Demographics and Fracture Patterns of Patients Presenting to US Emergency Departments for Intimate Partner Violence

Introduction: Orthopaedic surgeons are in a position to assist in identifying intimate partner violence (IPV) patients. It was the purpose of this study to analyze the demographics and fracture patterns of IPV patients in the United States. Methods: Data from the National Electronic Injury Surveillance System All Injury Program from 2005 through 2013 were used. Injuries due to domestic violence were identified, and statistical analyses accounted for the weighted, stratified nature of the data. Results: There were 1.65 million emergency department visits over nine years for IPV. The median age was 29.8 years, 83.3% were women, and 55.3% occurred at home. The major diagnoses were contusion/abrasions (43.4%), lacerations (16.9%), strain/sprains (15.6%), internal organ injuries (14.4%), and fractures (9.7%). The most common fracture involved the face (48.3%), followed by the finger (9.9%), upper trunk (9.8%), and hand (6.4%). The highest proportion of lower extremity fractures occurred in men, and upper extremity fractures increased with increasing age. The odds of fracture in an IPV patient were greatest in those sustaining an upper extremity injury (odds ratio [OR] = 6.62), lower extremity injury (OR = 6.51), upper trunk injury (4.28), and head/neck injury (OR = 3.08) compared with a lower trunk injury (referent), and women (OR = 1.80) compared with men (referent). Older patients sustaining IPV had higher odds of a fracture (the few patients 10-14 and >65 years old were excluded from this analysis). Conclusions: As this study encompasses the entire United States, these results are germane to all US orthopaedic surgeons. Knowing typical fracture patterns/locations is helpful in identifying IPV patients, although the victim may not fully divulge the history and details of the event. Identification is important for the physical and mental health of the victim, and abuse often continues if intervention does not occur. The odds of a fracture in an IPV patient are greatest when the injury involved the extremities and increased with increasing age of the patient.

abuse often continues if intervention does not occur. The odds of a fracture in an IPV patient are greatest when the injury involved the extremities and increased with increasing age of the patient. I ntimate partner violence (IPV) is a serious public health issue 1,2 and of notable concern to the orthopaedic surgeon [3][4][5][6][7][8][9][10] as well as all health care providers. 11,12 The recognition of IPV as a substantial problem among patients with orthopaedic injuries has been acknowledged by both the American Academy of Orthopaedic Surgeons (AAOS) 13 and the Canadian Orthopaedic Association. 14 Orthopaedic surgeons are positioned to identify IPV patients because of their involvement in the treatment of fractures, often as the initial provider outside of the emergency department (ED). This affords the orthopaedic surgeon the ability to provide appropriate care, but also referrals for other services to prevent additional harm. Although there is an increasing body of literature on the subject of IPV to orthopaedic surgeons, [3][4][5][6][7][8][9]15 there is little that describes the demographic characteristics, and especially fracture patterns, in IPV victims. 16,17 It was the purpose of this study to analyze the demographics of IPV patients presenting to the EDs in the United States, especially focusing on injury and fracture patterns. Such knowledge can assist orthopaedic surgeons in identification of these patients, in addition to questioning techniques and training programs. 7,[18][19][20][21][22][23]

Methods
The data for this study come from the National Electronic Injury Surveillance System (NEISS) All Injury Program (AIP). The NEISS is a data set managed by the US Consumer Prod-uct Safety Commission (USCPSC) which collects injury data from 100 hospitals in the United States and its territories having an ED. It was initially designed for injuries due to consumer products. However, not all injuries are from consumer products; thus, the USCPSC selected 65 of these hospitals to obtain data for all injuries, regardless of the association with consumer products. This has been designated as the All Injury Program (AIP). These data are in the public domain, housed by the Interuniversity Consortium for Political and Social Research (ICPSR), and can be accessed at https://www.icpsr. umich.edu/icpsrweb/ICPSR/search/ studies?q=all1injury1program. Use of this publicly available deidentified data was considered exempt by our local Institutional Review Board.
