Abuse in dependent relationships in adults between 18 and 65 years old refers to violence perpetrated by an individual upon whom the victim is dependent. This can include partners, ex-partners, family members, or family friends, often characterized by a significant power imbalance. When children are involved (even only as witnesses), the situation falls under the so called “Child-Check” of the KNMG reporting code and the child abuse directive must be followed. If the victim is 65 years or older, it is called elder abuse.

As a healthcare provider, it is crucial to recognize the signs of such violence promptly and initiate appropriate support to prevent further harm to both the patient and those around them. Currently, healthcare providers identify potential abuse based on a combination of signals including patient history, physical examination findings, risk factors, and other observations. Once abuse is suspected, the step-by-step protocol outlined in the Child Abuse and Domestic Violence Reporting Code is followed.

A key distinction between this type of violence and child abuse is the presence of volitional capacity. In cases of violence among adults, if the adult victim consciously and willingly refuses help and does not consent to further intervention, an assessment must be conducted to evaluate the risk of severe physical or psychological harm or death. Reporting is warranted only if the danger cannot be mitigated by other means.

There is currently no validated screening instrument available for this issue. Such a tool would assist professionals in considering, recognizing and identifying the possibility of abuse in dependent relationships among adults.

In the conclusions, the following cut-off values are used: ≥90% excellent, ≥80% <90% good, ≥70% <80% moderate, <70% poor.

Woman Abuse Screening tool (WAST)

Extended Hurt, Insulted, Threaten, Scream (E-HITS) Tool

Hurt, Insulted, Threaten, Scream (HITS) tool

Partner Violence Screen (PVS) tool

Ongoing Violence Assessment Tool (OVAT)

Slapped, Things, Threaten (STaT) tool

Ongoing Abuse Screen (OAS) tool

Abuse Assessment Screen (AAS) tool

Description of studies

Feltner (2018) performed a systematic review to assess the accuracy of screening questionnaires or tools to identify adults and adolescents with current, past, or increased risk for intimate partner violence (IPV). The review was an update of the 2011 evidence report for the US Preventive Services Task Force, so the literature search was performed from 2011 through October 4, 2017. Fifteen studies were included for the research question on accuracy of screening instruments. Screening instruments that were evaluated were:

• Woman Abuse Screening tool (WAST): 8 items assess physical and emotional IPV, no cutoff provided;
• Extended Hurt, Insulted, Threaten, Scream (E-HITS) Tool: 5 items, including all 4 HITS items and an additional sexual violence item, with a score of ≥7 representing positive screening for IPV;
• Hurt, Insulted, Threaten, Scream (HITS) tool: 4 items assess the frequency of IPV, with a score of ≥10 representing positive screening for IPV;
• Partner Violence Screen (PVS) tool: 3 items that assess physical IPV in the last year and current safety, with cutoff for IPV being affirmative response to ≥1 items (assuming person harming or making the respondent feel unsafe is a current or past partner);
• Ongoing Violence Assessment Tool (OVAT): 4 items assess ongoing physical and emotional IPV, with cutoff for IPV used affirmative response to items 1+H5, 2, or 4 and response of ≥3 for item 3;
• Slapped, Things, Threaten (STaT) tool: 3 items (2 assess physical IPV, 1 assesses threats), with a score of ≥1 representing positive screening for IPV;
• Ongoing Abuse Screen (OAS) tool: 5 items adapted from the AAS that assess ongoing physical, sexual, emotional IPV, and fear, with affirmative response to ≥1 item(s) used as cutoff;
• Abuse Assessment Screen (AAS) tool: 5 items assess physical, emotional, and sexual violence, with affirmative response to ≥1 item(s) used as cutoff.

Included studies used the following reference standards to determine the presence of IPV:

• Composite Abuse Scale (CAS): 43-item measure of partner abuse, including four dimensions: Severe Combined Abuse, Emotional Abuse, Physical Abuse, and Harassment;
• Conflicts Tactics Scale (CTS): 80 items to explore intrafamily conflict and violence;
• Revised Conflicts Tactics Scale (CTS-2): 39-item measure of individual engagement in or experience of physical or psychological violence with an intimate partner, based on the Conflict Tactic Scale;
• Index of Spousal Abuse (ISA): 30 item self-reported scale to measure physical and non-physical spouse abuse. It is used by clinicians to evaluate change in the degree or severity of both physical nonphysical abuse as perceived by female;
• Semi structured interviews.

