Intimate partner violence in orthopaedic trauma patients

Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: http://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.


Introduction
Injuries are the leading cause of death for females from one to thirty-four years old, and they are a major source of preventable morbidity and mortality in middle-aged and elderly women 1,2 . Motor vehicle-related injuries, falls, and violence are the most serious injury problems for women. Although morbidity is far greater than mortality, access to information about nonfatal injuries is extremely limited. Domestic violence is the most common cause of nonfatal injury to women in North America 1,2 . More than four in every ten women in North America are likely to have experienced one or more forms of violence, including child abuse (17.8%), physical assault (19.1%), rape (20.4%), and intimate partner violence (34.6%) 3 .
Intimate partner violence is a pattern of assaultive and coercive behaviors, including physical, sexual, and psychological attacks as well as economic coercion that men or women typically use against partners. In a review of women seen at a Minneapolis-based therapy and advocacy program for victims of intimate partner violence, investigators identified 144 injuries in 218 physically abused women 4 . The most prevalent injuries resulting from intimate partner violence were head and neck injuries 40%), and the second most common manifestation was musculoskeletal injuries (28%). The spectrum of injuries included sprains (twenty-one injuries), fracture-dislocations (seventeen injuries), and foot injuries (two injuries)all of which would have necessitated referral to an orthopaedic surgeon 4 .
Women who are victims of intimate partner violence have been known to use health services at higher rates than women who have not been abused 4,5 . Health-care providers of all specialties are in a unique position to identify and assist victims of intimate partner violence during routine or annual office visits. The prevalence of intimate partner violence in women presenting to emergency departments has been reported to range from 14% to 41% [5][6][7][8][9] ; however, only 12% to 17% of abused women have their experiences documented in a medical chart 6,8,10,11 .Women who have experienced intimate partner violence are seldom identified by emergency department physicians 11 . To our knowledge, there have been no reports on the incidence of intimate partner violence among women attending orthopaedic fracture and injury clinics.
Given the infrequent identification of intimate partner violence by emergency or primary care physicians 7 , orthopaedic surgeons and orthopaedic clinic personnel have a second opportunity to identify and provide necessary referral to a local agency or hospital service. Several specialty associations, including the American Medical Association, the American College of Obstetricians and Gynecologists, the American Academy of Pediatrics, the American College of Emergency Physicians, and the U.S. Centers for Disease Control and Prevention, have advocated screening for intimate partner violence. The American Academy of Orthopaedic Surgeons (AAOS) has recommended that ''orthopaedic surgeons should be aware of how harm from domestic violence and abuse may present to them.'' 12 We aimed to identify the perceptions, attitudes, and knowledge of Canadian orthopaedic surgeons about intimate partner violence.

Survey Methodology
The purpose of the survey was to assess the perceptions and knowledge of practicing orthopaedic surgeons about intimate partner violence. We performed a pilot test of a survey designed by members of the Violence Against Women Health Research Collaborative. Our 17 responsible for the abuse; and 20%, that victims have personalities (passive and/or dependent) that predispose them to abuse. Surgeons remained uncertain about whether some patients may get angry if asked about intimate partner violence in the clinic (especially if they were not being abused); about whether patients who do not openly tell their surgeon do not believe that it is the surgeon's business to know; and about whether it was their role as a health-care provider to interfere with the personal conflicts of a patient. (Table 3) Many Canadian surgeons (30%) were concerned about their personal safety when confronting a batterer in the clinic. In general, we identified considerable variability and a lack of consensus in responses (agree, disagree, and unsure) with regard to fear of placing the victim in greater danger by inquiring about abuse in front the batterer, the use of strategies to assist batterers with seeking help, and fear of retaliation by the batterer toward the health-care staff.

Roles and Need for Training
By the completion of the survey, the majority (91%) of Canadian surgeons believed that knowledge about intimate partner violence was relevant to their surgical practice. Only 9% of the respondents reported some training in intimate partner violence-related issues and 30% supported educational programs for orthopaedic surgeons.

Discussion
In this survey of 186 Canadian orthopaedic surgeons, our findings suggested that (1) surgeons believe that the victims of intimate partner violence are rarely seen (a prevalence of <1%) in their practice setting, (2) surgeons continue to have misperceptions about victims and batterers, and (3) surgeons believe that intimate partner violence is relevant to their field, but only one in three support formal educational programs for orthopaedic surgeons.
Contrary to the perception of the surgeons that intimate partner violence is rare, several investigations have shown an alarmingly high incidence in the community. In Canada, an estimated 653,000 women are involved in an abusive relationship yearly 15 . Among the women who reported such abuse, 54% stated that the violence was not an isolated incident but rather a recurring event in their lives 14 . Women in Canada are also more likely to report severe injuries, such as fractures, than are male victims of domestic violence (20% and 0%, respectively) 14 .
Estimates have ranged from 960,000 incidents of violence against a current or former spouse, boyfriend, or girlfriend per year to 3.9 million incidents of rape and/or physical assault on women by an intimate partner annually 16,17 . Nearly one third of American women (31%) have reported being physically or sexually abused by a husband or boyfriend at some point in their lives, according to a 1998 Commonwealth Fund survey 18 . While women are less likely than men to be victims of violent crimes overall, women are five to eight times more likely than men to be victimized by an intimate partner 16 . In one report, domestic violence often resulted in severe injury; 28% of the women interviewed required hospitalization for injuries, 13% required major surgical treatment, and about 40% had previously required medical care for abuse 15 . From 1987

Introduction
Intimate partner violence (IPV) is a serious public health concern that is receiving increasing attention in medical research 1 . The definition includes physical, sexual and/or psychological/ emotional forms of abuse between past or present heterosexual or homosexual partners 2 . Intimate partner violence occurs across all racial, ethnic, regional, and socioeconomic boundaries 3 . Women are more likely than men to be victims of IPV 4 , and it is estimated that one in four American women have been victims of IPV in their lifetime 5 . Richardson et al. found that only 17% of physical abuse victims have ever had it documented in a general practice medical chart, highlighting the serious problem of underreporting of IPV in healthcare 6 .
Intimate partner violence victimization has been reported to impact health and lead to increased use of healthcare services 7,8 . Intimate partner violence has been linked to mental health disorders such as depression, suicide, and post-traumatic stress disorder 8 . In a large multinational study by the World Health Organization (WHO), 24 000 women in 10 countries were interviewed about their experiences and beliefs surrounding IPV 9 . The study found that for all settings combined, women who reported physical violence at least once in their lifetime reported significantly more emotional distress, suicidal thoughts, and suicide attempts than non-abused women 7 .( Victims of IPV have a 50 to 70% higher chance of having gynaecological, central nervous system, and stress-related health problems 10 . Bonomi et al. found that currently or recently physically abused women have higher total annual health care costs and use more emergency, hospital outpatient, primary care, pharmacy, and specialty services than non-abused women. Mental health service utilization was found to be higher among women abused both physically and non-physically 10 . It is evident that the identification and treatment of IPV victims is highly relevant to healthcare, in which physicians have a key role to play. A national survey of US medical students identified, 91% of senior students as having had training on IPV, but only one third feeling highly confident in having discussions about IPV with patients 11 . We conducted a survey with the primary aim of determining medical students' and surgical residents' attitudes, beliefs, and perceptions regarding IPV screening, victims, and perpetrators. Secondary aims include examining the level of IPV education/training medical students and surgical residents have received, and exploring how gender and level of education (resident vs. medical student) are related to perceptions of IPV.

Survey Instrument
Due to the lack of literature on the views and/or knowledge of medical students and surgical residents regarding IPV, we chose to use a modified version of the Provider Survey for our study. The Provider Survey is an instrument intended to measure healthcare providers' attitudes, beliefs, and self-reported behaviours related to the identification and management of IPV. The Provider Survey is reliable and has been proven valid 12 . Wording modifications were made to the survey to make the questions applicable to medical students and surgical residents. Two versions of the survey were developed; one for medical students and the other for surgical residents. In addition to the Provider Survey, participants were asked to complete questions on their demographics as well as their current perceptions, knowledge and education on IPV. These questions were modified from ones used in recent IPV surveys of medical students, Canadian Orthopaedic Association members and chiropractors 11,13,14 . 29 There were 23 items in the medical student version of the survey, and 30 items in the surgical resident version. The residents' survey included questions relevant to their current and previous practice. The medical students' survey did not include these questions due to their lack of clinical practice experience. Questions were primarily either multiple choice or presented as a series of statements with an associated Likert Scale ranging from strongly disagree to strongly agree. Items were grouped into three categories: 1) demographic information, 2) attitudes, knowledge and education, and 3) clinical relevance of IPV.

Sampling Frame
The sampling frame included all medical students and surgical residents currently enrolled at McMaster University, Hamilton, Ontario. We chose to include surgical residents because of our interest in promoting IPV screening in surgical programs. We are unaware of any literature evaluating the attitudes of surgical residents toward IPV. We chose to exclude attending surgeons in our study because the attitudes of attending surgeons have been previously documented 13 . E-mail lists of McMaster University medical students and surgical residents from all years of study were obtained with permission from McMaster University's Undergraduate Medical Program Office and contacts in the Department of Surgery Residency Programs. Surveys were not sent to students or residents studying outside of McMaster University.

Survey Administration
We used SurveyMonkey, online survey software, to administer the survey and its cover letter in electronic form. We chose SurveyMonkey because it is easy to use for both administrators and participants. Following the initial emailing, three rounds of follow-up emails were sent out to the students. Participants were provided with the opportunity to withdraw at any time.

Statistical Analysis
For statistical significance to be reached, 193 trainees were needed for the study sample size. This was based on a population of approximately 700 medical students and surgical trainees at McMaster University, with an error level of 6% and a 95% confidence interval (http://www.custominsight.com/articles/random-sample-calculator.asp).
Survey data were analyzed using PASW version 18.0 (Chicago, IL). Descriptive analyses, including frequency counts and percentages, were performed for all collected data. We conducted Chi-squared tests to determine if there were differences in responses between the surgical residents and medical students using the Contingency Table Calculator 15 . We also conducted a subgroup analysis looking at differences in responses between males and females using chi-squared tests. Surveys with missing data were included in the analysis.

Response Rate
Two hundred trainees responded (29%), meeting the sample size requirements for this study. The response rate for medical students was 23% (127/542) and 49% (73/150) for surgical residents. No information was available about non-respondents, so we are unable to evaluate differences between those who did and did not participate. The survey was administered in the winter term. The lower medical student response rate may be due to possible interference with 30 exam time. Fifteen medical students and eleven surgical residents started the survey but did not complete it. All data collected from incomplete surveys was used in the analysis. There were no withdrawals from the study.

Respondent Characteristics
Respondents ranged in age from 20 to 45 years (mean age = 26± 4.5 years). The majority of the respondents were female (58.3%), which is approximately representative of the population of medical students and surgical residents at McMaster University, and 5% reported a history of IPV (personal history or family history). The top intended specialties for medical students were family medicine (30.8%) and surgery (12.3%). Over two thirds of the surgical residents were specializing in orthopaedics (41.1%) or general surgery (27.1%) ( Table 1).

Misperceptions about IPV
Most respondents (91.2% of medical students and 96.9% of surgical residents) estimated the IPV prevalence in their intended practice to be 10% or less (Figure 1). Respondents held misperceptions about the following issues: 1) victims must get something from the abusive relationships (18.2%), 2) physicians should not interfere with a couple's conflicts (21%), 3) asking about IPV risk offending patients (45%), 4) victims choose to be victims (11.1%), 5) it usually takes 'two to tango' (18.3%), and 6) some patients' personalities cause them to be abused (41.1%) (Appendix A).
Surgical residents were significantly more likely to hold misperceptions about a victim's role in their abuse ('it takes two to tango') compared to medical students (28.2% vs. 12.9%, p=0.038) ( Table 2). Males were significantly less likely to disagree with victim-blaming statements than females, such as "People are only victims if they choose to be" (77.8% vs. 96.2% strongly disagree/disagree, p<0.001) and "Women who choose to step out of traditional roles are a major cause of IPV" (75.3% vs. 93.4% strongly disagree/disagree, p=0.001) ( Table 3). Many respondents were concerned for their personal safety when asking a patient battering (32.6% strongly agree/agree), and nearly one quarter of respondents fear that they will offend patients if they ask about IPV (22.8% strongly agree/agree) ( Table 2). Many respondents has misconceptions about batterers, for example, the abuse would stop if the batterer stopped using alcohol (34.9% agree/strongly agree) (Appendix B).

Barriers to Assessment of IPV
Key barriers to IPV assessment perceived by respondents included lack of time (82.9%), lack of knowledge of what to ask (66.8%), lack of knowledge of community resources (69.9%), and personal discomfort (52.8%) ( Table 2). Medical students were significantly more likely to have issues with lack of time (89.5% vs. 71.0%, p=0.004), lack of knowledge of what to say (73.4% vs. 55.1%, p=0.010), lack of knowledge of what to do if a patient is abused (72.5% vs. 44.9%, p<0.001), and personal discomfort (62.1% vs. 36.2%, p=0.003) when compared to surgical residents ( Table 2). 42% of respondents reported that the main barrier to screening for IPV is a lack of training (Appendix C).

Need for Education and Training
Most respondents identified IPV identification as relevant in their practice (89%) but many were unsure or incorrect about legal reporting requirements (20.0% unsure, 12.1% incorrect) ( Table 4). Almost all respondents had little or no previous IPV training (99%), yet only 75% believed their education was inadequate. Most trainees supported additional training and educational initiatives in IPV (77.2%), and 42% cited their lack of education and training as the primary barrier to routine assessment ( Table 4). Medical students were significantly more likely to report a lack of training (83.9% vs. 60.3%, p=0.001) and desire for increased IPV education (87.9% vs. 58.0%, p=0.062).
Many surgical residents (48.4%) reported that they had identified a victim of IPV and 27.4% of residents reported that they had identified a batterer. Only 9.7% of surgical residents reported that their clinical setting has guidelines for detecting and managing IPV, and over one quarter (25.8%) were unsure if there are guidelines in their clinical setting (Appendix D). 35

Discussion
In this survey of 200 Canadian medical students and surgical residents, our findings suggest that medical students and surgical residents have multiple misperceptions about IPV and have not received adequate training on the identification and treatment of IPV. Most respondents underestimated the IPV prevalence in their intended practice. Despite having misconceptions about IPV and underestimating its prevalence, the majority of respondents believed identifying IPV was very relevant to clinical practice. In addition, most of the respondents acknowledged that their level of training on IPV was inadequate and over three quarters of respondents would like to receive additional education and training on IPV.
Medical student and surgical trainees incorrectly believed that the prevalence of IPV is 10% or less. These results are similar to the findings of two previous surveys of health care practitioners. In a recent survey of orthopaedic surgeons, most respondents indicated that the prevalence of IPV in their practice was rare (<1%) 13 . In a similar study of Ontario chiropractors, the majority of respondents indicated that the prevalence of IPV in their practice was between 0.1% and 1% 14 . In contrast to the survey findings, multiple prevalence studies have shown that the lifetime prevalence of IPV is much higher. For example, several American, Australian, and Canadian studies have found that IPV prevalence is well over 10% in both emergency medicine [16][17][18][19][20] and family medicine [21][22][23][24][25] . In addition, the PRAISE Investigators 26 recently found that the 12 month prevalence of IPV in orthopaedic fracture clinics, one of the most common specialties among our respondents, was over 30%.
Almost half of the respondents held the misconception that patients would be offended if they were asked about IPV. This finding is similar to a survey of Ontario chiropractors, which shows that almost half of chiropractors were afraid of offending patients when asking about IPV 14 . This finding contradicts other research by Hurley et al 27 who found that 86% of men and women presenting to Canadian emergency departments agreed that health care providers should screen for IPV. Similarly, Caralis et al 28 established that the majority of American survey respondents believe doctors should screen for abuse in their practices. In addition, Feder et al. 29 conducted a meta-analysis that showed women who have been abused support screening programs for IPV in a health care setting. Dispelling the misconception that women do not wish to be screened for IPV and providing additional education on IPV could help ensure additional screening for IPV among future health care practitioners.
Education in IPV was valued among respondents; however, most reported feeling that they have received an inadequate level of education and training on IPV and have a desire to receive additional training on the assessment and treatment of IPV. Both of these results were higher for medical students, which we speculate is attributed to their shorter time in the medical curriculum. These results are similar to a survey of American medical students that concluded that despite national interest in IPV issues, efforts in U.S. medical schools to increase IPV screening and prevention have not achieved saturation 11 . Similarly, a recent report from the Association of American Medical Colleges found that 20% of U.S. graduating physicians in 2004 believed that the curriculum time dedicated to IPV was inadequate 30 . Hamberger 31 has suggested that, although most medical schools educate students on IPV in some form, the teaching is mainly done in a basic science module as opposed to in a clinical setting. Edwardsen et al 32 have showed that a structured IPV training program with use of mnemonics and clinical role-playing can help medical students to ask their patients questions about their history of IPV. In addition, Chapin et al revealed that emergency medical personnel who received IPV training 36 from a domestic violence center were better informed about IPV services and the obstacles faced by victims 33 . One clinic-based IPV education program for pediatric residents has increased IPV screening from less than 1% to over 30% 8 months after program completion 34 . We suggest that IPV education be included in both medical and residency training and that it should focus on clinic-based practical exercises.
Nearly one third of respondents were either incorrect or unsure when asked if health care provider reporting of IPV is mandatory in Canada. Reporting of IPV is not mandatory in Canada 34 , but it is mandatory in some American states 35 . Most surgical residents reported screening for IPV only seldom or never, which is consistent with the findings of a recent survey of orthopaedic surgeons 13 and of U.S. medical residents 11 . In addition, it was very uncommon for surgical residents to screen patients with illnesses linked to IPV such as hypertension/ coronary artery disease, irritable bowel syndrome, or headaches, indicating a need for additional education on the identification and screening of IPV. The survey found that residents were more likely to screen patients with injuries for IPV. This is consistent with a study of U.S. residents 36 and of primary care physicians 37 .
Medical students were more likely to report that they felt time constraints, lack of knowledge, and personal discomfort compared to residents. Residents may have had more opportunities to come across abused women due to increased time in clinical settings compared to medical students. Perhaps some of the residents' knowledge comes not from formal academic training, as in medical school, but from experiential learning in a clinical setting.
In previous studies, IPV training has been proven successful in raising awareness as well as improving the ability of healthcare professionals to detect IPV. Warburton et al 38 reported that a brief IPV educational program improved dental hospital staff's attitudes and knowledge about IPV. As well, an educational program for internal medicine residents was both well received and effective at improving detection of IPV victims 39 . Three key questions aimed at raising suspicion of IPV were included in a questionnaire; 54% of the intervention group were able to answer at least two out of three questions correctly compared to 20% of the control group 39 .
This study has provided valuable insight into the knowledge, education, attitudes and perceptions of medical students and surgical residents regarding the topic of IPV. One of the strengths of the study is the survey instrument, which was created by domestic violence experts and validated in IPV surveys of health care providers. It has met basic face and basic content validity, although we did use a slightly modified version of the Provider Survey for students and residents which has not been validated. The survey also had an adequate sample size. However, the study also had some limitations. Our survey had a relatively low response rate, which may be a potential source of bias. It is possible that responders differed in some characteristics compared to non-responders. For example, non-responders could be more likely than responders to have experienced IPV, which would influence many of the results. Our survey was restricted to only medical students and surgical residents from McMaster University who could be contacted via email.
It remains unclear whether our findings are generalizable to other universities and jurisdictions. This study was a descriptive, cross-sectional study that can only be used to identify perceptions and barriers regarding IPV inquiry. Causal inferences about specific variables and outcomes cannot be made. 37

Conclusion
Misconceptions exist among one Canadian medical school's medical students and surgical residents about IPV and may be related to lack of education and low self-efficacy, or possibly to other factors such as gender. Curricula in medical schools and surgical training programs should appropriately emphasize educational opportunities in the area of IPV. Future research should explore the most optimal methods of disseminating IPV information among students and health care providers in order to increase awareness of IPV and reduce these misconceptions. It is anticipated that increased awareness of IPV among health care providers will motivate them to seek additional IPV knowledge and training, and ultimately increase screening and care of patients experiencing IPV in their practice. Intimate partner violence (IPV) against women has been identified as a serious health problem and is a leading cause of non-fatal injury in females in North America Dannenberg et al, 1994). Intimate partner violence is described by the American Medical Association as "a pattern of coercive behaviours that may include repeated battering and injury, psychological abuse, sexual assault, progressive social isolation, deprivation, and intimidation" . These behaviours are perpetrated by someone who is in a private and personal relationship with another individual, such as a victim's spouse, common-law partner, sexual partner, or dating partner.

