Access to domestic violence advocacy by race, ethnicity and gender: The impact of a digital warm handoff from the emergency department.

BACKGROUND: Racial/ethnic minority survivors of domestic violence (DV) referred from emergency departments (EDs) face barriers connecting with advocacy services due to systemic inequities. This study assesses the impact of Domestic Violence Report and Referral (DVRR), a digital mandatory reporting compliance system with integrated direct-to-advocacy referrals sent with patient consent, on survivors' likelihood of receiving advocacy by race/ethnicity and gender.
METHODS: We assessed ED charts between 2014-2018 in an urban, Level 1 trauma center for DV-related positive screening, chief concern, or documentation in medical/social work notes. We matched these visits by name to local DV advocacy agency records. Using logistic regression, we assessed survivor odds of reaching advocacy by DVRR receipt, race/ethnicity and gender. We used predicted probabilities to identify between-group differences in advocacy services received.
RESULTS: Of the 1366 visits, 740 received the DVRR intervention; 323 received advocacy services Without DVRR, compared to white, female survivors (n = 87), male survivors (n = 132) had 0.20 (95% CI: 0.07-0.56) times the odds of reaching advocacy compared to female survivors; Latinx survivors (n = 136) had 2.53 (95% CI: 1.58-4.07) times the odds of reaching advocacy compared to white survivors. With DVRR, the odds and predicted probabilities of reaching advocacy increased significantly for all survivors. White, female survivors (n = 74) who received DVRR experienced 2.60 (95% CI: 1.66-4.07) times the odds of connecting with advocacy. Compared to this reference group, Black survivors (n = 480) who received DVRR experienced 4.66 additional times the odds of reaching advocacy services (95% CI: 3.09-7.04) and male survivors (n = 84) experienced 8.96 additional times the odds (95% CI: 2.81-28.56). Overall, we predict 43% (95% CI: 38-53%) of Latinx survivors (n = 177), 36% (95% CI: 31-40%) of Black survivors (n = 480) and 23% (95% CI: 14-32%) of white survivors (n = 83); 37% (95% CI: 33-40%) of women (n = 656) and 29% (95% CI: 18-42%) of men (n = 84) received advocacy services with DVRR.
CONCLUSION: Direct-to-advocacy ED referrals facilitated by eHealth technology improve access to advocacy care for all survivors in this sample; marginalized racial and ethnic groups most often victimized by DV appear to have the highest rates of connection to advocacy.

Introduction

Approximately one in four women has been victimized by physical or sexual violence from an intimate partner during her lifetime, and many more have been harmed by family members, caregivers, or roommates [1]. Factors such as isolation, depression, post-traumatic stress disorder, internalizing responsibility for the perpetrator’s abuse, and fear for their safety or legal status have dissuaded domestic violence (DV) survivors from seeking help [2-4]. Nevertheless, nationally representative studies suggest that between 28,000 and 120,000 survivors were seen in emergency departments (EDs) for DV-related chief concerns annually and the authors noted that these were likely vast underestimates due to inconsistent screening and the underutilization of DV-related diagnostic codes [5, 6].

ED providers in multiple studies expressed concern that they were unable to address DV survivors’ underlying danger, noting that they lacked training and resources to effectively offer non-medical interventions to DV survivors [7-12]. Standard care for DV includes identification of DV, but rarely includes providing support, facilitating access to support, or following up to see if the patient ultimately received support [13, 14]. When it does, it typically consists of printed educational material or a phone number to a community-based advocacy agency that previous research suggests survivors feel a need to hide from the abuser [15]. In addition, survivors frequently experience post-traumatic stress that inhibits long-term planning [16, 17], and manipulation and coercion from the abuser [3] among other factors that compromise their ability to follow up with these resources. To address these challenges, some hospitals have implemented a strategy known as a warm handoff, in which the ED provider describes ED advocacy agency services personally transfers the survivor’s DV care to a DV advocate, typically via a phone call with the patient in the room or an advocate arriving in-person to the patient’s bedside to begin offering care [18, 19]. This approach, while resource-intensive, appears to successfully connect survivors to advocacy care [19, 20]. Warm handoff interventions for domestic violence have not been studied by race, ethnicity, gender, and other factors, although survivor experiences with DV victimization, help-seeking and intervention vary among these groups [21-24].

Inequities by race, ethnicity and gender

In 2017, the Centers for Disease Control and Prevention reported higher victimization rates among individuals who self-identified as Hispanic (all races) (8.6%), Black (9.4%), American Indian/Alaska Native (8.2%) or multiracial (12.5%), compared to non-Hispanic whites (5.7%) in the preceding 12 months [25]. Further, according to the National Health Interview Survey, Asian, Black, and Hispanic (Latinx)-white patients all experienced significantly less access to healthcare services than non-Hispanic white patients even after adjusting for insurance status [26]. Further, even when survivors of color access healthcare, they appear to experience differential quality of care: a cross-sectional study of 484 medical students found that they systematically discounted the pain, distress and discomfort of non-white survivors and adjusted their treatment recommendations accordingly [27]. Yet a systematic review of 36 studies focused on DV and health outcomes among racial and ethnic minority women found insufficient sample sizes to accurately represent racial or ethnic differences and that these studies often confounded race and ethnicity with environmental and other social determinants of health [28].

Social dynamics surrounding gender further affect survivors’ experiences of DV and receipt of medical and advocacy care. A cross-sectional survey of over 10,000 American adults found that 2.5 times more women than men experience DV [25], and 4 times more require medical care for an DV-related injury [29]. Women account for an even larger proportion of DV-coded ED visits: as many as 93–94%, according to nationwide U.S. surveys [5, 30]. Female DV survivors experience a greater likelihood of injury, sexual assault, fear, and depression than male survivors [31]. In addition, consistent barriers to accessing quality healthcare for women carry unique costs for survivors of DV. A landmark review of clinical and experimental research found that healthcare providers tend to systematically disbelieve or downgrade women’s self-reports of pain and distress by female patients [32]—a bias providers extend toward elderly and non-white patients as well [33]. Very little of this research has included the experiences of transgender and gender non-binary survivors of domestic violence, despite the additional barriers they face [34].

In contrast, while male survivors of DV have comparatively better access to medical care, they seek and receive DV advocacy care less frequently than female survivors [35]. Two systematic reviews of men’s experiences with DV, or help-seeking after DV, suggest this difference derives from men’s reluctance to acknowledge abuse victimization, beliefs that DV services are unavailable to male survivors of DV, and fears of professionals not believing them or that they might be accused of perpetrating abuse [36, 37]. In some cases, these fears may be founded. In a qualitative study, both male DV survivors and DV service providers express challenges to men’s help-seeking and providers’ offering supportive services [38]. In addition, a 2015 literature review of the DV experiences of men suggests many men seek help after experiencing retaliatory violence from a partner against whom they have committed DV [31].