Two studies were performed in primary care, so these are not reported in this literature analysis. In total, 4460 patients were included, and all studies enrolled adults, mostly women. The prevalence of current or recent IPV ranged from 11% to 29%, with a median of 24%. The quality of all studies was rated as fair.

Results

The studies reporting instruments used in primary care (Chen, 2005; Dubowitz, 2007; McMillan, 2006; Sohal, 2007; Zink, 2007), lifetime abuse (Paranjape, 2003), and predicting future abuse (Koziol-McLain, 2001) are not reported here.

Feltner (2018) reported that for screening instruments on past-year IPV, sensitivity ranged from 65% to 87% and specificity ranged from 80% to 95%. For instruments identifying ongoing or recurrent relationship violence, sensitivity ranged from 46% to 94% and specificity ranged from 38% to 95%. See Figure 1 for an overview. Because TP, TN, FP and FN were not reported by the authors, no other diagnostic accuracy outcomes such as PPV and NPV could be calculated.

Figure 1. Overview of sensitivity and specificity reported in the individual studies included in Feltner (2018)

_{Figure is directly retrieved from Feltner (2018)}

Woman Abuse Screening tool (WAST)

Wathen (2008), included in Feltner (2018), reported a IPV sensitivity of 88% (95%CI 85 to 90) and specificity of 89% (95%CI 88 to 90) for the WAST. The recall period was past year IPV exposure. The prevalence of IPV in the analysed population based on the reference standard (CAS tool) was 14%. The study was performed in primary, acute, and specialty care centres in Canada. Participants were English-speaking and reading woman aged 18-64 years with a male partner in the past year, and 5604 participants were included. All included participants were woman and the mean age in the screen group was 39 years. The quality of the study was rated as fair.

Extended Hurt, Insulted, Threaten, Scream (E-HITS) Tool

Iverson (2015), included in Feltner (2018), reported a IPV sensitivity of 75% (95%CI 55 to 95) and specificity of 82% (95%CI 72 to 90) for the E-HITS tool. The recall period was past year IPV exposure. The prevalence of IPV in the analysed population based on the reference standard (CTS-2 tool) was 25%. The study was performed using mailed surveys in the United States. Participants were female veterans aged ≥18 years found through Veterans Health Administration database who reported an intimate relationship in the past year, and 80 participants were included. All included participants were woman, and the mean age was 49 years. The quality of the study was rated as fair.

Hurt, Insulted, Threaten, Scream (HITS) tool

Iverson (2015), included in Feltner (2018), reported a IPV sensitivity of 75% (95%CI 55 to 95) and specificity of 83% (95%CI 73 to 92) for the HITS tool. The recall period was past year IPV exposure. The prevalence of IPV in the analysed population based on the reference standard (CTS-2 tool) was 25%. The study was performed using mailed surveys in the United States. Participants were female veterans aged ≥18 years found through Veterans Health Administration database who reported an intimate relationship in the past year, and 80 participants were included. All included participants were woman, and the mean age was 49 years. The quality of the study was rated as fair.

Iverson (2013), included in Feltner (2018), reported a IPV sensitivity of 78% (95%CI 63 to 89) and specificity of 80% (95%CI 71 to 87) for the HITS tool. The recall period was past year IPV exposure. The prevalence of IPV in the analysed population based on the reference standard (CTS-2 tool) was 29%. The study was performed using mailed surveys in the United States. Participants were female veterans aged ≥18 years found through Veterans Health Administration database who reported an intimate relationship in the past year, and 160 participants were included. All included participants were woman, and the mean age was 48 years. The quality of the study was rated as fair.

