Article type : Research Methodology: Instrument Development

Accepted Article
Validation and Adaptation of the Danger Assessment-5 (DA-5): A Brief Intimate Partner
Violence Risk Assessment

RUNNING HEAD: Validation & Adaptation of the DA-5

Jill Theresa MESSING, PhD, MSW

Associate Professor
School of Social Work
Arizona State University
411 N Central Ave., Suite 800
Phoenix, AZ 85004
Jill.Messing@asu.edu
Phone: 602-496-1193
Fax: 602-496-0960

Jacquelyn C. CAMPBELL, PhD, RN, FAAN

Professor & Anna D. Wolfe Chair
School of Nursing
Johns Hopkins University

Carolyn SNIDER, MD, MPH

Associate Professor
Department of Emergency Medicine,
Max Rady College of Medicine
University of Manitoba

Conflict of interest:

No conflict of interest has been declared by the authors.

Funding statement:

This research was funded by the National Institute of Justice (WG-BX-2008-002).

This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1111/jan.13459
This article is protected by copyright. All rights reserved.
 ABSRACT
Accepted Article
Aims: To assess the predictive validity of the DA-5 with the addition of a strangulation item

in evaluating the risk of an intimate partner violence (IPV) victim being nearly killed by an

intimate partner.

Background: The DA-5 was developed as a short form of the Danger Assessment for use in

healthcare settings, including emergency and urgent care settings. Analyzing data from a

sample of IPV survivors who had called the police for domestic violence, the DA-5 was

tested with and without an item on strangulation, a potentially fatal and medically damaging

IPV tactic used commonly by dangerous abusers.

Design: A heterogeneous sample of 1081 women recruited by police between 2009-2013 at

the scene of a domestic violence call and interviewed by researchers at baseline; 619 (57.3%)

were contacted and re-interviewed after an average of 7 months.

Methods: The predictive validity of the DA-5 was assessed for the outcome of severe or near

lethal IPV re-assault using sensitivity, specificity and ROC curve analysis techniques.

Results: The original DA-5 was found to be accurate (AUC=.68), equally accurate with the

strangulation item from the original DA substituted (AUC=.68) and slightly more accurate

(but not a statistically significant difference) if multiple strangulation is assessed.

Conclusion: We recommend that the DA-5 with the strangulation item be used for a quick

assessment of homicide or near homicide risk among IPV survivors. A protocol for

immediate referral and examination for further injury from a strangulation should be adopted

for IPV survivors at high risk.

Keywords: Emergency; healthcare settings; healthcare professionals; nursing; strangulation;

domestic violence; intimate partner homicide; femicide; partner abuse

This article is protected by copyright. All rights reserved.

 SUMMARY STATEMENT
Accepted Article
Why is this research or review needed?

• This study assesses the predictive validity of the DA-5, a shortened version of the

Danger Assessment, an intimate partner violence risk assessment, for use in

healthcare settings.

• The DA-5 was adapted to incorporate assessment for strangulation, a potentially fatal

and medically damaging form of intimate partner violence that survivors may not

disclose unless it is specifically assessed.

What are the key findings?

• The original DA-5 can predict severe or near lethal repeat violence with a medium

effect (AUC=.68).

• Substituting an item on strangulation or multiple strangulation for IPV during

pregnancy on the DA-5 resulted in predictive ability that is the same as the original

DA-5.

• Incorporating attempted strangulation enhances the sensitivity of the DA-5 and

provides healthcare practitioners an opportunity to screen for strangulation among

intimate partner violence survivors.

How should the findings be used to influence policy/practice/research/education?

• The DA-5 and accompanying protocol helps practitioners triage survivors of intimate

partner violence while providing appropriate and timely medical care for their

injuries.

• Using a version of the DA-5 that includes a strangulation item provides a means for

strangulation screening, informing health care practitioners of the need to assess for

further medical trauma due to strangulation.

This article is protected by copyright. All rights reserved.

 • All survivors of IPV who disclose victimization in healthcare settings must be
Accepted Article provided resources and referred to knowledgeable practitioners.

INTRODUCTION

Intimate partner violence (IPV) and intimate partner femicide (the killing of women) are

global threats to the health and safety of women. In this manuscript, we primarily use the

terms “intimate partner violence” and “intimate partner homicide” when referring to violence

or homicide committed by one partner within an intimate relationship (e.g., romantic, sexual,

dating, spousal, ex-partner, or child-in-common). Domestic violence, a term that

encompasses violence between intimate partners and those living in the same household who

are not in a past or current intimate relationship, is used when referring to police reporting of

intimate partner violence or services for intimate partner violence survivors, as this is

common in the criminal justice and social service practice communities. It is estimated that

worldwide, 30.0% of women (15 years of age and over) are victimized by an intimate partner

in their lifetimes (Devries et al., 2013). Rates of IPV are slightly higher in the USA, with

estimates suggesting that 35.6% of women experience IPV and 24.3% of women experience

severe IPV in their lifetimes (Black et al., 2011). Also in the USA, physical IPV results in 1.8

million injuries per year, with 28.8% of those requiring medical attention (NCIPC, 2003).

Multiple international studies have firmly established that serious and long term physical and

mental health sequelae are associated with IPV; repeated injuries and stress responses

combine to affect multiple systems negatively, including gastrointestinal, cardiac,

reproductive and neurological (Devries et al., 2013; Ellsberg et al., 2008; Kwako et al., 2011;

Ruiz-Perez et al., 2007; Stockman et al., 2015; Tadegge, 2008). These problems are

exacerbated for women who are victimized by severe IPV (Campbell et al., 2003).

This article is protected by copyright. All rights reserved.

 Worldwide, murder by an intimate partner is the largest homicide risk for women; a
Accepted Article conservative global estimate is that 38.6% of women killed are murdered by an intimate

partner (Stöckl et al., 2013). In the USA, when the relationship between the victim and

offender is known, the proportion of femicide victims killed by a current or former intimate

partner ranges up to 50% (Campbell et al., 2007; Catalano et al., 2009; Rennison &

Welchans, 2000; Puzone et al., 2000; VPC, 2014). In 66.9-80.8% of cases, physical IPV is a

precursor to intimate partner femicide (Campbell et al., 2003; Campbell et al., 2007;

Moracco, Runyan & Butts, 1998), suggesting that appropriate IPV interventions informed by

risk assessment may prevent femicide (Stöckl et al., 2013).

Given the risk of injury and negative health consequences of IPV, nurses and other

practitioners (physicians, social workers) encounter IPV survivors across settings. IPV risk

assessment instruments provide practitioners with information about the risk of re-assault or

homicide faced by survivors of IPV. The Danger Assessment (DA) is a 20-item risk

assessment (Campbell et al., 2003) that is unique because it is the sole IPV risk assessment to

ask questions only of the survivor of IPV, intended to predict lethality and envisioned as a

collaborative effort between a survivor of violence and a practitioner with the goal of

promoting safety behaviors (Messing & Thaller, 2015).