McClennan S, Worster
Some studies have reported that as high as 54 percent of all women have experienced some form of IPV during the course of their life (Reisenhofer et al, 2007). Women who are victims of IPV have been known to utilize health care services at higher rates than women who have not been abused Hazen et al, 2004). They may present to a number of different medical specialists including emergency room physicians, orthopaedic or trauma surgeons, family physicians, and specialists in obstetrics and gynecology during routine appointments or for appointments specific to their injuries. Researchers and advocates have argued that health care providers can play a vital role in detecting IPV (Davidson et al, 2001).
Despite these recommendations for universal screening, the response from the medical community to address the issue of domestic violence has been characterized as "slow and inconsistent" . Intimate partner violence is an under-recognized, recurrent part of trauma, present in at least one in five women who present to the Emergency Department, and physicians frequently fail to make this diagnosis . The failure to detect IPV contributes to recidivism and long-term health problems ). In addition, when injuries resulting from IPV are not recorded in medical records, abused women are denied documentation of their injuries for future references in court cases . Additionally, opportunities to provide education about prevention, lethality assessment, safety planning, and options for escaping an abusive situation and referrals to resources within the community may also be missed ).
The reported rates of IPV vary across different subspecialties and may be underreported. While several studies have reported the prevalence rates of IPV in the health care setting, these estimates vary substantially. Best estimates of rates across various subspecialties are lacking. We therefore undertook a systematic review and meta-analysis of the published literature to determine the prevalence rates of IPV reported to different medical subspecialties including family medicine, emergency medicine, obstetrics and gynaecology, internal medicine, addiction recovery clinics, and adults with child patients in paediatric clinics. The study has important implications for healthcare professionals in various settings. It will direct attention to medical practitioners in subspecialties where IPV is overlooked and will impact their interaction with and treatment of IPV victims as well as victims' willingness to disclose. Also, identification of health care settings with the highest prevalence rates will allow targeted research and focus towards improvement of the quality of life of IPV victims.

Methods
This systematic review adheres to the reporting guidelines of the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) Statement (Stroup et al, 2000).

Eligibility Criteria
We identified articles in all languages that met the following eligibility criteria: 1) the study was published between January 1995 and July 2009, 2) the target population consisted of adult women between the ages of 16 and 65 years presenting to physicians of any medical specialty, and 3) the primary objective of the study was to determine the prevalence rates of IPV. We included all study designs including cross-sectional designs, surveys, and chart or medical record reviews. We excluded studies focusing on children, males, and special populations such as pregnant women, elderly patients, and patients suffering from a specific illness (i.e. depression, chronic gastrolo-enterological disease). We also excluded review articles, letters, comments, case reports, and guidelines.

Identification of Studies
Two reviewers (C.H., S. dM.) independently conducted a search of the electronic database PubMed for relevant articles published from 1995 to July 2009 in all languages with the help of a professional librarian. Additional strategies used to identify studies included consultation with experts, a review of reference lists from articles that fulfilled our eligibility criteria, and use of the "related articles" feature in PubMed for all studies meeting our entry criteria.

Assessment of Study Eligibility
Two of the authors (C.H. S. dM.) independently assessed all studies identified for full evaluation and resolved disagreements through discussion towards consensus. The eligibility was verified by two additional authors (S.S., N.S.). Agreement between observers for study eligibility with regard to abstract and full text review was examined using the (kappa) statistic. The inter-observer agreement in methodological quality scores was evaluated using the intraclass correlation coefficient. We chose an a priori criterion of 95% confidence intervals were calculated using the generic formula: Estimate ± 1.96*(Standard Error).

Assessment of Methodological Quality
Two reviewers (N.S., S.S.), both with methodological expertise and one with content expertise (S.S.) independently graded the methodological quality of each included study using questions adapted from the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement designed for cross-sectional studies (von Elm et al, 2008 Kelley et al, 2003) and we did not find that any of the previously developed checklists were appropriate for this study. Therefore, we adapted the STROBE statement to assess study quality. A priori we determined that studies that met at least 10 of the 49 13 reporting criteria within our modified STROBE criteria would be considered high quality, studies that met seven to nine reporting criteria would be considered medium quality, and studies that met six or fewer criteria would be classified as low quality.

Data Extraction
A structured data extraction form was developed and relevant data was extracted from each eligible study in duplicate (S.S., N.S.) to ensure accuracy. Pertinent data included study characteristics (year and location), patient characteristics (age), type of medical sub-speciality (emergency medicine, family medicine, obstetrics and gynaecology, internal medicine, addiction recovery, pediatrics, and public health), method of data collection (survey, interview, chart review, and other design), timeframe (lifetime, one-years, two-years, and other), sample size, and response rate. We documented how the authors defined IPV, physical, emotional, and sexual abuse. Finally, we recorded the reported rates of IPV, physical abuse, emotional, and sexual abuse as reported by the authors.

Data Analysis
We provide pooled rates of lifetime and one year physical abuse, emotional abuse, and sexual abuse for emergency medicine, family medicine, and other specialties. Analyses were performed with Predictive Analytics SoftWare (PASW) Statistics version 18.0 (SPSS Inc. Chicago, IL, USA).

Evaluation of Heterogeneity
Before analyzing the data, we hypothesized that there would be a large degree of heterogeneity between the studies. Differences such as study methodology (e.g. survey versus interview versus other study designs), timelines of reported IPV (e.g. lifetime versus one-year), differences in screening tools used (e.g. AAS, CTS, PVS etc.), variability within and between populations (e.g. socioeconomic status, age, jurisdiction), differences in the definition of IPV perpetrator (e.g. male perpetrator versus perpetrator of any gender, or spouse versus any intimate partner), or variable study quality could contribute to the heterogeneity.

Study Identification
Our literature search identified 894 potentially relevant citations, of which 32 were considered for inclusion after full text review (Figure 1). Eighteen of these 32 studies were excluded after verification by additional reviewers (NS and SS) because of insufficient data on IPV prevalence, or they did not meet the eligibility criteria. Of 80 studies considered for inclusion based on review of bibliographies, 20 were included in this study. Three studies were identified and included after consultation with a content expert. In total, 37 studies were included in this review. The weighted kappa on overall agreement between reviewers when choosing articles for the full text review was 0.70 (95% confidence interval, 0.59 to 0.82), and for final study inclusion there was only one disagreement.

Study Characteristics
Details of the selected studies are shown in Table 1. The mean age of participants in the included studies was between 28 and 65 years, with 17 studies not reporting mean age. Response rates ranged from 9% to 100% with most studies (24/37 64.9%) reporting a response rate of 70% or greater. Twenty-five studies (67.6%) took place in the United States, three studies took place in each of Australia and the United Kingdom, and other locations (Canada, Spain, Iran, Israel, South Africa, and China) each have one study. Most studies took place in family medicine clinics (15/37, 40.5%) and emergency departments (12/37, 32.4%). Of the remaining studies, three took place in obstetrics and gynecology clinics, three took place in internal medicine clinics, and four studies had more than one specialty. Nineteen studies (51.4%) used written questionnaires to assess rates of IPV, fourteen studies (37.8%) used an inperson interview, one study (2.7%) used a telephone survey, one study (2.7%) used a computerbased survey, and two studies (5.4%) used more than one method.

Study Quality
We judged 26 studies to be of high methodological quality, 10 studies to be of moderate quality, and the remaining 1 study to be of low quality ( Table 1). The reviewers of methodological quality achieved moderate agreement (intraclass correlation coefficient, 0.65; 95% confidence interval, 0.57 to 0.71.

Definition of Intimate Partner Violence
The definition of intimate partner violence varied greatly between studies. The two most commonly used screening tools among included studies were the Abuse Assessment Screen (13 studies, 35.1%) and the Conflict Tactics Scale (5 studies, 13.5%), or modified versions of these tools. Many studies used gender-neutral language, such as "partner", but others specified that the abuser must be male (8 studies, 21.6%). Most studies included only female victims, but a small number (3 studies, 8.1%) also included male victims.

Intimate Partner Violence Prevalence
Twenty-nine of the included studies (78.4%) presented an overall rate of IPV. Overall pooled lifetime prevalence rates were similar across family medicine and emergency medicine (38% and 40%, respectively). Rates reported for other specialties, however, were considerably higher (59%). Prevalence rates of intimate partner violence in the past year were also similar across studies in family practice and emergency clinics (19.9% versus 19.5%, respectively) (Tables 2 and 3). Between 2% and 4.3% of female patients presenting to emergency departments were reported to have injuries caused by a current or former intimate partner ( Table 2). Based on a single study in each of pediatric emergency medicine and public health clinics, these specialties have lower rates of physical IPV than other specialties in the reviewed articles ( Table  4). Please see Figure 2 for pooled prevalence rates of IPV across specialties.

Discussion
Our systematic review and meta-analysis found that the pooled prevalence rates of IPV in emergency medicine and family medicine are similar for both lifetime and one year timelines. Pediatric emergency medicine and public health clinics had lower rates of physical IPV than other specialties, although there were fewer studies in these specialties than others. Physical IPV was more commonly reported than both emotional IPV and sexual IPV, and the rates of sexual abuse tend to be lower than physical and emotional abuse, with emotional abuse being the most common type of IPV.
Several strengths contributed to the quality of this study. A comprehensive search of the literature as well as data abstraction from eligible studies was conducted by two independent reviewers to ensure accuracy, and the articles were screened thoroughly and systematically for inclusion criteria. There was a high degree of agreement between the two reviewers, and many studies of general high quality are included in the review. Also, the results can be highly generalized due to the broad eligibility criteria.
Despite these strengths, our study is limited by a large degree of heterogeneity across the included studies. Heterogeneity was likely a product of many factors including different methodology, populations, and definitions of IPV utilized across the eligible studies. The included studies had varying population characteristics and utilized a variety of different screening tools, which likely contributed to the wide variations in prevalence rates reported. Despite the large variations, our pooled data shows that the prevalence of IPV in emergency medicine and family medicine are similar. Another limitation is the moderately low intraclass correlation coefficient for agreement on methodological quality. This is likely due to the chance for greater variability with an extensive 13-item checklist.
The definition of IPV used is essential to compare studies of IPV prevalence. Differences in definitions include the timeline reported, screening tool used, type(s) of IPV included, and definition of perpetrator and victim. A total of nine different timelines were presented by included studies (lifetime, one year, acute, one month, three months, most recent relationship, current relationship, past relationship, and past five years). The included studies also used a variety of screening tools that have been validated to differing degrees, or adaptations of screening tools that have not undergone validation studies.  reported that even the most commonly studied screening tools are only validated in a small number of studies. Since no "gold standard" exists to screen for IPV in a clinical setting, comparing results across screening tools is a challenge. These differences in definition make IPV a difficult subject to fully understand without some sort of standardization.
Despite similar definitions differences still occurred. There were three studies that used the AAS (the most common screening tool among included articles), similar definitions of physical IPV, and included a family medicine population (Kramer et al, 2004;Marais et al, 1999;Johnson D, 1997). All three studies had sample sizes of over 1,000 patients, used a lifetime timeline and were considered moderate or high quality studies. However, Kramer et al (2004) found that the prevalence of physical IPV was far greater than the other two studies (over 50% vs. 14.6% and 13.9%). The population may account for this variation. In addition, the studies included in this review employed a variety of methods of data collection. MacMillan et al (2006) found that the prevalence of IPV depends on the type of administration (written, verbal, or computer-based) and there was an interaction between method of administration and screening tool used. Some studies suggest that computer-based screening can increase disclosure (Turner et al, 1998), and others suggest that face-to-face interviews increase disclosure over written surveys (McFarlane et al, 1991). One of the included studies used computer-based screening, and obtained an overall prevalence of 36.0% in the past year in emergency medicine . This is the highest included prevalence for the one year timeline in emergency medicine. Results vary for written surveys versus in person interviews in this review.
We did not identify published studies on the prevalence of IPV across all subspecialties. There were no included studies in populations undergoing surgery, although surgeons likely encounter multiple patients who are victims of IPV. Orthopaedic and trauma surgeons, for example, may frequently treat victims of physical IPV in their practice ), yet none of the included studies measured prevalence of IPV in orthopaedic clinics. A large multicentre cross-sectional prevalence study is presently underway that will investigate the prevalence of IPV in patients presenting to orthopaedic fracture clinics (PRAISE Investigators, 2010). None of the included studies reported the prevalence of IPV among psychiatric patients, which could be another high-risk population. It may be important to study other medical specialties, even those expected to identify lower rates of IPV, because victims of IPV do not readily disclose their history of abuse to health care providers (Spangaro et al, 2010).
In conclusion, prevalence rates have been reported in the literature for various medical specialties including family medicine, emergency medicine, and others. However, the studies show a great degree of heterogeneity and are, therefore, difficult to compare. It would be easier to compare prevalence rates across studies if the definition of IPV remained constant. Based on our pooled data, best estimates of IPV prevalence in emergency medicine and family medicine appear to be similar. Future studies could investigate the prevalence of IPV in other specialties, and should use similar definitions of IPV.

ABSTRACT
Background: Intimate partner violence (IPV) is described by the American Medical Association as "a pattern of coercive behaviors that may include repeated battering and injury, psychological abuse, sexual assault, progressive social isolation, deprivation, and intimidation." The long-term consequences of IPV include health risks, posttraumatic stress disorder, depression, and staggering economic costs for health care of victims. Intimate partner violence is often underreported among women who seek medical attention. The current study seeks to address the issue of possible underreporting of IPV in orthopaedic fracture clinics by establishing prevalence rates of IPV among women seeking treatment for musculoskeletal injuries.
Methods/Design: We propose a cross-sectional multicenter study wherein 3,600 women will complete a self-reported written questionnaire across clinical sites in North America, Europe, and Australia. Recruitment of participants will take place at orthopaedic fracture clinics at each clinical site. The questionnaire will contain a validated set of questions used to screen for IPV, as well as questions that pertain to the participant's demographic, injury characteristics, and experiences with health care utilization. Female patients presenting to the orthopaedic fracture clinics will complete two validated self-reported written questionnaires (Woman Abuse Screening Tool (WAST) and the Partner Violence Screen (PVS)) to determine the prevalence of IPV in the past 12 months and in their lifetime. The two questionnaires were designed for rapid assessment of IPV status in emergency departments, family practice, and women's health clinics that we believe are similar to our intended setting of an orthopaedic clinic.

Discussion:
If the prevalence of IPV among women attending orthopaedic clinics is greater than the current perceptions of orthopaedic surgeons, this study will serve to advocate for the continued education of medical professionals to better recognize probable IPV cases and offer existing services to enhance the care of these patients.

Background
Intimate partner violence (IPV) is described by the American Medical Association as "a pattern of coercive behaviors that may include repeated battering and injury, psychological abuse, sexual assault, progressive social isolation, deprivation, and intimidation" [1]. The longterm consequences of IPV include health risks, posttraumatic stress disorder, depression, and staggering economic costs for health care of victims [2].
The cumulative lifetime prevalence of domestic violence for women admitted to the Emergency Department is approximately 54 percent [2]. However, IPV is underreported among women who seek medical attention [2][3][4]. The American College of Surgeons position statement on IPV states that surgeons have the responsibility to identify IPV and appropriately treat women at risk of further harm [5]. In July 2009, the Canadian Orthopaedic Association (COA) released a position on the role of the orthopaedic surgeon in the care of abused women. The statement affirms that the "COA recognizes that IPV is a significant social determinant of morbidity and mortality, and that orthopaedic surgeons are well positioned to identify patients living with IPV and initiate an intervention. Therefore, the COA encourages its members to educate themselves further about IPV and considers it good medical practice to take steps to identify and offer assistance to its victims" [6]. In accordance with this position statement, the COA is working towards raising awareness of domestic violence as well as educating orthopaedic surgeons over the next year. This shows the timeliness of the issue in orthopaedic surgery.
Despite these positive initiatives, there is currently no data in orthopaedic literature to support the hypothesis that the prevalence of IPV in orthopaedic trauma clinics warrants additional resources to identify and manage victims [7,8]. The current study seeks to address the concern that IPV is underreported in orthopaedic fracture clinics by establishing prevalence rates of IPV among women seeking treatment for musculoskeletal injuries across multiple centers in different jurisdictions. This will be the first multi-center study to evaluate this critical issue in the field of orthopaedic trauma.
The proportion of women in Canada 15 years of age and older who have experienced physical or sexual violence in a marital or common-law union in the last 5 years is between 6 percent and 8 percent, affecting approximately 300,000 women in the province of Ontario [9]. This is likely an underreporting of the true rate since detection is hindered by the reluctance of respondents to talk openly about their IPV experience [10]. In documented cases of IPV in Canada, 25 percent of women reported that they were beaten, 20 percent reported choking, and 20 percent were sexually assaulted. Of the documented cases, 40 percent of women in Canada who have experienced IPV suffer a physical injury, and 15 percent of these cases are serious enough to warrant medical attention [10].
Several studies have found that IPV is underreported among women who seek medical attention [2,3,4]. Dearwater et al noted that women who were treated at 11 community hospital emergency departments secondary to IPV-related acute trauma were identified in only 44 percent of all cases presented [3]. Davis et al. reported that of the victims who presented to a level I trauma center because of confirmed or probable IPV, only 26 percent received referrals to social services and 63 percent were discharged without any investigation into their safety at home [11]. The investigators further note that the number of IPV victims was underestimated because of the propensity of conflicting information in the patients' history, lack of information to adequately rule out IPV, and failing to actively screen for IPV among all cases who presented to their trauma service. Additionally, a recent study found that sprains, dislocations, fractures, and foot 68 injuries accounted for 28 percent of all clinical manifestations of IPV among women who were identified in a 2-year period by the Minnesota Domestic Abuse Program [12]. Despite these findings, IPV remains underemphasized in this medical field [7].
The majority of orthopaedic surgeons (87%), however, in a Canada-wide survey believed that female victims of IPV accounted for less than 1 percent of patients in their care [7]. The findings of this survey suggested a misperception between surgeons' beliefs about the prevalence of IPV in their fracture clinics and reported rates of IPV in the community [7]. Current guidelines suggest that orthopaedic surgeons should play an active role in the identification of IPV victims and their timely referral to local agencies [7,13,14].
To explore prevalence rates of IPV in the fracture clinic setting, we conducted a pilot screening study of injured women across two trauma centers in Ontario. We found that one third of women have been victims of IPV (including physical, emotional, and sexual abuse) in the past 12 months. Furthermore, we found that 2.5 percent of women presented with injuries directly resulting from IPV [15]. While our findings were compelling, the generalizability between two level I trauma centers remained unknown. The current proposal aims to expand the generalizability and significantly increase the sample size to provide more precise estimates of IPV in the fracture clinic setting.

Research Objectives and Hypotheses
The primary objective of this observational study is to determine the proportion of women who have experienced IPV in the past 12-months among women who present to orthopaedic fracture clinics for treatment of orthopaedic injuries. We define IPV as physical, emotional, or sexual abuse that is caused by an intimate partner such as a victim's spouse, common-law partner, or dating partner.
Secondary objectives include determining the proportion of women presenting to orthopaedic fracture clinics for the treatment of orthopaedic injuries who have experienced IPV in their lifetime; determining the proportion of women who present to orthopaedic clinics for treatment of orthopaedic injuries that present with an orthopaedic injury that was the direct result from IPV from a current and ongoing relationship; and investigating patients' previous experiences, knowledge, and perceptions with regards to approaching healthcare professionals about IPV.
We hypothesize that the prevalence of IPV among women who present to orthopaedic fracture clinics for treatment of orthopaedic injuries is sufficiently high to warrant targeted programs to assist orthopaedic surgeons in the clinic with identification and appropriate referral to specialists.