Mills (2006), included in Feltner (2018), reported a IPV sensitivity of 30% (95%CI 13 to 54) and specificity of 88% (95%CI 71 to 96) for psychological aggression and a sensitivity of 46% (95%CI 18 to 75) and specificity of 88% (95%CI 74 to 96) for physical violence for the HITS tool. The recall period was past year IPV exposure. The prevalence of IPV in the analysed population based on the reference standard (CTS-2 tool) was 39% for psychological aggression and 20% for physical violence. The study was performed in Emergency departments in the United States. Participants were men aged ≥18 years triaged to the medical or trauma sections, and 53 participants were included. All included participants were men, and the mean age was 40 (SD = 11) years. The quality of the study was rated as fair.

Partner Violence Screen (PVS) tool

Feldhaus (1997), included in Feltner (2018), reported a IPV sensitivity of 64% (95%CI 51 to 76) and specificity of 80% (95%CI 74 to 86) for the PVS tool with the ISA tool and sensitivity of 71% (95%CI 59 to 82) and specificity of 84% (95%CI 78 to 90) with the CTS tool. The recall period was past year IPV exposure. The prevalence of IPV in the analysed population based on the reference standard ISA was 24% and based on the CTS 27%. The study was performed in emergency departments in the Unites States. Participants were English-speaking woman aged ≥18 years who were noncritical, and 255 participants were included in the ISA group and 230 in the CTS group. All included participants were woman, and the mean age was 36 (SD = 16) years. The quality of the study was rated as fair.

Mills (2006), included in Feltner (2018), reported a IPV sensitivity of 35% (95%CI 16 to 59) and specificity of 84% (95%CI 67 to 94) for psychological aggression and a sensitivity of 46% (95%CI 18 to 75) and specificity of 83% (95%CI 68 to 92) for physical violence for the PVS tool. The recall period was past year IPV exposure. The prevalence of IPV in the analysed population based on the reference standard (CTS-2 tool) was 39% for psychological aggression and 20% for physical violence. The study was performed in Emergency departments in the United States. Participants were men aged ≥18 years triaged to the medical or trauma sections, and 53 participants were included. All included participants were men, and the mean age was 40 (SD = 11) years. The quality of the study was rated as fair.

Ongoing Violence Assessment Tool (OVAT)

Ernst (2004), included in Feltner (2018), reported a IPV sensitivity of 86% (95%CI 75 to 93) and specificity of 83% (95%CI 78 to 88) for the OVAT. The recall period was current IPV exposure. The prevalence of IPV in the analysed population based on the reference standard (ISA tool) was 20%. The study was performed in emergency departments in the Unites States. Participants were English-speaking patients, and 306 participants were included. In total, 70% was woman, and the mean age was 34 (SD = 10) years. The quality of the study was rated as fair.

Slapped, Things, Threaten (STaT) tool

Paranjape (2006), included in Feltner (2018), reported a IPV sensitivity of 95% (95%CI 90 to 100) and specificity of 37% (95%CI 29 to 44) for the STaT tool. The recall period was current IPV exposure or most recent relationship. The prevalence of IPV in the analysed population based on the reference standard (ISA tool) was 33% during most recent relationship and 15% currently. The study was performed in urgent care in the Unites States. Participants were English-speaking women aged 18-65 years, and 240 participants were included. All included participants were woman, and the mean age was 38 (SD = 10) years. The quality of the study was rated as fair.

Ongoing Abuse Screen (OAS) tool

Weiss (2003), included in Feltner (2018), reported a IPV sensitivity of 60% (95%CI 52 to 67) and specificity of 90% (95%CI 87 to 92) for the OAS tool. The recall period was current IPV exposure. The prevalence of IPV in the analysed population based on the reference standard (ISA tool) was 19%. The study was performed in emergency departments in the Unites States. Participants were patients with a current partner who were not too ill to participate, and 856 participants were included. In total, 62% was woman, and the mean age was 36 years. The quality of the study was rated as fair.

Abuse Assessment Screen (AAS) tool

Weiss (2003), included in Feltner (2018), reported a IPV sensitivity of 92% (95%CI 87 to 96) and specificity of 55% (95%CI 52 to 59) for the AAS tool. The recall period was current IPV exposure. The prevalence of IPV in the analysed population based on the reference standard (ISA tool) was 19%. The study was performed in emergency departments in the Unites States. Participants were patients with a current partner who were not too ill to participate, and 856 participants were included. In total, 62% was woman, and the mean age was 36 years. The quality of the study was rated as fair.