The database includes date of ED visit, sex/race/age of the injured patient, diagnosis, disposition from the ED, incident locale, body part injured, perpetrator and type of assault, and hospital size (strata). Age is also categorized into 12 different groups (#4, 5 to 9, 10 to 14, 15 to 19, 20  Those who were sexually assaulted were on the average younger (28.5 versus 32.7 years-P , 10 24 ) and nearly always female (99.6% versus 82.8%-P , 10 24 ). Although the majority of the patients were white (52.3%), the percentage of white patients was greater for the sexual assault group (66.4% versus 52.0%-P = 0.015). The most common anatomic area of injury for the sexual assault group was the lower trunk (81.7%) and the head/neck for the nonsexual assault group (81.7%) (P , 10 24 ). There were no differences in any of the other demographic variables between the sexual and nonsexual IPV groups. There were many differences by sex; however, since most IPV patients were women, detailed analyses by sex are given in Supplemental File 2, http://links.lww. com/JG9/A64. The major differences were that women (1) were younger than men (31.6 versus 33.4 years), (2) more frequently white (54.6 versus 41.1%), (3) sustained more injuries to the head/neck (60.5% versus 49.7%), and (4) less commonly admitted to the hospital (3.4 versus 6.4%). Regarding race (in addition to the differences by assault intent and sex noted above), there was an increasing percentage of white patients and concomitant decreasing percentage of black and Amerindian patients with increasing age (Figure 1). Detailed analyses by race are shown in Supplemental File 3, http://links.lww. com/JG9/A65. Table 2 shows the differences between the five major diagnoses, which accounted for 97.5% of all the injuries. These five diagnoses were contusions/abrasions (43.4%), lacerations (16.9%), strains/sprains (15.6%), internal organ injuries (14.4%), and fractures (9.7%). Lacerations were more frequent in men and blacks (Figure 2, A). Patients sustaining fractures and internal organ injuries were more commonly admitted (Figure 2, B).  Differences by race and age group (P , 10 24 ) in IPV patients. The number of patients is shown in the column boxes. IPV = intimate partner violence.  (Table 4). There were notable differences by sex, age group, race, and type of assault. Men sustained a higher proportion of lower extremity fractures compared with women ( Figure 3, A). Head/neck fractures were more common in the younger age groups and decreased with increasing age; upper extremity fractures increased with increasing age (Figure 3, B). Head/ neck fractures were most common in Amerindian patients with more upper extremity and trunk fractures in white patients (Figure 3, C). A fracture that occurred during a sexual assault was nearly twice as likely to be located in the lower extremity compared with the trunk. Table 5 shows the results of multivariate logistic regression analyses for demographic predictors of a fracture. The parameters included in the model were simple ones that any health care provider would have available from a simple history and included age, sex, race, anatomic area of injury, and type of assault. The odds of a fracture in a patient sustaining IPV were greatest in those sustaining a nonsexual assault (OR = 4.8) compared with a sexual assault (referent); an upper extremity injury (OR = 6.62), lower extremity injury (OR = 6.51), upper trunk injury (4.28), and head/neck injury (OR = 3.08) compared with a lower trunk injury (referent); whites (OR = 1.33) compared with blacks (referent); and women (OR = 1.80) compared with men (referent). Older patients sustaining IPV had higher odds of a fracture (45 to 54 years, OR = 2.07; 55 to 64 years, OR = 1.96), ( Finally, we studied temporal variation. There was a notable difference in the month of ED visit between the sexual and nonsexual assault IPV patients (Figure 4, A). The nonsexual assault group demonstrated a mild increase in the summer months. The sexual assault group demonstrated marked peaks in May and August. For both groups, more of the ED visits occurred on Saturday and Sunday (Figure 4, B).   injury in 281 physically abused women; the anatomic location of the injury involved the head/neck in 40%, less than the 58.7% in this study. There were a total of 39 fractures, or 27% of the injuries, which is much higher than the 9.7% in this study (Table 2). This could be due to the fact that the patients in their study had already been referred to a nonprofit organization (Domestic Abuse Project). We surmise that the simpler strains/sprains and contusions/abrasions might not have been deemed appropriate for referral. The location of the 39 fractures in their study was the head/neck in 17 (44%), upper extremity in 13 (33%), lower extremity in 1 (3%), and trunk in 8 (20%). This is different than the results in this study, where the head/neck comprised 49.4% of the fractures, upper extremity 29.0%, lower extremity 9.8%, and trunk 12.0%. Spedding et al 26 in the United Kingdom studied 103 female assaults due to domestic violence; of these 103, fractures occurred in 18%, more than the 9.7% in this study. The fractures in their study were located in the head/neck in 4 (21%), upper extremity in 9 (47%), and trunk in 4 (21%). Interestingly, five of the 103 women initially stated they had fallen down the stairs, but later volunteered that domestic violence was the cause of the injury. Thus, knowing typical fracture patterns/ locations is helpful in identifying IPV patients, although the victim may not fully divulge the history and details of the event. Identification of IPV is important for many reasons. The first is the physical and mental health of the victim, 12 as well as the mental health of the abuser and both the physical and mental health of children in the relationship. Abuse often continues if intervention does not occur, 2,27 and such violence can potentially result in homicide and/or suicide, not only to the victim but also children of the IPV victim, 28 as well as cruelty to animals. 29 Financial costs to society are also increased, as IPV patients consume more health care resources than those without IPV. 12,30,31 One in 50 women presenting to an orthopaedic fracture clinic is a victim of IPV, 10 with 64% sustaining fractures. Owing to this high prevalence of IPV-related injuries seen in orthopaedic clinics, programs are now becoming established to guide Differences in fracture location in IPV patients. The number of patients is shown in the column boxes. A, By sex (P = 0.0032). B, By age group (P , 10 24 ). C, By race (P = 0.044). IPV = intimate partner violence.