Methods and Design
We propose a cross-sectional multi-center study wherein 3,600 women will complete a validated self-reported written questionnaire across ten clinical sites in North America, Europe, and Australia. Recruitment of participants will take place at the fracture clinic at each clinical site. The questionnaires will contain a validated set of questions used to screen for IPV, as well as questions that pertain to the participant's demographic, fracture characteristics, and experiences with health care utilization. A flow diagram of the study is presented in Figure 1.

Rationale for an Observational Study Design
We explored the possibility of addressing our research questions by other less costly or less time-consuming study designs. Other methods considered were a retrospective cohort and a cross-sectional study using a self-reported written questionnaire sent by mail. The advantages and disadvantages of these study designs when compared to the proposed design are discussed below.
A retrospective cohort study would be less costly and time consuming than the prospective cohort study or case-control study proposed. A medical chart review can be undertaken at each participating clinic, where data collectors can review the charts for reported cases of IPV among patients who visited the clinic within a certain time period. Such a study would have a greater likelihood of undergoing an expedited ethics review than a study that requires active participation from the respondent. However, given that IPV is not properly recognized among health professionals [3,4,16], such a review would likely be unable to detect a significant proportion of IPV victims. Moreover, the non-standardization of the definition of IPV may also affect the internal validity of the study and result in an over reporting or underreporting of cases of IPV.
The use of a mailed questionnaire that can be completed and sent back by the participant is a potential strategy that would save costs over a study that requires the respondent to be present at the clinic to be enrolled. Female patients over the age of 18 years can be identified from the registries of the orthopaedic fracture and injury clinics and be mailed a written questionnaire and information letter about the study. Informed consent would be implied if the questionnaire is sent back to the study investigators. However, while such a study would be easier to implement than a prospective study, the administration of a mailed questionnaire has 70 been met with limited success. In their study of abuse prevalence among patients visiting gynaecology clinics, Wijma et al noted response rates as low as 67 percent among recruiting clinics that mailed questionnaires to eligible participants [4]. Campbell et al, in their attempt to assess the physical consequences of IPV with a questionnaire that was mailed, reported a response rate of only 12 percent [17].
In contrast to the use of mailed questionnaires, written self-reported surveys such as the questionnaire suggested in the proposed study design may have less likelihood of response bias. Use of such questionnaires has resulted in response rates between 62 percent and 95 percent among women who are invited into the study personally by research personnel and then administered the questionnaire if they chose to participate [1,4,18,19].

Inclusion and Exclusion Criteria
All women who present to a recruiting orthopaedic or trauma clinic will be screened for eligibility. Our inclusion criteria for this observational study are: 1) The patient presents to the fracture clinic for her own appointment; 2) The patient is 16 or 18 years of age or older; 3) The patient is able to read and write well enough to complete the study questionnaires; 4) The patient is being seen at the fracture clinic for the treatment of an orthopaedic injury; and 5) The patient is able to separate herself from anyone who accompanied her to the fracture clinic to complete the questionnaire in privacy. Clinical sites may include patients who are over the age of 16 years in the study if it is permitted by their Research Ethics Board. Some sites may not allow patients under the age of 18 to participate. Lowering the age requirement to 16, where possible, will improve the generalizability of the study and will allow us to determine the rates of IPV in a younger demographic.
The exclusion criteria for this study are: 1) The patient is considered too ill or injured to participate in the study; and 2) The patient is cognitively impaired and unable to participate in the study. Similar inclusion and exclusion criteria have been used in other studies that sought women's experience with IPV [1,3,8,16]. Once a patient is deemed eligible, she will be invited to participate in the study by a female study coordinator. We will track the number of patients screened and their reasons for ineligibility.

Primary and Secondary Outcome Measures
To measure the prevalence of IPV, our questionnaire will ask patients if their intimate partner had abused them physically, emotionally, or sexually in the past 12 months and throughout their lifetime. We believe that it is important that our study attempt to quantify levels of emotional and sexual abuse among our intended study sample because the various forms of IPV -physical, emotional, and sexual abuse -are concomitant with each other, and orthopaedic surgeons may serve as a second line of detection of IPV if victims are undetected in healthcare settings antecedent to their presentation to the orthopaedic clinic, such as the emergency department. Moreover, we take the stance that orthopaedic surgeons should be concerned about the holistic care of the individual instead of only being concerned with what is immediately treatable in their area of expertise. Therefore, to estimate the overall prevalence of IPV, we will combine the positive answers to the questions on physical, emotional, or sexual abuse.
We will also ask participants to complete two validated questionnaires that were designed for rapid assessment of IPV status in emergency departments, family practice, and women's health clinics that may be similar to our intended setting of an orthopaedic fracture clinic. We selected the Woman Abuse Screening Tool (WAST) and the Partner Violence Screen (PVS) for their psychometric properties, reliability, and specificity in identifying partner abuse. Moreover, a study that was published in 2006 reported that 94 percent of women who were administered both the PVS and WAST concurrently, as written questionnaires, considered them "easy" to complete [18]. Both questionnaires are widely used in IPV screening studies [18,[20][21][22][23][24][25][26].
We chose to request that both questionnaires be administered to each participant to ensure that we are identifying all probable cases of IPV among women who attend orthopaedic clinics. In addition to the elucidation of physical abuse, the WAST contains questions that assess the levels of emotional and sexual abuse that we also feel are important to ascertain in our study, as mentioned previously.
Another research question that will be addressed in the study is what are patients' previous experiences and perceptions about discussing IPV with health care professionals. To that end, the questionnaire will also query the participant about her age, income, education, race/ethnicity, marital status, sexual orientation, and length of relationship. Additionally, participants will be queried about perceptions and previous experiences with reporting IPV in health care settings.
The questionnaire will also ask participants to record the characteristics of the injury that they are being seen for in the fracture clinic including: type of injury, how the injury occurred, locations of injury, and date of injury.

Data Collection
The self-report component of the data collection will involve a written self-completed questionnaire. This method of data collection has been shown in a randomized trial to provide the least amount of missing data by the respondent and is generally favored over the use of a face-to-face interview or computer-based self-completed questionnaire [18].
A female study coordinator at each participating clinical site will approach all female patients who present to the fracture clinic for participation in this study. Due to the sensitive nature of the study, the informed consent process and completion of the questionnaire will take place in a private location by the female study coordinator. A screening form will be completed for all patients that will document their eligibility and whether they agree to participate in this study. After providing informed consent, the female study coordinator will provide the participant with the questionnaire to complete. When the participant has completed the questionnaire, she will place it in a sealed envelope and return it to the study coordinator. This process is described in Figure 2.

72
Step 1: Approaching Patients The study coordinator will approach every female patient who presents to the fracture clinic at a participating centre for participation in the study. We suggest using the following wording when approaching a potential participant: "Hello, I am the study coordinator working with Dr. XXX. I am helping to conduct a women's health study of all women who attend fracture clinic. The study consists of completing a brief questionnaire that will take approximately 10 to 15 minuteof your time. Would you be interested in participating in the study?" If a patient responds positively, the study coordinator will request that they go to a private location so that they can complete the questionnaires in private.
Step 2: Screening Once at the private location, the study coordinator determines whether the patient is eligible for the study and then completes a screening form. If the patient is ineligible, the study coordinator records the reason for ineligibility on the screening form and thanks the patient for her time. If the patient is eligible, the study coordinator will explain the study to the patient and obtain informed consent. If the patient is not interested in participating at this time, the study coordinator completes a screening form, and will check off the appropriate box for the reason that she did not participate. A screening form needs to be completed for all female patients who present to the fracture clinic.

Step 3: Informed Consent
The study coordinator will provide the potential participant with a copy of the informed consent form to review. The study coordinator will discuss the study with the potential participant and address any questions the potential participant has. If the patient declines to participate, the study study coordinator will record the reason on the screening form.

Step 4: Questionnaire Completion
The study coordinator will ensure that the participant completes the questionnaire in private. Once the participant has signed the consent form, the study coordinator provides her with the questionnaire and a pen as well as an envelope to place the completed questionnaire in. we suggest providing the following instructions: "The questionnaire in anonymous. Once you have completed it please place it into the envelope provided and seal it. Please answer honestly. At the pack of the questionnaire is an information sheet that you may take with you if you wish." When the participant has completed the questionnaire, the study coordinator will instruct her that she should not reveal the topic of the study, due to its sensitive nature, to other patients within the fracture clinic. The study coordinator will also ask her to use discretion if she chooses to tell the person she came with about the questionannire that she completed.

Protecting Against Sources of Bias
A bias towards under-reporting IPV is our primary concern. Considerations with respect to confidentiality will be addressed during data collection to reduce bias when participants are completing the questionnaire. Participants will be approached by a female study coordinator, and the consent process and the completion of the questionnaire will take place alone in a private location so as to reduce influence from others. Additionally, participants will be told that the survey is anonymous and will be instructed to not discuss the nature of this study with other participants, so as to increase reporting of IPV, if it exists, when completing the survey (participants may answer the questionnaire differently if they knew that they would be identified as IPV victims or non-IPV victims by the individual administering the questionnaire). Questions pertaining to the participant's demographics should be subject to minimal bias because the questions are categorical and are not intended to be subjective. An under-reporting bias is certainly possible; however, if our study finds an appreciable prevalence of IPV among respondents, we will have confidence that the likely "true" estimate is even higher.
Another source of potential bias in this study is that the questionnaire is self-administered as opposed to interview administered. This may result in some missing and inconsistent data. Due to the extremely sensitive and private information, as well as taking the women's safety into consideration, we believe that an interview-administered questionnaire is not appropriate. Due to the sensitive content of the questionnaire, the study coordinator will not check it over following completion therefore she will not be able to ask the patient about missing questions or inconsistencies.
An inherent limitation of this study which may produce bias is that non-participants may differ from participants in terms of demographics and abuse prevalence. Patients may be less likely to participate if they are a victim of IPV, thus resulting in a lower IPV prevalence rate. Also, some patients may decline to participate because they are not victims of IPV and as such feel that the issue is not relevant to them. Another limitation which may lower the reported IPV prevalence is that we may miss women who are severely injured as a result of IPV because we will not approach women who are taken to the fracture clinic by Emergency Medical Services. However, we may approach these patients at subsequent follow-up visits.

Ontario PRAISE Pilot Study
In preparation for the multi-center definitive P.R.A.I.S.E. study, we successfully completed a pilot study in Ontario, Canada at two clinical sites (Hamilton Health Sciences -General Site, Hamilton and St. Michael's Hospital, Toronto). The purpose of the pilot study was to provide estimates of patient enrollment, inform study logistics, and prove feasibility for the definitive study. The patients from the pilot study will not be included in the larger prospective study, as we have made multiple changes to the protocol and questionnaire based on our experience with the pilot study.
Our preliminary pilot work suggests that women presenting to the fracture clinics are experiencing IPV to much greater extents than previously recognized by orthopaedic surgeons. In a sample of 282 women at the two participating sites, we found that over the past 12 months 8.5% of women were physically abused, 30.5% were emotionally abused and 3.2% were sexually abused [15]. One third of women in the pilot study sample had experienced IPV (including physical, emotional and/or sexual abuse) in the past 12 months. Our pilot study also found that 2.5% of women presenting to the fracture clinics with orthopaedic injuries suffered these injuries as a direct result of IPV. The generalizability of our pilot study results to all North 74 American, Australian, and European centers requires confirmation from our proposed larger study.

Feasibility of the Multicenter Definitive Study
The pilot study helped to inform our enrollment rate as well as inform the study logistics. We enrolled 204 patients over five months at the site in Hamilton, which equates to approximately 40 patients per month. St. Michael's Hospital enrolled 78 patients over 2.5 months, or approximately 30 patients per month. Therefore we believe that sites will enroll between 30 and 40 patients per month and that it will take approximately 10 months of enrollment at each site to complete the pilot study. Therefore, it is feasible to complete the definitive multi-center study within the allotted timeframe.
We also made multiple revisions to the protocol and questionnaire based on our experience with the pilot study. We believe these changes will make the questionnaire easier to complete for participants and also enhance the quality of the data obtained.

Sample Size
According to a recent survey of orthopaedic surgeons in Canada [7], 87 percent of all respondents believe that the prevalence of IPV within their practice is less than 1 percent, with almost all of the remaining respondents believing that the prevalence of IPV in their practice is between 5 percent and 10 percent. Using an estimated IPV prevalence of 5 percent within orthopaedic clinics and standard statistical formulae for estimating sample size of prevalence studies, we have calculated that a sample size of 278 women is necessary for our study to provide an estimate of IPV prevalence with a 95 percent confidence interval between 2.78 percent and 8.31 percent. If the point-estimate of IPV prevalence within orthopaedic clinics is higher than 5 percent, the proportion of the margin of error relative to the estimated prevalence will be lower. To be adequately powered within each region or demographic, each site will therefore recruit 300 participants. The total sample size for the study will be 3,600 participants across the 12 participating centers.

Data Analysis
Data will be analyzed using SPSS Version 17.0 [SPSS, Chicago, IL] and will be stratified according to clinical center, at which point descriptive statistics can be reported for each site and overall. Dichotomous data will be reported as number of participants and proportions, with corresponding confidence intervals to estimate precision. Continuous data will be presented as means and medians with standard deviations. We will provide descriptive statistics describing the patient demographics and injury characteristics across each clinical site. We will also report the results of the WAST and PVS across each clinical site for IPV prevalence in the past 12 months. These two screening tools will be scored according to the developers guidelines. Finally, we will report descriptive statistics, with 95% confidence intervals, across each clinical site on patient's previous experiences, knowledge, and perceptions with regards to approaching health care professionals about IPV. We will use Chi-square tests (dichotomous variables) or t-tests (continuous data) to determine if there are differences in IPV prevalence across different clinical sites and jurisdictions.

Methods Center
The Project Manager at the McMaster University Methods Center will be responsible for the overall day to day coordination of the study. The Project Manager will be responsible for communicating with the clinical sites, providing the clinical sites with the necessary case report forms and validating the data. The Project Manager will also work with the Research Coordinators at the clinical sites to ensure that the protocol is followed and that 3,600 participants are recruited for this study in the allotted timeframe.

Participating Centers
The necessity of a multicenter study is twofold. First, as there are currently no prevalence reports of IPV in women who attend orthopaedic trauma clinics, a cross-sectional study will provide the first report of IPV prevalence in orthopaedic trauma clinics. Second, the collection of data from clinics across North America, Europe, and Australia will enhance the generalizability of our study. We will include clinical sites from a variety of trauma populations and settings (i.e. inner city versus suburban versus rural). Additional centers may be added to allow for further comparisons of the prevalence of IPV across different settings and to increase the generalizability of the findings.

Recruitment Rate
For the multi-center definitive P.R.A.I.S.E. study we aim to recruit 3,600 participants across 12 sites in North America, Europe, and Australia. The pilot study helped to inform our enrollment rate proving feasibility of the definitive trial whereby both clinical sites in the pilot study enrolled at least 30 patients per month (St. Michael's Hospital with 30 participants per month and Hamilton Health Sciences -General Site with 40 participants per month). Consecutive patients will be recruited from fracture clinics across North America, Europe, and Australia and based on the results of our pilot study we anticipate that each site will be able to enroll 300 patients within ten months.

Discussion
If the prevalence of IPV among women attending orthopaedic clinics is greater than the current perceptions of orthopaedic surgeons, this study will serve to advocate for the continued education of medical professionals to better recognize probable IPV cases and offer existing services to enhance the care of these patients. This is especially important because healthcare providers who receive education on screening and ways to care for IPV victims detect them more readily [27]. Furthermore, this study may encourage more open communication between orthopaedic surgeons and their patients, as two major barriers to IPV detection are either the patient is never asked [28] or the healthcare provider is reluctant to inquire [7,8,16]. A positive study will also inform the Canadian Orthopaedic Association's position statement on the role of the orthopaedic surgeon and domestic violence.

Introduction
Intimate partner violence (IPV) is described by the American Medical Association as "a pattern of coercive behaviors that may include repeated battering and injury, psychological abuse, sexual assault, progressive social isolation, deprivation, and intimidation" 1 . From 1999 to 2004, an estimated 653,000 women in Canada were either physically or sexually abused by their current or previous intimate partner 2 . Of those women who seek medical attention, only 24% disclose IPV in the emergency room 3 .
Injuries associated with physical abuse often require referral to orthopaedic surgeons. A recent study reported that sprains, dislocations, fractures, and foot injuries accounted for 28% of all clinical manifestations of abuse among women who were identified over a two-year period by the Minnesota Domestic Abuse Program 4 . The importance of orthopaedic surgeons' roles in identifying IPV has recently been recognized by national organizations including the Canadian Orthopaedic Association 5 .
Our primary objective was to estimate the prevalence of IPV within a twelve month period among women who attend orthopaedic fracture clinics for orthopaedic injuries, asking direct questions about physical, sexual, and emotional abuse. Secondary objectives were threefold: 1) to estimate the prevalence of IPV within a 12 month period among women who attend orthopaedic fracture clinics for the treatment of orthopaedic injuries with the use of two previously developed questionnaires, 2) to determine what proportion of women presented to orthopaedic clinics for the treatment of an injury that was the direct result from intimate partner violence from a current or previous relationship, and 3) to investigate patients' previous experiences, knowledge, and perceptions with regards to approaching health care professionals about intimate partner violence.

Materials and Methods
We conducted a cross-sectional observational study at two Level

Participants
All women who presented to the two participating centers were screened for eligibility by a female study coordinator. Patients may have presented to the fracture clinic for immediate treatment of an injury (after being seen in the emergency department) or for follow-up of an injury. Our inclusion criteria included: (1) patients presenting to the fracture clinic for their own appointment; (2) had to be eighteen years of age or older; (3) had to be able to read, understand, and write in English; (4) had to be seen at the fracture clinic for the treatment of an orthopaedic injury, and (5) had to be able to separate themselves from anyone who accompanied them to the clinic to ensure that she could complete the questionnaire in privacy. We excluded patients who were too ill, injured, or cognitively impaired to participate, although we did interview some patients who sufficiently recovered and consented at the time of follow-up visits. We also excluded patients who presented to the clinics because of chronic pain, arthritis, or wear and tear issues. We recorded the number of patients who were screened and the reasons for ineligibility.
Once a patient was deemed eligible, the study coordinator obtained informed consent and provided the participant with the self-reported written questionnaire which was completed anonymously (Figure 1). The patient was provided with an intimate partner violence resource sheet for the local area if she chose to take it.

Primary Outcome Measure
To measure the overall prevalence of IPV, our anonymous questionnaire directly asked each patient if her partner had abused her physically, emotionally, or sexually in the past twelve months. The three direct questions were taken from the Woman Abuse Screening Tool (WAST), developed by Brown et al 6 . The questions "Have you been abused physically/emotionally/ sexually by an intimate partner?" had possible responses of "Often", "sometimes", and "never." Responses of "sometimes" or "often" constituted a positive screen on these questions. These direct questions have not been independently validated, but were chosen because they are intuitive and easy to understand.

Secondary Outcome Measures
We also asked all participants to anonymously complete two validated questionnaires that were designed for assessment of IPV status in health care settings. We selected the WAST and the Partner Violence Screen (PVS) for their psychometric properties, reliability, and specificity in identifying IPV 6, 7, 8 . 83 The WAST consists of eight questions pertaining to a woman's experience in her intimate relationship for the past twelve months 6 ( Table 1). We analyzed the WAST according to a 2008 revision of the original WAST 7 , which was originally created and validated by Brown et al 6 . MacMillan et al reported a sensitivity of 47% and a specificity of 96% when compared to the Composite Abuse Scale (CAS), which is commonly used to screen for IPV 9 .
The PVS consists of three questions. The first question asks if the woman has been hit, kicked, punched, or otherwise hurt by someone in the past year and if so, by whom. The second two questions assess a woman's perception of safety 8 ( Table 1). The PVS was scored according to the developers' guidelines 8 . The sensitivity of the PVS in detecting partner abuse was 64.5% and the specificity was 80.3% when compared with the Index of Spouse Abuse (ISA) 8 . When compared with the Conflict Tactics Scale (CTS), sensitivity of the PVS was 71.4% and the specificity was 84.4% 8 . 3. Is there a partner from a previous relationship who is making you feel unsafe now? * The present study involved the use of an adapted version of the WAST, as described in 2008 by Yut-Lin and Othman 7 . The overall WAST score is derived by summing the responses of all eight WAST items. The scores on each question range from 1 to 3 (e.g., a lot of tension to no tension). The final score is derived by were included in the WAST analysis. † A positive response to any of the three PVS questions is indicative of a positive screen for partner violence. For Question 1, the perpetrator had to be a person in a current or previous intimate relationship in order for the response to be considered positive.