Level of evidence of the literature

Woman Abuse Screening tool (WAST)

The level of evidence regarding the outcome measure sensitivity was downgraded by one level because of study limitations of the individual study (risk of bias) to MODERATE.

The level of evidence regarding the outcome measure specificity was downgraded by one level because of study limitations of the individual study (risk of bias) to MODERATE.

Extended Hurt, Insulted, Threaten, Scream (E-HITS) Tool

The level of evidence regarding the outcome measure sensitivity was downgraded by two levels because of study limitations of the individual study (risk of bias) and low number of included participants (imprecision) to LOW.

The level of evidence regarding the outcome measure specificity was downgraded by two levels because of study limitations of the individual study (risk of bias) and low number of included participants (imprecision) to LOW.

Hurt, Insulted, Threaten, Scream (HITS) tool

The level of evidence regarding the outcome measure sensitivity was downgraded by two levels because of study limitations of the individual studies (risk of bias) and low number of included participants (imprecision) to LOW.

The level of evidence regarding the outcome measure specificity was downgraded by two levels because of study limitations of the individual studies (risk of bias) and low number of included participants (imprecision) to LOW.

Partner Violence Screen (PVS) tool

The level of evidence regarding the outcome measure sensitivity was downgraded by two levels because of study limitations of the individual studies (risk of bias) and low number of included participants (imprecision) to LOW.

The level of evidence regarding the outcome measure specificity was downgraded by two levels because of study limitations of the individual studies (risk of bias) and low number of included participants (imprecision) to LOW.

Ongoing Violence Assessment Tool (OVAT)

The level of evidence regarding the outcome measure sensitivity was downgraded by two levels because of study limitations of the individual study (risk of bias) and low number of included participants (imprecision) to LOW.

The level of evidence regarding the outcome measure specificity was downgraded by two levels because of study limitations of the individual study (risk of bias) and low number of included participants (imprecision) to LOW.

Slapped, Things, Threaten (STaT) tool

The level of evidence regarding the outcome measure sensitivity was downgraded by two levels because of study limitations of the individual study (risk of bias) and low number of included participants (imprecision) to LOW.

The level of evidence regarding the outcome measure specificity was downgraded by two levels because of study limitations of the individual study (risk of bias) and low number of included participants (imprecision) to LOW.

Ongoing Abuse Screen (OAS) tool

The level of evidence regarding the outcome measure sensitivity was downgraded by one level because of study limitations of the individual study (risk of bias) to MODERATE.

The level of evidence regarding the outcome measure specificity was downgraded by one level because of study limitations of the individual study (risk of bias) to MODERATE.

Abuse Assessment Screen (AAS) tool

The level of evidence regarding the outcome measure sensitivity was downgraded by one level because of study limitations of the individual study (risk of bias) to MODERATE.

The level of evidence regarding the outcome measure specificity was downgraded by one level because of study limitations of the individual study (risk of bias) to MODERATE.

A systematic review of the literature was performed to answer the following question:

Which validated screening instrument can be used (by all caregivers in the emergency department) for signalling and recognizing domestic violence in adults in secondary care?

Relevant outcome measures

The guideline development group considered sensitivity as a critical outcome measure for decision making; and specificity as an important outcome measure for decision making.

A priori, the working group did not define the outcome measures listed above but used the definitions used in the studies.

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms until 7 December 2023. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 2082 hits. Due to the high amount of hits, we decided to only select systematic reviews, which were 145 hits. Studies were selected based on the following criteria: systematic reviews about violence, adults, and screening instruments. Fourteen studies were initially selected based on title and abstract screening. After reading the full text, 13 studies were excluded (see the table with reasons for exclusion under the tab Methods), and one study was included.

Results

One study was included in the analysis of the literature. Important study characteristics and results are summarized in the evidence table. The assessment of the risk of bias is summarized in the risk of bias table.