In 2009, a shortened version of the DA, the DA-5, was developed to assist

professionals in assessing femicide risk among IPV survivors in healthcare settings such as

Emergency Departments (Snider et al., 2009). In this original research, the predictive validity

of the DA-5 as measured by the Receiver Operating Characteristic (ROC) Area Under the

Curve (AUC) was .79 (CI=.73-.85; Snider et al., 2009), corresponding to a large effect size

(Rice & Harris, 2005. To understand the reliability of a risk assessment, these instruments are

commonly tested multiple times among different samples (Messing & Thaller, 2013). Thus,

the first aim of this paper is to extend the research on the creation of the DA-5 by examining

This article is protected by copyright. All rights reserved.

 the predictive validity of the instrument among a sample of women interviewed after being
Accepted Article visited by the police for an incident of IPV. We will also examine the impact on predictive

validity of incorporating strangulation as a risk factor on the DA-5. As is described in the

following section, strangulation is a particularly lethal form of IPV; it is hypothesized that

including strangulation as a risk factor on the DA-5 may enhance the predictive validity of

the instrument.

Strangulation

Strangulation occurs when an abuser places external pressure on a victim’s neck

leading to reduced blood flow to the brain and/or airway closure (Sauvageau & Boghossian,

2010). In cases of IPV, manual strangulation (pressure to the neck that is applied by hands,

forearms or other limbs) is the most common form of strangulation (Wilbur et al., 2001;

McClane, Strack & Hawley, 2001; Shields et al., 2010). Strangulation is often described by

women as being “choked” (Thomas, Joshi & Sorenson 2013; Joshi, Thomas & Sorenson

2012); as such, “choking” is the term often used in assessment instruments, including the DA,

to refer to strangulation.

Limited previous research has reported that between 9.6-74.3% of IPV survivors have

been strangled (Dobash et al., 2007; Glass et al., 2008; Mcquown et al., 2016; Messing et al.,

under review; Wilbur et al., 2001). Strangulation is associated with sexual assault, death

threats, stalking, violence or threats with a weapon, being physically assaulted (Thomas,

Joshi & Sorenson 2013; Messing et al., under review) and femicide (Campbell et al., 2003;

Dobash et al., 2007). In cases where men killed or attempted to kill their partners,

strangulation was 6.7-7.5 times more likely than among men who abused their partners but

did not inflict fatal or near-fatal violence (Glass et al., 2008).

This article is protected by copyright. All rights reserved.

 Although there is limited research, it appears that strangulation as a tactic of IPV is
Accepted Article often repetitive (Messing et al., under review; Vella, 2013; Wilbur et al., 2001). Repeated

strangulation results in a statistically significant increase in physical (e.g., memory problems)

and mental health symptoms (e.g., nightmares; Smith, Mills & Taliaferro, 2001). In a study

conducted by Messing and colleagues (under review), women who had been strangled

multiple times by an intimate partner were more likely than women who had not been

strangled to report additional risk factors for femicide including an increase in the frequency

and/or severity of IPV, a belief that her partner can kill her, partner’s use of a weapon and

sexual assault.

Rates of medical care after an incident of strangulation range widely across studies,

perhaps dependent on the timeframe examined, recruitment methods and populations

investigated. Shortly after a strangulation incident, rates of medical care were found to be as

low as 5% (Strack, McClane & Hawley, 2001); perhaps because women do not know that

medical treatment is indicated and because injuries are usually not visible. Rates of medical

care are higher among women with high levels of symptom development (28.6%; Wilbur et

al., 2001), among women strangled more than 5 times (39.1%; Smith, Mills & Taliaferro

2001) and vastly higher among strangled women meeting the criteria for attempted femicide

and referred by trained law enforcement for forensic evaluation (69%; Shields et al., 2010).

Correct diagnosis and treatment is imperative to mitigate the negative physical health effects

of strangulation, including stroke and long term neurological problems. For this same reason,

it is important that women seeking treatment for IPV are routinely assessed for strangulation

(Smith, Mills & Taliaferro, 2001; Joshi, Thomas & Sorenson, 2012).

This article is protected by copyright. All rights reserved.

 The Danger Assessment
Accepted Article The DA (www.dangerassessment.org) is an IPV risk assessment that is intended to

predict femicide among IPV survivors (Campbell et al., 2003; Campbell, 1986). Survivors of

IPV are asked to recall incidents of past IPV with the help of a 12-month calendar and

complete 20 yes/no questions about risk factors present in the relationship. Items on the DA

are then weighted and scored (range=0-37), placing women in one of four danger levels:

variable danger (<8), increased danger (8-13), severe danger (14-17) and extreme danger

(>18; Campbell, Webster & Glass, 2009). The DA is intended to be used by survivors of IPV

in collaboration with a healthcare professional or advocate; after scoring the DA, a trained

advocate or healthcare professional should work with a survivor to safety plan around issues

of risk and encourage the use of community resources as appropriate.

The DA was originally created in 1986 and revised in 2003 based on evidence from a

case-control study that examined differences between women who were killed (n=220) or

nearly killed (n=143) by an intimate partner and 356 IPV survivors (Campbell et al., 2003).

Hypothesized risk factors were tested for their ability to predict femicide compared with IPV.

The DA has an AUC of .92 when examining attempted femicides (Campbell, Webster &

Glass, 2009). The AUC can be interpreted as the probability that any randomly selected case

would have a higher score on the instrument than any randomly selected non-case (Rice &

Harris, 1995; Rice & Harris, 2005). Thus, the AUC can be interpreted as a 92% chance that a

randomly selected victim of attempted femicide would have a higher score on the DA than a

randomly selected victim of assault (Campbell, Webster & Glass, 2009). A meta-analysis

examining the ability of IPV risk assessment instruments to predict re-assault found that the

weighted average AUC of the DA was .62, which is similar to other IPV risk assessment

tools (Messing & Thaller, 2013).

This article is protected by copyright. All rights reserved.

 Aims
Accepted Article The DA-5 is intended to be a brief screening tool that can be used in a fast-paced

health care environment to provide information about which survivors of IPV are at risk for

future severe and near fatal violence to refer them for further assessment and services (Snider

et al., 2009). It is important to understand the predictive validity of the instrument across

settings and samples to understand whether use of the instrument will, in fact, correctly refer

those at high risk of future severe and near lethal violence and correctly fail to refer those not

at risk. Thus, this research aims to, first, examine the predictive validity of the DA-5 in a

different sample from that on which is was developed (Snider et al., 2009). Second, given the

information presented above on the prevalence, dangerousness and lack of consistent and

appropriate medical treatment for strangulation, we will examine the impact of replacing the

item assessing abuse during pregnancy with an item assessing attempted strangulation and/or

multiple strangulation on the predictive validity of the DA-5. Substituting this item has the

potential to create a more broadly applicable form of the DA-5 that is also an evidence-

informed assessment for strangulation, an injurious and deadly form of IPV.

