Original contribution
Improving the Emergency Department detection rate of domestic violence using direct questioning

https://doi.org/10.1016/S0736-4679(00)00204-3Get rights and content

Abstract

The purpose of this study was to compare the domestic violence (DV) rate identified with simple direct questioning to a historical cohort of patients receiving routine emergency department (ED) care. One thousand ED charts of female patients were retrospectively reviewed. Each patient in the prospective cohort was asked five DV specific questions. The historical cohort revealed a DV prevalence rate of 0.4%. The prospective study group of 302 patients identified 11 (3.6%) patients who admitted to acute DV on direct questioning. Ten of these patients accepted help. Twenty (6.6%) were identified as probable DV and 12 (4%) admitted to past violence. The total number of victims of DV, past, present, and probable was 43 (14.2%). This increase in detection from 0.4% (4/1000) to 14.2% (43/302) is significant at p < 0.001. Only 1.3% of patients refused to participate in the DV specific questions. The conclusion of the study indicated that the use of simple, direct questioning significantly improves the detection rate of DV in the ED.

Introduction

Experts in the field have estimated that at least 10% of all females are victims of domestic violence (DV) (1). Most would agree this number does not reflect the true prevalence because detection and reporting are suboptimal 2, 3, 4, 5. Strauss estimated an incidence rate of 3.8% in a U.S. national survey published in 1980 (6). McLeer extrapolated, using this rate, and a medium frequency of 2.4 assaults per year, and a U.S. population of 47 million married couples, to conclude that every 7.4 s a woman is beaten by her husband (7). A national representative sample of women surveyed by telephone in 1996 revealed an overall 1 year prevalence rate of physical abuse as 7.3% (8).

The emergency department (ED) is an entry point into the health care system used by some of the more severely abused women who may be at risk for serious physical impairment or death (9). There is a high incidence of domestic violence and child abuse in the families of batterers and victims of battering, suggesting that witnessing violence as a way of dealing with frustration and anger has a negative impact on children 10, 11. By not asking directly about the risk of DV, the physician misses the underlying precipitant of the patient’s medical condition and the opportunity to prevent further violence. The failure to diagnose DV is an important factor in the development of psychopathology (6). Studies have documented that DV escalates in frequency and severity, and morbidity and mortality rates rise without recognition and intervention 6, 10. The personal and social morbidity and related mortality rates are too high to ignore this diagnosis and fail to treat appropriately. In any other clinical situation, it is unacceptable to discharge a patient with a potentially life-threatening illness without making the diagnosis, offering appropriate therapy, and allowing the patient to make an informed decision about her care (12).

A survey of 1000 women conducted by Bower in 1987 revealed that abused women ranked health care professionals the lowest in effectively addressing their abuse (13). Abused women report poor communication and dissatisfaction with physician care more frequently than non-abused women (14). Stark et al. reported that physicians detected only 1 in 25 DV cases (15). In 1985, Goldberg and Tomlanovich reported that Emergency Physicians had a 5% documentation rate of DV (16). Hilberman and Munson reported that four out of 120 cases referred for a psychiatric assessment were identified, by the referring physician, as victims of domestic violence. In contrast, with direct questioning, 60 answered affirmatively (17). Using direct questioning, Downey et al. improved the domestic violence detection rate from 7% to 20% in families visiting the local social service agency (18). Abused women typically suggest that specific direct screening questions are more effective than indirect generic questions (14).

Historically, most commonly used validated instruments (Combat Tactics Scale and the Index of Spouse Abuse) measure violence as opposed to screen for victims of violence and are long and cumbersome to apply to the ED setting 19, 20. Recently, a shorter interview instrument, the Partner Abuse Interview, was shown to have good inter-rater agreement (kappa 0.77–1.0) and strong internal consistency (alpha of 0.82) for Englishspeaking participants with a committed relationship for at least 6 months. This instrument screens for physical abuse only and consists of 11 questions. It reportedly takes 3 min to complete when there is no abuse and 10–15 min with evidence of physical abuse (21). It has not been validated for universal application in the ED clinical setting. The need for Emergency Physicians to recall verbatim or remember to use an instrument consisting of 11 questions may limit the universal clinical application of this screening tool.

Feldhaus et al. published a diagnostic test evaluation of the Partner Violence Screen (PVS) using the Combat Tactics Scale and the Index of Spouse Abuse as the standardized measures. The PVS is comprised of 3 questions (Table 1 ) and identifies a similar prevalence rate as the two standardized tests with substantial confidence limit overlap on the point estimate. The sensitivity was 64.5% and 71.4% and the specificity was 80.3% and 84.4%, respectively, relative to the standard of comparison (22). The feasibility of use in the ED and the reliability, internal consistency, and concurrent validity measures of the instrument have not been published.

It was the intention of this study group to evaluate the feasibility of using a minimum number of questions focused directly on all forms of abuse as a universal screening tool for DV. The challenge was to design the questions such that they could be easily incorporated into every routine ED patient encounter and generic enough to allow for variations in physician style and comfort level. This prospective observational study was done to determine if the detection rate of victims of violence in the ED could be improved using this simple, direct line of questioning compared to a historical control group.

