Randomized Comparison of 3 Methods to Screen for Domestic Violence in Family Practice

The article by Chen and colleagues is a useful addition to an emerging evidence-base regarding the prevalence of intimate partner violence (IPV) in different health care settings, and ways to identify women who have experienced IPV. The finding that the prevalence of IPV in a sample of mainly African-American and Hispanic women is quite comparable to that previously found in mainly white women demonstrates that IPV is a cross-cultural issue affecting women of diverse ethnic, cultural and economic backgrounds. The growing body of evidence in this area should highlight the importance of IPV as a public health problem experienced by individuals across all sectors of society.

We also consider it noteworthy that the Woman Abuse Screening Tool (WAST) – a tool that ‘leads-in’ to the topic in a gentler way than some screening instruments – was shown not only to be good at identifying abuse but is also the tool preferred by physicians. In our earlier study [MacMillan, Wathen, Jamieson, et al. A randomized trial of approaches to screening for intimate partner violence in health care settings. JAMA 2006, 296(5), 530-536] the WAST was deemed to have the best combination of comprehensiveness and acceptability, when compared to the Partner Violence Screen, and is currently the instrument being used in a large multi-site RCT of screening effectiveness [http://clinicaltrials.gov/ct/show/NCT00182468?order=1]. Another finding of our testing trial was that women preferred written or computer-based methods over face-to-face (FTF) screening by a physician or nurse. This preference for self-complete methods was even greater among women who reported IPV in the past year compared with those negative for IPV, a finding not reported in the JAMA paper due to space limitations.

Much faith has been placed in the ‘simple’ act of asking about abuse – that asking in itself will lead to reductions in violence and will de- stigmatize the issue, and therefore change societal norms and ultimately individual behavior. This remains unknown; the benefits and harms of early identification, through universal screening or any other form of inquiry, have not been established. Nor has the link been made between routine inquiry in clinical settings and a more aware, informed and less violent society.

Clearly, ignoring the problem is not acceptable, but we must address the issue of potential harms as well as potential benefits of intervention, individually and at a population level. For these reasons both the US Preventive Services Task Force and the Canadian Task Force on Preventive Health Care advocate approaches to identification based on clinical indicators of abuse, rather than universal screening. Recent work by our group [Wathen, Jamieson, Wilson et al. Risk indicators to identify intimate partner violence in the emergency department. Open Med 2007 1(2), e113-22. Free at: http://www.openmedicine.ca/article/view/63/62], adds to a growing body of evidence regarding clinically important indicators that can alert clinicians to the presence of IPV and thus trigger inquiry. It is hoped that the question of effectiveness of IPV screening will be answered in the near future; whatever the outcome of ongoing studies, we should make policy and practice decisions based on the best available evidence.

Competing interests:   None declared

Domestic violence (DV) is widely recognized as an important health care topic, yet there is no consensus on the best method of screening patients. Health care providers clearly face numerous barriers to screening for domestic violence, including fear of offending patients, fear of opening Pandora’s box, inadequate training, lack of resources, underestimating the prevalence of DV in their practice, and very real time constraints. Given the difficulty and discomfort that often accompany discussions of domestic violence, there is a need to improve screening in routine practice. Therefore, we wholeheartedly support Chen and colleagues’ research on screening methods.

The authors found equivalent levels of disclosure and comfort for all three methods that they investigated—self-report, medical staff interview, or physician interview. One limitation of this study, however, is the failure to examine other, technologically advanced screening methods. With increasing availability and acceptance of computerized, video, or telephone interviews, they have become promising new adjuncts to support and simplify health care providers’ efforts.

In our previous research, for example, we found that technologically advanced methods consistently achieved significantly higher disclosure of sexual risks, illicit drug use, and smoking than traditional face-to-face interviews or written questionnaires. [Gerbert, Bronstone, Pantilat, McPhee, Allerton, Moe. When Asked, Patients Tell: Disclosure of Sensitive Health-Risk Behaviors. Medical Care. 1999; 37(1):104-111.] The greater sense of privacy afforded by these methods may increase patients’ willingness to disclose sensitive and potentially stigmatizing behaviors. Furthermore, we found that there were no significant differences in patterns of disclosure when patients were told that their answers would be shared with their health care provider rather than remaining with the researchers. This indicated that patients understood the connection these behaviors had to their health care. When we examined disclosure of domestic violence by various methods, we found equivalent levels of disclosure for both traditional and technologically advanced methods, indicating that unlike sex and illicit drug use all methods would result in similar levels of disclosure.

This finding led us to further investigate the topic. We interviewed providers who had dealt with domestic violence in their practices [Gerbert, Caspers, Bronstone, Moe, Abercrombie. A Qualitative Analysis of How Physicians with Expertise in Domestic Violence Approach the Identification of Victims. Annals of Internal Medicine. 1999; 131(8):578 -584.] as well as survivors of domestic violence [Gerbert, Abercrombie, Caspers, Love, Bronstone. How Health Care Providers Help Battered Women: The Survivor’s Perspective. Women and Health. 1999; 29(3):115-135.]. We concluded that, perhaps counter-intuitively, identification of DV should not be the goal. Instead, providers would do better to simply ask about DV in a compassionate manner and provide validating statements when appropriate, even when no disclosure takes place. Survivors frequently described a complicated “dance of disclosure,” indicating that full, direct disclosure is just one point on a continuum of disclosure. Women also described substantial relief and comfort when providers offered simple validating statements (e.g. “nobody deserves to feel unsafe at home”).

Given the pervasive, sensitive nature of domestic violence, we encourage ongoing research into better methods to address it as a part of routine care. We urge other researchers not to neglect the potential benefits of technologically advanced screening methods, but we also stress the quiet and powerful effects of compassionate inquiries and validating statements.

Competing interests:   None declared

Accolades to Chen et. al. for their contribution to the growing body of literature that looks at methods for identifying domestic violence (DV) in family practice and patient and provider comfort related to these efforts. In our work: [Zink, Levin, Putnam, Pabst, Beckstrom. The Accuracy of Five Domestic Violence Questions with Non-graphic Language For Use in Front of Children. Clinical Pediatrics. 2007; 46(2):127-134] we found no difference in the prevalence of DV identified using written, computer or verbal formats. This is good news and clinicians have options about how to initially assess for DV. But as Chen concludes, clinicians still need to conduct further assessment and resource sharing.

It is also important to consider that often with DV in family medicine we are not screening but case finding; therefore, the US Preventive Service Guideline conclusions about inadequate evidence may not be terribly helpful. Identifying DV is part of comprehensive and quality care. Studies show us that DV is commonly associated with depression, anxiety, chronic pain conditions such as headaches, pelvic pain, irritable bowel, etc. With these patients, clinicians need to ask or administer a screen much like we might use the PHQ-9 to assess depression. Both the HITS and WAST, used in Chen’s study, are instruments that have been tested in a number of studies and might be helpful to busy clinicians who want to administer a written or computer tool, and then return to discuss the results with their patients.

Competing interests:   None declared