Domestic Violence Report and Referral (DVRR) intervention

A novel digital intervention, Domestic Violence Report and Referral (DVRR), offers a digital warm handoff for survivors between ED providers and community-based DV advocates via a web-based platform. It is completed by doctors, nurses, social workers, or other care team members in a private setting with no ED visitors present and is available to providers at their discretion as an alternative to faxed, paper-based mandatory reports. DVRR includes body maps to record the nature, images and treatment of injuries; it also includes the 20-question Danger Assessment, a validated questionnaire that predicts a survivor’s risk of being killed by their intimate partner [39]. Answers to all Danger Assessment and other DVRR questions are required prior to form submission. These features are guided to bridge any gaps caused by ED providers’ lack of DV training [9, 40]. Upon completion, DVRR sends a digital referral to local law enforcement in compliance with California’s mandatory reporting requirement, which stipulates that medical professionals who encounter injuries caused by DV must report them to law enforcement [41]. Alongside this mandatory report, DVRR offers an optional automated referral to a local domestic violence advocacy agency. With survivor consent, DVRR sends their Danger Assessment score and referral information to a local DV advocacy agency, including a safe phone number at which the survivor can be reached by the advocate. DVRR is one of very few DV interventions to include a warm handoff to advocacy or to enable advocates to initiate contact with survivors in the days or weeks after their ED visit [42]. In addition, DVRR is the only mandatory reporting intervention to include a referral or warm handoff to DV advocacy.

A previous analysis of data at three hospitals suggests DV-affected ED patients who received DVRR were over three times as likely to receive subsequent advocacy services [43]. In this paper, we assess the impact of this digital warm handoff referral on survivors’ odds and predicted probabilities of receiving advocacy services after an ED visit for DV by race, ethnicity and gender. We compare these findings between groups to determine any differential impact of this intervention on survivors’ receipt of advocacy services by race, ethnicity and gender.

Methods

Data

This study draws on data collected between February 2014 and April 2018 from a Level 1 trauma center ED and a large DV advocacy center. The Committee for the Protection of Human Subjects at the University of California, Berkeley, as well as the institutional review board of the hospital, approved the study protocol. Trained research assistants at the hospital and trained advocates at the advocacy agency collected data from survivors’ electronic health records (EHR) and agency records using a standardized abstraction form. A built-in EHR search program identified all ED visits in which a patient met one of three inclusion criteria: they answered “yes” to the DV screening question “are you being physically hurt or threatened by someone close to you in your living situation?,” they received an ICD-9 or ICD-10 (diagnostic) code related to DV, or they stated a chief concern related to DV (e.g., arm broken in fight with boyfriend). These survivors were considered to have “Known DV;” this included violence perpetrated by current or former intimate partners as well as first-degree family members (e.g., siblings, grandparents) and roommates. We included both intimate and non-intimate partner domestic violence as both types are included in California’s mandatory reporting law, making both groups equally eligible for DVRR intervention. We documented the number of visits by patients who qualified for study inclusion more than once. These survivors may have required separate intervention for distinct DV episodes, or their visit may have been a follow-up or continuation of an earlier visit; because this distinction could not be determined for repeat survivors, we document the number of visits by patients who qualified for study inclusion more than once. We also noted the number of DV-related visits that included sexual assault, as these individuals received an intervention protocol specific to sexual assault that included referrals to services and advocacy intervention independent of DVRR. Research assistants reviewed all qualifying medical records and abstracted information including the survivor’s gender (male/female), race (white, Black), ethnicity (Latinx, non-Latinx), the relationship between perpetrator and victim (e.g., boyfriend, partner, spouse), ED visit date, and whether the DVRR intervention was administered. The researchers provided the advocacy agency with a list of the survivors’ full names; the list was grouped by the survivor’s race and/or ethnicity, gender and hospital visit date. The agency searched advocacy records using first and last name to determine whether the survivor had received agency services within six months of their hospital visit. To protect the confidentiality of the clients, the agency provided researchers with aggregate client information by race, ethnicity, gender and the date range of their visit. For this reason, age could not be included in the final aggregated dataset.

We excluded the records of survivors who did not identify in the included racial (white/Black), ethnic (Latinx/non-Latinx) and gender (male/female) categories (n = 111) due to the small sample sizes of these groups. We documented DVRR administration based on a completed report in the online system; we also considered that survivors with a partial, incomplete DVRR record had not received DVRR. We detected no missing or incomplete data within the hospital dataset; due to its aggregate nature the quality of the advocacy data is impossible to verify. Our final sample included 1366 survivor ED visits. Fig 1 outlines the chart selection process.

Fig 1

Selection flow of medical and advocacy records included for analysis.

Analysis

We used chi-square tests to detect significant differences in receipt of DVRR by race, ethnicity, gender, sexual assault victimization and repeat visit status. We used logistic regression to assess the association between receiving DVRR and the odds of receiving advocacy services. The logistic regression model included indicator variables for survivor race, ethnicity (Black, Latinx, white) and gender (female, male). We also included two interaction terms, one between DVRR administration and gender and the other between DVRR administration and racial/ethnic group. Stata’s lincom command was used to estimate the linear combination of coefficients. For each category of race, ethnicity and gender, we assessed the impact of DVRR on the survivor’s eventual connection to advocacy services using predicted probability with 95% confidence intervals. This offers a straightforward way to compare the differences between group likelihood of receiving advocacy with vs without DVRR. To accomplish this, we measured differences in predicted probability between groups, and tested those differences for significance using chi square tests. We conducted statistical analyses using Stata 14.2.

Results

Characteristics of the study population

Between February 2014 and April 2018, medical records for 1366 survivor visits to the ED documented a diagnosis, chief concern or medical/social work note that indicated a DV episode precipitated the ED visit. The majority of survivors were female (84.2%), identified as Black or African American (63.5%), did not experience sexual assault during the DV episode (85.1%), and had no other DV-related visits to the trauma center during the study period (74.2%). Providers administered DVRR in 54.2%, or 740, of the study visits (Table 1). Chi-square tests comparing the characteristics of survivors who received DVRR and those who did not revealed that the groups were similar in terms of race, ethnicity and repeat visit status. However, significantly more female than male survivors received the intervention. Providers were also significantly less likely to administer the intervention to survivors who had also experienced sexual assault during the DV episode.

Table 1

Descriptive characteristics of study ED visits.

	(b) Total (n = 1366)		(a) DVRR administered (n = 740)		(a) DVRR not administered (n = 626)
	%	N	%	N	%	N
Gender
Female**	84.2%	1150	88.6%	656	78.9%	494
Male	15.8%	216	11.4%	84	21.1%	132
Race and Ethnicity
Black	63.5%	868	64.9%	480	62.0%	388
White	13.5%	185	11.2%	83	16.3%	102
Latinx	22.9%	313	23.9%	177	21.7%	136
Sexual Assault
Yes**	14.9%	203	12.2%	90	18.1%	113
No	85.1%	1162	87.8%	650	81.8%	512
Multiple DV visits
Yes	25.7%	351	24.5%	181	27.2%	170
No	74.2%	1014	75.5%	559	72.7%	455

*p<0.05;

