|
|
| |||||||||||||||||||||||||||||||||||||||||
Background Domestic violence is the most common cause of nonfatal injury to women in the United States. To identify risk factors for such injuries, we examined the socioeconomic and behavioral characteristics of women who were victims of domestic violence and the men who injured them.
Methods We conducted a case–control study at eight large, university-affiliated emergency departments. The 256 intentionally injured women had acute injuries resulting from a physical assault by a male partner. The 659 controls were women treated for other conditions in the emergency department. Information was collected with a standardized questionnaire; no information was obtained directly from the male partners.
Results The 256 intentionally injured women had a total of 434 contusions and abrasions, 89 lacerations, and 41 fractures and dislocations. In a multivariate analysis, the characteristics of the partners that were most closely associated with an increased risk of inflicting injury as a result of domestic violence were alcohol abuse (adjusted relative risk, 3.6; 95 percent confidence interval, 2.2 to 5.9); drug use (adjusted relative risk, 3.5; 95 percent confidence interval, 2.0 to 6.4); intermittent employment (adjusted relative risk, 3.1; 95 percent confidence interval, 1.1 to 8.8); recent unemployment (adjusted relative risk, 2.7; 95 percent confidence interval, 1.2 to 6.5); having less than a high-school education (adjusted relative risk, 2.5; 95 percent confidence interval, 1.4 to 4.4); and being a former husband, estranged husband, or former boyfriend (adjusted relative risk, 3.5; 95 percent confidence interval, 1.5 to 8.3).
Conclusions Women at greatest risk for injury from domestic violence include those with male partners who abuse alcohol or use drugs, are unemployed or intermittently employed, have less than a high-school education, and are former husbands, estranged husbands, or former boyfriends of the women.
As part of the Violence against Women Act of 1994, a panel created by Congress noted that "the development of effective preventive interventions will require a better understanding of the causes of violent behavior against women" and recommended "the examination of risk factors."4 We examined socioeconomic and behavioral characteristics of women and their male partners to identify risk factors for injury to women as a result of domestic violence.
Methods
Study Design
We conducted a multicenter case–control study at eight large, university-affiliated emergency departments throughout the United States. The emergency departments were in suburban, urban, and inner-city areas and had a diverse clientele, ensuring a broad representation of women with acute injuries due to domestic violence. The institutional review board at each study site approved the study. All subjects provided oral consent, and all were guaranteed anonymity. The use of written consent was avoided in order to prevent a link identifying subjects to the questionnaires administered.
Study Population
The study population consisted of intentionally injured women and a control group of women who had not been intentionally injured who sought medical care at the participating emergency departments during the study period and who had current or recent (within one year) male partners. We defined an intentionally injured woman as any woman, 18 to 64 years of age, who had been assaulted and injured within the preceding two weeks by a current or recent male partner (boyfriend, husband, ex-boyfriend, or ex-husband). Specially trained physicians or research assistants identified eligible women with use of a standardized questionnaire administered to women with a history of trauma or signs of injury. The questionnaire was designed for use in emergency departments to identify episodes of domestic violence.7,8 Eligible women were included only if they reported or acknowledged being physically assaulted by their male partners. Women were excluded if they had also been sexually assaulted.
The control group was made up of women 18 to 64 years of age who were seen in the emergency department and who were selected in order to represent the distribution of study variables in the source population. At each study site, for periods of up to 15 months, research assistants identified, selected, and interviewed all consecutive eligible female patients. Women without a current or recent (within one year) male partner and women with a history of injury from domestic violence within the preceding year were excluded. Emergency-department–based controls were considered more appropriate than population-based controls because some victims of domestic violence in the source population may not have been treated for their injuries.9,10,11 This method of nonrandom sampling was preferable to population-based sampling because it considered the selection factors that brought the controls to the emergency departments.9,10,11,12,13,14 The spectrum of initial diagnoses among the control patients was very wide; we did not document their final diagnoses.