METHODOLOGY

This is a secondary analysis of data collected between 2009 - 2013 as part of the

National Institute of Justice Oklahoma Lethality Assessment (OK-LA) Study (2008-WG-BX-

0002). The original study was a quasi-experimental field trial that examined the effectiveness

of the Lethality Assessment Program (LAP), a risk-informed collaborative intervention

developed by the Maryland Network Against Domestic Violence (Messing & Campbell

2016; Messing et al., 2015a). Data for the OK-LA study were collected via structured

telephone interviews from female survivors of IPV who were referred to the study by police

officers called to the scene of domestic violence incidents in 7 participating police

This article is protected by copyright. All rights reserved.

 jurisdictions in Oklahoma. Women who agreed to be contacted by the research team were
Accepted Article subsequently assessed for eligibility by researchers, provided informed consent and

interviewed at two time points approximately 7 months apart (for additional information, see

Messing, Campbell & Wilson, 2015; Messing et al., 2015a)

Participants

Women were referred to researchers by police officers if they were the survivor of

IPV and this was either a repeat call for service or the officer believed that the survivor was

in danger. If a woman was willing to be contacted by researchers after a standardized referral

script was read to her, the police officer recorded 1-2 safe telephone numbers and a safe time

for researchers to call. Over the course of 3.6 years, 3,159 women agreed to be contacted by

researchers. Of these referrals, researchers were unable to contact 1,527 women (48.3%) due

to unanswered, disconnected or wrong numbers. An additional 90 women (2.8%) were

ineligible (e.g., not an IPV survivor, underage). Of the 1,542 eligible women contacted by

researchers, 445 (28.9%) declined to participate. The remaining 1,097 women completed a

baseline interview; subsequently, sixteen duplicate interviews were removed from the

sample. Of the 1,081 women interviewed at baseline, 619 (57.3%) could be located and

interviewed approximately 7 months later. Women who completed the follow-up interview

were 2 years older on average (t=-3.3 [df=1064], p<0.001), more likely to be employed

(X2=6.8, p<0.01) and more likely to have a high school or higher education (X2=15.0,

p<0.001) than women who did not complete the follow-up interview. Measures of violence

and risk at baseline were the same for women who completed the follow-up interview and

those who did not; there were no significant differences in the mean weighted DA score

(Campbell et al., 2003) or the severity-times-frequency weighted score of the revised Conflict

Tactics Scale (CTS-2; Straus, 2004).

This article is protected by copyright. All rights reserved.

 Data collection ceased when researchers had sufficient power conduct the main study
Accepted Article analysis (see Messing et al., 2015a). Sample size estimates were calculated by iterative

approximation (Hanley & McNeil, 1982) and indicate that 619 participants provides

sufficient power to detect an area under the curve of .60 (standard error=0.02), which

corresponds to a small effect (Rice & Harris, 2005).

Measures

Women participated in two structured telephone interviews, each lasting 45 minutes

on average. Questions that were included were part of standardized instruments, such as the

CTS-2 (Straus et al., 1996) or adapted from the Risk Assessment Validation Study (RAVE;

Campbell, O’Sullivan, Roehl & Webster, 2005). The CTS-2 has an internal consistency

reliability of .79-.95 and test-retest reliability of .70-.76; concurrent and discriminant validity

have been found to be acceptable (Straus, Hamby & Warren, 2003).

Demographic and Relationship Characteristics (Baseline). At the baseline interview,

participants were asked to report their educational achievement, employment status,

racial/ethnic background and age in years. Participants reported their legal marital status as

well as their relationship status and current level of involvement with their abusive partner.

Risk Factors (Baseline). Most risk factors were taken from the DA (Campbell et al.,

2003), described in more detail above. For this analysis, we were particularly interested in the

original DA-5 items (Snider et al., 2009): (1) an increase in the frequency and severity of

IPV; (2) sexual jealousy; (3) survivor’s belief that her partner could kill her; (4) IPV during

pregnancy; and (5) use of or threats with a weapon. We were also interested in the DA item

assessing attempted/actual strangulation (“Does he ever try to choke you?”). Because of

previous literature indicating that multiple strangulation (compared with a single incident)

leads to more physical and mental health symptoms and is associated with risk factors

This article is protected by copyright. All rights reserved.

 predictive of femicide (Messing et al., under review; Smith, Mills & Taliaferro, 2001), one
Accepted Article item from the CTS-2 (Straus et al., 1996) was included assessing multiple strangulation

(“[your partner] choked you… 2 or more times”).

Near Lethal Violence (Follow-up). Following the example of the original DA-5

research, the outcome measured at follow-up was whether, since the baseline interview, the

participant’s abusive partner or ex-partner had inflicted severe injury or a potentially fatal

assault (see Table 1).

Ethical considerations

The original research was approved by the institutional review boards at Johns

Hopkins University, Arizona State University, the University of Oklahoma Health Sciences

Center, the Oklahoma State Department of Health, the Cherokee Nation and the National

Institute of Justice. In addition, a privacy certificate was obtained from the National Institute

of Justice. If a woman agreed to participate in the research after an interviewer explained the

purpose of the study and what participation entailed, informed consent was obtained verbally

and recorded in writing by the interviewer; a written copy of the informed consent was

available for participants who preferred. The interview was conducted at a time safe and

convenient for the participant. The interviewer’s first priority during all contacts with

participants was to ensure participant safety. Interviewers were trained using telephone safety

precautions first developed for the Canadian Domestic Violence survey and used in the

RAVE study; these precautions are described elsewhere (Wilson et al., 2011). The IRB of

Arizona State University approved this secondary analysis of data.

This article is protected by copyright. All rights reserved.

 Data analysis
Accepted Article Data were cleaned in SPSS and transferred to Stata for analysis. Missing data on

demographic and relationship characteristics are reported in Table 2. When data were missing

on risk factors, the participant was coded as not experiencing that risk factor. This decision

was made to be consistent with the information that would be available to a practitioner

asking questions about risk in a practice setting. For example, if a woman refused to answer

one or more questions included on a risk assessment, the practitioner can only use the

information available to her/him (i.e., the number of “yes” response) to score the risk

assessment; this is a more conservative measure of risk for femicide. Participants were

included as being victimized by severe or near fatal violence on follow-up if they responded

affirmatively to one or more items in the outcome measure; no participants refused to answer

all outcome measures. Nevertheless, consistent with the more conservative approach

identified above, if a participant did not indicate that they had experienced at least one form

of violence included in the outcome measure, she was not coded as having been victimized

by severe or near fatal violence.

Participant and relationship characteristics, risk factors (defined above in the

“Measures” section) and the outcome (or dependent variable) of severe injury or near lethal

violence are described using univariate statistics. Unadjusted odds ratios are used to examine

the association between the measured risk factors and the dependent variable. Predictive

validity, or the correct prediction of future events, is the best measure of the accuracy of a

risk assessment instrument. This includes the correct classification of cases (sensitivity) and

the correct classification of non-cases (specificity); a risk assessment generally aims to

maximize both sensitivity and specificity (Douglas et al., 2005). The ROC AUC has been

suggested as a standard measure across disciplines for analyses of the predictive validity of

risk assessment instruments (Rice & Harris, 2005). The ROC is a curve (shown on a graph)

This article is protected by copyright. All rights reserved.

 that takes into account both the sensitivity and the specificity of the instrument (Rice &
Accepted Article Harris, 1995). The AUC is the proportion of the graph that lies under the plotted ROC curve

and provides the overall predictive accuracy of the instrument with .50 indicating that the

instrument predicts no better than chance and 1.0 indicating that every case was predicted

accurately (Rice & Harris, 1995; Douglas et al., 2005). No assumptions of the statistical tests

were violated in the analysis of data.

RESULTS

As shown in Table 2, at baseline, participants in this study were 33 years old on

average (Range=18-79; SD 10.3). The largest racial/ethnic group was White (41.8%),

followed by African American (29.4%) and small (<10.0% each group) proportions of

Latina, American Indian and Multiracial survivors. Most women in this sample attended

some college (55.4%) and nearly half (44.3%) were working full- or part-time at the time of

the baseline interview. Most women reported their marital status as single (58.6%), their

partner as a former partner (49.8%) and the relationship as having ended (69.3%).