Demographics, Injury Characteristics, and Perceptions about IPV
The questionnaire included demographic questions and questions about the characteristics of the participants' injury. The questionnaire also asked about patients' previous experiences and perceptions about discussing and reporting IPV with health care professionals.

Sample Size
According to a recent survey of orthopaedic surgeons in Canada 10 , 87% of all respondents believe that the prevalence of IPV within their practice was <1%, with almost all of the remaining respondents believing that the prevalence of IPV in their practice is between 5% 84 and 10%. Using an estimated IPV prevalence of 5% within orthopaedic clinics and standard statistical formulae for estimating sample size of prevalence studies, we have calculated that a sample size of 278 women would be necessary for our study to provide an estimate of IPV prevalence with a 95% confidence interval between 2.78% and 8.31%, an interval that should bolster support for a more comprehensive study. If the point-estimate of IPV prevalence within orthopaedic clinics is higher than 5%, the proportion of the margin of error relative to the estimated prevalence will be lower.

Statistical Analysis
Data are reported as number of participants and proportions, with corresponding confidence intervals to estimate precision. Continuous data are presented as means and medians. The answers and overall score for both the PVS and the WAST are presented using descriptive data. We conducted chi-squared tests to determine if there were any differences in reported outcomes between the two clinical sites. Patients who completed part of the questionnaire were included in our analyses. The PVS has a positive cutoff score of one and the WAST has a positive cutoff score of 13. If these cutoff scores were reached, the patient scored positive on that questionnaire regardless of whether she completed all questions on the questionnaire. No statistical extrapolation process was carried out for missing information.

Recruitment
Six-hundred and ninety patients were screened for participation in this study and 295 patients were deemed ineligible. Of the 395 patients who met our eligibility criteria, 282 patients provided informed consent and completed all or part of the questionnaire (Figure 1).

Characteristics of Included Women
The majority of women in this study were Caucasian (85.1%; 240 of 282), over age 40 years (62.8%; 177 of 282), and had an annual income of less than $40,000 (59.6%l 168 of 282) (Appendix). Demographic characteristics were similar across participating sites except for annual income (p<0.05), which was higher for the St. Michael's Hospital site. We did not collect data on the characteristics of women who did not participate in this study.

Injury Characteristics
Fractures were the most common type of injury (73%; 206 of 282). The most common reported cause of injury was falls (47.9%; 79 of 165,) and most injuries involved the lower extremity (59.6%; 168 of 282) (Appendix).

Prevalence of Intimate Partner Violence in the Past 12 Months: Primary measure
The overall prevalence of intimate partner violence (emotional, physical, and sexual abuse) on the basis of direct questions was 31.6% (eighty-nine of 282) ( Table 2). Twenty-four women (8.5%) experienced physical abuse, eighty-six (30.5%) experienced sexual abuse. The eighty0nine women who screened positive for intimate partner violence when directly asked spanned different ages, ethnicities, income levels and education levels (Appendix). The lengths of their current relationships also varied form less than one year to more than forty years. No specific injury patterns were noted among women disclosing abusive relationships.

Prevalence of Intimate Partner Violence in the Past 12 Months: Secondary measures
Thirty-five women (15.4%) screened positive for intimate partner violence with the use of the WAST ( Table 3). Twenty-six women (9.2%) screened positive for intimate partner violence with the use of the PVS tool ( Table 4). Seven women (2.5%) indicated that the reason for their visit to the orthopaedic clinic was directly related to a physical injury caused by their intimate partner. Five of these women were being treated for a fracture, one was unsure about her diagnosis, and one was being treated for an unspecified injury. These women spanned all ages, ethnicities, and educational levels.  * Chi-square tests were performed to see if differences existed between the two trauma centers. The level of significance was set at p<0.05. No differences were found. † The values are given as the number of patients, with the percentages in parentheses. ‡ The values are given as the number of patients, with the percentage in parentheses and the 95% confidence interval in brackets. § These women responded positively to one or more of the PVS items pertaining to violence exhibited by an intimate partner. # These women did not respond positively to any of the PVS items pertaining to violence exhibited by an intimate partner Previous Screening Twelve (7.6%) of 157 women responded that they had been previously screened by health-care professionals other than their treating surgeon. Of the twenty-four victims of intimate partner violence who had experienced physical abuse, four (16.7%) had been asked by another health-care professional about intimate partner violence. In addition, of the twenty-four women who had experienced physical abuse, nine (37.5%) indicated that they believed that health-care providers should ask about intimate partner violence. Of all women screened, 36.2% indicated that they believed that health-care providers should ask about intimate partner violence.

Discussion
Our principal findings suggest that 31.6% women who presented to fracture clinics had experienced some form of intimate partner violence within the last year, that 8.5% of women had experienced physical abuse in the last year, and that 2.5% of women reported that the reason for their fracture clinic visit was directly related to a serious injury caused by an intimate partner. We did not identify any patient characteristics unique to those women who had experienced intimate partner violence.

Strengths and Limitations
Our study has several important strengths including the use of previously developed screening questionnaires and direct questions, broad eligibility criteria, the use of female study coordinators in order to maximize enrolment, the completion of self-administered questionnaires in a private location, and the assurance of anonymity of the respondents. The use of multiple screening tools may help to identify patients who have experienced forms of abuse that may not be screened for with the use of only one tool. However, there is a possibility that broadening the definition of intimate partner violence creates a higher risk of false-positive results. Female study coordinators were used to recruit and survey the patients so that abused women would feel safer and more at ease. Many women have reported that it is difficult to talk to male health-care providers about their experiences with intimate partner violence 11 . The decision to use a self-administered questionnaire is supported by studies that have suggested improved acceptability by women 9,10 . However, the use of a self-administered questionnaire led to some incomplete responses. For example, a large number of the women who had been hit, kicked, punched or otherwise hurt in the past year declined to indicate how their current injury was caused. Since the survey was anonymous, we were unable to link the surveys to a patient chart to obtain the missing information. This finding may indicate that intimate partner violence was underreported.
Despite our efforts to maximize enrolment, 113 women declined to participate in the study. It remains plausible that non-participants differed from participants in terms of the prevalence of abuse. Patients may have been less likely to participate if they were a victim of intimate partner violence, resulting in a lower prevalence of intimate partner violence in our study. Alternatively, some patients may have declined to participate because they were not victims of intimate partner violence and as such believed that the issue was not relevant to them.
In Ontario, in 2006, the median age was thirty-nine years 12 , the median household income was $62,400 13 , and 77.2% of the population was white 14 . Our population was, on average, older than the general Ontario population and had a higher percentage of low-income and white individuals. The generalizability of our findings may be limited only to environments with similar processes inherent in the triage and referral of patients in our study. Level I trauma centres in other locations may have different referral procedures.

Relevant Literature
A recent survey of 186 Canadian orthopaedic surgeons suggested that 87% believed that the prevalence of intimate partner violence within their practice is <1% 10 . Our findings support a much higher prevalence of intimate partner violence among women attending orthopaedic fracture clinics. The finding of most concern was the 2.5% prevalence of severe orthopaedic injuries, such as fractures, that were directly reported as the reason for the woman's visit to the 89 surgical fracture clinic. Our findings support those of an American-based audit of musculoskeletal injuries sustained by women attending a domestic abuse treatment and counseling center, the Minneapolis Domestic Abuse Project 4 . That study demonstrated that head and neck and musculoskeletal injuries were the most common manifestations of intimate partner violence and that the spectrum of musculoskeletal injuries included sprains, fractures, dislocations and food injuries 4 . Similarly, Spedding et al. reported that multiple injuries (especially those involving the head and neck), fractures, loss of consciousness, abdominal injuries, and injuries occurring on stairs were markers of domestic violence in women presenting to the emergency department 15 .
Our overall estimate of intimate partner violence was higher than previously reported rates of intimate partner violence in other medical specialties, including emergency medicine, obstetrics and gynecology, primary care, internal medicine, and pediatrics (Figure 2) 16,17 . However, the reported prevalence of intimate partner violence in patients with musculoskeletal injuries was similar to the prevalence of intimate partner violence in a hospital-based addictionrecovery unit 16 . The varying prevalences across medical disciplines may also reflect differing definitions of intimate partner violence. The high prevalence of intimate partner violence in orthopaedic fracture clinics as compared with other medical specialties confirms the need for orthopaedic surgeons to be aware of the issues related to intimate partner violence and to recognize opportunities to assist women experiencing intimate partner violence. Our study included women who varied in terms of age, marital status, annual income, and education level. There was an array of injury types, causes of injury, and locations of injury. Similar to previous research 18,19,20 , our study demonstrated that women of all ages, ethnicities, socioeconomic status levels, and injury patterns may experience intimate partner violence. In a previous study, two of us (M.B. and S.D) and colleagues identified several risk factors associated with increased frequency of physical abuse, including young age; a shorter duration of the relationship; the coexistence of emotional, psychological, and/or sexual abuse; and drug or alcohol dependency 4 .
Only one of seven women who indicated that the reason for the visit to the clinic was an injury directly resulting from intimate partner violence indicated that she had been previously asked about intimate partner violence by a health-care professional. These findings are consistent with those of other studies. A recent survey of >500 emergency department patients demonstrated that 86% of the patients believed that it is appropriate for all women to be asked if they had experienced violent or threatening behaviour from someone close to them 21 . Similarly, Glass et al. found that the majority of both abused and non-abused women presenting to the emergency department supported routine screening for intimate partner violence 16 . The same study also demonstrated that the screening rates for women who had reported abuse in the prior year was only 13% 16 .
The most common barriers to reporting intimate partner violence include a lack of awareness of intimate partner violence resources, a perception that the situation was not major enough to seek help, a fear of retaliation from the abusive partner, concerns regarding the custody of children, financial issues, lack of trust in the health-care provider, shame and embarrassment, concern that physicians cannot help, social and language barriers, and the reluctance to talk to a male doctor 11 . A recent survey of Canadian orthopedic surgeons found that almost one-third of the orthopaedic surgeons who responded felt personal discomfort about intimate partner violence and that more than half of the respondents indicated that there was a lack of knowledge and education regarding intimate partner violence 10 . In the same survey, 91% of surgeons believed that knowledge about intimate partner violence was relevant to their surgical practice and 30% supported educational programs for orthopaedic surgeons 10 .
The surgeon's role in identifying intimate partner violence has not gone unrecognized by national organizations. The American College of Surgeons position statement on intimate partner violence states that surgeons have the responsibility to identify intimate partner violence and to appropriately treat women who are at risk of further harm 22 . The Canadian Orthopaedic Association takes a similar stance, and its position statement declares that "orthopaedic surgeons are well positioned to identify patients living with IPV and initiate an intervention." 5

Influence of Screening Tools on Prevalence of Intimate Partner Violence
In the present study, we used three direct questions as our primary measure of prevalence of intimate partner violence. A substantially lower prevalence was obtained with the WAST and PVS as compared with the three direct questions. We believe that the difference in prevalence obtained with the use of the different tools can be accounted for by looking at which type of abuse each tool is screening for. The majority of the women who had a positive score for intimate partner violence on the direct questions had experienced emotional abuse. A common challenge faced with intimate partner violence research is that emotional abuse is difficult to define and experiences are often interpreted differently across patients. The PVS focuses primarily on experiences with physical violence and feelings of safety. Women who experienced 91 emotional or sexual abuse but not physical abuse may screen negatively on the PVS. Although the WAST asks about the three types of abuse, women who are experiencing only one type of abuse may answer the questions in such a way that they do not meet the positive cutoff score of 13. Given the limitations of the PVS and the WAST, we believe that our primary measure of directly asking about intimate partner violence provides the most accurate assessment of the prevalence of intimate partner violence.

Clinical Relevance
If our findings are generalizable to the practices of most general orthopaedic surgeons, the opportunities to identify and assist victims of IPV are numerous. Based on general clinic volumes, an orthopaedic surgeon sees approximately forty-five female patients per week (approximately 2,340 female patients per year), half of whom will be new patients (1170 new female patients per year). Our findings suggest that screening all women would result in ninetynine women (8.5%) disclosing physical abuse within the 12 months of that year, and approximately twenty-nine women (2.5%) presenting to the orthopaedic surgeon's clinic for the care of injuries directly resulting from intimate partner violence.
We support the idea that primary contact health care professionals should be suitably trained and aware of intimate partner violence to help women presenting to the emergency and urgent care clinics across North America. There has been considerable research and focus to understand how women presenting to the first contact health care professional can be identified and assisted. As we understand, several challenges have emerged with this approach. When patients present to the emergency room, they are often frightened and in a great deal of pain. Previous literature has shown that this is not the optimal time to ask about intimate partner violence 5 . As orthopaedic surgeons see patients multiple times, can create trust, and are more likely to see a patient in a more calm state than emergency department staff, they are in a good position to identify victims of intimate partner violence 5 . Our study shows that, in some cases, the first time a woman discloses intimate partner violence can be in orthopaedic fracture clinics. This provides an opportunity to integrate hospital resources to ensure that appropriate referral from an orthopaedic clinic is possible. While not directly related to the musculoskeletal complaint a woman may have when seeing her orthopaedic surgeon, her emotional well-being, compliance, and health will be impacted by intimate partner violence . The focus on quality of life in orthopaedic outcomes, and the common finding that psychological factors are important prognostic factors in recovery could have direct relevance to the outcome of orthopaedic care. Finally, if surgeons ask about abuse, it may potentially prevent the abuse from escalating to the point of physical violence in serious traumatic injury. This strategy would be relevant to the field for injury prevention.
Orthopaedic surgeons can help by having materials available on local shelters and tollfree help lines, screening patients for intimate partner violence in a private setting without the partner present, using direct questions to elicit a direct response and respecting a women's choice to not disclose if she is not ready to do so 5 . After a woman has disclosed abuse, orthopaedic surgeons should be non-judgmental and supportive, assess her safety, respect her decisions, take clear notes regarding the injury or injuries, and provide resources and referrals for help 5 .
While recent reports question the utility of universal screening, no reports to date have demonstrated harm from a practice of universal screening of women [23][24][25][26][27] . In addition, Davis notes that "the long-term sequelae [of intimate partner violence] include health risks, posttraumatic stress disorder, depression, and the economic costs for health care and the 92 economic loss to victims" and also that the failure to diagnose intimate partner violence and intervene may have a detrimental outcome as 44% of domestic violence related homicide victims had presented to an emergency department within 2 years of their deaths 28 .

Need for a Multinational Study
The current study was the first to look at prevalence rates of intimate partner violence in women presenting to orthopaedic fracture clinics in Ontario. This study included only two clinical centers in Ontario, both in urban settings, limiting the generalizability of our findings. To confirm local findings nationally and internationally a larger, multi-national study is required.
Given the high prevalence of intimate partner violence at two orthopaedic fracture clinics, these results have implications for implementing intimate partner violence screening and providing further education for medical professionals. McCloskey et al reported that screening or assessment for intimate partner violence prompted 3-fold more disclosure than might be volunteered spontaneously 17 . This study also found that abused women who were asked about partner violence by a provider, 32% received help contacting services compared to 5% who were not asked 17 . Contrary to these findings, MacMillan et al recently found that there continues to be a lack of evidence that universal screening alone improves health outcomes for survivors of intimate partner violence 24 . In an editorial in response to theses findings, Moracco and Cole argue that universal screening with passive referrals to community services is an inadequate response to intimate partner violence 29 . They continue to stress that specific interventions to prevent the recurrence of abuse for women at risk of violence should be implemented and rigorously tested without further delay. This confirms the need for further studies on outcomes of screening for intimate partner violence In conclusion, our study confirms that there is a high prevalence of intimate partner violence among female patients with injuries who are seen at orthopedic clinics in Ontario. Similar to previous research, our study demonstrated that women of all ages, ethnicities, socioeconomic status levels, and injury patterns may experience intimate partner violence. Surgeons should consider screening all injured women for domestic violence in their clinics. ) Declined to answer 1 (0.5%) 0 (0.0%) 1 (0.4%) *Chi-square tests were performed to see if differences existed between the two trauma centers. The level of significance was set at p < 0.05. A significant difference was found in terms of annual income (p = 0.04). †The values are given as the number of patients, with the percentage in parentheses. to see if differences existed between the two trauma centers. The level of significance was set at p < 0.05. A significant difference was found between the two clinic sites in terms of injury type (p = 0.048). †The N values pertain to the total number of patients for whom information was available. The values that follow each N value are given as the number of patients in each category, with the percentage in parentheses. ‡Patients may have had more than one injury, in multiple locations; therefore, percentages may not add to 100%.  6.1%] *Chi-square tests were performed to see if differences existed between the two trauma centers. The level of significance was set at p < 0.05. No significant differences were found. †The values are given as the number of patients, with the percentage in parentheses. ‡The values are given as the number of patients, with the percentage in parentheses and the 95% confidence interval in brackets.
Interpretation: PRAISE is the largest prevalence study conducted to date in orthopaedics. Orthopaedic surgeons should be confident in the assumption that one in six women have a previous history of physical abuse, and that one in 50 injured women will present as a direct result of IPV. Our findings warrant serious consideration for optimizing fracture clinics to identify, respond to, and provide referral services for, victims of IPV.

Introduction
Intimate partner violence (IPV) -also known as domestic violence, spousal abuse, and battering -refers to physical, emotional, sexual, psychological or financial abuse between intimate partners. 1,2 The list of short and long-term physical and psychological consequences of IPV is exhaustive and well established in the literature. 3,4 In a large-scale, multi-country study, 5 WHO reported that the lifetime prevalence of IPV among women ranged from 15 to 71% and in a systematic review, 6 investigators reported that the prevalence ranged from less than 20% to more than 50%. This review also emphasized the scarcity of studies in which the prevalence of IPV within orthopaedic fracture clinics was investigated. 6 We have shown that the 12-month prevalence of IPV in fracture clinics at two trauma centers in Ontario, Canada 7 was 32%, which was much higher than anticipated. 7 In the present multi-centre study, we aimed to establish the prevalence of IPV in the past year among women who present to orthopaedic trauma fracture clinics in different countries.

Methods
We did a cross-sectional multi-centre study at 12 orthopaedic trauma fracture clinics at medical centres in Canada (seven sites), the USA (one site), the Netherlands (two sites), Denmark (one site), and India (one site). We selected sites on the basis of feasibility for successful completion of the study (i.e. previous collaboration, interest in participation, and availability of research personnel) and to represent a range of geographic areas.
The Methods Centre received approval from the McMaster/Hamilton Health Sciences Research Ethics Board (REB #08-369), and from each local research ethics board. We will summarize only the methods used, since a full report of the methodology of the PRAISE study has already been published (reference 8, and appendix). Subsequent to the publication of the protocol, we added one site from India and removed our site in Australia to improve feasibility.

Study Design and Participants
Patients presenting at participating fracture clinics (for Canadian and American sites) or trauma clinics (for Dutch, Danish, and Indian sites), collectively known as injury clinics, were screened for participation in this study by a female study coordinator. Patients at these injury clinics are seen by an orthopaedic surgeon about an orthopaedic injury that needs either surgical or conservative management. Acute injuries are typically seen in emergency departments at our sites and are followed up over several months in the injury clinic, although some patients with acute injuries are first seen in the injury clinic.
To be included in the study, patients must have presented to the fracture clinic for their own appointments (i.e. they must not have accompanied someone else to the clinic for their appointment for a musculoskeletal injury); been at least 16 or 18 years of age (dependent on the site ethics policies); been able to read, understand, and write in English, Dutch, Danish, Hindi, or Marathi (dependent on the study site); been seen at the fracture clinic for treatment of an orthopaedic injury; and separated themselves from anyone who accompanied them to the fracture clinic in order to complete the questionnaire in private. We excluded patients who were judged too ill or injured to participate and those who were cognitively impaired and unable to participate. All patients provided written informed consent.