Variables
The study variables were selected on the basis of findings from our preliminary study15 and prior biomedical, psychological, sociological, and population-based investigations.16,17,18,19,20,21,22,23,24,25,26,27,28,29 Variables in the analysis were classified on the basis of a woman's responses to specific questions concerning herself and her partner. Spousal and next-of-kin surrogates have been shown to be accurate sources of information in case–control studies.30,31,32 Study site was included in the statistical analyses so as to limit the possibility of confounding by this variable.
Definition of Variables
The male partner was categorized as a husband, boyfriend, or former partner. A former partner was defined as a former husband, estranged husband, or former boyfriend.
The partner was categorized as working full time; working part time, defined as working at a regular job that was less than full time; working intermittently (e.g., as a day laborer); long-term unemployed; or recently unemployed. The woman's employment status was not assessed.
A partner's drug use was based on the woman's response to the following question: "During the last year, has your partner used illegal drugs?" Alcohol use by both the woman and her partner was categorized according to an abridged version of the Alcohol Use Disorders Identification Test,33 which included only the three questions on the frequency and quantity of alcohol consumption. The questions, responses, and scores assigned to the responses were as follows: "During the past year, how often did you (or your partner) have a drink containing alcohol?" The possible responses were never (a score of 1), monthly or less (2), two to four times a month (3), two or three times a week (4), or four or more times a week (5). "How many drinks containing alcohol do you (or does your partner) have on a typical day when you are (or your partner is) drinking?" The possible responses were never drink (a score of 1), 1 or 2 drinks (2), 3 or 4 drinks (3), 5 or 6 drinks (4), 7 to 9 drinks (5), or 10 or more drinks (6). "How often do you (or does your partner) have six or more drinks on one occasion?" The possible responses were never (a score of 1), less than monthly (2), monthly (3), weekly (4), or daily or almost daily (5). A total score of more than 8 of a possible 16 points was categorized as representing alcohol abuse, and a total score of 8 or less was categorized as not representing alcohol abuse. The wide range that we used to define alcohol abuse reflects a wide range of types of drinking behavior that may not be classified as abusive by others. The three questions were also evaluated independently as variables indicating alcohol use. In an independent evaluation, these three questions were found to be valid for use as a primary care screening test for excessive drinking and alcohol abuse.34
Description of Injuries
We recorded the type, location, and severity of the injuries sustained by the women. We documented the use of weapons to inflict the injuries but not the treatment given, the management of the case by the emergency department, or the long-term outcomes.
Data Collection
Data were collected for periods of 3 to 15 months, depending on the study site, from July 1997 through September 1998. Information regarding the variables was collected from all the women at the time of their medical evaluation with use of a standard data-collection instrument with structured and closed-ended questions.35,36 We did not review the women's charts. This approach enhanced the quality of information obtained from both groups of women.10,11 Interviewers at each study site were instructed with respect to selection criteria, interviewing techniques, and use of the questionnaire.
Statistical Analysis
We first examined the distributions of the socioeconomic and demographic variables in the two groups of women and the percentage of missing values for each variable.37 We used cross-tabular univariate analyses to estimate crude odds ratios for each categorical and ordinal variable.37 We used cross-tabular stratified analyses to identify confounding and interactions among the variables.9 We used multiple logistic-regression analyses to estimate adjusted odds ratios.9,38 We then conducted sensitivity analyses to assess the variability of the estimates of the odds ratios depending on the inclusion and exclusion of different variables and product terms in the models.9,38 We used odds ratios to estimate relative risks.9,38 We used residual analyses to assess the fit of the multiple logistic-regression models.38 The selected model included terms representing all the variables for the women's characteristics and those of their partners. All P values are two-sided.
Results
At the eight emergency departments, 282 intentionally injured women and 749 control women were identified as eligible to participate in the study, and 256 (90.8 percent) and 659 (88.0 percent), respectively, agreed to participate. The age and race or ethnic group of women who declined to participate were similar to those of the women who agreed to participate. We did not record specific reasons for nonparticipation. Among the various study sites, the number of intentionally injured women ranged from 12 to 61, and the number of controls ranged from 20 to 244. The percentages of missing values ranged from 0.2 percent to 6.0 percent (average, 2.0 percent). The base-line characteristics of the women and their partners are presented in Table 1.
|
|
|
The other characteristics of the women that we examined were not significant risk factors. In particular, alcohol abuse among the women was not a significant factor in the multivariate analysis, apparently because of confounding by alcohol abuse among the partners.