Risk factors for femicide at baseline were reported by 21.0% to 71.8% of women in

the sample, depending on the risk factor assessed (see Table 3). The least commonly reported

risk factor was IPV during pregnancy and the most common risk factor was attempted

strangulation. At the follow-up interview, 13.1% of women reported that their partner had

severely injured or nearly killed them (see Table 1). Each of the risk factors examined is

significantly, independently related to the outcome of severe injury and/or near lethal

violence with unadjusted odds ratios between 1.9-3.4.

This article is protected by copyright. All rights reserved.

 Predictive Validity of the DA-5
Accepted Article In this sample, the AUC for the DA-5 (with the item assessing IPV during pregnancy)

is .68 (95% CI=.62-.74) for prediction of severe/near lethal repeat violence. This corresponds

to a medium effect (Rice & Harris, 2005). This AUC is lower than reported by Snider and

colleagues (2009; AUC=.79), but it is typical to find higher predictive validity in the sample

used to create the instrument than in subsequent tests. Snider and colleagues suggested that

affirmative responses to 3 or more risk items result in a classification of high risk; 46.2% of

women in this sample responded affirmatively to three or more risk factors. Of these women,

nearly 60% were victimized by severe or near lethal violence on follow up (i.e., correct

classification). The sensitivity of the DA-5 at this cut-off is 71.6% and the specificity is

57.6% (see Table 4). Increasing the cut-off to 4 reduces by half the number screened in and

improves the percent correctly classified (75.3%), but results in a reduction of sensitivity

(44.4%) and a two-fold increase in false negatives.

Strangulation: Adapting the DA-5

Several adaptations of the DA-5 were tested. Given the potential for injury and long

term negative health consequences of strangulation and the lack of disclosure of this form of

IPV to healthcare professionals unless women are specifically asked about it, versions of the

DA-5 substituting attempted strangulation and multiple strangulation items for IPV during

pregnancy were tested. Neither modification resulted in a significantly different AUC from

the original DA-5. With the attempted strangulation item replacing the IPV during pregnancy

item, the AUC was .68 (95% CI=.62-.74). Although the AUC is the same, because a large

proportion of this sample reported attempted strangulation (71.8%), at the suggested cut-off

of 3, this version of the DA-5 screens in a higher proportion of IPV survivors (60.3%), has

higher sensitivity (81.5%) and lower specificity (42.9%) and percent correctly classified

This article is protected by copyright. All rights reserved.

 (48.0%; Table 4). Replacing the IPV during pregnancy item with the multiple strangulation
Accepted Article item resulted in an AUC of .69 (95% CI=.63-.75). At the cut-off of 3, the proportion screened

in (50.7%) is like the DA-5 with the IPV during pregnancy item. As shown in Table 4, the

sensitivity (74.1%) is slightly higher and the specificity (52.8%) and percent correctly

classified (55.6%) are slightly lower. Finally, we added the multiple strangulation item to the

DA-5 with the IPV during pregnancy item included (resulting in a total of 6 questions). The

AUC (.70, 95% CI was .64-.76) was significantly higher than the original DA-5 AUC, though

it remains a medium effect. A cut-off of 3, keeps the instrument most similar to the DA-5

with only the IPV during pregnancy item; 54.4% of IPV survivors screened in, the sensitivity

was 79.0%, the specificity was 49.3% and 53.2% of the participants were correctly classified.

Increasing the cut-off to 4, however, reduces the proportion of those screened in to 34.3%.

Although some sensitivity is lost (60.5%), specificity (69.7%) and percent correctly classified

(68.5%) remain high.

DISCUSSION

The DA-5 was originally created for faced paced health care settings (e.g., Emergency

Departments) to help health care practitioners quickly assess for femicide risk after a positive

IPV screening (Snider et al., 2009). This study validated the DA-5 in a large sample of IPV

survivors who were referred to researchers by police officers after being the victims of IPV,

further substantiating the predictive validity of the instrument. Because almost half (47%) of

women who were killed by partners in a national study were seen in health care settings

before they were killed compared with 4% who were seen by domestic violence service

organizations (Sharps et al., 2001), it is imperative that health care practitioners better assess

for IPV and recognize signs of lethality risk among IPV survivors. The DA-5 and the

protocol that accompanies it (see Figure 1) helps practitioners triage survivors of IPV to focus

This article is protected by copyright. All rights reserved.

 resources, such as immediate IPV safety services and help from the police, on those at high
Accepted Article risk for femicide.

Using a version of the DA-5 that includes an item on strangulation provides a means

for strangulation screening. Strangulation is a common, important and often overlooked

health risk, as well as a femicide risk factor, for IPV survivors (Messing et al., under review;

Sorenson, Joshi & Sivitz, 2014). Substituting the risk factor of attempted, actual or multiple

strangulation for the item specific to IPV during pregnancy retains the predictive power of the

DA-5 while, at the same time, informing health care practitioners of the need to assess for

further medical trauma due to strangulation. It has been recommended that alternative light

source (Holbrook & Jackson, 2013) and imaging (Christe et al., 2009; McClane, Strack &

Hawley, 2001; Yen et al., 2005) techniques be used to detect vascular, neurological and soft

tissue damage, as well as delayed responses to strangulation (such as blood/brain barrier

damage). If a survivor has been strangled, a protocol for further assessment and treatment

will ideally accompany the DA-5 (Smock & Sturgeon, 2016).

Substituting the attempted strangulation item in the DA-5 resulted an AUC that is not

statistically different than the DA-5 with the question on IPV during pregnancy while casting

a wider net by improving the sensitivity of the instrument. This maximizes sensitivity and

minimizes the number of false negatives as suggested for clinical prediction (Laupacis &

Sekar, 1997). Similarly, substituting the item on multiple strangulation (Figure 1, Question

4c: “Has your partner or ex ever tried to choke (strangle) you? Did it happen more than

once?”) in the DA-5 did not result in a significant difference in the AUC. At a cut-off of 3,

the DA-5 with an item on multiple strangulation (substituted for the item on IPV during

pregnancy) screens in 50.7% of participants, compared with 60.3% of participants screened

in using the item on attempted strangulation. Substituting the item on multiple strangulation

results in a false positive/false negative ratio of 12.1 at a cut-off of 3 (see Table 4), which is

This article is protected by copyright. All rights reserved.

 closer to the recommended ratio of 10.0 provided as guidance in criminal justice research
Accepted Article (Berk, Kriegler & Baek, 2006). As indicated in Figure 1, practitioners are able to modify the

suggested protocol to fit their practice settings and needs; this study provides the data needed

to inform these adaptations.

An added benefit of substituting the item assessing strangulation for the item

examining IPV during pregnancy is that it further includes women who have not been

pregnant while in a relationship with their abusive partner. In settings specific to maternity

(e.g., prenatal care, prenatal home visitation, labor and delivery), a practitioner may want to

continue using the IPV during pregnancy item on the DA-5 and add an item to assess for

strangulation (totaling 6 items).