Objectives
The primary objective of this study was to establish the proportion of women who had been subjected to IPV in the past 12-months among those who presented to orthopaedic fracture clinics for treatment of orthopaedic injuries. We defined IPV as physical, emotional, or sexual abuse that is caused by an intimate partner and asked participants by direct questioning if they were subject to any abuse. We also asked patients to complete two validated questionnaires (Women Abuse Screening Tool [WAST] and the Partner Violence Screen [PVS]), which were designed for the assessment of IPV in health care settings. [9][10][11] Secondary objectives were to determine the proportion of women who presented with an orthopaedic injury that was the direct result of IPV, assessment of which patient characteristics are associated with IPV, and investigation of patients' previous experiences with, and perceptions of, being asked about IPV by health care professionals.

Outcome Measures and Survey Methods
The 12 month prevalence of IPV was assessed with three direct questions (taken from WAST) about patients' experience of IPV in the past 12 months, PVS, and WAST. We used the three direct questions that were rephrased to assess lifetime IPV. These questions were: "Have you been physically abused?"; "Have you been emotionally abused?"; and "Have you been sexually abused?" with a preamble defining the time period of interest (i.e. "in the past 12 months" or "in your lifetime"). We chose to use more than one measure of IPV prevalence to minimize the number of missed cases. We chose to use the WAST and PVS because, according to a recent systematic review 12 about IPV screening methods, these two devices are among the most widely used and well-validated IPV screening tools.
We translated the questionnaires into Dutch, Danish, Hindi, and Marathi with the assistance of clinicians who are fluent in English and the language of interest. In each case, the translations were back-translated into English and the study team compared the wording to the original English version and made any corrections if necessary.

Data Collection
After obtaining informed consent, a female study coordinator provided each participant with a questionnaire to complete independently in a private location. The questionnaire was comprised of 37 questions concerning the patient's demographics and orthopaedic injury, direct questions about their experience with IPV in the past 12 months and over their lifetime, the WAST, the PVS, and several questions about their experiences and opinions on health care professionals asking about IPV. After completing the questionnaire, the participant returned it to the study coordinator in a sealed envelope.

Sample Size
We had initially aimed to include 300 women per site, across 12 sites, for a total sample size of 3,600 women. We used an estimated 12-month IPV prevalence of 25% within orthopaedic clinics and a standard statistical formula for estimating sample size of prevalence studies, 13,14 for estimation of sample size of prevalence studies and calculated that a sample size of 288 (rounded to 300) women per site was necessary for our study to provide an estimate of IPV prevalence with a confidence level of 95%.
In September 2012, we performed an analysis of the 2,945 women included in this study and the primary outcome estimate to ensure that our sample size assumptions remained valid. 102 The 12-month prevalence of IPV among the recruited women in this study was 16%. Assuming a 5% margin of error, our sample size was readjusted to 206 women per site, for a total of 2,478 across 12 clinical sites. This readjustment was done because the sample size increases as the prevalence approaches 50% (highest variability). 14 We also did a sensitivity analysis for clustered data by centre and found no difference in the primary outcome. Given that our recruitment had exceeded the revised estimate by over 450 women, and the low likelihood of recruiting an additional 600 women, with increased cost and delay to dissemination of key findings, the lead investigators agreed to stop enrolment.

Statistical Analysis
Surveys were entered into a study-specific database and descriptive analyses, including frequency counts and percentages, were calculated for all data collected. Continuous data are presented as means and standard deviations. We conducted Chi-squared tests, regression analyses, and a sensitivity analysis to determine if there were any differences in patients' perceptions about IPV or cluster effects between sites located in Canada and the USA, Denmark and the Netherlands, and India. No statistical imputation process was carried out for missing information. Based on the previous literature, we performed a multivariable logistic regression analysis on the following demographic characteristics -age (continuous), ethnicity, sexual orientation, income level, education level, marital status, length of relationship, having children, type of injury, location of injury, and continent of residence -to establish whether any of these variables were independently associated with experience of IPV and physical abuse in the patient's lifetime and the past 12 months. 15,16 We added in the type and location of injury because these may be of particular interest to orthopaedic surgeons. All data analyses were performed using SPSS version 20.0 (SPSS, Inc., Chicago IL).

Results
Ten sites screened 3853 patients for inclusion and 1094 women were ineligible. Of the 2759 patients who met the eligibility criteria, 2344 provided informed consent and completed all or part of the study questionnaire. Two additional sites in Canada (Foothills Medical Centre, Calgary, and St. Michael's Hospital, Toronto) included 601 patients but did not keep track of exclusions due to resource limitations and clinic structure; thus, a total of 2945 women from 12 centres were included in this study (Figure 1) and the response rate was 85·0% (range 73·3-99·3%) for the ten sites that kept screening records. Because of missing data, the sample size for the primary outcome was 2839.  Table 1 shows patient demographics, injury, and risk factors. Most women who participated in the study were of white ethnicity (>80%), older than 40 years (60%), married or in a common law relationship (>50%), had children (>60%), and had a college or university degree or diploma (two-thirds). The most frequently reported cause of injury was a fall (more than a third), and most injuries (>50%) affected the lower limb.
104  †All reported incomes were converted to Canadian dollars. Table 2 shows the overall prevalence of IPV in the past 12 months (about 16%) and during lifetime (about 35%) based on direct questioning. About 16% had experienced emotional abuse, 3% physical abuse, and 1% sexual abuse. The 12-month prevalence of IPV was higher in Canadian and US centers (18%) than in Dutch and Danish centers (12%, p<0.001), but was much the same as that for the centre in India (18%; p=0. 35). No significant differences between continents were recorded for the prevalence of physical abuse in the past 12 months (p=0.87; Figure 2). Both PVS and WAST identified two-times fewer women who had experienced IPV within the past 12 months (7·1%, 95% CI 6·2-8·1% and 7·4%, 6·5-8·4% respectively) than with direct questioning (appendix). IPV prevalence for either PVS or WAST (p=0·244 and p=0·322, respectively) did not differ across continents (Figure 2). The 12-month IPV prevalence estimate did not differ significantly when we omitted the two sites that did not record reasons for exclusions (16·7%, 5·0-18·0%), and no difference was noted in 12-month IPV prevalence when adjusted by site (16%, 13-18%).  *Patients presenting to the orthopaedic clinic with injuries caused by an intimate partner. 108 The overall lifetime prevalence of IPV based on direct questioning was about 35% ( Table 2). A third of women had experienced emotional abuse, about 18% physical abuse, and just under 10% sexual abuse in their lifetime ( Table 2). The prevalence of IPV was significantly higher in the North American centres (40%) than in the European centres (24%; p=0·001) or the Indian center (18%; p<0·016, Table 2). Women in Canada and the USA had a significantly higher rate of self-reported physical abuse in their lifetime (20%) compared to those in the Netherlands and Demark (13%) and India (2%; p<0·004; Figure 3).  Figure 3: Prevalence of acute, lifetime, and past year physical intimate partner violence across continents. Error bars represent 95% CIs. Lifetime physical abuse differed significantly between continents (P<0.0001). There were no differences between continents for acute physical abuse (p=0.27) or physical abuse in the past 12 months (p=0.87).
Just under 2% of women presented to fracture clinics with injuries as a direct result of IPV ( Table 2). 40 (80%) of the 49 women with acute injury due to IPV were treated for fractures. Most injuries were fractures of the foot or ankle (nine patients, 18%), hand or wrist (seven patients, 14%), or were at more than one location (nine patients, 18%). The proportion of women who presented as a direct result of IPV varied across the clinical sites, ranging from less than 1% to between 3 and 4% ( Table 2). Of the women who presented because of an IPV injury, only seven (14%) had ever been asked by a health-care provider in the medical system whether they were subject to IPV.
Married women, women with children, women with injuries of the limbs, and women living in the Netherlands or Denmark were at a lower risk of IPV over their lifetime than those in the other groups. Women who were older, in a short term relationship (<10 years), and who lived in the Netherlands or Denmark, and those who had no children were at a significantly decreased risk of any abuse in the past 12 months than those in the other groups ( Table 3).

Canada/US
The Netherlands/ Denmark  ** "Other" injuries include head/neck/face, chest, spine, pelvis, more than one injury, and "unknown". 110 Only 6% of the study participants and none from India had ever been asked by another healthcare professional about IPV ( Table 4). About 60% of participants were aware of specific resources for people who have been abused by an intimate partner; however, none at the site in India were aware of such resources. Almost three-quarters of participants believed that health-care providers should ask all women whether they are subject to IPV, and about 60% indicated that orthopaedic surgeons, in particular, should ask all women about IPV ( Table 4). Significant differences were noted between respondents in Canada and the USA, the Netherlands and Denmark, and India about whether health-care providers and orthopaedic surgeons should ask all women about IPV -patients in Canada (81%) and the USA (66%) were more open to such questions than those in the Netherlands or Denmark (18%, p<0.0001).

Discussion
The results of the PRAISE study show that one in six women presenting to orthopaedic clinics have a history of abuse in the past year, and that one in 50 attend their appointment as a direct consequence of IPV. These findings are similar to those of a recent meta-analysis of 37 studies in which the prevalence of IPV in patients presenting to different medical specialties was investigated, with the best estimates of the lifetime prevalence of any type of IPV being 40% in family medicine and 38% in emergency medicine (Figure 4). 6 Arguably, the severity of physical abuse in women presenting to orthopaedic clinics is higher than in those presenting to other specialties, in view of the high proportion of fractures caused by IPV (eg, 80% in this study).

112
Characteristics of women who were positive for IPV included being in a relationship for a short period (<10 years), being unmarried, and having children. This profile is consistent with Capaldi and colleagues' systematic review, 18 in which they showed that high income is a protective factor, whereas the effect of marital status as a risk factor was unclear. Differences between the USA and Canada and the Netherlands and Denmark could be real and attributable to cultural and political divergence between the continents, or they might be due to differences in how the participants interpreted the survey questions because of variations in culture or language.
This study has multiple strengths, such as the inclusion of several clinical sites, the use of female study coordinators for patient comfort and safety, broad eligibility criteria, the use of previously developed screening questionnaires and direct questions, and the completion of self-administered questionnaires in a private location. Our broad eligibility criteria and large sample size ensured that women with a range of characteristics were eligible for participation.
The broad definition of IPV used is a limitation of the study and might have categorized some women as victims whereas, in reality, they were not. However, since there is no widely accepted gold standard for assessment of IPV, the patients' answers to the direct questions about abuse were interpreted at face value. The use of several screening methods was helpful to maximize the number of cases of IPV detected; nevertheless, the prevalence of IPV was higher with the direct questions than with WAST or PVS. PVS contains questions that enable easy assessment of physical abuse, whereas the WAST contains questions that help detection of emotional and sexual abuse, which could help to explain some of the differences recorded. 6 Another limitation is that WAST and PVS are validated only for English-speaking populations, which might limit the accuracy and external validity of the results across all continents because of differential interpretation of the questions among non-English-speaking participants.
Our response rate was high (85%) and is much the same as that in prevalence studies in other health-care settings 19 . Despite efforts to maximize enrolment, 415 women refused to provide informed consent and participate in the study. The prevalence of IPV might differ between non-participants and participants, and the direction in which it could vary is unclear. Because of confidentiality and ethics requirements, we were unable to obtain information about patients who did not participant and those who were missed. Therefore, we cannot assess the effect of these patients on our data. Although sites aimed to include consecutive patients, this approach was not possible in a few cases because of logistics (ie, some clinics were missed because of research coordinator availability).
We used a self-administered questionnaire, which patients completed privately. The study coordinator was available to address any questions; however, to ensure patient privacy and confidentiality because of the sensitivity of the topic, the study coordinator did not review with the study participant the completed questionnaire for completeness or to ensure that the responses were logical and consistent. Consequently, several questionnaires had missing data. Other IPV prevalence studies faced similar challenges. 15,16 In our analyses, we used all available data and we did not exclude questionnaires because of missing data.
Despite a focus on emergency departments as case-finding opportunities for IPV, [20][21][22] we argue that women admitted to emergency departments with physical injuries, such as fractures, often do not disclose IPV to their emergency physicians. Further, musculoskeletal injuries are the second most common type of injury incurred as a result of IPV and invariably lead to referrals to an orthopaedic surgeon. 23 Additionally, healthcare professionals in injury clinics are well positioned to identify patients experiencing IPV, since they often develop long-term interactions with women during repeat clinic visits for follow-up of fractures and associated surgical procedures. 24 This information, combined with our finding that the prevalence of IPV is high in orthopaedics, leads us to believe that the development, and assessment of an IPV identification and support programme in injury clinics is warranted. Although IPV screening has been controversial historically, 25 the US Preventative Services Task Force has recently published a statement recommending that all clinicians screen women for IPV and provide referrals to appropriate support services. 26 Additionally, Liebschutz and Rothman 27 recommended that all primary care providers screen all women older than 12 years and suggest some simple steps for identification and referral.
The Canadian Orthopaedic Association and American Medical Association judge it as good medical practice to take steps to identify and offer assistance to IPV victims, 28,29 but health-care providers have been consistently shown not to screen for IPV routinely -eg, in our study, only 6% of women had previously been asked about IPV by a health-care professional. Of the women who attended a fracture clinic with an injury resulting from IPV, only 14% had ever been asked about IPV by a health-care provider. Health-care professionals in injury clinics have an opportunity to identify the patients who were missed in the emergency department.
Almost two-thirds of the women in our study suggested that orthopaedic surgeons should ask all women presenting after trauma whether they are experiencing IPV, indicating that patients with injuries are generally supportive of this notion. Nonetheless, orthopaedic guidelines for management of IPV victims are new and are not implemented widely in orthopaedic practices. 30 Consequently an intervention programme in orthopedic and trauma clinics is needed. Such a programme should be supported by an appropriate referral and support scheme and should also include professionals who are specifically trained to assist IPV victims 31 .
A recent survey 32 in India found that 31% of Indian women of childbearing age had been subject to physical domestic violence. The low prevalence recorded in our study could be related to differences in the way that we gathered data, the cultural acceptability of the questions asked, and the fact that participants form India tended to be young women who had never been in a relationship. Moreover, an earlier pilot screening study in India 33 identified several key cultural issues that may account for the low reported prevalence. These issues were an inherent belief that for health-care providers to ask about IPV in a fracture clinic is not appropriate, the fact that Indian women were shy, patients might be offended and never return to the hospital, and domestic violence and abuse is often regarded as normal and women might not recognize that they are being abused or might perceive the behavior to be standard 33 . Furthermore, the use of WAST and other checklists might not translate culturally in India 33 . Twenty-two women from the site in India declined to participate after they learned that the topic of the study was IPV, suggesting that participation bias was probably a factor. Additionally, the youth and absence of an intimate relationship in the women from India probably explains the lower estimate of IPV prevalence that has been reported previously. Another explanation for the 114 under-reporting in India could stem from a pervasive attitude in men and women in that country who do not view IPV as a contributor to health issues. 34 The results of our study and the lessons learned will help to inform our future planned initiative to further investigate the prevalence of IPV at clinical sites in that country.
Health-care professionals should be especially concerned about escalation of abuse to intentional physical injury with associated fractures, which suggests that victim identification and support to ensure their future safety should be a high priority. The escalation of physical violence is a key risk factor for intimate partner homicide 35 and is the first question on the widely used danger assessment scale. 36 Additionally, between 2002 and 2009, the top cause of death after domestic violence was trauma (42% of cases). 35 Orthopaedic fracture clinics, as our results suggest, are key to the identification of women subjected to severe IPV who might be at increased risk of further injury and homicide.

WOMEN'S HEALTH SURVEY
Thank you for your interest in completing this questionnaire. We greatly value your willingness to review the questions below and your responses will help orthopedic surgeons and nurses better understand the issues of women's health and intimate partner violence within orthopedics.
Some of the questions may be uncomfortable for you to answer, or you may be irritated if the question has no bearing on your life. However, we ask that you try your best in answering all of the questions. Your participation is important to us and for those whom would benefit from this research! Also, please feel free to keep the intimate partner violence resource page at the end of this survey should you feel that it would be of use to you.

What is your annual income in Canadian dollars?
Less than $20 000 $60 000 to less than $80 000 $20 000 to less than $40 000 $80 000 to less than $100 000 $40 000 to less than $60 000 $100 000 or more Intimate partner violence (IPV) against women has been reported to have a negative impact on both the physical and mental health of victims, which extends beyond the period of abuse (Zolotor, Denham, & Weil 2009). IPV has been defined by the American Medical Association as "a pattern of coercive behaviours that may include repeated battering and injury, psychological abuse, sexual assault, progressive social isolation, deprivation and intimidation" (McCloskey et al. 2007). These behaviors are directed toward a perpetrator's spouse, common-law partner, sexual or dating partner. The long-term consequences of IPV include health risks, post-traumatic stress disorder, depression and staggering economic costs for the healthcare of victims .
The lifetime prevalence of IPV has ranged from 20 to 40 percent among women in North America (Renker, 2008). In the documented cases of IPV in Canada alone, 40 percent of women who have experienced IPV suffered a physical injury, and 15 percent of these cases were serious enough to warrant medical attention . These women present to a wide variety of healthcare professionals including emergency room physicians, orthopedic or trauma surgeons, family physicians, and specialists in obstetrics and gynecology. These medical specialists play a vital role in the detection and treatment of IPV among their patients.
Despite this high prevalence and potentially serious consequences of IPV, screening for IPV is often contested in the literature, and several academics and organizations have expressed their lack of support for IPV screening in health care settings. Current guidelines from the U.S. Preventive Services Task Force (USPSTF) state that routine screening for IPV is not recommended (U.S. Preventive Services Task Despite recommendations for routine IPV screening in various medical settings, as well as the introduction of IPV screening and intervention protocols in emergency departments (Rhodes et al., 2011), healthcare workers do not routinely screen for IPV even when treating injuries (D'Avolio, 2011). Primary care and emergency room physicians screen about 1.5 to 13 percent of their patients while obstetricians, and gynaecologists screen about 10 percent of their patients for IPV . Failure to diagnose IPV and intervene may have detrimental outcomes as 44 percent of 127 domestic violence related homicide victims had presented to an emergency department within 2 years of their deaths . As a result of their relative inaction in identifying and providing support to women who experience IPV, health care providers have been criticized for letting these women fall through the cracks. According to Alpert (2007), "[t]he ability of most health professionals to effectively identify, assess, and respond to domestic violence has lagged far behind societal awareness and community responses".
In response to the low IPV screening rates, a number of researchers have conducted studies to investigate barriers to IPV screening across various medical settings. The reported barriers include time constraints, unpreparedness to screen for IPV and discuss the issue comfortably and thoroughly, lack of clarity about implementation of screening and inadequate referral resources (Sormanti & Smith 2010; Colarossi, Breitbart, Betancourt 2010). Although barriers to screening for IPV are widely reported, a systematic review across studies with an emphasis on a quantitative approach, and informing opportunities for education and research has yet to be conducted. The current authors therefore undertook a systematic review of the published literature to identify frequently reported barriers to IPV screening across different medical sub-specialties to improve the understanding of IPV screening barriers among different health care professionals.

Eligibility Criteria
The authors identified articles in English that met the following eligibility criteria: 1) the study was published before May 2011 (when the literature search was performed), 2) the study reported HCP barriers to screening for IPV and 3) the study was identified in our search process (see next section). Studies were excluded for the following reasons: 1) the article was a review study, meeting abstract, commentary etc.; 2) the paper reported on a duplicate study population; 3) the study did not ask HCPs about barriers to screening for IPV; and 4) the article did not present any quantitative data on barriers to screening (e.g. stating the percentage or proportion of those who reported certain barriers to screening),.

Identification of Studies
The authors conducted a search of MEDLINE for relevant articles published before May 2011 in all languages. They used the following key words in their search: (Spouse abuse (MeSH) OR Domestic violence (MeSH) OR Partner violence OR Wife abuse) AND (Challenge OR Challenges OR Barrier OR Barriers). To identify any studies that they did not capture with our MEDLINE search, they consulted with a content expert, and reviewed reference lists from articles that fulfilled our eligibility criteria.

Assessment of Study Eligibility
Two of the authors independently assessed the studies identified for full evaluation and resolved disagreements through discussion until consensus was reached. One author additionally verified inclusion of all studies.

