Preventing Intimate Partner Violence: Title and subTitle BreakScreening Is Not Enough

Physical, sexual, and psychological abuse of women by their intimate partners is common around the world.¹ In response to this widespread public health problem, organizations of health care professionals, including the American Medical Association² and the American College of Obstetricians and Gynecologists³ recommend that all adult female patients be asked routinely about abuse, regardless of their presenting symptoms. However, evidence-based guides to clinical preventive services, such as the US Preventive Services Task Force⁴ and the Canadian Task Force on Preventive Health Care,⁵ have concluded that there is insufficient evidence of health benefits to abuse survivors to recommend for or against screening for intimate partner violence (IPV), primarily due to methodological weaknesses of available studies.

In this issue of JAMA, MacMillan et al⁶ address this lack of evidence in the report of a trial that randomized women presenting for care in emergency departments, family practices, and obstetrics/gynecology clinics to be screened for abuse before seeing a clinician or to be seen by a clinician without being screened for abuse. It was then at the discretion of clinicians to discuss abuse (if present) or to make referrals for IPV services. To ensure a minimum standard of care for all women participating in the study, each participant, regardless of screening status, was given a printed card with names and telephone numbers of local agencies and telephone hotlines for women exposed to violence.

Screening took place in primary care facilities where clinicians had received standardized training in responding to IPV. After 18 months of follow-up, the difference in recurrence of IPV for screened vs nonscreened women was not statistically significant, and a slight improvement in quality-of-life scores for the screened group was no longer statistically significant after a multiple imputation technique was used to account for loss to follow-up of study participants. Therefore, universal screening, which involves routinely asking all patients about abuse, was not found to be beneficial in this study. Universal screening should be distinguished from assessing abuse as a diagnostic test, which was not addressed in this study. Assessing abuse in women at increased risk may not only detect violence but may also lead to more accurate diagnosis and treatment of co-occurring health problems.^{7 - 8}

MacMillan et al⁶ offer 2 possible interpretations for their study's failure to demonstrate health and quality-of-life benefits of universal screening for exposure to IPV. First, the lack of demonstrated benefits may be attributable to limitations of the study methods and data. These limitations included potential errors in the measurement of violence, enrollment of women during an escalating period of violence that was likely to decrease over time even in the absence of screening (regression to the mean), unmeasured benefits to screened and unscreened women associated with being followed up by study personnel over time (Hawthorne effect), distribution of information about community resources to both screened and unscreened women (the printed card with local agencies' information), approximately equal use of community resources by both groups, and loss to follow-up, which was 43% in the screened group and 41% in the nonscreened group. However, the authors conducted additional analyses to assess the sensitivity of their findings to these limitations, which confirmed that the most valid conclusion of their investigation was a lack of health benefits of universal screening.

The authors' alternative explanation is that the lack of efficacy of screening for IPV in this study may have been due to the lack of an evidence-based, effective intervention for IPV to accompany the screening. That is, if the clinicians had the opportunity to refer abused women to such an intervention, perhaps they would have been motivated to make more referrals. Similarly, if the study participants had perceived a potential benefit from an intervention for IPV, perhaps more women would have kept their appointments and remained in the study. Most important, if an intervention had been available that was effective in reducing violence and was acceptable to the study participants, measured violence subsequent to screening might have abated in the screened group more than the unscreened group.