The protocol used with a risk assessment instrument is as important as the predictive

validity of the instrument itself. It is recommended that the DA-5 be used in conjunction with

routine screening for IPV in all parts of the health care system; the DA-5 should be used as a

follow-up assessment for potential lethality with all women who disclose current (or past

year) IPV. In Figure 1, we have provided guidelines for levels of action depending on the

DA-5 score. In brief, using the DA-5 with the IPV during pregnancy, strangulation and/or

multiple strangulation item(s), we suggest that a score of 4 or 5 is treated as the highest

danger for lethality. At this level of danger, depending on survivor choice, the practitioner

should report the IPV to law enforcement, find the survivor space at a local shelter, or work

with the survivor to place a call to a confidential domestic violence hotline. A survivor with a

score of 3 should be referred to a social worker or trained professional in victim services for

further assessment with the full 20-item DA. A score of 2 should result in survivor education

of risk and encouragement to seek services. Finally, if 0-1 risk factors are present, the

setting’s standard procedures for a positive IPV screen should ensue. In addition, a positive

response to either of the strangulation items would ideally trigger a strangulation assessment

This article is protected by copyright. All rights reserved.

 and radiographic evaluation as needed; additional information
Accepted Article (https://www.strangulationtraininginstitute.com/resources/library/) and training

(https://www.strangulationtraininginstitute.com/training/) is available through the Training

Institute on Strangulation Prevention.

Limitations

Women who participated in this study were referred by police officers at the scene of

domestic violence incidents in seven jurisdictions in the Southwest USA and, therefore, may

not be representative of other locations or of women seeking medical care for IPV injury.

Women in this study were English speaking only; therefore, caution needs to be taken in

generalizing the findings beyond women who speak English (Messing et al., 2013). Further

research should entail translating and validating the tool among immigrant and refugee

populations, particularly those that do not speak English. Neither the 20-item DA nor the DA-

5 has been validated on male survivors of IPV even though most health care settings

universally screen adults for IPV. There is very little information in the literature on risk

factors for men being killed by their intimate partners (male or female). Two risk assessments

for IPV re-assault (the Domestic Violence Screening Instrument-Revised and the Ontario

Domestic Assault Risk Assessment) have been found to have similar predictive validity with

male and female offenders (Hilton & Harris, 2017). However, more research on risk,

especially risk of homicide with male survivors of IPV whether by male or female partners, is

needed.

This article is protected by copyright. All rights reserved.

 CONCLUSION
Accepted Article We recommend that the DA-5 be used as a brief assessment of IPV survivors’ risk of

femicide, near femicide, or serious injury. This research provides further evidence for the

predictive validity of the DA-5 as originally created (with an item on IPV during pregnancy;

Snider et al., 2009) and suggests that revisions of the DA-5 that incorporate or substitute

items on attempted strangulation or multiple strangulation result in similar predictive validity.

Versions of the DA-5 that incorporate strangulation can be further used to assess for

strangulation and injury due to strangulation, to guide appropriate medical care for this

specific type of IPV. Screening for IPV and femicide risk with a brief tool such as the DA-5

presents an opportunity to provide vulnerable survivors of IPV with intervention, resources

and referral to knowledgeable practitioners.

Author Contributions:

All authors have agreed on the final version and meet at least one of the following criteria

(recommended by the ICMJE*):

1) substantial contributions to conception and design, acquisition of data, or analysis and

interpretation of data;

2) drafting the article or revising it critically for important intellectual content.

* http://www.icmje.org/recommendations/

This article is protected by copyright. All rights reserved.

 REFERENCES
Accepted Article
Black, M.C., Basile, K.C., Breiding, M.J., Smith, S.G., Walters, M.L., Merrick, M.T., Chen,

J. & Stevens, M.R. 2011, "The national intimate partner and sexual violence survey:

2010 Summary Report", Atlanta, GA: National Center for Injury Prevention and

Control, Centers for Disease Control and Prevention, .

Brink, O. 2009, "When violence strikes the head, neck and face", The Journal of trauma, vol.

67, no. 1, pp. 147-151.

Campbell, J. C., O’Sullivan, C., Roehl, J., & Webster, D. W. 2005, Intimate partner violence

risk assessment validation study: The RAVE study, National Institute of Justice,

Washington DC.

Campbell, J. C., Webster, D., Koziol-McLain, J., Block CR, Campbell, D., Curry, MA, Gary,

F, Sachs, C. Sharps, PW, Wilt, S., Manganello, J., Xu, 2003, "Risk factors for femicide

in abusive relationships: Results from a multi-site case control study", American Journal

of Public Health, vol. 93, pp. 1089-1097.

Campbell, J.C., Webster, D.W. & Glass, N. 2009, "The Danger Assessment validation of a

lethality risk assessment instrument for intimate partner femicide", Journal of

Interpersonal Violence, vol. 24, no. 4, pp. 653-674.

Campbell, J.J.C. 2007, "Intimate Partner Homicide: Review and Implications of Research

and Policy", Trauma, violence & abuse, vol. 8, no. 3, pp. 246; 246-269; 269.

Campbell, J.C. 1986, "Nursing assessment for risk of homicide with battered women",

Advances in nursing science, vol. 8, no. 4, pp. 36.

Catalano, S., Smith, E., Snyder, H. & Rand, M. 2009, "Female victims of violence", .

Christe A, Thoeny H, Ross S, et al. Life-threatening versus non-life-threatening manual

strangulation: are there appropriate criteria for MR imaging of the neck? Eur Radiol.

2009;19(8):1882-1889.

This article is protected by copyright. All rights reserved.

 Devries, K.M., Mak, J.Y., Garcia-Moreno, C., Petzold, M., Child, J.C., Falder, G., Lim, S.,
Accepted Article Bacchus, L.J., Engell, R.E., Rosenfeld, L., Pallitto, C., Vos, T., Abrahams, N. & Watts,

C.H. 2013, "Global health. The global prevalence of intimate partner violence against

women", Science (New York, N.Y.), vol. 340, no. 6140, pp. 1527-1528.

Dobash, R.E., Dobash, R.P., Cavanagh, K. & Medina-Ariza, J. 2007, "Lethal and nonlethal

violence against an intimate female partner: comparing male murderers to nonlethal

abusers", Violence Against Women, vol. 13, no. 4, pp. 329-353.

Douglas, K.S., Guy, L.S., Reeves, K.A. & Weir, J. 2005, "HCR-20 violence risk assessment

scheme: Overview and annotated bibliography", .

Eden, K.B., Perrin, N.A., Hanson, G.C., Messing, J.T., Bloom, T.L., Campbell, J.C., Gielen,

A.C., Clough, A.S., Barnes-Hoyt, J.S. & Glass, N.E. 2015, "Use of online safety

decision aid by abused women: effect on decisional conflict in a randomized controlled

trial", American Journal of Preventive Medicine, vol. 48, no. 4, pp. 372-383.

Ellsberg, M. Jansen, H., Heise, L., Watts, C., García-Moreno, C. 2008. Intimate partner

violence and women’s physical and mental health in the WHO multi-country study on

women’s health and domestic violence: an observational study. Lancet 371: 1165–72

Glass, N., Laughon, K. & Campbell, J. 2008, "Non-fatal strangulation is an important risk

factor for homicide of women", The Journal of emergency medicine, vol. 35, no. 3, pp.

329.

Glass, N., Perrin, N., Hanson, G., Bloom, T., Gardner, E. & Campbell, J.C. 2008, "Risk for

reassault in abusive female same-sex relationships", Journal Information, vol. 98, no. 6.

