key: cord-341381-jzjhhn23
authors: Davis, M.; Gilbar, O.; Padilla-Medina, D.
title: Intimate Partner Violence Victimization and Perpetration among U.S. Adults during COVID-19: A Brief Report
date: 2020-06-09
journal: nan
DOI: 10.1101/2020.06.08.20125914
sha: 
doc_id: 341381
cord_uid: jzjhhn23

Importance: Anecdotal evidence such as increased calls to domestic violence (DV) hotlines across the globe suggest that there may be an increase of IPV prevalence in association with the COVID-19 outbreak; however, no study has investigated this phenomenon empirically. Objective: To evaluate the association between COVID-19 related conditions and recent use or experience of IPV (since the pandemic outbreak in the U.S). Design, Setting, and Participants: This cross-sectional study analyzed data collected online from a sample of noninstitutionalized adults (age 18+) in the U.S. (N=2,045). More than half of the sample self-identified as being in an intimate relationship at the time of the study. Main Outcomes and Measures: A four-item tool was used to assess IPV perpetration and victimization since the outbreak of COVID-19. The rapid tool inquired about two forms of IPV, psychological and physical. Participants self-reported demographic data and recent health histories, including COVID-19 tests results, related symptoms and degree of personal social distancing. We hypothesized that COVID-19 related factors would increase risks of IPV. Results: In this study, self-reported COVID-19 impacted respondents had an increased risk of IPV victimization and perpetration. Among those who reported having symptoms consistent with coronavirus, but were denied access to testing, psychological IPV victimization was 3 times greater than those who did not (Exp[B] =3.15, [1.19, 2.29] p <.05). For participants who reported testing positive to COVID-19, the odds of using psychological IPV (Exp[B] =3.24, [1.18, 8.89] p <.05) and physical IPV (Exp[B]=3.02, [1.12, 8.17] p <.05) against an intimate partner increased by more than 3 times. Conclusions and Relevance: Patient education and community outreach/health care system initiatives focused on IPV risk behaviors may help reduce the potential development of IPV. Continued surveillance is imperative to improve health and well-being along with effective intervention development and implementation.

As of May 20, 2020, more than 1.5 million cases of coronavirus disease 2019 were reported in the United States (U.S.) 1 . Patients with underlying comorbidities are excessively affected by COVID-19. 2, 3 Black/African-Americans, subjects of longtime systemic/institutional/structural racism are most disproportionately hit by the pandemic, along with Hispanic/Latinx populations. 1, 2, 3 Referred to as a "shadow" pandemic, intimate partner violence (IPV), a pre-existing public-health problem, has reportedly intensified and is expected to rise amid the crisis. 4 However, studies examining connections between the COVID-19 outbreak and IPV are skim. 5, 6 Social scientists, medical professionals, and public-health experts have called for urgent research into this area as a critical step in responding to mounting and/or anticipated IPV. 7, 8, 9, 10, 11 This study aimed to characterize risk of IPV in relation to .

This was a cross-sectional observational study of adults. Participants were recruited by a survey company (Qualtrics) from an online research panel using stratified quota sampling to ensure sample characteristics of sex, age, and race/ethnicity were representative of the U.S.

Inclusion criteria required respondents to: live in USA; be age 18+. Study procedures were approved by a university institutional review board. Informed consent was obtained from participants. For participation, each respondent was distributed points valuing $4.80. Pilot-testing started March 17, 2020; data collection ended May 1, 2020. Majority of participants completed the survey April 22 nd -April 30 th . Around 33750-37500 people were invited or may have seen an . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.08.20125914 doi: medRxiv preprint invitation; 3750 expressed interest in the study. After eligibility screening, refusals, and removal of poor-quality responses, 2,045 participants' data were collected.

IPV was measured using the Jellinek inventory for assessing partner violence (J-IPV). The J-IPV is a 4-item screening tool developed to assess IPV victimization and perpetration in patients entering substance abuse treatment 12 . Drawing on a recent review of brief IPV screening measures 13 , the J-IPV was selected for its strong psychometric properties and brevity.

Participants were asked about recent engagement or experience of IPV, e.g. : "since the coronavirus crisis started, the situation with your partner got so out of hand that you acted in a threatening way to your partner, or threatened to hurt him/her"? 'Yes'=1; 'No'=0. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.08.20125914 doi: medRxiv preprint

We calculated descriptive statistics to examine distributions of variables and generalized linear modeling (GLM) to examine covariate levels of cases by state. Spearman nonparametric correlation analyses were used to determine whether IPV victimization and perpetration were significantly associated with COVID-19 status and social distancing restrictions. Four binary logistic regressions within a GLM framework were performed to determine impact of coronavirus exposure, government restrictions, and social distancing differences on IPV probability, while controlling for sociodemographics. Covariates evaluated in univariate analysis (P < .10) were included in multivariable analysis. For dependent variables, missing data were low (0%-3.4%). Analyses were performed using SPSS version-25.