**p<0.01.

DVRR and receiving advocacy services by race, ethnicity, gender

DVRR administration was associated with a significant increase in all survivors’ odds of receiving advocacy services (Table 2). At baseline, white, female survivors’ odds of reaching advocacy services were 0.12 (95% CI: 0.07–0.20); Black survivors experienced similar odds. Latinx survivors experienced significantly higher odds of connecting with advocacy services than Black and white survivors (OR: 2.53, 95% CI: 1.58–4.07). Black survivors who did not receive the intervention had approximately the same odds of receiving services as white survivors. White women who received the intervention experienced an estimated 2.60-fold (95% CI: 1.66–4.07) increase in their odds of connecting with advocacy services. No significant or meaningful difference in odds was detected for Latinx survivors who received the intervention compared to white female survivors who received it; to preserve power in the model, no separate odds ratio was calculated for this group. A significant interaction (DVRR x Black) indicated that DVRR was associated with a greater change for Black survivors than for white women or Latinx survivors; Black survivors who received the intervention experienced an additional 4.67-fold (95% CI: 3.09–7.04) increase in the odds of connecting to advocacy services. Male DV survivors who did not receive the intervention had 0.20 times (95% CI: 0.07–0.55) the odds of connecting with advocacy services compared to female survivors. However, a significant interaction (DVRR x male) indicated that DVRR was associated with a greater change for male survivors than for female survivors. Male survivors who received DVRR experienced an additional 8.96-fold (95% CI: 2.81–28.56) increase in the odds of connecting with advocacy services.

Table 2

Odds ratios of connection to advocacy by patient receipt of DVRR, race/ethnicity and gender.

	No DVRR		DVRR		Interaction terms
	n	OR	n	OR	OR for DVRR = 1 vs DVRR = 0
	advocacy / no advocacy	(95% CI)	advocacy / no advocacy	(95% CI)
White, Female	9/78	Reference	19/55	2.60** (1.66–4.07)	N/A
Black	34/354	1.00 (0.54–1.82)	168/312	1.80* (0.99–3.26)	4.67** (3.09–7.04)
Latinx	28/108	2.53** (1.58–4.07)	76/101	Included in reference	N/A
Male	4/132	0.20** (0.07–0.56)	24/60	3.45* (1.09–10.87)	8.96** (2.81–28.56)

*p<0.05;

**p<0.01.

Next, we assessed the predicted probability of Black, Latinx, white, female and male survivors receiving advocacy services with and without the intervention. Without the intervention, 13% of Black, 20% of Latinx, and 10% of white survivors received follow-up services (Fig 2). When DVRR was administered, more than twice as many Black, Latinx and white survivors were predicted to reach advocacy, and these increases were significant within each group. This resulted in 29% of Black survivors, 43% of Latinx survivors and 23% of white survivors reaching advocacy. With the intervention, Latinx survivors were predicted to have a significantly higher probability of reaching advocacy services than white survivors (PP: 1.78; 95% CI: 1.17–2.71; p<0.01, results not shown).

Fig 2

Predicted probability (95% CI) of follow-up advocacy services by DVRR status and race and ethnicity.

Female survivors were significantly more likely than male survivors to subsequently connect with advocacy services (Fig 3). DVRR use appears to narrow that gap within this sample. Without DVRR, 16% of women and 4% of men subsequently received advocacy services, a significant difference (Likelihood difference: -9.46%; 95% CI: -13.08% to -5.82%; p<0.01). With DVRR administration, more than twice as many women and six times as many men received advocacy services, resulting in 33% of women survivors and 25% of men survivors receiving advocacy services (Likelihood difference: -9.16%; 95% CI: -18.89 to 0.57%; p = 0.06).

Fig 3

Predicted probability (95% CI) of follow-up advocacy services by DVRR status and sex.

(see S1 Appendix for detailed results).

Discussion

In this study, receipt of the DVRR digital warm handoff intervention increased patient odds and rate of connection to DV advocacy services for all survivors. It facilitated the highest rates of connection to advocacy among the groups most likely to be affected by DV: an estimated 43% of Latinx survivors and 29% of Black survivors reached advocacy when given DVRR [25]. Increased odds of reaching advocacy services were significantly greater for Black than white survivors due to the interaction between DVRR and Black identity.

DVRR was not developed to address inequities in DV care. However, many of the mechanisms through which this digital warm handoff intervention operates appear to correspond to strategies identified as culturally appropriate among Black and Latinx survivors. For Latinx survivors, several studies highlight the importance of interpersonal solidarity, cultural connections and family support in fostering safety and resilience [44-46]. Related themes of cultural solidarity and resisting victimization narratives are highlighted in research with Black survivors [22, 24, 47]. Warm handoff interventions such as DVRR facilitate personal connections between the advocate and survivor. In addition, warm handoffs may protect survivors against needing to embrace a stance of victimization and weakness that may conflict with cultural norms in order to reach advocacy services [22, 24, 48]. These factors may make digital warm handoff interventions uniquely applicable to Black and Latinx survivors and may at least partially explain the large effect size among these groups in this study.

Another mechanism for these changes may include a reduction in provider bias among delivery of the intervention to survivors. Historical scholarship and experimental research that suggest providers are more likely to stereotype and discount experiences of pain among racial and ethnic minorities (45,60). The structured questions in this intervention may bypass these providers’ unconscious biases and increase non-white survivors’ likelihood of reaching follow-up services. By including the Danger Assessment as the basis for its law enforcement report and advocacy referral, this digital warm handoff standardizes information-gathering and automates the content, delivery and destination of referrals. Currently, the 20-question Danger Assessment form, validated for women in heterosexual abusive relationships, is used for all DVRR recipients; future developments of DVRR include the brief, 5-question Danger Assessment tool that is validated across genders and racial and ethnic groups. This is an important step toward improving the strength of the Danger Assessment component of DVRR in reducing provider bias and providing culturally appropriate care. Any effect of DVRR in reducing provider bias would only apply to providers who chose to use DVRR. A shortened (5-question) Danger Assessment may shorten the provider time required to complete DVRR, thus increasing the likelihood of its use from 54% to an even higher value.

DVRR was also associated with a significant increase in receipt of advocacy services among both men and women. An estimated 33% of women and 25% of men reached advocacy services when given DVRR. This suggests that the intervention was associated with a reduction in the barriers that prevented both men and women survivors from reaching care. For women, the standardized Danger Assessment questionnaire may limit provider bias that discounts the severity of women’s experiences [32]. While the Danger Assessment questionnaire has not been validated in men, the structured, direct-to-advocacy referral may circumvent a self-reported reluctance to report or seek help and it may bypass potential fears that they may be ineligible for services [48, 49]. Providers administered DVRR at their discretion, so providers who chose to administer DVRR may also have been more likely to proactively connect survivors to services.

Strengths and limitations of the study

This study is the first to examine the impact of a digital warm handoff to DV advocacy on survivor outcomes by race and ethnicity and by gender. It contributes to the small bodies of literature exploring eHealth interventions for DV in the ED and warm-handoffs for DV survivors [50]. Additionally, it is one of the few to explore how digital interventions facilitate a warm handoff for DV survivors [42, 43]. Our findings highlight digital warm handoffs’ potential to provide additional benefit to the survivors most negatively impacted by DV. This study is further unique within DV research as its methods confidentially link hospital and advocacy records of DV survivors.

