involved corresponded closely to those of all women presenting at the same centers in 2019 and also of women presenting at all centers in the northeast macro-region in Italy. In our sample, 70% of respondents had children, and 74% were Italians; these percentages were 70% and 67%, respectively, in the macro-region (Istat, 2020b). Non-Italian respondents represented one-quarter of the sample, with nonsignificant difference between cohabitating and noncohabitating women, but we lack further information on their nationality or citizenship status. Involving women who experienced violence before the lockdown may seem a limitation; however, the studies in Spain and Bangladesh suggest that IPV was augmented during confinement rather than starting then.

Conclusions

At the time these conclusions were written, Italy and other countries were experiencing a new wave of COVID-19, and people were again in a lockdown. Results from our study confirm that women who share a residence with a violent man are at increased risk, as are their children, of extremely high levels of violence. All measures should be put in place to strengthen the services, such as the AVCs, that respond to their specific needs. Health professionals, the police, teachers, and also the general public should be, more than usually, alerted to the possibility that some women were victims of violence, and information about the available resources should be widespread.

That noncohabitating women experienced significantly less violence during confinement leads to recommendations that go beyond this specific period: Confinement represented a sort of natural experiment, showing that by applying rigorous distancing measures, IPV can be reduced.

In Italy, legislative measures concerning gender-based violence—such as protection orders, the removal of the perpetrator from the family home, pre-trial detention—do exist, but are applied in an uneven manner (GREVIO, 2020; Senato della Repubblica, 2018). At a time when concern about women victims of violence is high, pressures should be put in place to improve the training of the professionals involved, the application of these measures, and their monitoring.

During the lockdown, frequent messages concerning the possibility of getting help were directed to potential victims of IPV. Messages should also be directed to potential perpetrators and to family or friends who support them. In these difficult times, State institutions as well as authoritative figures have stressed the need to respond to the epidemic not only as individuals, but above all, as a community (Ruxton & Burrell, 2020). It could be a fruitful moment to convey to men who are violent that they need to change and the value of doing so.

	Increased	Remained the same	Decreased	NA
	1	2	3	0
Psychological violence (insults, denigration)
Controlling behaviors (telephone, Internet, etc.)
Threats of hurting/killing you
Threats to take the children away from you, to hurt them
Assisted violence (children witnessing violence)
Violence on children
Physical violence
Sexual violence
Stalking (physical)
Violence via telephone, web (insults, denigrations, threats, revenge porn or cyber sexual harassment)
Economic violence
Threats of suicide