practitioners in appropriate questioning techniques regarding the potential of IPV. 3,20,22,[32][33][34][35][36] Orthopaedic surgeons are thus in a unique position to identify IPV victims. However, it is well known that an IPV victim may not always disclose the abuse/violence. 18 Understanding the demographics and fracture patterns of IPV victims is additional knowledge that orthopaedic providers can use when discussing the possible issue of IPV with orthopaedic patients. Resources that are available to orthopaedic surgeons if an IPV victim is identified are given in a recent review. 3 A telephone hotline, for both the United States and Canada, is 1-800-799-SAFE, which can be given to a patient and they will immediately assist the patient.
This study found that the odds of a fracture in an IPV patient are greatest when the injury involved the extremities (OR = 6.62 for the upper extremity and 6.51 for the lower extremity), followed by the upper trunk (OR = 4.28) and head/neck (OR = 3.08), compared with the lower trunk (reference). The odds of fracture increased with increasing age of the IPV victim, were greater in women (OR = 1.8) compared with men (reference), and were greatest in white, compared with Amerindian or black IPV patients. It has been previously noted that injuries to the head/neck are the most common in IPV patients. 17,25,37 We noted the same in this study, where 58.7% of all IPV injuries involved the head/neck. Similarly, 48.3% of the fractures occurred in the face. Although the absolute number of fractures involved the face, the odds of a fracture are twice greater if the IPV victim sustained an injury to the extremities compared with the face. This new information is especially important to the orthopaedic surgeon when evaluating a fracture patient where the history seems suspect and could guide the orthopaedic surgeon to further explore/discuss the injury circumstances with the patient.
Most IPV patients in this study were women; however, men can also be victims of IPV. In this study, 16.7% of the IPV patients were men (0.4% of the sexual assaults and 17.2% of the nonsexual assaults). Men were more likely to sustain lacerations compared with other injuries (Table 2 and Figure 2, B). This 16.7% is very similar to the 17% male IPV victims in a study of 29 patients from Finland. 38 However, our prevalence of fracture was less in men than women. When men did sustain fractures, they more commonly occurred in the extremities compared with women ( Figure 3, A).
The major strength of this study is that it encompasses the entire United States, all ages, racial groups, and both sexes for patients visiting the ED for IPV injury care. As such, it reflects the entire US cohort, and these results are germane to all orthopaedic surgeons whether practicing in an urban or rural location or an academic or private situation. Many of the previously published studies regarding IPV originate from academic institutions and may not be applicable to the entire US cohort or the private practice orthopaedic surgeon.
Limitations of this study must be acknowledged. Large data sets inherently possess inaccuracy. However, the NEISS data collection protocols have 89% to 98% accuracy. 39,40 Second, the NEISS only captures those Assault, confirmed, or suspected: Injury from an act of violence where physical force by one or more persons is used with the intent of causing harm, injury, or death to another person or an intentional poisoning by another person. This category includes perpetrators as well as intended and unintended victims of violent acts (e.g., innocent bystanders). This category excludes unintentional shooting victims (other than those occurring during an act of violence), unintentional drug overdoses, and children or teenagers "horsing" around.
Assault-sexual: An assault as defined above that also involves • the use of physical force to compel another person to Temporal variation in IPV ED visits. A, By month. The differences between the sexual and nonsexual assault IPV patients were highly significant (P , 10 24 ). B, By weekday. There were no notable differences between the sexual and nonsexual assault IPV patients. ED = emergency department, IPV = intimate partner violence.
engage in a sexual act against his or her will, whether the act is completed or not, attempted or completed sex act involving a person unable to 1) understand the nature of the act, 2) decline participation, or 3) communicate unwillingness to participate for whatever reason. • abusive sexual contact: intentional touching, either directly or through the clothing, of the genitalia, anus, groin, breast, inner thigh, or buttocks of any person against his or her will or of a person who is unable to consent (e.g., because of age, illness, disability, or the influence of alcohol or other drugs) or refuse (e.g., due to the use of guns or other nonbodily weapons, or due to physical violence, threats of physical violence, real or perceived coercion, intimidation or pressure, or misuse of authority). This category includes rape, completed or attempted; sodomy, completed or attempted; and other sexual assaults with bodily force, completed or attempted.