Only one hospital was included in this study, so the primary limitation of this study stems from its lack of generalizability. In addition, the matching technique between hospitals and advocacy records relies on first and last names, which could give groups with a higher frequency of identical given names (e.g., Latinx, male) a higher risk of false positive matches within this sample. Because records were matched based on survivor first and last names, spelling variability between hospital and advocacy records may also compromise the validity of the matched records.

To protect client privacy at the advocacy agency, covariate data collection was restricted a priori to variable combinations with cell sizes of 2 or greater. As a result, researchers were limited to six major covariate indicators (DVRR administration, male or female gender, Black, Latinx or non-Latinx white race/ethnicity). This required excluding certain populations with small sample sizes (e.g., Asian, Pacific Islander, Indigenous), and not specifying other relevant characteristics either due to a small sample size (e.g., LGBTQ+ identity) or non-categorical data structure (e.g., age). State-level studies, nationally representative research, and systematic reviews of smaller studies suggest many of these groups experience heightened barriers to DV response or services. These include survivors who are Asian, Pacific Islander, and Indigenous [28, 50] queer, transgender, and non-binary [51, 52] elderly [53] or who individuals live in neighborhoods with violence, limited access to services and other forms of community trauma [29, 54]. Future research may examine the efficacy of digital warm handoff interventions in connecting these DV survivors to advocacy services. In addition, the hospital in this study saw a majority of Black IPV patients (63.5%); this proportion is likely higher than in many other hospitals and may be associated with a higher level of culturally appropriate healthcare. The impact of DVRR for racial and ethnic minority patients in hospitals with a less heterogenous patient population may differ from the findings of those in this study due to potential differences in care. Future research should investigate the role of DVRR in reducing inequities among a broader sample of hospitals to determine the generalizability of the present study’s findings. As DVRR administration is contingent on provider discretion, future research should determine if providers differentially administer DV care, including DVRR, and any role this may play in survivor outcomes.

Finally, though rates of advocacy contact are significantly higher with a digital warm handoff, they are still quite low. Despite a two-fold or higher effect size for each group, fewer than half of all DV survivors in this sample ultimately reached contact with advocacy. Reasons for this may range from non-working phone numbers to unanswered phone calls. These missed connections may have resulted from survivors changing their mind about receiving advocacy care, the abuser monitoring the survivor’s phone, or inadvertent typos or missed calls. Future research should investigate the outcomes of survivors who do not connect with advocacy, including their preferences, needs, and interventions that may help them achieve safety, as well as the impact of connection to advocacy services. Any differences between survivors harmed by an intimate partner versus a family member or roommate should also be explored.

Conclusion

The present study found that a digital warm handoff referral to DV advocacy improved access to care for women and men and among Black, Latinx, and white survivors, with a significant additional increase in the odds of advocacy connection among Black and male survivors. These findings suggest that a digital warm handoff provides meaningful assistance to all DV survivors; they further suggest that such an intervention can be particularly meaningful for members of groups at the greatest risk of DV and inequities in care. Given the disproportionate burden of DV on these vulnerable groups and the additional barriers they face in accessing adequate DV and healthcare, DVRR represents a meaningful step toward adequate support for these vulnerable survivors as they seek DV care in an ED setting.

Standardized data abstraction form.

(PDF)

Click here for additional data file.

Related manuscript—Main effects.

Domestic Violence Report and Referral: A Multiple Baseline Study of an eHealth Warm Handoff for Emergency Department Patients affected by Domestic Violence.

(PDF)

Click here for additional data file.

Minimum underlying dataset.

(XLSX)

Click here for additional data file.

27 May 2021

PONE-D-21-05014

Access to domestic violence advocacy by race, ethnicity and gender:

The impact of a digital warm handoff from the emergency department

PLOS ONE

Dear Dr. Brignone,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Thank you for submitting this manuscript on a very important topic related to IPV and racial differences in receiving and accessing advocacy services. This manuscript fills an important gap, but needs some additional editing and clarity before publication. 

Please submit your revised manuscript by Jul 11 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at  https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Michelle L. Munro-Kramer, PhD, CNM, FNP-BC

Academic Editor

PLOS ONE

Additional Editor Comments:

Both of the reviewers provided detailed comments and suggestions that should be attended to prior to resubmission. I would specifically encourage the authors to pay attention to:

1) Streamlining the background section to focus specifically on the topics related to this study. Please see Reviewer #1's suggestion for how to do this.

2) Reviewer #1 & Reviewer #2's questions related to clarification in the methods section (e.g., which Danger Assessment was used - 5 versus 20 items; what EHR search program; how referral data was linked; how are repeat visits handled; measurement/definitions used for race and gender).

3) Discussion - As noted by Reviewer #2 - the study included a high proportion of Black patients. How did this impact the results? Please provide more details/discussion around the fact that about half of the sample did not receive the intervention and of those who did only 23-43% accessed services.

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating the following in the Financial Disclosure section:

"LB received funding for this research from Gilead Inc. Grant #00495. The funders played no role in the study design, data collection or analysis, decision to publish or preparation of the manuscript. Their website can be accessed here: " ext-link-type="uri" xlink:type="simple">https://www.gilead.com/purpose/giving/funding-requests"

We note that you received funding from a commercial source: Gilead Inc.

Please provide an amended Competing Interests Statement that explicitly states this commercial funder, along with any other relevant declarations relating to employment, consultancy, patents, products in development, marketed products, etc.

Within this Competing Interests Statement, please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests).  If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your amended Competing Interests Statement within your cover letter. We will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

3. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is an important study of ED patients with DV, and their receipt of a digital warm handoff intervention called DVRR, and their receipt of advocacy services. I have several questions and comments to improve clarity.

Abstract, Methods:

1. Can you describe briefly how 1366 ED visits are related to 323 DV advocacy visits? Meaning are the DV advocacy visits from DVRR performed in ED?

2. Is the intervention DVRR? Meaning odds and predicted probability are comparing ED visits for DV where ED patients received DVRR or not? Did all ED patients with DV receive DVRR?

3. Can you explain more what predicted probabilities adds to odds ratios? Is it to show disparities by race/ethnicity of receiving DVRR?

Abstract, Results:

1. Latinx odds reference group is Black and white combined? In a logistic regression? Did you use a multinomial logistic regression with Latinx compared to Black compared to white

2. Is this odds of 2.53 for Latinx survivors who did not receive DVRR?

3. When you write compared to white, female survivors, is Black survivors’ 4.66 odds of reaching advocacy in those without DVRR? Is this a logistic regression (binary outcome) or multinomial outcome?

4. Male survivors experienced 12.8 times the odds compare to who? Female survivors with DVRR?

Manuscript text:

Introduction:

1. The section on race and ethnicity first sentences describes DV race/ethnicity differences. I would keep that. The 5 paragraphs describe differences in use of or service-provider creation of DV advocacy services and shelter. However, the length of text distracts from the focus on ED patients, DVRR those patients do or don’t receive, and eventual services those patients do or don’t use. I would remove the remaining material except for the last paragraph on “examining ED services”.

2. I would keep the section on gender as is, as it’s concise at 2 paragraphs.

3. DVRR section: can you describe more about DVRR? This in many ways is the crucial part of your introduction. I don’t know how to find reference 51, which is listed as a poster. Is it https://urapprojects.berkeley.edu/projects/detail.php?id_list=Soc0795? That describes data to collect, not data about DVRR. See my next set of questions below.