For the partners of intentionally injured women, factors that were associated with an increased risk of inflicting injury as a result of domestic violence included alcohol abuse (adjusted relative risk, 3.6; 95 percent confidence interval, 2.2 to 5.9), drug use (adjusted relative risk, 3.5; 95 percent confidence interval, 2.0 to 6.4), intermittent employment (adjusted relative risk, 3.1; 95 percent confidence interval, 1.1 to 8.8), recent unemployment (adjusted relative risk, 2.7; 95 percent confidence interval, 1.2 to 6.5), and having less than a high-school education (adjusted relative risk, 2.5; 95 percent confidence interval, 1.4 to 4.4). The race or ethnic group of the partner was not associated with the risk of inflicting injury as a result of domestic violence.
The adjusted relative risks of injury from domestic violence according to the responses to the three questions on alcohol use by both the women and their partners are presented in Table 4. For the women, only the response to the question about the number of drinks containing alcohol that were consumed on a typical day during periods of drinking was significantly associated with the risk of injury from domestic violence. The estimates, however, may be imprecise because of the very few subjects in some of the categories. For male partners, all three measures of alcohol consumption were associated with the risk of inflicting injury as a result of domestic violence.
|
Discussion
In a study at eight university-affiliated emergency departments, we examined risk factors for injury to women as a result of domestic violence. We placed special emphasis on evaluating the effect of alcohol use by both the woman and her male partner. Despite the strong association between alcohol use and injury from domestic violence in biomedical, pharmacologic, psychological, and sociological studies,21,26,39,40,41,42,43,44 there is still controversy about the precise effect of alcohol use in this setting.
We found that alcohol abuse by men was associated with an increased likelihood of inflicting injury as a result of domestic violence. In addition, there was a clear dose–response effect for the three measures of alcohol consumption that we evaluated. The precise mechanism by which alcohol acts to increase the risk of injury from domestic violence is not clear, but physiologic, psychological, and environmental factors may all be important. In addition, many women in our study noted that their partners did not abuse alcohol and had not been drinking just before the assault. Thus, alcohol use by men cannot account for all injuries from domestic violence and cannot be considered either a necessary or a sufficient condition for domestic violence.
More important, the use of alcohol does not explain the underlying pathological relationship between intimate partners that results in physical assault. This relationship has been characterized as one in which there is an imbalance of power, with one partner exerting coercive control. Psychologists and sociologists believe that this type of asymmetric power structure is the main determinant of conflict in intimate relationships.22,27,28,29,45,46 Alcohol use by men, therefore, may increase the risk that such men will assault their partners. If this is the case, then decreasing the use of alcohol may reduce, but not eliminate, this risk.47,48
The association between alcohol use and domestic violence was not as apparent for women as it was for men. Specifically, the effect of alcohol abuse by women appeared to be confounded by alcohol abuse by their partners. This is consistent with the finding of a high degree of concordance between spouses with respect to alcohol dependence and heavy drinking.49 In addition, women may turn to alcohol use in response to the recurrent episodes of physical and emotional abuse inflicted by their partners.
Drug use by men was also associated with an increased risk of inflicting injury as a result of domestic violence. This finding corroborates those of earlier studies.20,26,50 The women in our study reported a much lower frequency of drug use than of alcohol use by their partners just before the assault. We did not assess the effect of specific types or quantities of drug use.
We also found that intermittent employment and unemployment (both recent and long term) of the partner were risk factors. Possibly, the stress of finding work or of unemployment (alone or in combination with other factors) increases the risk that a man will physically abuse his partner.
Lower levels of education for men were also associated with an increased risk of inflicting injury as a result of domestic violence. This factor may be closely associated with other risk factors in men, such as unemployment or alcohol abuse. Low levels of education may also be an indicator of poor communication skills, which have likewise been linked to a history of domestic violence among men.46
The man's status as a partner was also associated with the risk of inflicting injury as a result of domestic violence. Specifically, women with a former partner appeared to be at greater risk for injury, and the risk increased substantially if the woman was still living with her former partner.