Although interventions for IPV have not yet consistently been demonstrated to be effective in randomized trials,⁹ at least 3 approaches hold promise for ameliorating the deleterious effects of IPV and preventing recurrence of violence. First, the most widely used intervention for IPV survivors is referral to community resources, such as counseling, legal services, shelters, and other clinical and social services. In the United States, community-based domestic violence organizations usually serve as the hub for service provision for IPV survivors, offering direct services such as 24-hour telephone hotlines, support groups, counseling, emergency shelter, and court advocacy, as well as case management, referral, and coordinated delivery of housing, child care, and other community services.¹⁰ There are few evaluations of such multifaceted interventions, in part because of the inherent methodological difficulties.⁷ However, research suggests that individualized case management and follow-up with IPV survivors may lead to a decreased risk of abuse recurrence, increased effectiveness in accessing needed resources, and improved quality of life.^{11 - 13} For example, in one of the few longitudinal studies of this type of intervention,¹¹ battered women who had spent at least 1 night in a domestic violence shelter were followed up for 2 years. Half of the women were provided with counseling services by trained advocates who worked with the women for 10 weeks, 4 to 6 hours per week, to improve their access to needed community resources. At 24-month follow-up interviews, the group of women who received advocacy services reported less physical violence by a partner, better access to community resources, and improved quality of life and social support compared with women who did not receive advocacy services.

Second, perinatal home visiting is another potential approach to providing services to abuse survivors. The US Preventive Services Task Force has concluded that perinatal home visiting for high-risk mothers is effective in decreasing the incidence of child maltreatment and maternal drug use.⁴ Home-visiting programs typically provide anticipatory guidance, education regarding child development and parenting skills, social support, and referral to community resources.¹⁴ Although perinatal home-visiting programs have not been specifically designed to address IPV, a systematic review found that half of the home-visiting programs assessed mothers for exposure to IPV and that past-year prevalence of IPV among women in those programs ranged from 14% to 52%.¹⁴ However, none of the home-visiting programs integrated IPV-related content as part of their curriculum, and most made outside referrals only when IPV was identified. Given the high prevalence of IPV among women in perinatal home-visiting programs and the promising initial results of home visiting for abused women, this approach may help improve outcomes for abused women but needs to be studied in a controlled trial.

The critical element for both case management and home-visiting programs may be the provision of social support to IPV survivors. Social support refers to the various types of assistance that people receive from other people via their social networks. Theories of social support define 4 major categories of support: emotional (caring, support), instrumental (tangible assistance), informational (provision of education, information, facts), and appraisal (advice).¹⁵ Increased social support for survivors of IPV has been associated with multiple positive health and quality-of-life outcomes and has been found to buffer against stressors, including exposure to trauma.^{16 - 17} Social support may be an important mediator of intervention effects for abuse survivors.^{11 ,17} Given that a common tactic of battering is isolation from family and other members of women's social networks, enhancing social support through case management and home-visiting interventions holds promise for decreasing IPV.

Third, substance abuse treatment is another promising approach to reducing IPV. In intimate relationships, the abuse of alcohol or other drugs by either partner has consistently been linked to more severe male-perpetrated IPV.¹⁸ Addiction may increase a woman's vulnerability to IPV, lessen her ability to seek help, and increase her dependence on an abusive partner (for example, many shelters will not admit women with active substance problems¹⁰ ). Therefore, effective interventions for substance abuse also may be effective in decreasing IPV. Similarly, interventions designed to address other conditions that commonly occur with IPV, such as posttraumatic stress disorder and depression,^{17 ,19} also may decrease the risk of IPV. The relationship between IPV and substance abuse is complex, and more work needs to be done to identify opportunities for intervention. As an initial step in this research, randomized trials of substance abuse and treatment interventions could include measures of the frequency and severity of IPV. However, information must be obtained from study participants at risk of abuse in such a way as to ensure their safety, as was done in the study by MacMillan et al.⁶

In summary, there continues to be a lack of evidence that universal screening alone improves health outcomes for IPV survivors. It is certainly understandable that clinicians and health care facilities have implemented universal screening programs, given the prevalence and potential severity of IPV. However, the results of the study by MacMillan et al⁶ should dispel any illusions that universal screening with passive referrals to community services is an adequate response to violence in intimate relationships. Specific interventions to prevent the recurrence of abuse for women at risk of violence should be implemented and rigorously tested, preferably in randomized trials, without further delay.