Methods

Intervention

1. I would move this paragraph to introduction. This material says DVRR is used at the discretion of the ED provider. Can you tell us more about this discretion? This gets at provider bias, and in many ways may determine if DVRR is useful or not.

2. Is the Danger Assessment the full 20 questions? Or the shorter versions like at https://pubmed.ncbi.nlm.nih.gov/28921610/? A short vs long questionnaire (5 questions vs. 20 questions) may mean providers use or don’t use DVRR, as an extra 15 questions will increase provider time in working with patients.

3. Describe more about CA mandatory reporting law. It is unlike other state mandatory reporting laws (https://www.futureswithoutviolence.org/userfiles/file/HealthCare/Compendium%20Final.pdf).

4. Do you know how many ED patients with DV receive DVRR, as opposed to provider who calls police directly to make report (for example in other states providers must make a call)?

5. Who (nurse, social worker, doctor) completes DVRR during a clinical encounter? Does that influence if DVRR is used?

6. Do patients not receive other DV assistance during DV encounter (safety planning for positive danger assessment screen) if DVRR is completed (where an advocate later may or may not call patient)?

7. Are ED’s completing DVRR without other ED visitors/potential perpetrators present?

8. Can survivors voluntarily decide not to receive DVRR? I know you may not have answers to many of my questions on DVRR, but the answers all influence how we interpret your results. DVRR is a novel process that I think is great, but it deserves explaining so that providers in other states can think about if a similar program is worth developing.

Data

1. Can we see a copy of the standardized abstraction form? Maybe put it as an appendix?

2. Tell me more about the EHR search program that identified all ED visits in which a patient screened positive for DV. What screening tool was used? Was the answer to only 1 DV question (out of however many DV questions were asked) considered a positive screen?

3. How do you define chief complaint (I recommend calling it a chief concern; providers may feel that patients are complaining, but patients express important concerns about their health) related to DV? Does the patient have to say “I am a victim of DV” to be considered having a chief concern related to DV?

4. Do ED patients with repeated visits for DV receive multiple DVRR?

5. Were visits by survivors of sexual assault by a partner considered to have experienced sexual assault or DV?

6. How is gender measured? Does it include non-binary identifying patients?

7. How is race defined? By patient or provider? What about a patient who has more than 1 race?

8. When you describe “the relationship between perpetrator and victim”, aren’t all DV cases relationships between partners? Is this referring to sexual assault cases?

9. Do you have a measure of how well DVRR was administered? Like full 20 question Danger Assessment, or 5-item Danger Assessment screen described above? Was there safety planning conducted in the ED for positive Danger Assessment screens?

10. Were there duplicates for survivor full names and hospital visit dates at DV agency?

Figure 1: How do you define known DV? What are your inclusion criteria? Do we know why some patients did not receive DVRR? Shouldn’t they all have?

Data analysis

1. Does DVRR administration mean that DVRR was documented to have occurred during the ED visit? Or that a DVRR report existed in the online system? Did patients receive multiple DVRR over multiple visits?

2. Describe more about what predicted probably adds to the logistic regression models

Characteristics of the study population

1. 74.2% had no other visits to the trauma center during the study period. Does that mean no other ED visits? Or no other visits for DV?

2. You write that survivors who experienced sexual assault were less likely to have administered DVRR. Do those providers receive a different intervention?

3. I’d recommend moving this paragraph to results.

Table 1: what are the **denoting? A level of p value? Or just a statistically significant difference in % values in that characteristic?

Results

1. Can you explain how the interaction terms increased the odds of connecting to advocacy services?

2. You write odds of accessing advocacy services, and odds of connecting to advocacy services. Are those the same things?

3. Explain what the predicted probability adds to the odds ratio analyses already presented. Isn’t it showing similar results to DVRR associated with increased odds of receiving advocacy services?

Table 2: is this a multinomial logistic regression? Of a series of logistic regressions shown in different columns? What is the reference group for Baseline odds, No DVRR white female? Why in the DVRR female white and latinx are white and latinx grouped together?

Discussion:

1. 2nd paragraph: You write that DVRR is an automatic referral. Given that only 54% of patients received it (from Methods/Characteristics of the study population), and your later description of providers choosing to administer DVRR may be more likely to connect survivors to services, I think you can describe this as a digital referral with lethality assessment.

2. Is there literature on Danger Assessment being validated in men who experienced violence?

Strengths and limitations

1. You write that “it confidentially links the hospital and advocacy records of DV survivors.” Do you mean DVRR, or this research study on DVRR?

2. Do you have data on how many identical given names (duplicates) occurred in your data? That would allow you to determine if duplicates affected intervention and non-intervention groups equally. I otherwise do not think you can conclude this.

Reviewer #2: Access to domestic violence advocacy by race, ethnicity and gender:

The impact of a digital warm handoff from the emergency department

Accept minor revisions

This study fills an important gap in our evidence base regarding intervention for IPV, and its focus on racial differences is particularly timely. It is also very nice to see an intervention that differentially and positively impacts Latinx and Black women. The intervention itself is well-conceived to address challenges in IPV referral in the ED, and in particular that the hand-off is to an advocacy organization rather than only to law enforcement (as required by CA law).

Specific comments:

1. The introduction is far too lengthy and reads more like an essay. Only the salient points for this analysis, summarized, are needed.

2. The authors state in their methods that they detected no missing or incomplete data, but with only aggregate data from the advocacy organizations, the quality of the data is impossible to verify. This should be noted.

3. The process for linking referrals to follow-up visits is not described. Linkages based on name, because of spelling variability, are particularly problematic and generally have some validity challenges.

4. Given strong evidence that BIPOC patients report challenges in their interactions with healthcare facilities, the generalizability of this study needs to be addressed. For example, many EDs might not have a majority of Black IPV patients (in this sample, 63.5%), and perhaps the staff at this hospital are particularly adept in culturally appropriate and ethnicity-centered healthcare. How would this potentially impact EDs with different patient profiles?

5. Although these findings are very promising, nearly half of the IPV victims in the ED did not receive the intervention, and even with the improved access to advocacy services with the intervention, only between 23% and 43% of those referred accessed services. The discussion should address these issues by providing some insight into why these proportions are low and how they might be improved.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

16 Jun 2021

Both of the reviewers provided detailed comments and suggestions that should be attended to prior to resubmission. I would specifically encourage the authors to pay attention to:

1) Streamlining the background section to focus specifically on the topics related to this study. Please see Reviewer #1's suggestion for how to do this.

The background section has been streamlined and substantially shortened as requested.

2) Reviewer #1 Reviewer #2's questions related to clarification in the methods section (e.g., which Danger Assessment was used - 5 versus 20 items; what EHR search program; how referral data was linked; how are repeat visits handled; measurement/definitions used for race and gender).

These clarifications have been provided and are described in detail in the reviewer responses.

3) Discussion - As noted by Reviewer #2 - the study included a high proportion of Black patients. How did this impact the results? Please provide more details/discussion around the fact that about half of the sample did not receive the intervention and of those who did only 23-43% accessed services.

The discussion section has been expanded to include these details.