Mean age of the sample was 46.63 years (SD = 17.19, range = 18-91 years); 49.9% (n = 1020) were 'female' and 1.5% (n = 30) of participants were 'other'. Majority of individuals indicated they were in an intimate relationship (58.2%, n = 1183); 87.4% completed a collegedegree; 47.2% of household incomes were $49k or above. Racial identity was: 62.6%

White/European-American, 11.9% African-American, 3.3% Hispanic/Latino, 2% Asian-American, 20.1% other. Sexual orientation was: 80.1% heterosexual, 3.8% gay/lesbian, 5.9% bisexual, 5.5% other.

Spearman nonparametric correlation analyses, presented in Table 1 , examined items as quoted in the survey. One set of items were positively correlated with IPV, while another set . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.08.20125914 doi: medRxiv preprint were negatively correlated. We examined group differences by state-level intensity of positive COVID-19 cases in relationship to IPV using GLM (see Table 2 ). Differences between coronavirus intensity groups were estimated for Psychological Victimization: We then employed a group mean ranking, based on post-hoc pairwise comparisons between expected violence rates. These comparisons were subject to the Bonferroni correction for multiple comparisons. As expected, group 4 had significantly higher rates of IPV victimization than group 1.

IPV usage and experiences were regressed on COVID-19 conditions. Prior to running regression models, we examined independent variables for multicollinearity. Collinearity diagnosis revealed small collinear effects between independent indicators. Tolerance was nearly 1.00 for each independent indicator; variance inflation factor (VIF) was slightly above 1.00 for all indicators (following recommendations 15 ; VIF<10). . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 9, 2020. 

People reporting COVID-19 positive were nearly 3 times more likely to experience IPV and 3 times more likely to use IPV. Residents in states with low-levels of COVID-19 were impacted with IPV at lower rates than states with higher virus spread. However, state level of COVID-19 outbreak was not found to be a significant IPV risk factor. People who lost their jobs due to COVID-19 were 2.5-3 times more likely to perpetrate IPV and experience physical IPV.

The results must be interpreted with caution due to study limitations such as utilizing selfreported data from a cross-sectional survey. An additional limitation is that the models in this brief report did not include key variables such as race/ethnicity. A more comprehensive model is expected to improve explanatory power and yield more precise results. These findings should not be interpreted as an endorsement to disengage in social distancing practices. Rather, these results emphasize need for an ongoing public-health response to IPV, including further investigation, . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.08.20125914 doi: medRxiv preprint funding for development/evaluation of interventions, and implementation of evidence-based practices.

This study provides initial evidence the pandemic is associated with increased likelihood of IPV, which seems to support emerging evidence 16 . The relationship between IPV and COVID-19 has important implications for scientific study, medical, nursing, public-health and social work practice.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 9, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.08.20125914 doi: medRxiv preprint .15 ** .12 ** .15 ** .12 ** . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 9, 2020. * p<.05 Note: IPV-intimate partner violence. Group 1 States: '1-5000 positive COVID-19 cases' (n=165); Group 2 States: '5000-10000 positive COVID-19 cases' (n=229); Group 3 States: '10000-24000 positive COVID-19 cases' (n=228); Group 4 States: '24000 or more positive COVID-19 cases' (n=1136). Latin letters were used to rank sub-group mean from the smallest mean ("a") and upward based on the pairwise post hoc comparisons with the Bonferroni correction for multiple comparisons. This was in effect in the case probability of experiencing IPV differed across the four state groups. Table 3 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 9, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.08.20125914 doi: medRxiv preprint

Coronavirus (COVID-19)

Characteristics of hospitalized adults with COVID-19 in an integrated health care system in california

Variation in COVID-19 Hospitalizations and deaths across new york city boroughs

Danger in danger: interpersonal violence during COVID-19 quarantine

COVID-19: Reducing the risk of infection might increase the risk of intimate partner violence

Using social and behavioural science to support COVID-19 pandemic response. Nat Hum Behav

COVID-19: a public health approach to manage domestic violence is needed

The pandemic paradox: the consequences of COVID 19 on domestic violence

Adel Elkbuli Alarming trends in US domestic violence during the COVID-19 pandemic

Impact of social distancing during COVID-19 pandemic on crime in Indianapolis

Note: IPV-intimate partner violence, N=1024 (only in any type of intimate relationship)