Indicators of assault-related injuries among women presenting to the emergency department

Abstract

Study objective: We sought to determine whether women presenting for treatment of assault-related injuries at a public hospital emergency department differed from those presenting for unintentional injuries with regard to a variety of demographic and presentation characteristics, nature and anatomic site of injury, and admission or follow-up treatment for injury.

Methods: We conducted a random-sample retrospective medical record review of women aged 15 years and older who presented at either of 2 24-hour public-hospital emergency departments in Auckland, New Zealand. The characteristics of women identified as presenting with assault-related injuries on the basis of the record review were compared with those of women who presented for treatment of unintentional injuries. We also assessed the sensitivity and predictive value of nature and anatomic site of injury as markers of assault.

Results: We reviewed 8,051 records, of which 2,966 (37%) involved an injury at presentation. Two hundred sixty patients (9%) were identified as victims of assault. Of those women who presented with assault-related injuries and had known assailants, most were likely injured by a partner or former partner. Women with assault-related injuries were more likely to be younger and of Maori or Pacific Islands origin. They were also more likely to present between the hours of 6 PM and 6 AM on Friday, Saturday, or Sunday and to have a greater history of prior presentations to the emergency department. Compared with patients who presented with unintentional injuries, women with assault-related injuries had a greater likelihood of presenting with contusions (odds ratio, 3.54; 95% confidence interval, 2.57 to 4.88); ill-defined signs and symptoms (odds ratio, 3.20; 95% confidence interval, 1.95 to 5.24); internal injuries (odds ratio, 2.48; 95% confidence interval, 1.46 to 4.18); fractures of the head, spine, or trunk (odds ratio, 2.09; 95% confidence interval, 1.23 to 3.53); and open wounds (odds ratio, 1.90; 95% confidence interval, 1.39 to 2.61). Assault-related injuries most commonly involved the head (odds ratio, 12.8; 95% confidence interval, 9.33 to 17.68). Despite the strength of these associations, however, with regard to nature of injury the sensitivity and positive predictive value of these indicators were limited (sensitivity < or = 26.5%, positive predictive value < or = 24.3%). The maximum sensitivity for anatomic site as a marker for assault was found for injuries to the head (63.7%), but the positive predictive value was still low at 35.7%. Women with assault-related injuries were more likely than women with unintentional injuries to be discharged from the emergency department without referral for follow-up treatment and were more likely to leave the department without completing treatment.

Conclusion: Women identified as presenting with assault-related injuries differ from those who present with unintentional injuries in terms of their demographic and presentation characteristics, as well as the nature, anatomic site of injury, and follow-up treatment for injuries. Although some of this information has implications for service delivery to abused women, the use of clinical indicators such as nature and anatomic site of injury have limited predictive value. Therefore we recommend that health care providers routinely screen patients for assault, particularly assault by intimate partners, so that they may respond appropriately by providing better treatment and referral.