Journal Requirements:

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_ title_authors_affiliations.pdf

PLOS guidelines have been reviewed and submission materials have been updated accordingly

2. Thank you for stating the following in the Financial Disclosure section: "LB received funding for this research from Gilead Inc. Grant #00495. The funders played no role in the study design, data collection or analysis, decision to publish or preparation of the manuscript. Their website can be accessed here: https://www.gilead.com/purpose/ giving/funding-requests." We note that you received funding from a commercial source: Gilead Inc. Please provide an amended Competing Interests Statement that explicitly states this commercial funder, along with any other relevant declarations relating to employment, consultancy, patents, products in development, marketed products, etc. Please include your amended Competing Interests Statement within your cover letter. We will change the online submission form on your behalf.

This statement has been amended as requested and declared in the cover letter.

Within this Competing Interests Statement, please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests). If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

This statement has been declared in the cover letter.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

The authors have declared in the cover letter that we have no other competing interests to disclose.

3. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability. Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized. We will update your Data Availability statement to reflect the information in your cover letter. Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

The minimum underlying dataset is attached as a Supporting Information file. This has been noted in the cover letter.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

This information has been updated as requested.

Reviewer #1: This is an important study of ED patients with DV, and their receipt of a digital warm handoff intervention called DVRR, and their receipt of advocacy services. I have several questions and comments to improve clarity.

Abstract, Methods:

1. Can you describe briefly how 1366 ED visits are related to 323 DV advocacy visits? Meaning are the DV advocacy visits from DVRR performed in ED?

The abstract text has been updated to read: “323 subsequent visits to a local DV advocacy agency.” The Methods now clarify that DV advocates respond to the referral within days or weeks after their ED visit.

2. Is the intervention DVRR? Meaning odds and predicted probability are comparing ED visits for DV where ED patients received DVRR or not? Did all ED patients with DV receive DVRR?

The intervention is DVRR; about half of ED patients with DV received DVRR. This has been clarified in the abstract.

3. Can you explain more what predicted probabilities adds to odds ratios? Is it to show disparities by race/ethnicity of receiving DVRR?

Yes; this has been clarified in the abstract.

Abstract, Results:

1. Latinx odds reference group is Black and white combined? In a logistic regression? Did you use a multinomial logistic regression with Latinx compared to Black compared to white?

This has been clarified in the abstract; white female survivors constituted the odds reference group in this logistic regression. Without DVRR, Black and white survivors reached advocacy at rates that were not significantly different.

2. Is this odds of 2.53 for Latinx survivors who did not receive DVRR?

Yes. This has been clarified in the abstract.

3. When you write compared to white, female survivors, is Black survivors’ 4.66 odds of reaching advocacy in those without DVRR? Is this a logistic regression (binary outcome) or multinomial outcome?

Logistic regression; this has been clarified in the abstract.

4. Male survivors experienced 12.8 times the odds compare to who? Female survivors with DVRR?

Yes. This has been clarified in the abstract.

Manuscript text:

Introduction:

1. The section on race and ethnicity first sentences describes DV race/ethnicity differences. I would keep that. The 5 paragraphs describe differences in use of or service-provider creation of DV advocacy services and shelter. However, the length of text distracts from the focus on ED patients, DVRR those patients do or don’t receive, and eventual services those patients do or don’t use. I would remove the remaining material except for the last paragraph on “examining ED services”.

The section has been edited per this suggestion and now consists of two paragraphs, comparable in length to the section on gender.

2. I would keep the section on gender as is, as it’s concise at 2 paragraphs.

Done, thanks.

3. DVRR section: can you describe more about DVRR? This in many ways is the crucial part of your introduction. I don’t know how to find reference 51, which is listed as a poster. Is it https://urapprojects.berkeley.edu/projects/detail.php?id_list=Soc0795? That describes data to collect, not data about DVRR. See my next set of questions below.

This paragraph has been combined with the first paragraph of the Methods section and its detail has been expanded considerably.

Methods

Intervention

1. I would move this paragraph to introduction. This material says DVRR is used at the discretion of the ED provider. Can you tell us more about this discretion? This gets at provider bias, and in many ways may determine if DVRR is useful or not.

This paragraph is now in the introduction. Details about provider discretion are also included in this paragraph.

2. Is the Danger Assessment the full 20 questions? Or the shorter versions like at https://pubmed.ncbi.nlm.nih.gov/28921610/? A short vs long questionnaire (5 questions vs. 20 questions) may mean providers use or don’t use DVRR, as an extra 15 questions will increase provider time in working with patients.

Right now, the program uses the full 20 questions (added to the Methods). In addition, the discussion now notes this complication and states that future DVRR development is replacing this with the 5-question version for a more efficient, streamlined approach.

3. Describe more about CA mandatory reporting law. It is unlike other state mandatory reporting laws (https://www.futureswithoutviolence.org/userfiles/file/HealthCare/Compendium%20Final.pdf).

Added: Upon completion, DVRR automatically sends a digital referral to local law enforcement in compliance with California’s mandatory reporting requirement, which stipulates that medical professionals who encounter injuries caused by DV must report them to law enforcement.

4. Do you know how many ED patients with DV receive DVRR, as opposed to provider who calls police directly to make report (for example in other states providers must make a call)?

Because it is hospital policy (and a legal requirement) that all DV cases seen in the ED be reported to police, I assume that all survivors who did not receive DVRR did have a phone call made to police on their behalf (and I’ve added that assumption to the Methods). However, we did not measure compliance with the mandate.

5. Who (nurse, social worker, doctor) completes DVRR during a clinical encounter? Does that influence if DVRR is used?

Any member of the care team can complete DVRR (added to methods). Right now we don’t know if that influences whether or not DVRR is used – this is noted in the discussion and a question for the next manuscript.

6. Do patients not receive other DV assistance during DV encounter (safety planning for positive danger assessment screen) if DVRR is completed (where an advocate later may or may not call patient)?

This was not measured; existing hospital protocols don’t suggest that treatment of DVRR recipients would differ from non-DVRR patients.

7. Are ED’s completing DVRR without other ED visitors/potential perpetrators present?

Added: DVRR is completed without ED visitors for the survivor’s privacy.

8. Can survivors voluntarily decide not to receive DVRR? I know you may not have answers to many of my questions on DVRR, but the answers all influence how we interpret your results. DVRR is a novel process that I think is great, but it deserves explaining so that providers in other states can think about if a similar program is worth developing.

Survivors can decide to not answer questions and can decline to have the report sent to advocacy; they can’t prevent the mandatory report from being sent. This has been clarified in the text.

Data

1. Can we see a copy of the standardized abstraction form? Maybe put it as an appendix?

Attached.

2. Tell me more about the EHR search program that identified all ED visits in which a patient screened positive for DV. What screening tool was used? Was the answer to only 1 DV question (out of however many DV questions were asked) considered a positive screen?

Only one screening question was used at this hospital (“are you being physically hurt or threatened by someone close to you or in your living situation?”). This question has been added to the text.

3. How do you define chief complaint (I recommend calling it a chief concern; providers may feel that patients are complaining, but patients express important concerns about their health) related to DV? Does the patient have to say “I am a victim of DV” to be considered having a chief concern related to DV?