Glass, N., Clough, A., Case, J., Hanson, G., Barnes-Hoyt, J., Waterbury, A., Alhusen, J.,

Ehrensaft, M., Grace, K.T. & Perrin, N. 2015, "A safety app to respond to dating

violence for college women and their friends: the MyPlan study randomized controlled

trial protocol", BMC public health, vol. 15, pp. 871-015-2191-6.

This article is protected by copyright. All rights reserved.

 Glass, N., Eden, K.B., Bloom, T. & Perrin, N. 2010, "Computerized aid improves safety
Accepted Article decision process for survivors of intimate partner violence", Journal of Interpersonal

Violence, vol. 25, no. 11, pp. 1947-1964.

Goodman, L.A., Dutton, M.A. & Bennett, L. 2000, "Predicting Repeat Abuse Among

Arrested Batterers Use of the Danger Assessment Scale in the Criminal Justice System",

Journal of Interpersonal Violence, vol. 15, no. 1, pp. 63-74.

Hanley J.A. & McNeil B.J. (1982). "The meaning and use of the Area under a receiver

operating characteristic (ROC) curve." Radiology 143:29-36

Heckert, D.A. & Gondolf, E.W. 2004, "Battered women’s perceptions of risk versus risk

factors and instruments in predicting repeat reassault", Journal of Interpersonal

Violence, vol. 19, no. 7, pp. 778-800.

Hilton, N.Z., Harris, G.T., Rice, M.E., Houghton, R.E. & Eke, A.W. 2008, "An indepth

actuarial assessment for wife assault recidivism: The Domestic Violence Risk Appraisal

Guide", Law and human behavior, vol. 32, no. 2, pp. 150-163.

Hilton, N.Z., Harris, G.T., Rice, M.E., Lang, C., Cormier, C.A. & Lines, K.J. 2004, "A brief

actuarial assessment for the prediction of wife assault recidivism: the Ontario domestic

assault risk assessment.", Psychological assessment, vol. 16, no. 3, pp. 267-275.

Holbrook DS, Jackson MC. Use of an alternative light source to assess strangulation victims.

J Forensic Nurs. 2013;9(3):140-145

Joshi, M., Thomas, K.A. & Sorenson, S.B. 2012, "“I Didn't Know I Could Turn Colors”:

Health Problems and Health Care Experiences of Women Strangled by an Intimate

Partner", Social work in health care, vol. 51, no. 9, pp. 798-814.

Kwako, L.E., Glass, N., Campbell, J., Melvin, K.C., Barr, T. & Gill, J.M. 2011, "Traumatic

brain injury in intimate partner violence: a critical review of outcomes and mechanisms",

Trauma, violence & abuse, vol. 12, no. 3, pp. 115-126.

This article is protected by copyright. All rights reserved.

 Laupacis, A. & Sekar, N. 1997, "Clinical prediction rules: a review and suggested
Accepted Article modifications of methodological standards", Jama, vol. 277, no. 6, pp. 488-494.

Lin, P. & Gill, J.R. 2011, "Homicides of pregnant women", The American journal of forensic

medicine and pathology, vol. 32, no. 2, pp. 161-163.

Strack, G.B., McClane, G.E. & Hawley, D. 2001, "A review of 300 attempted strangulation

cases part i: criminal legal issues, The Journal of emergency medicine, vol. 21, no. 3, pp.

303-309.

McClane GE, Strack GB, Hawley D. A review of 300 attempted strangulation cases Part II:

clinical evaluation of the surviving victim. J Emerg Med. 2001;21(3):311-315.

Mcquown, C., Frey, J., Steer, S., Fletcher, G.E., Kinkopf, B., Fakler, M. & Prulhiere, V.

2016, "Prevalence of strangulation in survivors of sexual assault and domestic violence",

The American Journal of Emergency Medicine, .

Messing, J.T., Amanor-Boadu, Y., Cavanaugh, C.E., Glass, N.E. & Campbell, J.C. 2013,

"Culturally competent intimate partner violence risk assessment: Adapting the danger

assessment for immigrant women", Social work research, vol. 37, no. 3, pp. 263-275.

Messing, J.T. & Campbell, J. 2016, "Informing collaborative interventions: Intimate partner

violence risk assessment for front line police officers", Policing, , pp. paw013.

Messing, J.T., Campbell, J., Webster, D.W., Brown, S., Patchell, B. & Wilson, J.S. 2015a,

"The Oklahoma lethality assessment study: a quasi-experimental evaluation of the

Lethality Assessment Program", Social Service Review, vol. 89, no. 3, pp. 499-530.

Messing, J.T., Campbell, J. & Wilson, J.S. 2015, "Research designs in the real world: testing

the effectiveness of an IPV intervention", National Institute of Justice Journal, vol. 275,

pp. 48-56.

Messing, J.T. & Thaller, J. 2015, "Intimate partner violence risk assessment: A primer for

social workers", British journal of social work, vol. 45, no. 6, pp. 1804-1820.

This article is protected by copyright. All rights reserved.

 Messing, J.T., Campbell, J., Sullivan Wilson, J., Brown, S. & Patchell, B. 2015b, "The
Accepted Article Lethality Screen: The Predictive Validity of an Intimate Partner Violence Risk

Assessment for Use by First Responders", Journal of Interpersonal Violence, .

Messing, J.T. & Thaller, J. 2013, "The average predictive validity of intimate partner

violence risk assessment instruments", Journal of Interpersonal Violence, vol. 28, no. 7,

pp. 1537-1558.

Moracco, K.E., Runyan, C.W. & Butts, J.D. 1998, "Femicide in North Carolina, 1991-1993 A

Statewide Study of Patterns and Precursors", Homicide Studies, vol. 2, no. 4, pp. 422-

446.

National Center for Injury Prevention and Control. Costs of Intimate Partner Violence

Against Women in the United States. Atlanta, GA: Centers for Disease Control and

Prevention; 2003. http://www.cdc.gov/violenceprevention/pub/IPV_cost.html. Accessed

May 24, 2017.

Puzone, C.A., Saltzman, L.E., Kresnow, M., Thompson, M.P. & Mercy, J.A. 2000, "National

Trends in Intimate Partner Homicide United States, 1976-1995", Violence Against

Women, vol. 6, no. 4, pp. 409-426.

Rennison, C.M. & Welchans, S. 2000, "Bureau of Justice Statistics special report: Intimate

partner violence", Bureau of Justice Statistics Special Report.NCJ, vol. 178247.

Rice, M.E. & Harris, G.T. 2005, "Comparing effect sizes in follow-up studies: ROC area,

Cohen's d and r", Law and human behavior, vol. 29, no. 5, pp. 615-620.

Rice, M.E. & Harris, G.T. 1995, "Violent recidivism: Assessing predictive validity.", Journal

of consulting and clinical psychology, vol. 63, no. 5, pp. 737.

Ruiz-Perez, I., Plazaola-Castano, J. & Del Rio-Lozano, M. 2007, "Physical health

consequences of intimate partner violence in Spanish women", European journal of

public health, vol. 17, no. 5, pp. 437-443.

This article is protected by copyright. All rights reserved.

 Sauvageau, A. & Boghossian, E. 2010, "Classification of asphyxia: the need for
Accepted Article standardization", Journal of forensic sciences, vol. 55, no. 5, pp. 1259-1267.

Sharps, P. W., Koziol-McLain, J., Campbell, J. C., McFarlane, J., Sachs, C., & Xu, X.