Assessment of Methodological Quality
Four reviewers independently graded the methodological quality of included studies, with two reviewers grading each paper. At least one reviewer with methodological expertise and one reviewer with content expertise graded each study. Each of the methodological experts who were involved had a graduate degree in clinical research methods and a great deal of experience in conducting and reporting systematic reviews. The reviewers with content expertise each had an extensive background in IPVrelated research through participation in both qualitative and quantitative studies for graduate theses.
Questions , and the current authors found that none of the previously developed checklists were appropriate for this study. Therefore, they adapted the STROBE statement to assess study quality (Appendix 1). A priori they determined that studies that met at least ten of the eleven reporting criteria within their modified STROBE criteria would be considered high quality, studies that met seven to nine reporting criteria would be considered moderate quality, and studies that met six or fewer criteria would be classified as low quality.

Data Extraction
Two authors developed a structured data extraction form, and four reviewers (two reviewers per study) extracted data from each eligible study. This structured form ensured that all data was being extracted and recorded consistently and completely by all reviewers. They extracted data on study characteristics, including location of study, type(s) of HCPs involved in the study, method of data collection etc. They extracted data on what HCPs perceived were barriers to screening for IPV, including what proportion of HCPs in the studies identified particular barriers.

Data Analysis
The authors produced an extensive list of barriers that HCPs in the included studies identified. They observed similarities between specific barriers and condensed the list by grouping the barriers based on their common themes. They proceeded to merge notably similar themes into fewer, more general themes. When there was disagreement 129 between reviewers regarding what the coding of a particular barrier, the reviewers would discuss their reasoning until consensus was reached. Afterward, they constructed a table to identify how many and which studies in particular reported each barrier. They grouped the barriers into five separate categories including personal barriers, resource barriers, attitudes and perceptions, fears, and patient-related barriers.
To identify any changes in the perceived barriers over time, they later grouped the studies by the time in which they were published into two ranges. 1995-2005 and 2006-2010. They analyzed the data from these groups to see any trends in the older versus more contemporary perceptions of barriers to screening for IPV.

Evaluation of Heterogeneity
Before analyzing the data, they hypothesized that a large degree of heterogeneity would be observed between the studies. Differences such as study methodology (e.g. mail survey versus in-person survey), HCPs surveyed (e.g. physicians versus nurses versus dentists), differences in survey tools used or questions asked, variability within and between populations (e.g. socioeconomic status, age, location), or variable study quality could contribute to the heterogeneity.

Study Identification
The authors identified 496 articles through their MEDLINE search that were possibly relevant: 421 of which were excluded after review of titles, either because they were irrelevant to the subject matter or because they were not in English and thereby did not meet the inclusion criteria. An additional 43 studies were excluded after review of abstracts (Figure 1). Fourteen of the remaining (19) studies were excluded after full-text review. Nine of these studies were excluded because they did not include proportion data, four studies were excluded because they did not ask HCPs about barriers to screening for IPV, and one study was excluded due to its use of a duplicate study population, which left 18 articles included from the MEDLINE search (Figure 1). The authors also identified and included two studies after consultation with a content expert, and they included two studies after reviewing the references of other included studies. They thus included a total of 22 studies in this review from all examined sources.

Study Characteristics
Six of the included studies reported mean age of respondents, ranging from 38 to 46 years; nine studies reported a range of ages ( Table 1). Overall ages of respondents ranged from 20 to over 80 years. The remaining seven studies did not report ages. The response rates varied widely from 14.6% to 98%, with two studies failing to report response rate and almost half of the studies (9/22, 40.9%) reporting a response rate, of 70% or greater. The percentage of female respondents in the included studies ranged from 9% to 100%. Most studies (17/22, 77.3%) were conducted in the United States.

Study Quality
Questions adapted from the STROBE statement for cross-sectional studies were used to evaluate the quality of the 22 selected studies. We categorized five studies as high quality, nine studies as moderate quality, and eight studies as low quality ( Table 1).

Personal Barriers
In more than half of the included studies (12/22, 54.5%), HCPs reported personal discomfort with discussing the topic of IPV as a barrier (Figure 2). This was followed by concern for personal safety (5/22, 22.7%) and concern of misdiagnosis (3/22, 13.6%). Forgetting to ask about abuse, personal history of abuse, and lack of confidence to refer victims was each reported as a barrier to screening for IPV in two studies.

Attitudes and Perceptions
The most frequently reported perception was that it is not the HCP's role to screen for IPV (10/22, 45.5%) followed by the perception that HCPs have more pressing issues to address (4/22, 18.2%), that abused women should be blamed for the abuse (4/22, 18.2%), and that abuse is rare (2/22, 9.1%) (Figure 2). In addition, the perception that battered patients do not want a referral and that what HCPs view as abuse, the patients view as normal, were each reported in one study.

Fears
Fear of invading the patient's privacy and fear of offending patients who were not abused were each reported as barriers to screening in half of the included studies (11/22, 50%) (Figure 2). This was followed by fear of the partner's reaction and making life more difficult for the victim (4/22, 18.2%). Fear of police involvement and fear that the patient will stop seeing the HCP if he or she asked about abuse were each reported in one study.

Patient-related Barriers
The most frequently reported patient-related barrier was that the patient's language interfered with effective screening (3/22, 13.6%), followed by the perception that patients with psychosocial issues and/or difficult personalities were difficult to screen (2/22, 9.1%), and that the abused victim would stay with the abuser anyway (2/22, 9.1%) (Figure 2). The perception that patients would deny battering as a cause of injury, that patients feared repercussions of being identified, that patients would not mention abuse in their medical history, and that patients would not be aware of their rights were each reported in one study.

Changes in Perceived Barriers Over Time
Nine of the twenty two of the studies included in our review were published between 2006-2010, while thirteen studies were published within the 1995-2005 range. Most of the perceived barriers were reported proportionately between studies from these ranges, including time constraints, lack of office protocol, the perception that it is not within a HCP's role, language barriers, and the fear of making patients angry if they were not abused.
There were some discrepancies in the barriers that were reported between studies published in the earlier versus later range. Seven of the nine (87.5%) studies from the more recent range cited personal discomfort with the issue of IPV as a barrier to screening, as compared to five of the thirteen (38.5%) studies published between 1995-2005. Inadequate follow-up resources/support staff to assist in victim education, safety planning, legal advocacy and referral and feelings that HCPs cannot help patients who are abused were cited as a barriers in four of the nine (44.4%) studies published between 2006-2010 as compared to ten of the thirteen (76.9%) of the studies from the earlier range. Three of the nine (33.3%) of studies published within 2006-2010 demonstrated that HCPs' perceptions that they had more pressing issues as a barrier, as compared to one of thirteen (7.7%) studies from the earlier range. Lastly, studies from the five of the thirteen (38.4%) studies included from the 1995-2005 range indicated that the HCP's fear of making the patients angry if they were not abused as a barrier to IPV screening, while six of the nine (66.7%) studies from 2006-2010.

Consistency of Barriers Across Specialties
Barriers reported by HCPs in the included studies were similar across various medical specialties, except that musculoskeletal-and dental-related specialties did not report any barriers characterized as patient-related. Each of the four following barriers was reported by all specialties: personal discomfort with the issue, lack of knowledge, time constraints, and fear of making patients angry if they were not abused.

Discussion
The authors' systematic review found a wide range of barriers to screening for IPV perceived by HCPs in various medical subspecialties. The major barriers to screening for IPV reported by HCPs were attributable to a lack of resources. The most frequently reported resource barriers were time constraints; lack of knowledge, education, or training regarding screening of abuse; and inadequate follow-up resources and support staff to assist victims. Of the personal barriers, fears, and perceptions, the most frequently reported barriers to screening were personal discomfort with the issue of IPV, fear of invading the patient's privacy or offending patients if they were not abused, and the belief that it is not the HCP's role to screen, respectively. Patient-related barriers were much less often reported as barriers to screening for IPV than HCP-related barriers. The current findings are consistent with those of Waalen and colleagues in their 2000 review of 12 studies addressing barriers to screening for IPV by HCPs and interventions to increase HCP screening behaviors, in which they found that lack of provider education and lack of effective interventions were the most frequently reported barriers in both open-ended interviews and written surveys; limited time was most frequently reported in open-ended interviews, and fear of offending or endangering the patient was reported most frequently in written surveys . The current review improves upon Waalen's review by including more recent studies and by considering a greater number of studies. Furthermore, within the current review, authors examined the differences in reported barriers over time.
The current authors also found that the barriers reported by HCPs in the included studies were similar across various medical specialties. This is comparable to the findings of the same study by Waalen et al. in which the lists of barriers reported by health professionals were similar despite differences in methods and provider population .
Although a number of studies have suggested that patients support IPV screening ; McCauley, Yurk, Jenckes, & Ford 1998), the current findings indicated that several attitudes and perceptions held by HCPs act as barriers to screening, which may be related to lack of education and personal discomfort with the issue. Education, and training for HCPs should address such perceptions in an attempt to remove these barriers . In addition, these resources should clearly outline the role and responsibilities of HCPs in addressing IPV and highlight the importance of screening in their practice . According to Erickson, Hill and Siegel (2001), family practitioners who were more educated and experienced with the issue of domestic violence were less likely to report lack of time as a screening barrier. This suggests that the perceived lack of time to screen for IPV, which was the most frequently reported barrier in this review, may be a result of lack of knowledge, and that with increased training and experience, time constraints may become less of an issue.
The 22 studies included within the review were published over the span of 1995-2010. Due to this considerable range, the authors expected to see some differences in the reported barriers to IPV screening in earlier versus later studies. They hypothesized that these differences might be attributed to system-level changes including policy physician training. The fact that a greater proportion of the more recent studies indicated that HCPs' personal discomfort was a barrier in screening, as compared to earlier studies reflects a need for more training on the matter of IPV so that HCPs may feel more comfortable and confident in their ability to screen for it in their patients. Conversely, the authors also found that a lesser proportion of more contemporary studies cited inadequate follow-up resources/support staff to assist in victims and feelings that HCPs cannot help patients who are abused as barriers, which may reflect recent improvements in the area of physician education and resources in the area of IPV.
Three of the nine (33.3%) studies published within 2006-2010 demonstrated that HCPs' perceptions that they had more pressing issues as a barrier, as compared to one of the thirteen (7.7%) studies from the earlier range. This could be because of increased workload demands on the contemporary HCP.
The results of this review were limited by several factors, one of which included small sample sizes and low response rates in some of the included studies. In addition, the heterogeneity across the included studies as a result of variations in study design, methodology, provider characteristics and specialties, limited the direct comparison of studies. Another limitation was the high number of low-and moderate-quality studies that were included in the review, and the inclusion of articles that were only published in English. The authors did not include grey literature in our review, which may introduce bias because published articles are more likely to show positive results.
Additionally, although most of the included studies reported demographic information on their respondents (e.g. age range, mean age, sex), the vast majority of studies pooled their data and did not report their findings by age and/or sex. This limited the authors' ability to report any trends in perceived barriers in male versus female and younger versus older health care providers. This would be valuable information in developing strategies to engage a wide range of health care providers. Similarly, it would be beneficial to explore discrepancies in reported barriers to IPV screening based on the economy of the countries included in our review. However, only two of the studies included in the review took place in non-industrialized nations, while the remaining twenty took place in either Canada, the United Kingdom, or the United States (all industrialized nations). The authors feel that it would be biased to identify differences in reported barriers between non-industrialized versus industrialized nations with such a small sample of studies from non-industrialized nations. They would recommend that future reviews take into account potential variation in reported barriers to IPV screening between health care providers of different sex, age, as well as socioeconomic status. As the literature base on IPV screening grows, it would be also be valuable to learn of occupation-specific barriers to IPV screening which may indicate which health care providers are best suited to successfully identify and respond to cases of IPV. The authors recommend that this is examined in future reviews. Some bias may have been introduced in grouping older versus more recent studies based solely on their date of publication. Frequently, results of a study are published several years after the study was initially conducted. Unfortunately, the authors were limited to the information that the studies reported with regards to the time frames of their study, from data collection to publication. Seven of the twenty two studies included in the review (31.1%) reported the time period that data was collected, making it impossible to compare differences in reported barriers over time based on grouping studies by the time they were conducted.
Despite these limitations, the review had several strengths, including a thorough and systematic search of the literature, followed by thorough and systematic screening of the studies for inclusion, and completion of data abstraction from the included studies by multiple independent reviewers. This review was also strengthened by its broad inclusion criteria. The results of this review indicated numerous barriers to IPV screening that need to be addressed to increase IPV screening by HCPs. This review has demonstrated that provider-related barriers were reported more often than patient-related barriers in the current literature. The authors feel that this finding is promising and that these barriers are easier to address and can be done so systematically through change in policy and education of health care providers. It serves to advocate for the continued training of medical professionals to recognize and treat IPV cases, which will ultimately increase the quality of life for victims. Education and training for HCPs should focus on increasing awareness and exploring perceptions on IPV screening to remove barriers. Future research should explore optimal methods of disseminating IPV information among HCPs and developing efficient methods of screening.

Background
Intimate Partner Violence (IPV) is described by the American Medical Association as "a pattern of coercive behaviors that may include repeated battering and injury, psychological abuse, sexual assault, progressive social isolation, deprivation and intimidation" [1]. Injuries associated with IPV often require treatment provided by orthopaedic surgeons [2]. Canadian orthopaedic surgeons may not recognize the extent that IPV affects the patients seen in their clinics; an overwhelming 87 percent who participated in a national study believed that female victims of IPV accounted for less than one percent of patients in their care [3]. A recent prevalence study found that one third of women attending two fracture clinics for an orthopaedic injury had experienced physical, emotional, and/or sexual abuse within the last 12 months [4]. This rate of IPV is much greater than the orthopaedic surgeons estimated [3], and provides a rationale for IPV screening and support programs in orthopaedic clinics.
To address the low rates of screening, previous studies have explored barriers to IPV screening among various health care professionals such as emergency department health care workers, obstetricians/gynecologists, family physicians, internists and health care staff in family planning organizations [5,6,7,8]. A recent systematic review of 22 studies investigating barriers to IPV screening reported by health care professionals described five categories of barriers: 1) patient-related barriers; 2) health care provider fears; 3) lack of resources; 4) personal barriers; and 5) health care provider misconceptions [8]. Across the included studies, the most commonly cited barriers to screening for IPV were personal discomfort with the issue of IPV, lack of time, and lack of knowledge about IPV [8]. This review did not find any studies that comprehensively examined the perceived barriers to IPV screening among health care professionals who treat patients in the orthopaedic fracture clinic setting [8].
This research aimed to address this gap in the literature by exploring perceived barriers to IPV screening in the orthopaedic fracture clinic and by identifying potential facilitators for addressing these barriers among orthopaedic surgeons and surgical trainees (senior and junior orthopaedic residents).

Qualitative Method and Rationale
This research was conducted using the qualitative descriptive approach, a qualitative research method which aims to provide a descriptive summary of the research organized in a way that best reflects the data. This method is described by Sandelowski [9] as being valuable when straight descriptions are required to provide answers to questions of special relevance to practitioners and policy makers.
The authors are qualified to conduct this study because we have an interdisciplinary mix of expertise in qualitative methods, intimate partner violence, orthopaedic surgery, and trauma surgery. Some authors had preconceived ideas of what some of the barriers may be and others did not. The mixture of areas of expertise contributed to the multiple perspectives needed to effectively analyze the information of interest in an appropriate and holistic context.

Data Collection
Data were collected through focus groups comprised of orthopaedic surgeons and orthopaedic surgical trainees. Focus groups are generally recommended for qualitative descriptive studies as they typically provide a broad range of information about experiences [9]. The opportunity for interactive discussion during the focus groups enhanced the ability to collect in-depth data on the perceived barriers to and facilitators for screening for IPV in the orthopaedic fracture clinic. We conducted three separate focus groups -one with orthopaedic surgeons, one with senior surgical trainees, and one with junior surgical trainees. When transcribing data, we did not record the names of any participants to preserve their privacy and confidentiality. We assigned each participant a code, such as "Surgeon 1" or "Junior resident 2". The codes were kept in a secure location and only one member of the team had access to the codes.

Sampling
Sampling for this study was purposeful with an emphasis on maximum variation sampling. The goal of purposeful sampling is to select participants who provide "information-rich" cases, that is, participants who provide data that will allow us to learn in-depth about the phenomenon of interest [10]. In this study, we purposefully sampled to include variation in health care professional type (i.e. surgeons and surgical trainees) because we believe that this variable has the potential to influence experiences with and perceptions about screening for IPV. The research coordinator (KM) emailed invitations to the individuals that were selected to participate until we reached our intended sample size.
To facilitate an open and comfortable environment for discussion, the focus groups themselves were homogeneous in terms of health care professional type. This approach allowed us to analyze our data within profession type (i.e. surgeon versus surgical trainee) and to compare our findings between the two groups. Any common themes that emerge from variation are particularly relevant to the research question [10].

Sample Size
We included six or seven participants in each focus group for a total sample size of 20 participants plus two interviewees. Our sample provided enough saturation [11] in the data to adequately describe the perceptions about barriers to and facilitators for IPV screening by orthopaedic surgeons and surgical trainees.

Recruitment
A recent prevalence study conducted at the Hamilton Health Sciences General Site found that the prevalence of IPV (physical, emotional, sexual) among female patients in orthopaedic fracture clinics during a period of 12 months was alarmingly high; 31.6% [4]. As a result, participants for this research were recruited from this fracture clinic, as well as two other fracture clinics in Hamilton, Ontario, all which are affiliated with an academic teaching institution. Individuals who agreed to participate in the focus groups were likely to have an interest in IPV and yield the "information-rich" cases that Patton describes as important for qualitative research [10]. Potential participants were sent an email invitation which briefly outlined the subject, purpose, agenda and expected outcomes of the focus group. Prior to proceeding with this study, ethics approval was 147 obtained from the local Research Ethics Board (Project Number: 11-491). We asked each participant to sign an informed consent form before proceeding with the focus group or interview. Participants were not financially compensated for participating in this study and they were assured that their participation or lack thereof would not impact their employment or residency status.

Focus Groups
We used a semi-structured focus group guide to structure the discussion about health care professionals' experiences with and perceptions about screening women for IPV. This approach is useful as it provides participants with some guidance on what to discuss while also enabling exploration of issues that may not have been considered by the researchers [12]. Participants were also asked to complete a brief demographic questionnaire.
The focus groups were conducted in private location at a time convenient for participants and were facilitated by an experienced focus group facilitator (MS). The focus groups were digitally recorded and the recordings were transcribed verbatim.

Structured Interviews
To address institutional and personal barriers such as those identified by the focus group participants, we recognize that it is important to have institutional support and changes to policies at a level above the fracture clinic. We chose to interview two opinion leaders in the field of orthopaedic surgery, who are well-versed and active in health policy in the field of orthopaedic surgery. The purpose of the interviews were to help us to better understand the barriers at the policy level and to identify facilitators for making changes to policy to better assist IPV victims within the fracture clinic setting. The independent, semi-structured telephone interviews were conducted after the completion of the three focus groups which allowed for a more detailed exploration of the themes identified during the focus group discussion. A more experienced interviewer (MS) conducted the two interviews which were digitally recorded and transcribed verbatim.

Data Analysis
In qualitative research, data collection and data analysis usually occur simultaneously to allow for new themes in the early data to be incorporated into collection of later data. In this study, we used field notes from the focus groups to identify new themes to explore in future focus groups and we began coding after each focus group. As data collection proceeded, new data and new insights about the data were incorporated into the data analysis, making it reflexive and interactive.
Four investigators (SS, MS, RS, and KM) participated in the data coding and analysis of the focus groups and interviews. When all of the transcripts from the focus groups were coded, the four investigators met to organize the codes into meaningful clusters [13,14] and discussed potential relationships between the categories, a process known as axial coding [15]. The analysis resulted in an organized and comprehensive summary of orthopedic surgeons' and surgical trainees' experiences and perceptions related to IPV screening in orthopaedic fracture clinics. The transcripts from the two interviews were coded by the four investigators following the procedures described above.
Transcripts and field notes from the focus groups and interviews were analyzed using conventional qualitative content analysis, as recommended for qualitative descriptive studies [9]. In conventional qualitative data analysis the coding categories are derived directly from the data rather than using preconceived categories [13,14]. The qualitative software program N'Vivo (version 10.0; QSR International, Melbourne, Australia) was used for data management and analysis.