A chief concern (this change has been incorporated throughout the manuscript) related to DV was considered to be any health concern that could reasonably be inferred to have been a direct result of DV (e.g., broken arm in fight with a boyfriend).

4. Do ED patients with repeated visits for DV receive multiple DVRR?

They may, per the provider’s discretion (e.g., if the second visit is two years later with a different partner, they will more likely receive DVRR than if the second visit a follow-up the next day). Available data did not allow us to distinguish meaningful criteria for these differing types of concerns so this was not measured. This has been noted in the text.

5. Were visits by survivors of sexual assault by a partner considered to have experienced sexual assault or DV?

Both; that’s why they were noted in data collection. This has been reworded for clarity in the Methods.

6. How is gender measured? Does it include non-binary identifying patients?

Binary. Only one DV-related record was identified as having a gender non-binary patient – and out of thousands of survivors that itself is interesting– but that obviously wasn’t sufficient for analysis. This has been noted in the text.

7. How is race defined? By patient or provider? What about a patient who has more than 1 race?

By patient, and racial and ethnic information were recorded separately. As a result, the hospital records on race and ethnicity offered very granular information. However, for analysis, only Black and white patient groups had sufficient sample sizes for analysis, as well as survivors of any race who identified their ethnicity as “Hispanic/Latinx.” When an included racial and ethnic label were recorded for a survivor, the ethnic label was given preference (e.g., a survivor who identified as “race: white, ethnicity: Latinx” was considered “Latinx” for this study). Survivors who identified with any racial or ethnic category not included in “white” “Black” or “Latinx” groups, including multiracial survivors, were not included for analysis.

8. When you describe “the relationship between perpetrator and victim”, aren’t all DV cases relationships between partners? Is this referring to sexual assault cases?

Added: For the hospital DV screening question and in this study, DV included violence perpetrated by current or former intimate partners, first-degree family members (e.g., siblings, grandparents) and roommates.

9. Do you have a measure of how well DVRR was administered? Like full 20 question Danger Assessment, or 5-item Danger Assessment screen described above? Was there safety planning conducted in the ED for positive Danger Assessment screens?

Added/edited: it also includes the 20-question Danger Assessment, a validated questionnaire that predicts a survivor’s risk of being killed by their intimate partner (52). Answers to all Danger Assessment and other DVRR questions are required prior to form submission.

10. Were there duplicates for survivor full names and hospital visit dates at DV agency?

Not to our knowledge, though this is possible and noted in the limitations.

Figure 1: How do you define known DV? What are your inclusion criteria? Do we know why some patients did not receive DVRR? Shouldn’t they all have?

Added: These survivors were considered to have “Known DV;” this included violence perpetrated by current or former intimate partners, first-degree family members (e.g., siblings, grandparents) and roommates.

Added: Within this hospital, DVRR was available to providers at their discretion as an alternative to faxed, paper-based mandatory reports.

Data analysis

1. Does DVRR administration mean that DVRR was documented to have occurred during the ED visit? Or that a DVRR report existed in the online system? Did patients receive multiple DVRR over multiple visits?

Added: We documented DVRR administration based on a completed report in the online system; we also considered that survivors with a partial, incomplete DVRR record had not received DVRR.

Some repeat survivors received DVRR more than once, as described in the following clarification: [Repeat] survivors may have required separate intervention for distinct DV episodes, or their visit may have been a follow-up or continuation of an earlier visit and not required additional DV intervention; because this distinction could not be determined for repeat survivors, we document the number of visits by patients who qualified for study inclusion more than once.

2. Describe more about what predicted probably adds to the logistic regression models

This paragraph has been expanded as requested.

Characteristics of the study population

1. 74.2% had no other visits to the trauma center during the study period. Does that mean no other ED visits? Or no other visits for DV?

DV-related visits. This has been clarified in the text.

2. You write that survivors who experienced sexual assault were less likely to have administered DVRR. Do those patients receive a different intervention?

Yes. From line 175: We also noted the number of DV-related visits that included sexual assault, as these individuals received a separate intervention protocol that included referrals to services and advocacy intervention Added: independent of DVRR.

3. I’d recommend moving this paragraph to results.

The paragraph has been moved.

Table 1: what are the **denoting? A level of p value? Or just a statistically significant difference in % values in that characteristic?

A level of p-value (*p0.05; **p0.01). This was a typographical error and has been corrected in the text.

Results

1. Can you explain how the interaction terms increased the odds of connecting to advocacy services?

Revised text (comparable for both interaction terms): A significant interaction (DVRR x Black) indicated that DVRR was associated with a greater change for Black survivors than for white or Latinx survivors; Black survivors who received the intervention experienced an additional 4.66-fold (95% CI: 3.09-7.04) increase…

2. You write odds of accessing advocacy services, and odds of connecting to advocacy services. Are those the same things?

Yes. The language has been standardized.

3. Explain what the predicted probability adds to the odds ratio analyses already presented. Isn’t it showing similar results to DVRR associated with increased odds of receiving advocacy services?

Yes. Predicted probability more effectively illustrates the change in between-group differences with/without DVRR. This also allows the multidimensional relationship to be graphically depicted clearly.

Table 2: is this a multinomial logistic regression? Of a series of logistic regressions shown in different columns? What is the reference group for Baseline odds, No DVRR white female? Why in the DVRR female white and latinx are white and latinx grouped together?

The headings in this table have been adjusted to address these concerns.

Discussion:

1. 2nd paragraph: You write that DVRR is an automatic referral. Given that only 54% of patients received it (from Methods/Characteristics of the study population), and your later description of providers choosing to administer DVRR may be more likely to connect survivors to services, I think you can describe this as a digital referral with lethality assessment.

The text has been amended.

2. Is there literature on Danger Assessment being validated in men who experienced violence?

Not to my knowledge; this has been noted in the text.

Strengths and limitations

1. You write that “it confidentially links the hospital and advocacy records of DV survivors.” Do you mean DVRR, or this research study on DVRR?

Revised text: This study is further unique within DV research as its methods confidentially link hospital and advocacy records of DV survivors.

2. Do you have data on how many identical given names (duplicates) occurred in your data? That would allow you to determine if duplicates affected intervention and non-intervention groups equally. I otherwise do not think you can conclude this.

These data were not collected; the conclusion has been deleted.

Reviewer #2: Access to domestic violence advocacy by race, ethnicity and gender:

The impact of a digital warm handoff from the emergency department

Accept minor revisions

This study fills an important gap in our evidence base regarding intervention for IPV, and its focus on racial differences is particularly timely. It is also very nice to see an intervention that differentially and positively impacts Latinx and Black women. The intervention itself is well-conceived to address challenges in IPV referral in the ED, and in particular that the hand-off is to an advocacy organization rather than only to law enforcement (as required by CA law).

Specific comments:

1. The introduction is far too lengthy and reads more like an essay. Only the salient points for this analysis, summarized, are needed.

The introduction has been substantially shortened.

2. The authors state in their methods that they detected no missing or incomplete data, but with only aggregate data from the advocacy organizations, the quality of the data is impossible to verify. This should be noted.