(2001). Missed opportunities for prevention of femicide by health care providers.

Preventive Medicine 33, 373-80.

Shields, L.B., Corey, T.S., Weakley-Jones, B. & Stewart, D. 2010, "Living victims of

strangulation: a 10-year review of cases in a metropolitan community", The American

journal of forensic medicine and pathology, vol. 31, no. 4, pp. 320-325.

Smith, D.J., Mills, T. & Taliaferro, E.H. 2001, "Frequency and relationship of reported

symptomology in victims of intimate partner violence: the effect of multiple

strangulation attacks", The Journal of emergency medicine, vol. 21, no. 3, pp. 323-329.

Smith, S.G., Fowler, K.A. & Niolon, P.H. 2014, "Intimate partner homicide and corollary

victims in 16 states: National Violent Death Reporting System, 2003–2009", American

Journal of Public Health, vol. 104, no. 3, pp. 461-466.

Smock, B., & Sturgeon, S. (2016). Recommendations for the medical/radiographic evaluation

of acute adult, nonfatal strangulation. Retrieved from:

https://www.strangulationtraininginstitute.com/wp-

content/uploads/2015/07/Recommendations-for-Medical-Radiological-Eval-of-Non-

Fatal-Strangulation-v17.9.pdf

Snider, C., Webster, D., O’sullivan, C.S. & Campbell, J. 2009, "Intimate partner violence:

Development of a brief risk assessment for the emergency department", Academic

Emergency Medicine, vol. 16, no. 11, pp. 1208-1216.

Sorenson, S.B., Joshi, M. & Sivitz, E. 2014, "A systematic review of the epidemiology of

nonfatal strangulation, a human rights and health concern", American Journal of Public

Health, vol. 104, no. 11, pp. e54-e61.

This article is protected by copyright. All rights reserved.

 Stöckl, H., Devries, K., Rotstein, A., Abrahams, N., Campbell, J., Watts, C. & Moreno, C.G.
Accepted Article 2013, "The global prevalence of intimate partner homicide: a systematic review", The

Lancet, vol. 382, no. 9895, pp. 859-865.

Stockman, J. K., Hayashi, H., Campbell, J. C. (2015). Intimate partner violence and its health

impact on ethnic minority women. Journal of Women’s Health, 24, 1-18.

http://dx/DOI:10.1089/jwh.2014.4879

Straus, M.M.A. 1996, "The Revised Conflict Tactics Scales (CTS2): Development and

Preliminary Psychometric Data", Journal of Family Issues, vol. 17, no. 3, pp. 283; 283-

316; 316.

Straus, M.A., Hamby, S., Warren, W. , The conflict tactics scales handbook: Revised conflict

tactics scale (CTS2) and CTS: Parent-child version (CTSPC). 2003, Los Angeles, CA:

Western Psychological Services

Tadegge, A.D. 2008, "The mental health consequences of intimate partner violence against

women in Agaro Town, southwest Ethiopia", Tropical doctor, vol. 38, no. 4, pp. 228-

229.

Thomas, K.A., Joshi, M. & Sorenson, S.B. 2013, "“Do you know what it feels like to

drown?” Strangulation as coercive control in intimate relationships", Psychology of

Women Quarterly, , pp. 0361684313488354.

Vella, S.A. 2013, Cognitions and Behaviors of Strangulation Survivors of Intimate

Terrorism, ALLIANT INTERNATIONAL UNIVERSITY.

Violence Policy Center 2011, "When men murder women: An analysis of 2009 homicide

data", .

W. Sharps, Jacquelyn Campbell, Dori,Phyllis P.W. 2001, "The Role of Alcohol Use in

Intimate Partner Femicide", The American journal on addictions, vol. 10, no. 2, pp. 122;

122-135; 135.

This article is protected by copyright. All rights reserved.

 Weisz, A.N., Tolman, R.M. & Saunders, D.G. 2000, "Assessing the Risk of Severe Domestic
Accepted Article Violence The Importance of Survivors' Predictions", Journal of Interpersonal Violence,

vol. 15, no. 1, pp. 75-90.

Wilbur, L., Higley, M., Hatfield, J., Surprenant, Z., Taliaferro, E., Smith, D.J. & Paolo, A.

2001, "Survey results of women who have been strangled while in an abusive

relationship", The Journal of emergency medicine, vol. 21, no. 3, pp. 297-302.

Wilson, J.S., West, J.F., Messing, J.T., Brown, S., Patchell, B. & Campbell, J.C. 2011,

"Factors related to posttraumatic stress symptoms in women experiencing police-

involved intimate partner violence", ANS.Advances in nursing science, vol. 34, no. 3, pp.

E14-28.

Yen K, Thali MJ, Aghayev E, et al. Strangulation signs: initial correlation of MRI, MSCT

and forensic neck findings. J Magn Reson Imaging. 2005;22(4):501-510.

Table 1. Outcome: Serious Injury or Potentially Lethal Assault at Follow-up (n=619)

Since the last interview: Number (%)
of respondents
Your partner used a knife or gun on you (one or more times) 22 (3.6)
Your partner choked you (one or more times) 62 (10.0)
Your partner burned or scalded you on purpose (one or more times) 9 (1.5)
Have you suffered internal injuries to vital organs because of a fight 6 (1.0)
with your partner?
Lost consciousness due to him choking you? 19 (3.0)
Have you lost consciousness for more than 1 hour due to head injuries? 4 (0.7)
Have you been hospitalized or in rehab for more than 4 days because of 11 (1.8)
injuries?
Did he try to kill you? 30 (4.9)
Total number of participants experiencing severe injury or potentially 81 (13.1)
fatal assault during follow-up*
*Does not equal sum as some participants responded affirmatively to multiple items.

This article is protected by copyright. All rights reserved.

 Table 2. Participant demographic and relationship characteristics
Accepted Article Mean (SD)
Demographic & Relationship Characteristics
or Number (%)
Age 33.1 (10.3)
Race / Ethnicity
Black / African American 182 (29.4)
White 259 (41.8)
Latina 44 (7.1)
Native American 58 (9.4)
Multiracial 47 (7.6)
Other / Declined 29 (4.7)
Foreign Born / Immigrant 26 (4.2)
Education
Not a high school graduate 110 (17.8)
High school graduate / GED 161 (26.0)
Some college or more 343 (55.4)
Declined 5 (0.8)
Employment
Working full or part time 274 (44.3)
Other (e.g., unemployed, homemaker, student) 345 (55.7)
Marital Status
Single 363 (58.6)
Married 156 (25.2)
Separated 28 (4.5)
Divorced 69 (11.1)
Declined 3 (0.4)
Relationship with the abusive partner
Boyfriend 117 (18.9)
Husband 106 (17.1)
Common law 21 (3.4)
Estranged husband 31 (5.0)
Ex-boyfriend 221 (35.7)
Ex-husband 45 (7.3)
Ex-common law 11 (1.8)
Other / Declined 67 (10.8)
Current level of involvement with the abusive partner
Cohabiting 105 (17.0)
Not cohabiting but intimately involved 33 (5.3)
On again, off again relationship 30 (4.8)
Not cohabiting, no intimate relationship 429 (69.3)
Declined 2 (0.3)

This article is protected by copyright. All rights reserved.