Rigor
The study's credibility was ensured by documented evolution of coding and analysis as well as coding decisions [16]. Rigor was also achieved through the use of multiple coders, many with a strong knowledge of the IPV literature, and coding consensus meetings. All categories were firmly grounded in the data by identifying sections of the transcripts from which they originated [17] and quotes were used to illustrate the codes which further demonstrated a good fit between the data and the analytic results.

Participants
We invited ten surgeons, ten junior residents, and ten senior residents to participate in the focus groups. Each of our three focus groups included six or seven participants, with a total of 20 surgeons or surgical trainees participating across the three focus groups. The mean age of the focus group participants was 33.9 ± 8.6 years and the majority of the focus group participants were male (75%) ( Table 1). The mean length of time in practice for the orthopaedic surgeons was approximately ten years.

Themes
Four main themes were identified: contextual thoughts on IPV management, barriers to screening for IPV, facilitators for screening for IPV, and policy implementation for fracture clinics. Within the contextual thoughts on IPV, the following two subthemes were identified: perceptions and observations, and comparison to external models. The focus group discussions yielded additional subthemes that reflected barriers to screening for IPV including fracture clinic barriers, perceived barriers for patients, perceived barriers specific to surgical trainees, and perceived barriers for surgeons. The following subthemes under the overall category of facilitators for IPV screening were identified: system-level characteristics, fracture clinic processes, and personnel resources. The policy implementation theme emerged from the analysis of the two interview transcripts.

Contextual Thoughts about IPV Screening the Fracture Clinic Setting
The focus groups began by having participants share their initial thoughts on IPV and some of their personal experiences with IPV in the fracture clinic (Figure 1). One focus group began with the following example, "I have a face of intimate partner violence and murder… Two patients, one tried to have his wife killed. She became my patient because they were unsuccessful in killing her". Most participants recognized and understood the importance of screening for IPV in the fracture clinic setting with one participant noting "I think the screening is hugely important because it probably opens up a door to a certain percentage of women that would then open up and tell you about it" and another describing "I think we are the first access point for those women." The participants also discussed the need for caution when screening to ensure the safety of the patient.

150
When discussing perceptions about common presentations of IPV within the fracture clinic setting one participated noted, "… identification is key… that's the part that I am most worried about is that I am missing people"; while another participant described: "...the biggest problem is picking it [intimate partner violence] up 'cause I think so much of it is silent." Another participant summarized current IPV screening practices: "It seems like if we are going on our intuition and our sense that something is wrong, we are doing a bad job with that from the evidence." Surgeons and surgical trainees agreed that they are more comfortable with screening for and caring for patients who are victims of child abuse as a result of policies being in place and having received appropriate training. One participant compared knowing the steps to take after screening for child abuse with not knowing the steps to take after screening for IPV: "With children we sort of know what to do, who to call. There's like a social work team we call but, you know, if you screen and then they do sort of come forth with "yes it was intimate partner violence" I wouldn't really know what to do next and so it's a little intimidating".
Participants also discussed selective screening for IPV (asking people about IPV based on pre-determined risk factors) versus universal screening for IPV (asking everyone in the fracture clinic about IPV) within the fracture clinic setting. For example, one participant said: "I think it's more ideal for screening everybody…because if not we are gonna screen people based on our assumptions and I don't think that's appropriate." Most participants recognized the limitations and challenges with selective screening and agreed that universal screening for IPV may be appropriate, citing the success of universal screening for osteoporosis within the fracture clinic setting.

Fracture Clinic Barriers to Screening for IPV
Participants described how the layout and organization of many fracture clinics makes it challenging for the orthopaedic surgeon to have privacy with their patients (Figure 2). Patients are often accompanied by someone to their appointments and it is difficult to separate the patient from this person. In addition, within the academic setting, orthopaedic surgeons rarely see their patients alone as they are usually accompanied by surgical learners when seeing their patients. One participant explained: "There's six other people, at least six plus learners so probably twelve people listening to every single conversation I have with patients; it's not the appropriate place". In addition, many fracture clinics follow an open concept model, with curtains separating exam rooms. One participant made the following analogy: "The fracture clinic is the equivalent of a family doctor seeing patients in their waiting office".

Figure 2: Perceived Barriers for Screening for IPV in the Fracture Clinic
Focus group participants also identified that adequate patient histories are not readily available or quickly accessible and noted that this is a barrier. Since the patient is often initially seen at the emergency department or by a surgical trainee, the surgeon may not be aware of their full history including previous injuries and the nature of these injuries.
Orthopaedic surgeons and surgical trainees described that they spend a very limited time with individual patients and have a high number of patients to see within their fracture clinic, which makes it difficult to find the time to appropriately screen for and address IPV. One participant described: "The biggest thing is time for orthopaedic surgeons. We see 70 to 80 people in a five hour clinic so if you do the math that's less than a few minutes a person." Both the surgeons and surgical trainees identified that there is a lack of policies on screening for and addressing IPV within the fracture clinic setting. The following exchange between the facilitator and several surgeons exemplifies this concern: Facilitator 1-"So am I right in understanding that the clinics you work in have no policies or sort of integrated mechanisms ... for screening for intimate partner violence?" Several Surgeons-"None that I'm aware, that's correct."

Perceived Barriers for Injured Patients Regarding Screening for IPV
The focus group participants were concerned that universal screening may be challenging if the patients are unable to speak, read, and/or write English. They were also concerned about responding appropriately to possible cultural differences and distinguishing cultural practices from signs of IPV.
Another perceived patient barrier identified during the focus groups was the brief interaction that the patient has with the surgeon, and that the short amount of time spent with the patient is focused on the patient's orthopaedic problem. As one participant explained, "It's hard to develop a feeling of trust in a short period of time." The focus group participants acknowledged that IPV victims may be more comfortable disclosing to females than males, and the field of orthopaedic surgery is predominately male.
Focus group participants recognized that a patient-surgeon power imbalance may be present which could deter patients from disclosing. One participant described this imbalance: "There's already a power balance right there and then there's the body language alone and we almost always wear white coats in there. There's just a lot of overlay that would probably inhibit the patient from disclosing some information." There was concern that patients may not want to disclose IPV, as revealed by the following comment: "If someone comes into the fracture clinic they may just want care for their physical injuries". However, one participant disagreed, stating that "patients tend to trust medical professionals and may want to open up to someone". Participants also discussed that women may not disclose due to fear of the consequences of disclosing.

Perceived Barriers Specific to Surgical Trainees Screening for IPV
In general, the surgical trainee participants felt pre-occupied with their learning and clinical activities and felt that they would not have the focus to screen for IPV. This is demonstrated by the following comment, "You are trying to do a good job at what you are, you know, at becoming a good surgeon and that's taking a lot of our energy away".
Surgical trainees have multiple demands on them, as illustrated by the following quote, "We're busy with our … patient load and our next exam coming up and our … evaluation and our fellowships and our no jobs and our everything else that we think about all the time and … that you are going into the OR the next day and you have to read this tonight…." The surgical trainees expressed concern about IPV screening influencing their relationship with the preceptor. The focus group participants clearly indicated that their preceptor would have to be supportive and encourage IPV screening.
The focus group participants also expressed concern about the lack of education and training that they received on IPV and consequently they were unsure of what to do with a positive screen. A few participants felt fearful about being held accountable due to lack of ability to identify victims as indicated in the following comment, "you saw Mrs. Smith ... she had this injury, it was obviously domestic violence cause she's dead now and this was her fourth occurrence and you didn't pick up on it, why not doctor?".

Perceived Barriers Specific to Orthopaedic Surgeons Screening for IPV
The participants felt that the orthopaedic culture, which includes short term interaction with patients and a certain personality-type, is an important barrier. One participant explained, "I think it's a lot of the perceptions of the surgeon, you know, not that orthopaedics isn't a caring profession but there are some people that, you know, don't have the best bedside manner. So I can see people feeling very reluctant to kind of share something personal". The orthopaedic surgeons also described the belief that screening for IPV falls outside of the role of the orthopaedic surgeon and that they tend to treat the injury in isolation, as opposed to treating the social and other medical issues.
Focus group participants indicated that there is lack of a champion for promoting IPV screening and effective social support for IPV victims, as described in the following comment, "There's no champion right now who actually works at the fracture clinic."

System-Level Characteristics that are Facilitators for Screening for IPV in the Fracture Clinic Setting
The participants indicated that trust of the medical profession was an important facilitator, as patients often open up to medical professionals (Figure 3). One of the participants commented, "I'm sometimes surprised at how open and forthcoming patients are in the short time you get to know them the things that they'll tell you. I mean I think there is a sort of inherent trust in the medical profession".
The focus group identified education for surgeons and surgical trainees as an important facilitator, as described in the following comment, "I think the next step would be towards education and make it not just completely academic but with these things I think you have to have case scenarios…"

Changes to Fracture Clinic Processes That Would Facilitate Screening for IPV in the Fracture Clinic
The focus group participants suggested numerous changes to the fracture clinic environment to facilitate IPV screening and care for IPV victims. Examples described included having a private area where surgeons could talk to patients about confidential issues and having posters and pamphlets on IPV available within the fracture clinic to potentially help patients feel more comfortable with the topic.
Participants noted that having a prepared information package for surgeons would greatly assist in screening for IPV. One participant suggested "having something like a preprinted order pack, an abuse care package so that I know, not just academically what resources are available … it has resources for the patient that has a set of orders." Another idea that was raised during the focus groups with having a screening form which the patient would complete upon presentation to the fracture clinic which could help to identify high risk patients for the orthopaedic team to follow-up with.
The focus group participants suggested that they could flag the charts of potential IPV victims. The following participant describes this suggestion: "They can stick like a red dot on the chart or something and then we go okay maybe that red dot means that we should ask them …" The focus group discussions included suggestions for changing the infrastructure within the fracture clinic to facilitate screening for IPV. These suggestions include ensuring that there is adequate support from all fracture clinic staff and the hospital.

Personnel Resources for Facilitators for Screening for IPV in the Fracture Clinic Setting
Focus group participants suggested that surgical trainees have the opportunity to spend more time with patients so that they can effectively screen for IPV, as suggested in the following comment: "I think sometimes …patients might actually disclose [IPV] to surgical trainees more than surgeons because surgical trainees tend to spend more time with the patient in the fracture clinic in terms of taking the history and doing the physical exam".
There was a great deal of discussion on the need for a dedicated person within the fracture clinic setting to screen for and manage IPV. The following quote from a surgical trainee demonstrates this: "When we were doing the screening for the IPV Prevalence study we had, you know, students coming in and asking every patient "can I talk to you in private?" That did it quite effectively." Facilitators for caring for IPV victims included having a simple plan in place for the fracture clinic including the availability of a crisis team. A participant noted, "You should give them a simple plan that this is a person to contact when these are the problems in the fracture clinic or this is the person coming to the clinic every day three four hours and in case you find some clue you don't need to waste much time because you are busy".

Policy Implementation
Both opinion leaders were asked about the need for local and national champions to promote the need for IPV screening and care for victims within the fracture clinic setting (Figure 4). Both believed that a champion would be invaluable, as evident from the following quote: "I think, you always need a champion not only to develop the policy but also to implement the policy." The interviewees identified a need to increase awareness about IPV among all health care professionals working within the fracture clinic setting. One opinion leader explained, "…the data that shows in the last year 30% of women that present to a fracture clinic have been subject to some type of IPV and I'm not so sure that that people even know those numbers so I think, you know, getting the information out there that it impacts a huge segment of the population, that the impact is substantial …" The interviewees agreed that local policy on IPV would be beneficial in assisting patients who experience abuse. As one opinion leader stated, "...it's easier taking a grassroots approach, you know, go to a smaller entity like a hospital or clinic...if you have enough you know clinics that a or hospitals that are sort of buying into this program which is being run more at a grassroots level then maybe at the sort of provincial or systemic level someone might so oh this is working really well or seems to be really important, what we need to do is formalize this".
The need for research to inform policy was also discussed and both opinion leaders believed that the implementation of an IPV screening program should be evidence-based and they supported future research to better inform the decision makers.

Discussion
The present study aimed to identify barriers to and facilitators for screening and caring for IPV victims in an orthopaedic fracture clinic setting. Our findings are consistent with the published literature looking at barriers to screening for IPV in other medical specialties. Briefly, Sormanti & Smith and Colarossi et al. conducted focus groups with emergency department surgical trainees and health care professionals at an urban family planning centre to examine the perceived barriers to screening for IPV in these specific health care settings. Similar to our findings, they reported barriers to IPV screening that included time constraints, unpreparedness to screen for IPV and discuss the issue comfortably and thoroughly, lack of clarity about implementation of screening and inadequate referral resources [5,7]. Waalen et al. conducted a review of published studies reporting on barriers to IPV screening among various health care professionals [18]. They found the following provider-related barriers: a lack of provider education regarding IPV, lack of time, and lack of effective interventions. The authors also identified the following patient-related factors: patient nondisclosure and fear of offending the patient. This review also reported that barriers to screening for IPV are documented to be similar among health care professionals across diverse specialties and medical settings. Waalen et al. did not include any studies in the field of orthopaedic surgery [18]. A recent systematic review investigating barriers to IPV screening reported by a variety of health care professionals found many barriers that were also identified in the current study such as a lack of time, language and cultural barriers, lack of training, and lack of institutional protocols [8]. The current study identified several barriers that were special to orthopaedics that the systematic review did not report in other specialties. These included lack of a "champion", the male-dominated nature of orthopaedics, orthopaedic culture, and the perception of a patient-surgeon power imbalance. Interestingly, none of our focus group participants noted that personal discomfort with IPV, fear of offending patients, or the perception that abuse is rare are barriers to IPV screening, as were commonly noted in other specialties [8].
The published literature demonstrates an exploration of the factors that influence IPV screening in multiple areas of health care, although none of the previous literature focuses on orthopaedic fracture clinics. Our previous research in the field of IPV and orthopaedics has found that injuries requiring the consultation of an orthopaedic surgeon account for 28 percent of clinical manifestations of IPV, but IPV is underemphasized in this medical specialty [3]. A recent survey of orthopaedic surgeons reported that nine percent of the respondents believed that inquiring about IPV was an invasion of the patient's privacy, and eleven percent believed that ruling out IPV as the cause of injury was not part of their duty [3]. The current study further explored these biases and identified barriers to screening for and addressing IPV in the orthopaedic fracture clinic. Many of the barriers identified in the current study can be addressed with education and the appropriate resources and infrastructure. For example, surgeons and trainees can learn how to approach the topic with sensitivity and confidence in their knowledge of the appropriate plan to take in the event that a patient discloses. Additionally, fracture clinics can be designed or renovated to reflect a more private environment that is conducive to discussing sensitive topics such as a patient's experience with IPV.
Our focus groups found that the participants were, in general, supportive of implementing practices to improve IPV screening and care within the orthopaedic fracture clinic setting.
They identified numerous facilitators, which could be implemented into orthopaedic fracture clinics to help effectively identify and help IPV victims. After analyzing our codes from the three focus group transcripts, it was apparent that there were many shared opinions, attitudes, and experiences among participants. For example, a common theme identified in each focus group discussion was a lack of time and privacy. However, some of the barriers identified by the surgical trainees were unique because they did not apply to the practicing orthopaedic surgeons (e.g. surgical trainee preceptor relationship). There were no other differences in the responses between groups. Our interviews with opinion leaders confirmed the findings from the focus groups and provided insight into and suggestions for implementing policies for IPV screening and providing health care professionals with the structure and support to effectively care for IPV victims.
Our methodology aimed to reduce bias and yield a purposeful sample. The qualitative descriptive method chosen for this research minimizes researcher biases as this method involves minimal interpretation of data and focuses on presenting and organizing the data in the language used by participants [19]. Recruiting volunteers to participate likely led to a bias of having individuals who are interested in the topic of IPV screening participate in our research. Since the goal in purposeful sampling is to include participants who can provide "information-rich" data [10], this bias is favorable as participants who have an interest in IPV screening are likely able to provide detailed information, experiences and perceptions about perceived barriers to and facilitators for IPV screening. In addition, the questions within the focus group guide were worded to ensure that they were neutral and did not lead the participants. The focus group facilitator used neutral probes and reflective statements to clarify what had already been discussed and to encourage more discussion among participants. The use of homogenous groups of professionals ensured that participants within each focus group had similar education levels and clinical expertise which helped to minimize bias related to conforming and suppressing minority views that might be different from views of dominant participants within the group. This study is limited by a relatively small sample size and by having all focus group participants affiliated with one academic institution. We are also limited by having only interviewed two opinion leaders and therefore, the results of these interviews should be interpreted cautiously. The two opinion leaders were from different academic institutions which may improve the generalizability of our findings, and it should be noted that there are few opinion leaders in the field of orthopaedics in Canada who are well-versed on the topic of IPV and health policy. It is important to note that there may have been bias in deriving out focus group sample exclusively from one academic institution because this community of health care professionals and trainees has had previous exposure to IPV through education initiatives to increase awareness and sensitivity to the issue. Additionally, researchers at this institution have published research focusing on IPV within the field of orthopaedic surgery. This unique culture may have influenced the attitudes and opinions of the orthopaedic surgeons and surgical trainees and this may limit the external validity of our results. This study warrants replication in other jurisdictions and in a community setting to further determine the 158 generalizability of the findings. Future research could also investigate patient's opinions of screening and caring for IPV victims within the orthopaedic fracture clinic setting.

Conclusions
There are a number of perceived barriers to screening women in the fracture clinic for IPV, many of which can be addressed through increased education and training, and additional resources in the fracture clinic. Orthopaedic health care professionals were supportive of implementing an IPV screening program in the orthopaedic fracture clinic.

Introduction
Over the past several decades, intimate partner violence (IPV) has increasingly been recognized as a substantial cause of mortality and morbidity 1 . IPV is prevalent among diverse populations, affecting both men and women regardless of their socioeconomic status, sexual orientation, age, and ethnicity [2][3][4] . The myriad of long-term physical and emotional consequences of present and past IPV are well established [4][5][6] .
Several studies have found that IPV is severely underreported among both male and female patient populations 2,4,7 . A recent study assessing the prevalence of IPV among women presenting to fracture clinics in two Level-1 trauma centers found that the overall prevalence of IPV (emotional, physical, and sexual abuse) within the previous twelve months was 32% 8 . The high prevalence of IPV among health-care users demonstrates the need for improved approaches for identifying patients in order to provide them with the opportunity for referral.
One method of identifying victims of IPV is through systematic screening programs implemented across health-care settings. Screening programs seek to encourage victims to feel comfortable speaking about their experience with IPV, as this disclosure is rarely spontaneous 9 . The multitude of physical and psychological outcomes associated with IPV frequently lead to victims' increased use of health-care services, including emergency, primary, and specialty care services 10 . These visits offer several opportunities for identification of IPV victims. One systematic review of twenty studies focusing on the effectiveness and acceptability of screening women for IPV in medical settings revealed that screening programs directly increased the rate of identification of victims and that most women had no objections to being screened 11 . Various medical fields including obstetrics, gynecology and family practice have successfully introduced screening programs for IPV 12 .
The American College of Surgeons has stated that surgeons have a responsibility to identify IPV victims and help them to reduce or prevent further harm 13 . In 2009, the Canadian Orthopaedic Association (COA) stated that it is "good medical practice" for health-care providers to actively identify and offer assistance to victims of IPV 9 . To our knowledge, no previous studies have sought to identify whether patients in fracture clinics are open to being screened for IPV. The purpose of the present study was to explore patients' opinions and preferences regarding IPV screening programs in orthopaedic fracture clinics.

Study Design
We conducted a cross-sectional study to evaluate patients' perceptions and opinions on screening for IPV in the fracture clinic setting. The primary objective of the study was to determine whether patients in the fracture clinic thought it was acceptable for orthopaedic surgeons, orthopaedic nurses, and/or social workers in the fracture clinic to screen for IPV. Secondary key objectives included identifying (1) patients' preferences regarding the location, timing, mode, and duration of screening; (2) the type, age, sex, and race of the health-care provider by whom patients would feel most comfortable being screened; (3) differences between the attitudes and opinions held by men and women regarding IPV screening; and (4) differences in the reported preferences and opinions among the study sites. Four clinical sites in Ontario, Canada and one clinical site in the Netherlands participated in the study. Research Ethics Board approval was obtained from the McMaster University/Hamilton Health Sciences Research Ethics Board and from the Research Ethics Boards at each participating site.