Section snippets

Materials and methods

The study was conducted at the Royal Victoria Hospital in Montreal, Canada. This is a downtown tertiary hospital affiliated with McGill University. The ED annual volume at the time of the study was 36,000, including 15,000 to 17,000 adult female patients. The historical cohort consisted of a random sample of 1000 charts of women who presented to the ED during the two-month period just before the onset of the prospective study. The charts were selected as per the temporal frequency of patient

Results

The retrospective chart review of 1000 charts identified four patients (0.4%) as past or present victims of domestic violence by the physicians, consultants, nurses, or ambulance staff.

The prospective cohort consisted of 343 women. A questionnaire was completed on all the female patients including the 41 excluded patients (12%). Excluded patients were unstable, incoherent or unresponsive. The final sample size was 302. Included in the final sample were four patients who left before being seen

Discussion

Abbott et al. (1995) reported a cumulative lifetime prevalence rate of domestic violence in female patients presenting to the ED to be 54.2% (95% CI, 50.2% to 58.1%) and an incidence of ADV to be 11.7% (23). Only six of all victims of ADV were directly questioned, or told the nurse or MD about DV, for a detection rate of 2.6%. A review of charts revealed that ADV was documented on two records (23).

In our study, limited direct questioning in the ED improved the detection rate of victims of

Acknowledgements

This study was funded by peer reviewed grants from the Quebec Association of Emergency Physicians and the Sunnybrook Research Foundation.

References (40)

  • C. Warshaw

    Limitations of the medical model in the care of battered women

    Gender Soc

    (1989)
  • A.I. Kellermann

    Domestic violence and the internist’s responseadvocacy or apathy

    J Gen Intern Med

    (1990)
  • M. Strauss et al.

    behind closed doorsviolence in the American family

    (1980)
  • S.V. McLeer

    Spouse abusethe handbook of marital therapy

    (1981)
  • L. Walker

    Battered women, psychology and public policy

    Am Psychol

    (1984)
  • R.E. McAfee

    Physicians and domestic violencecan we make a difference?

    JAMA

    (1995)
  • Physicians and domestic violenceethical considerations

    JAMA

    (1992)
  • L.H. Bowker et al.

    The medical treatment of battered wives

    Women Health

    (1987)
  • L.A. McNutt et al.

    Reproductive violence screening in primary careperspectives and experiences of patients and battered women

    J Am Med Womens Assoc

    (1999)
  • E. Stark et al.

    Wife abuse in the medical settingan introduction for health personnel

    (1981)
  • Cited by (46)

    • Teen Dating Violence: Old Disease in a New World

      2019, Clinical Pediatric Emergency Medicine
    • Enhanced Versus Basic Referral for Intimate Partner Violence in an Urban Emergency Department Setting

      2017, Journal of Emergency Medicine
      Citation Excerpt :

      In addition, both victims reporting current and previous victimization contacted referral resources at similar rates, indicating that the basic and enhanced referrals demonstrated consistent efficacy across current and past victims. Research has shown that direct screening in health care settings may be an effective way to identify individuals who have experienced IPV (8,21). Specifically, direct questioning to assess for IPV victimization can improve detection rates within an ED population (21).

    • A comparison of the training needs of maternity and sexual health professionals in a London teaching hospital with regards to routine enquiry for domestic abuse

      2010, Public Health
      Citation Excerpt :

      The UK Department of Health recommend routine enquiry for DA, defined as asking all women if they are experiencing domestic abuse, regardless of whether or not they show any indications.1 Studies from the UK and elsewhere have shown that this policy, coupled with other measures such as clinical guidelines and referral pathways for victims, significantly increases the rate of identification of abused women.6–8 Sexual health and maternity professionals are strategically placed to identify and support victims of DA.

    • A Comparison of 2 Protocols to Detect Intimate Partner Violence

      2009, Journal of Oral and Maxillofacial Surgery
      Citation Excerpt :

      The net effect of having predominantly one type of misclassification error (more false negatives than false positives) is an underestimation of the true sensitivity of the DP, and it should have a minimal effect on the specificity estimates. The concept of early detection of women who may or may not have IPV-related injuries or complaints is the first step in successful management strategies to deal with the immediate- and long-term effects of violence.10,14-20 The results of this study suggest that a DP composed of injury location and verbal responses to a limited questionnaire in comparison to the SOP of the ED may more often aid in the early identification of women injured because of IPV.

    • Prevalence of Past Year Assault Among Inner-City Emergency Department Patients

      2009, Annals of Emergency Medicine
      Citation Excerpt :

      In a similar manner, nonpartner violent assault aggression was assessed by asking, “In the past year, have you hit, kicked, punched, or otherwise hurt friends, strangers, neighbors, people in bars, coworkers, bosses?” Although this 1-item screen has not been validated, the full-length Partner Violence Screen has adequate sensitivity (54.5 to 71.4) and specificity (80.3 to 84.4) compared with more lengthy measures for use in ED settings.42 The Short Form-12 Health Survey, which has been validated in an ED setting,43 was used to assess physical and mental health functioning during the previous 4 weeks.44-46

    View all citing articles on Scopus
    View full text