 Table 3. Prevalence of examined risk factors at baseline and their independent associations
with serious injury or potentially lethal assault during follow-up (n=619)
Accepted Article
Odds Ratio
Risk Factor Number (%)
(95% CI)
1
Has the physical violence increased in frequency or severity 357 (57.7) 1.9 (1.1-3.1)
over the past year?
1
Is your partner (or ex-) violently and constantly jealous of 377 (60.9) 2.1 (1.3-3.7)
you?
1
Do you believe your partner (or ex-) is capable of killing 365 (59.0) 2.5 (1.5-4.3)
you?
1
Have you ever been beaten by your partner while you were 130 (21.0) 2.1 (1.3-3.5)
pregnant?
1
Has your partner (or ex-) ever used a weapon or threatened 200 (32.3) 2.7 (1.7-4.4)
you with a weapon?
2
Has your partner ever tried to choke (strangle) you? 440 (71.8) 2.6 (1.4-4.9)
3
[Your partner] choked you… twice or more in the past 6 231 (37.3) 3.4 (2.1-5.5)
months.
1
Included on the original DA-5 (author, 2009)
2
Included on the Danger Assessment (Campbell et al., 2003)
3
Included on the Conflict Tactics Scale (CTS-2; Straus et al., 1996)

This article is protected by copyright. All rights reserved.

 Table 4. Predictive Validity
Accepted Article %
No. False False Correctl Positive Negative
Sensitivit Specificit FP/FN
Positiv n (%) Positiv Negativ y Predictiv Predictiv
y y Ratio
e e (FP) e (FN) Classifie e Value e Value
d
DA-5
542
116.2
>=1 (87.6 95.1% 13.6% 465 4 24.2% 14.2% 94.8%
5
)
428
>=2 (69.1 86.4% 33.5% 358 11 32.5 40.4% 16.4% 94.2%
)
286
>=3 (46.2 71.6% 57.6% 236 21 11.2 59.5% 20.3% 93.5%
)
144
>=4 (23.3 44.4% 79.9% 108 45 2.4 75.3% 25.0% 90.5%
)
29
>=5 11.1% 96.3% 20 72 0.28 85.1% 31.0% 87.8%
(4.7)
DA-5 with strangulation replacing abuse during pregnancy
573
>=1 (92.6 98.8% 8.4% 493 1 493 20.19% 14.0% 97.8%
)
482
>=2 (77.9 91.4% 24.2% 408 7 58.3 33.0% 15.4% 94.9%
)
373
>=3 (60.3 81.5% 42.9% 307 15 20.5 48.0% 17.7% 93.9%
)
222
>=4 (35.9 59.3% 67.7% 174 33 5.3 66.6% 21.6% 91.7%
)
89
>=5 (14.4 29.6% 87.9% 65 57 1.14 80.3% 27.0% 89.2%
)
DA-5 with multiple strangulation replacing abuse during pregnancy
544
>=1 (87.9 96.3% 13.4% 466 3 155.3 24.2% 14.3% 96.0%
)
430
>=2 (69.5 87.7% 33.3% 359 10 35.9 40.4% 16.5% 94.7%
)
314
>=3 (50.7 74.1% 52.8% 254 21 12.1 55.6% 19.1% 93.1%
)
177
>=4 (28.6 55.6% 75.5% 132 36 3.7 72.9% 25.4% 91.9%
)

This article is protected by copyright. All rights reserved.

 65
>=5 (10.5 24.7% 91.6% 45 61 0.7 82.9% 30.8% 89.0%
Accepted Article )
DA-5 with abuse during pregnancy and strangulation
547
>=1 (88.4 96.3% 12.8% 469 3 156.3 23.8% 14.3% 95.8%
)
444
>=2 (71.7 88.9% 30.9% 372 9 41.3 38.5% 16.2% 94.9%
)
337
>=3 (54.4 79.0% 49.3% 273 17 16.1 53.2% 19.0% 94.0%
)
212
>=4 (34.3 60.5% 69.7% 163 32 5.1 68.5% 23.1% 92.1%
)
99
>=5 (16.0 37.0% 87.2% 69 51 1.4 80.6% 30.3% 90.2%
)
21
>=6 9.9% 97.6% 13 73 0.2 86.1% 38.1% 87.8%
(3.4)

This article is protected by copyright. All rights reserved.

 Figure 1
Accepted Article
DANGER ASSESSMENT FOR CLINICIANS
(DA- 5)

This brief risk assessment identifies women who are at high risk for homicide or severe
injury by an intimate partner or former intimate partner.1,2

Mark Yes or No for each of the following questions.

1. Has the physical violence increased in frequency or severity over the past year?

2. Has your partner (or ex) ever used a weapon against you or threatened you with a
weapon?

3. Do you believe your partner (or ex) is capable of killing you?

4. Has your partner or ex ever tried to choke (strangle) you?

a. If yes, did he ever choke you? _______
b. About how long ago? __________
c. Did it happen more than once? ______
d. Did you ever lose consciousness or think you may have? ______

5. Is your partner or ex violently and constantly jealous of you?

1 This is a brief adaptation of the Danger Assessment (2003). It is designed for use by a health care
provider following a positive screen for intimate partner violence. The full Danger Assessment with
weighted scoring provides the most accurate assessment of risk.
2 Snider, C., Webster, D., O’Sullivan, C. S. and Campbell, J. (2009), Intimate Partner Violence:
Development of a Brief Risk Assessment for the Emergency Department. Academic Emergency Medicine,
16: 1208–1216.

This article is protected by copyright. All rights reserved.

 Protocol suggestions for use of
Accepted Article
DANGER ASSESSMENT-5
• Use 5-item version in the Emergency Department and other healthcare
settings; at protective order hearings, child custody etc. once intimate
partner violence has been identified.
• If 4 or 5 yes responses, tell the victim s/he is in danger, allow the victim to
choose reporting to the police &/or to domestic violence advocacy program
&/or confidential hotline (e.g., 800-799-7233). Follow through by calling
with the victim &/or with an in-person hand-off to a knowledgeable
advocate.
• If 3 of 5 yes responses, do full Danger Assessment (DA) with the
calendar and weighted scoring if victim is female; inform the victim of level
of danger and do safety planning based on DA or refer to someone
certified in administrating the DA and proceed based on results and best
practice.
• If 2 of 5, tell the victim there are 2 of 5 risk factors for serious
injury/assault/homicide present and recommend further advocacy. If the
victim agrees, follow through with a referral and hand-off to a
knowledgeable advocate. An in-person or voice-to-voice hand-off on the
phone (e.g. 3-way-call or speaker phone) is preferable.
• If 0-1 of 5, proceed with normal referral/procedural processes for domestic
violence.

Brief strangulation protocol

If yes to 4a. If strangulation was a week ago or less, examine inside of throat,
neck, face and scalp for physical signs of strangulation. See strangulation
assessment and radiographic evaluation information at
www.strangulationtraininginstitute.com. Proceed with emergency medical care
for strangulation, especially if loss of consciousness or possible loss of
consciousness (victims are often unsure, but if the victim lost consciousness,
s/he will have become incontinent – can ask if the victim “wet her/himself”). If
the victim reports more than one strangulation, conduct neurological exam for
brain injury or refer for examination and inform her/him of increased risk of
homicide. Notify police and/or prosecutors if the victim wants this action (know
state/local law on strangulation and mandatory reporting so that the victim can
be informed).

This article is protected by copyright. All rights reserved.