One limitation of our study is the possibility of selection bias. Although the control women were selected from the same emergency-department populations as the intentionally injured women, factors related to injury from domestic violence may have influenced their selection. To limit this type of bias, the controls were selected from among all eligible female patients in the emergency department (without regard to the reason for their visit), so that no one disease was unduly represented in the control group.10,11 Selection bias was also limited by the use of the same criteria for eligibility and exclusion for the two groups, by the high degree of participation in both groups of eligible women, and by the absence of demographic differences between participants and nonparticipants.9
Misclassification may also have biased our results. Since most of the study variables (e.g., age, race or ethnic group, level of education, income, and employment status) were not subject to substantial errors in recall or intentional misclassification, the degree of bias resulting from misclassification for these variables is likely to be very low. Misclassification of alcohol use and drug use was more likely. The assessment of alcohol use may have been inaccurate, depending on which group a woman was in, resulting in a misclassification bias toward or away from the null effect. For example, as compared with the controls, intentionally injured women may have underreported their alcohol consumption to deflect any assumptions about their responsibility for the domestic violence. These women may also have overreported alcohol consumption or drug use by their partners. Since intentionally injured women may be more concerned than other women about their partners' alcohol and drug use, recall bias may have resulted in an overestimation of the effect of these factors.
Potential misclassification of the variables was limited by our use of a validated questionnaire. We limited misclassification of intentionally injured women by including only women who reported or acknowledged being injured as a result of domestic violence. We limited misclassification of controls by excluding women with a history of physical abuse.
Because we examined potential risk factors among women who sought care in eight geographically diverse emergency departments that served a broad variety of people, our findings can be extrapolated (at least to a limited degree) to women who seek care in emergency departments. Even with this degree of heterogeneity, however, the women we studied may not represent victims of domestic violence in general. Risk factors for injury from domestic violence may differ greatly among women with higher socioeconomic status and women with injuries that do not require emergency medical care. This uncertainty makes it difficult to generalize our findings.
Our findings underscore the multifactorial nature of injuries from domestic violence. Future studies should consider additional factors, including psychological disorders, physiologic disorders, and social-learning determinants, such as a history of child abuse or witnessing domestic violence as a child. They should also include an adequate period of follow-up and an assessment of the past experiences of both the women and the men involved.
Supported by a Career Development Grant from the Emergency Medicine Foundation and by financial support and logistical assistance from the UCLA Southern California Injury Prevention Research Center (R49/CCR 903622-08).
We are indebted to Kathryn Brown Schaffer and Ani Grigorian for their assistance with this project.
Source Information
From the Department of Emergency Medicine, Olive View–UCLA Medical Center, Sylmar, Calif. (D.N.K.); the Southern California Injury Prevention Research Center, Department of Epidemiology, UCLA School of Public Health, Los Angeles (D.N.K., J.F.K.); the Department of Emergency Medicine, Los Angeles County/University of Southern California Medical Center, Los Angeles (D.A.); the Division of Emergency Medicine, Southwestern Medical Center, Dallas (E.T.); Emergency Medical Services, New York University–Bellevue Hospital, New York (S.S.); the Department of Emergency Medicine, University of Mississippi Medical Center, Jackson (T.T.); the Department of Emergency Medicine, Boston Medical Center, Boston (J.A.L.); the Department of Emergency Medicine, Truman Medical Center, Kansas City, Mo. (R.M.); and the Department of Emergency Medicine, Brigham and Women's Hospital, Boston (E.B.).
Address reprint requests to Dr. Anglin at the Department of Emergency Medicine, Los Angeles County/University of Southern California Medical Center, Rm. 1011, 1200 N. State St., Los Angeles, CA 90033.
References
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Domestic Violence
Zalar R. W., Harris R. B., Kyriacou D. N., Anglin D., Minow M.
Extract |
Full Text
N Engl J Med 2000;
342:1450-1453, May 11, 2000.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
