Quick Screening for Partner Violence in the ED

Abstract & Commentary

Source: Feldhaus KM, et al. Accuracy of three brief screening questions for detecting partner violence in the emergency department. JAMA 1997; 227:1357-1361.

Feldhaus and colleagues sought to devise a brief screening tool for use in the ED to detect partner violence and validate it against two detailed "gold standard" screening exams previously described. The prospective study was designed to be administered to all noncritical, English-speaking women who presented to the ED. The screening test (Partner Violence Screen or PVS) being studied consisted of three questions that Feldhaus et al said can be asked in 20 seconds (they remind us that this is less time than it takes to check vital signs). The questions include one exploring physical abuse directly and two addressing a woman’s perception of her safety: 1.) "Have you been hit, kicked, punched, or otherwise hurt by someone within the past year? If so, by whom?" 2.) "Do you feel safe in your current relationship?" 3.) "Is there a partner from a previous relationship who is making you feel unsafe now?" Participants were asked to complete the two more extensive screening exams, the Index of Spouse Abuse (ISA) and Conflicts Tactics Scale (CTS), as part of their ED stay. Part of the study included other critical questions regarding partner violence, centering on the relationship of their present ED visit to partner violence (injuries or stress), as well as past events relating to partner violence.

Three hundred twenty-two women (76% of eligible patients) completed the PVS, and 255 of those women also completed the ISA. Two hundred seventy-seven of the 322 completed the CTS. Using the PVS, 21% of women felt unsafe from a current or past partner. Nineteen percent reported physical abuse from a partner within the past year. Overall, the PVS identified 95 women (29.5%) as positive for partner abuse. In response to the additional questions, 14% of the ED visits were associated with acute partner violence. Several of these women scored as negative on one or more of the screens. Depending upon the actual prevalence of partner violence, the positive predictive value of the PVS was 51-63%, and the negative predictive value was 88-89%. The single question regarding physical abuse on the PVS was nearly as sensitive as the entire three-question set and had better specificity.


This study highlights the simplicity and the complexity of partner violence. The simplicity is that one simple question (supplemented by a careful HPI and assessment of past domestic violence), asked of all women in the ED, can detect an impressive number of women who may be at risk for increased violence and even death. The complexity is that while screening is an important first step—and we are well-placed as emergency physicians to ask these important questions—it is just one of many aspects of successful intervention in partner violence. We must have referral and protective resources, sometimes immediate, as well as effective prevention to detect and treat women who are victims of partner violence.

Despite the complexity, we as emergency physicians are obligated to include careful and routine screening for partner violence for virtually all of our female patients. The mere presence of a woman in some EDs puts her at roughly a 30% chance of being a partner-violence victim. A perceptible number of these victims will ultimately die from escalating partner violence. The resource you give to the identified victim may save a life.