Intimate partner violence has an annual prevalence of 2 to 12 percent. Health consequences include injury and death, as well as physical, social, and emotional problems, many of which extend to the infants of women experiencing violence during pregnancy. Recent attention to partner violence has emphasized screening programs. Wathen and MacMillan performed a systematic review of the evidence for strategies to identify and treat abused women presenting to a primary care setting.

A database search yielded 22 studies describing interventions meeting the investigators' selection criteria. Intimate partner violence was defined as physical and psychologic abuse of women by their male partners, including sexual abuse and abuse during pregnancy. Outcome measures included decrease in incidence of self-reported abuse, amount of accessible social support, use of safety behaviors or safety planning, and use of community resources.

Two interventions are generally available to primary care physicians: screening and referral. No studies are available to determine the effectiveness of screening in improving outcomes in women. Physicians can refer women to a safe place, to counseling, or to other community-based resources, and can refer men to treatment programs for abusive partners. Eleven studies described four referral interventions: advocacy counseling after at least one night's stay at a shelter; staying at a shelter; personal and vocational counseling; and prenatal counseling.

In a study that evaluated advocacy counseling following a stay in a shelter for at least one night, reabuse occurred less frequently at the two-year follow-up in the intervention group than in the control group (76 percent versus 89 percent). Physical violence decreased and quality of life increased in the intervention group. No studies with a quality rating of good or fair exist to measure the effectiveness of shelters. In addition, there were no studies of sufficient quality to determine the effectiveness of personal and vocational counseling. Studies looking at prenatal counseling using information cards, counseling, and more intensive interventions had design flaws as well. Of the studies that evaluated interventions aimed at men or couples, only one was rated as good, and that study concluded that three types of interventions were not effective in reducing further violence against women. In general, there was a low recidivism rate in both intervention and control groups, attributable perhaps to the deterrent setting provided by the military, the context in which the study took place.

Screening tools exist to identify women experiencing intimate partner violence, but to date these tools have been evaluated only to see if they do identify such women. There are no studies to determine whether screening is effective in preventing abuse. Intervention studies are of poor quality and flawed design. So far, only one study provides evidence for referring a woman to a shelter, followed by advocacy counseling.

The authors conclude that it is premature to recommend widespread screening programs for partner violence in health care settings. However, screening may be justified in determining the cause of symptoms and signs of abuse and preventing unnecessary work-ups when the etiology is, in fact, partner violence.