Added: We detected no missing or incomplete data within the hospital dataset; due to its aggregate nature the quality of the advocacy data is impossible to verify.

3. The process for linking referrals to follow-up visits is not described. Linkages based on name, because of spelling variability, are particularly problematic and generally have some validity challenges.

This is true, and is discussed in the manuscript. Added: Because records were matched based on survivor first and last names, spelling variability between hospital and advocacy records may also compromise the validity of the matched records.

4. Given strong evidence that BIPOC patients report challenges in their interactions with healthcare facilities, the generalizability of this study needs to be addressed. For example, many EDs might not have a majority of Black IPV patients (in this sample, 63.5%), and perhaps the staff at this hospital are particularly adept in culturally appropriate and ethnicity-centered healthcare. How would this potentially impact EDs with different patient profiles?

Added: In addition, the hospital in this study saw a majority of Black IPV patients (63.5%); this proportion is likely higher than in many other hospitals and may be associated with a higher level of culturally appropriate healthcare. The impact of DVRR for racial and ethnic minority patients in hospitals with a less heterogenous patient population may differ from the findings of those in this study due to potential differences in care.

5. Although these findings are very promising, nearly half of the IPV victims in the ED did not receive the intervention, and even with the improved access to advocacy services with the intervention, only between 23% and 43% of those referred accessed services. The discussion should address these issues by providing some insight into why these proportions are low and how they might be improved.

Added: Reasons for this may range from non-working phone numbers to unanswered phone calls. These missed connections may have resulted from survivors changing their mind about receiving advocacy care, the abuser monitoring the survivor’s phone, or inadvertent typos or missed calls.

Submitted filename: Reviewer responses.docx

Click here for additional data file.

28 Oct 2021

10 Dec 2021

The authors would like to thank the reviewers/editor for their thoughtful, constructive comments. Their suggestions have been implemented as described below.

Reviewer #1: This is an important study of ED patients with DV, their receipt of a digital warm handoff intervention called DVRR, and their receipt of advocacy services. I have several questions and comments to improve clarity.

Abstract:

1. Background: I suggest changing first sentence to “racial/ethnic minority survivors of domestic violence (DV) referred from emergency departments (Eds) face barriers connecting with advocacy services due to systemic inequalities.” Their inequities in healthcare access may not be the same as those found in advocacy services. I also suggest describing DVRR as an electronic reporting system to fulfill mandatory reporting requirements to law enforcement, with report also going to DV advocacy agency.

The proposed changes have been implemented.

2. Methods: Describe time frame (2014-2018) for study. Briefly describe how DV-associated ED visits were defined. I recommend not describing sample sizes (1366, 740, 323) in methods section as those are results. You could state something like “We assessed 2014-2018 ED visit chief concern for DV, measured if DVRR occurred, and if patient names matched DV advocacy agency records of survivors using services”. I would suggest giving sample size for number used in logistic regression analyses and number used in predicted probability analyses.

Each of these proposed changes has been implemented.

Manuscript text

Introduction

1. 2nd paragraph: You write that “standard of care for DV often does not extend beyond hospital screening; when it does, it typically consists of printed educational material or a phone number to a community-based advocacy agency…”, and I think you are meaning that standard of care for DV includes identification of DV but does not necessarily include providing support and resources, or follow-up to see if patient used advocacy agency services. I would clarify this.

The paragraph has been clarified per this exceptionally helpful comment.

2. 2nd paragraph: You write “…warm handoff, in which the ED provider personally transfers the survivor’s DV care to a DV advocate.” A transfer of care in medical care more typically means physical movement of the patient location. A warm handoff might be better described as a healthcare provider describing DV advocacy agency services and possibly calling the DV advocacy agency while in the patient room. This may include a DV advocacy agency staff member coming to the patient location for an in-person provision of support and resources. I would clarify this.

The paragraph has been revised to clarify this information.

3. Race and ethnicity, and Gender sections: I recommend condensing this into 1 paragraph. The focus of the paper is ED visits, DVRR and DV advocacy visit. The first 2 paragraphs describe ED visits and DV advocacy, and the last paragraph describes DVRR.

These two sections have been condensed into a single section with three paragraphs. The key findings in this paper highlight differential results by race, ethnicity and gender and this section has been condensed to more concisely presage those findings.

Results

1. I recommend changing “Analysis” header (between first and second paragraphs) to “DVRR and receiving advocacy services by race/ethnicity and gender”. Otherwise, the reader will think this is the methods section analysis plan

The heading has been revised per this section.

2. Last paragraph: what does likelihood difference assist reader in knowing?

This paragraph has been revised to read “Female survivors were significantly more likely than male survivors to subsequently connect with advocacy services (Fig 3). DVRR use appears to narrow that gap within this sample.”

Table 2:

1. I would put 95%CI with odds ratio in 1 row.

This table has been reformatted and the preceding paragraph has been updated accordingly

2. Why is the reference group baseline odds ratio (white, female) 0.12? Shouldn’t it be 1.0 as it’s the reference group?

This information has been removed from the table and inserted in the preceding paragraph

Figures 1 and 2: what does statistical test for % increase add? I think you can remove it. It’s otherwise too much information. It’s obvious from the table that there is a visually large increase by DVRR.

These tests have been removed.

S1 Appendix: can you explain more why this appendix is needed? Doesn’t Figure 2 adequately show differences in DVRR status by race and ethnicity?

This appendix has been removed.

Discussion: Paragraph 3: the authors describe how 5-question Danger Assessment “is an important step toward improving the strength of the Danger Assessment component of DVRR in reducing bias and providing culturally appropriate care.” I think the using 5-question DA will help provider time in completing the DA, and therefore completing the full DVRR assessment. It may increase DVRR administration from 54% to something higher.

This consideration has been added.

Strengths and limitations: Paragraph 3: can you describe the “data restrictions that made it impossible to study additional groups that…experience heightened barriers to DV response or services”. Can you explain more about those data restrictions?

The sentence described has been revised and expanded to state the following:

To protect client privacy at the advocacy agency, covariate data collection was restricted a priori to variable combinations with cell sizes of 2 or greater. As a result, researchers were limited to six major covariate indicators (DVRR administration, male or female gender, Black, Latinx or non-Latinx white race/ethnicity). This required excluding certain populations with small sample sizes (e.g., Asian, Pacific Islander, Indigenous), and not specifying other relevant characteristics either due to a small sample size (e.g., LGBTQ+ identity) or non-categorical data structure (e.g., age).

Submitted filename: Reviewer Responses - round 2.docx

Click here for additional data file.

18 Feb 2022

Access to domestic violence advocacy by race, ethnicity and gender:

The impact of a digital warm handoff from the emergency department

PONE-D-21-05014R2

Dear Dr. Brignone,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Michelle L. Munro-Kramer, PhD, CNM, FNP-BC

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Thank you for the revisions and congratulations on the acceptance of this article. It will make a significant contribution to the literature.

Reviewers' comments:

10 Mar 2022

PONE-D-21-05014R2

Access to domestic violence advocacy by race, ethnicity and gender: The impact of a digital warm handoff from the emergency department

Dear Dr. Brignone:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Michelle L. Munro-Kramer

Academic Editor

PLOS ONE