Participants
Male and female patients who presented to the participating orthopaedic fracture clinics were screened for eligibility. Patients may have presented for immediate treatment of an injury, for follow-up of an injury, or for a chronic problem. In order to be considered for inclusion in the study, patients had to (1) present to the orthopaedic fracture clinic for their own appointment; (2) be at least sixteen years of age; (3) be able to read, understand, and write in English (in Canada) or Dutch (in the Netherlands); and (4) be able to separate themselves from anyone who accompanied them to the clinic, in order to ensure that they could complete the questionnaire in privacy. Patients were excluded if they were too ill or injured to participate or if they were cognitively impaired. Once patients were deemed eligible, informed consent was obtained and they were provided with the self-reported written questionnaire, which was completed anonymously. Participants were also provided with an IPV resource sheet for the local area if they were interested.

Study Questionnaire
The questionnaire comprised forty-five questions and was developed for the purposes of this study. The questionnaire provided patients with a definition of IPV and included questions on patient demographics, attitudes toward IPV in general, acceptability of IPV screening in an orthopaedic fracture clinic setting, and opinions on how, when, and by whom IPV screening in the orthopaedic fracture clinic should be conducted. The questionnaire was reviewed by content experts for face validity and piloted by a small subset of orthopaedic surgeons to ensure that the questions were unambiguous and appropriately targeted the domains of interest.

Study Sample Size
To ensure a ±5% margin of error around our estimates under the assumption of a total clinic population size of 20,000 and an anticipated 50% endorsement of screening across all centers over the recruitment period, we projected the need for at least 370 eligible and included study participants (http://www.raosoft.com/samplesize.html). To limit the risk of under powering of the study, we approximately doubled this value. The resulting sample size of 750 patients (150 from each of the participating study sites) was larger than those in previously published surveys on IPV screening [14][15][16] and was sufficient to provide meaningful results.

Data Analysis
Completed questionnaires were entered into a study-specific database, and descriptive statistics (including frequency counts and percentages) were calculated for all collected data. Continuous data are presented as the mean and the standard deviations. Chi-square tests were performed to determine whether any significant differences existed 164 among the clinical sites. Chi-square tests were also performed to determine whether any differences existed between the attitudes and preferences of men and women regarding IPV screening. For the series of questions inquiring about patient opinions, descriptive statistics were used to analyze the frequency of each answer.
The responses to each of the twenty questions related to the patient's openness to being screening for IPV were coded on a scale of one (least open) to five (most open), and the scores for the twenty questions were added to obtain an overall score (possible range, 20 to 100) for each participant. A stepwise regression analysis was performed to determine whether the following demographic characteristics related to a patient's openness to being screened for IPV: age (as a continuous variable), sex, income level, education level (up to high school compared with post-secondary), marital status, the length of the relationship, country (a Canadian compared with a Dutch clinical site), and whether or not the patient had children.
Patients who completed part of the questionnaire were included in the analysis; no statistical exploration process was carried out for the missing information. All data analyses were performed with the use of SPSS software (version 20.0; IBM, Armonk, New York).

Recruitment (Figure 1)
Each participating site enrolled 150 eligible, consenting patients, resulting in a total of 750 patients included in the overall study cohort. Three of the five participating sites collected data on recruitment, including the number of patients excluded and reasons for exclusion. A total of 238 patients at these three sites were excluded from the study.

Patient Demographics
The majority of participants in the study were white (81%; 610 of 750), and the mean age (and standard deviation) of the participants was 45 ± 17 years ( Table 1). Just over one-half (56%) of the participants were male. Reasons for reporting to the fracture clinic included treatment of a fracture (40%), a sprain or strain (10%), a soft-tissue problem (8%), and arthritis (7%).

Acceptability of Screening for IPV in the Fracture Clinic
The majority of respondents (74%) indicated that the fracture clinic was a good place for health-care providers to ask patients about their experiences with IPV ( Table 2). Nearly all respondents (94%) either agreed or strongly agreed that orthopaedic surgeons should look for the cause of a patient's injury, and 61% agreed that orthopaedic surgeons could help with IPV. The role of nurses and social workers in supporting victims of IPV within the fracture clinic setting was also endorsed. Nearly 80% of respondents agreed that it would be easier for victims of IPV to get help if health-care providers in the fracture clinic asked about abuse. Participant responses were similar across the five clinical sites.

Patient Opinions on the Logistics of Screening
Nearly one-quarter of all respondents indicated that health-care providers in the fracture clinic should ask all women and all men about IPV (26% and 24%, respectively) ( Table 3). The majority of respondents thought that both women and men with suspicious injuries should be asked about IPV (73% and 68%, respectively). Talking face-to-face was the method of IPV screening endorsed by the vast majority of respondents (90%), followed by written questionnaires (22%). The overwhelming majority of respondents believed that screening should occur in a private location (91%) and that health-care providers should ask questions written by experts (72%).
Respondents also varied greatly in their opinions on what health-care providers in the fracture clinic should do when patients report that an intimate partner has abused them. The most common response involved asking the patients what they wanted to do (68%), followed by referral to hospital resources (60%) and to community resources (58%). Thirty-six percent indicated that the health-care provider should respond to such a situation by notifying the police. Less than 1% of respondents believed that the healthcare provider should do nothing when patients disclosed that they had been abused. Participant responses were similar across the five clinical sites.

Comfort Level with Screeners
Seventy-three percent of the participants indicated that they would be comfortable discussing IPV with a social worker in the fracture clinic; 57%, with an orthopaedic surgeon; and 51%, with an orthopaedic nurse ( Table 2). Few (8%) of the respondents reported they would be uncomfortable discussing IPV with any health-care provider in the fracture clinic. Respondents demonstrated a clear preference for having an older compared with a young health-care provider ask them about IPV (Figure 2). Participant responses were similar across the five clinical sites.

General Opinions on IPV and IPV Screening According to Sex
There were significant differences between male and female participants' opinions on IPV and IPV screening ( Table 4). For example, more female than male participants (74% compared with 55%) perceived IPV committed against men as a serious health issue (p < 0.001). Additionally, 57% of female respondents compared with 45% of male respondents disagreed or strongly disagreed with the statement that it would be embarrassing to be asked about IPV by a health-care provider (p = 0.004). There were no significant differences between male and female participants' opinion that IPV was a serious health in general.

Openness to Screening
The stepwise regression analysis yielded a significant model (F 3595 = 21.950, p < 0.001). The following characteristics were associated with increased openness to screening: being female (p < 0.0001), having higher income (p < 0.0001), and having higher education (p = 0.035). These characteristics predicted 9.5% of the variation (adjusted r² = 0.095) on the twenty questions regarding patient's openness to being screened for IPV. Age, marital status, having children, being at a Canadian compared with a Dutch clinical site, and the length of relationship were not significant predictors in the model. The participants were generally very open to being screened, as the mean score for the twenty questions regarding openness to screening was 75.1 ± 9.4.

Discussion
The findings of this multicentre study suggest strong patient endorsement of IPV screening within the surgical fracture clinic setting. Patients indicated that surgeons, as the primary providers of care, should investigate the root cause of injuries. In response to questions 175 regarding which health-care provider should screen for IPV in the fracture clinic, patients indicated that a variety of health-care providers could have a critical role in detecting and managing cases of IPV. Although the responses suggested that patients felt social workers to be the most helpful in addressing the needs of IPV victims, the majority of participants felt that orthopaedic surgeons and nurses working in the fracture clinic could also help with IPV screening. Our findings imply that a team of multidisciplinary health-care providers in the fracture clinic setting could play a major role in identifying and helping IPV victims.
Approximately three-fourths of respondents questioned the need for universal screening of patients. Universal screening would involve health-care providers asking all patients (male and female) in their practice about their experiences with IPV (e.g., as a standard question in patient visits). In the current study, respondents also questioned the need for screening all female patients who present to the fracture clinic. Several previous studies have supported the implementation of universal screening programs for IPV (in contrast to incident-based or selected screening programs) 6,14,[17][18][19][20] . Rodriguez et al. 17 cited several reasons for universal screening for IPV, including the high prevalence, the array of acute and chronic health problems that are associated with IPV, the low level of suspicion and inquiry on the part of physicians, the general unwillingness of abuse victims to volunteer information, and the high level of patient acceptance of direct physician inquiry. In addition, multiple instances of IPV may be missed in incident-based or selected screening initiatives 21,22 . Furthermore, clinician inquiry has been shown to be the most important determinant of a patient's disclosure of IPV to health-care providers 16,17 . Our study suggests that patients may need additional education on the benefits of, and rationale for, a universal IPV screening program within the fracture clinic setting. It is imperative that health-care providers foster a relationship grounded in trust and openness with their patients in order to make sensitive discussions (such as ones focused on IPV) more comfortable and honest.
We found that patients preferred screening to be performed face-to-face with their healthcare provider and in a private location. Our findings on the preferred mode of screening directly contradict the findings of a 2006 randomized controlled trial that found face-to-face screening to be the least preferred approach among patients, with computerized and written methods being the most favored 23 . Most of the participants in the present study felt that, during screening, healthcare providers should ask questions written by experts and should ask about abuse directly. This indicates that patients would consider a standard screening tool and favor a direct method of screening. The ethnicity of the health-care provider screening for IPV was a less important factor to participants in the present study than the provider's age or sex was. These preferences should be taken into account when developing a successful screening program in a health-care setting, as patients who are comfortable discussing IPV with their health-care provider may be more open to disclosing whether they have been or are being abused.
There were significant differences between the attitudes of men and women regarding general questions about IPV and IPV screening. This was evident in questions regarding whether IPV against men was a serious health issue, whether health-care providers who ask women about IPV should ask men as well, and whether it would be embarrassing to be asked about IPV by a health-care provider. These discrepancies could be explained by the widespread notion in our society that men are perceived as physically superior to women and may be ridiculed if they disclose abuse 2. However, despite the differences in responses between the sexes, the majority of all respondents indicated that they would not be embarrassed or offended if asked about IPV by a health-care provider. This finding serves as direct evidence against the misperception outlined in Bhandari et al. 24 , who reported that 19% of the surgeons surveyed believed that patients would get angry if asked about IPV. Multiple studies investigating patient opinions on IPV screening in other medical specialties have also revealed that most patients did not mind being asked about IPV and would be glad if someone showed interest in helping 16,25,26 . A slightly greater proportion (26.8%) of men in our sample strongly agreed or agreed that being asked about IPV would be embarrassing. This suggests that health-care providers should take a different approach with male patients than with female patients when broaching the topic of IPV.

Strengths and Limitations
Our results were strengthened by the use of trained female research coordinators to maximize enrollment and recruitment of both men and women sixteen years of age or older. The study cohort varied greatly in many characteristics, including age, sex, yearly household income, educational attainment, marital status, the length of the current relationship, and the injury or condition that was being treated at the fracture clinic. This diverse collection of participants is highly representative of individuals who may experience IPV at some point in their lifetime, increasing the external validity of our findings. Our inclusion of male fracture clinic patients offers a unique perspective on the issue of IPV screening. The majority of studies that have examined patients' preferences and opinions regarding IPV have only evaluated female patients' perspectives.
Because participants completed the survey independently (without their accompanying friend, family member, or partner present) and anonymously, we believe that the responses are an accurate depiction of the opinions and values of fracture clinic patients in general.
There are a few limitations to this study. First, only English-speaking patients were included from the Canadian sites and only Dutch and English speakers were included from the site in the Netherlands. This exclusion of patients on the basis of language may have resulted in selection bias. Culture may influence attitudes on IPV and IPV screening, and consideration of the perceptions of non-English-speaking individuals might have added to this study. Selection bias may have also been introduced by the exclusion of critically ill patients. Furthermore, the use of the self-completed questionnaire in a private location resulted in some missing data, which we were unable to retrieve later because the questionnaires were anonymous in design. Additionally, because of resources limitations, two of the participating sites were unable to collect information on patient screening. However, the data regarding screening and patient exclusion at these sites would presumably not be drastically different from the patterns exhibited at the other sites.
Our results demonstrate that fracture clinic patients were amenable to, and supported, active screening approaches for IPV. There is a need for increased awareness in the surgical community regarding IPV, with focused educational initiatives to prepare and engage surgeons to appropriately screen patients.
performed cross-tabulations for all frequency data. No extrapolation processes were carried out for missing data. We included partially completed questionnaires in the study. The WAST has a positive cut-off score of 13, and the PVS has a positive cut-off score of one. Once the cut-off score was reached, we identified the patient as a victim of IPV regardless of whether she completed the remainder of the questions. In addition, we carried out analyses using only fully completed questionnaires for comparison. We analyzed all data using PASW Statistics version 18.0 (SPSS Inc. Chicago, IL).

Results
Out of 395 eligible patients, 282 gave informed consent and completed the questionnaire (response rate 71.4%). The majority of women in this study were white (85.1%), over forty years old (62.8%), and had an annual income of less than $40,000 (59.6%). Most participants were reporting for the treatment of a fracture (73%). Additional participant details are reported in Bhandari et al, 2011.

Prevalence of IPV
Ninety-four of 282 participants (33.3%) screened positive for IPV on our composite of three screening methods (direct questioning, WAST, and PVS). Eighty-six women (30.5%) screened positive using the direct questioning approach, 35 (12.4%) using the WAST, and 26 (9.2%) using the PVS. The direct questioning approach identified 85 victims of emotional abuse (85/282, 30.1%), 24 victims of physical abuse (24/282, 8.5%) and 9 victims of sexual abuse (9/282, 3.2%). Of the 94 women identified as victims of IPV, half screened positive only using direct questioning and not on either the WAST or PVS (47/94, 50.0%). Refer to Tables 2, 3, and 4 for comparisons of those who scored positive on the WAST vs. direct questioning, the PVS vs. direct questioning, and the WAT vs. the PVS. Please see Table 5 for characteristics of participants who screened positive on the PVS versus the WAST versus the direct questioning method.
When we excluded participants with missing responses, 74 women out of 208 (35.6%) screened positive for IPV. Seventy women out of 208 (33.7%) screened positive using the direct questioning approach, 27 (13.0%) using the WAST, and 20 (9.6%) using the PVS. Of the women who screened positive for IPV, 38 of 74 (51.4%) screened positive using only direct questioning and not on either the WAST or PVS.

Validation Measures
Compared to the direct questioning method, the PVS has a sensitivity of 25.3% and a specificity of 97.9%. We could not calculate sensitivity and specificity of the WAST compared to the direct questioning approach because the direct questions appear on the WAST, which could lead to misleading sensitivity and specificity data. When incomplete questionnaires were excluded, the PVS has a sensitivity of 24.6% and a specificity of 97.8% compared to the direct questioning method.

Discussion
Our study found surprising differences in the prevalence rates of IPV in orthopaedic injury patients across three different screening methods. Specifically, we found that the prevalence of IPV was 30.5% when a direct questioning approach was utilized, 12.4% using the WAST, and 9.2% using the PVS. There are a number of plausible explanations for the difference in prevalence rates identified by these three methods. 187 The first explanation for the differences in prevalence rates is that different screening tools define IPV in a different manner, and they are therefore not identifying the same women as victims of IPV. This is a problem because the WAST and the PVS were both developed for the same purpose, and they are often used for the same purpose in research and clinical practice. Our direct approach included three questions that asked about physical, sexual, and emotional abuse, incorporating three types of abuse into our definition of IPV. The WAST also includes these three questions, however, answering positively to any of these three questions does not necessarily result in a positive score. For example, a woman may report no tension and no difficulty in her relationship, no emotional or sexual abuse, but she may have been punched by her partner on a few occasions after he came home intoxicated. Her WAST score would likely be negative, but we argue that she is still a victim of abuse. The direct questioning method would likely detect this while the WAST would not. The PVS asks questions relating to physical violence and threats of violence. A woman who is sexually abused but not physically abused by her partner may screen negative on the PVS if she reports feeling safe at home. Two-thirds of women in this study who said they had been sexually abused reported feeling safe at home. Many women experience more than one form of violence; however, women experiencing emotional abuse only may not be classified as IPV victims on the PVS, and those who experience emotional abuse only sometimes may not screen positive on the WAST. The WAST also includes questions about threats of violence. This is an advantage the WAST has over the direct questioning approach that we used.
Another plausible explanation for the differences that we identified in IPV prevalence is that patients are misinterpreting the questions or are not comfortable answering the questions. For example, women who suffered emotional abuse comprised a large percentage of those who screened positive for IPV using all methods. Since emotional abuse is a subjective term, each woman may have her own definition and threshold of what is considered emotional abuse. These differences in definitions may lead to a wide variation in who reports that they have been emotionally abused and who does not. Since the WAST asks the direct question "Has your partner ever abused you emotionally?" and the PVS does not ask a similar question, a patient's interpretation of what is and is not emotional abuse may skew the results. Women who have been abused may not admit that they have been abused if they feel uncomfortable discussing their abuse.  found that more abused women were uncomfortable with the physical and sexual abuse questions than non-abused women. We chose to include emotional abuse in our definition of IPV because emotional abuse has a tendency to escalate to physical abuse, and emotional abuse often co-occurs with physical and/or sexual abuse (Ahmad, 2007).
Another possible explanation is that one or more of the measures are incorrectly classifying women who have not experienced IPV as victims of IPV. While it is true that falsely identifying women as victims of IPV could be embarrassing and uncomfortable, we believe that failing to identify victims can be more harmful.  recommend that IPV screening tools have maximal sensitivity to avoid missing victims. The direct approach that we used is an attempt to err on the side of inclusion when screening for IPV. After a woman screens positive for IPV using a direct approach such as this, the health care professional can initiate a more in-depth discussion of the patient's situation to determine which actions, if any, are necessary for her specific situation. Even though the direct questioning approach is likely to increase the number of false positives, initiating an in-depth conversation with patients who have screened positive can help to reduce the number of false positives.
We chose to use direct questioning about IPV because the questions are simple, easy to understand, and intuitive. The three-question approach is also quick to administer and easy to score, which is an important issue in a busy health care environment. Medical organizations such as the American Medical Association (AMA, 1992) and the Canadian Orthopaedic Association recommend using direct questions because "it tends to elicit direct answers" (COA, 2009).

Previous Literature
In this study we asked direct questions in the form of "Have you been abused [physically/emotionally/sexually]?" Several other studies have also used a direct approach. Ruiz-Perez et al (2006) asked "Have you ever been abused by your current partner physically (hit, slapped, kicked, pushed)?" They estimated the overall prevalence of any type of IPV as 22.8%. McCord-Duncan et al (2005) used the direct question "Have you ever suffered domestic violence?" and estimated the overall rate of IPV to be 39%. Using the direct question "Have you ever been in a relationship and intentionally injured by an intimate male partner?" to determine the prevalence of IPV,  found that 37% of women were positive for IPV. Our study showed a similarly high prevalence using the direct questioning approach; however, these studies examined the lifetime prevalence of IPV while our study examined one year prevalence. Although patients in our study were explicitly asked to disclose 'IPV within the past year' at the beginning of the questionnaire, the language a few of the questions (use of the word "ever") could have caused some confusion and they may have disclosed IPV in their lifetime, leading to an overestimation of prevalence. Whether our findings do represent an inflated estimate remains debatable. Our findings are not that dissimilar from another Canadian study which found that the overall prevalence of IPV in the past year was 25% .
Few studies have compared the WAST to the PVS for the purposes of screening for IPV in a health care setting. Vivilaki et al (2010) compared a Greek translation of the WAST to the PVS using a sample of women at a perinatal care clinic in Greece. Using the PVS as a criterion standard, the sensitivity of the WAST was 99.7% and the specificity was 64.4% (Vivilaki et al, 2010). MacMillan et al (2009) found that, in English-speaking emergency, family medicine, and women's health clinic patients, the PVS and WAST had similar sensitivities and specificities when compared to the CAS. This high degree of correlation between the two screening tools is not seen in our study. Orthopaedic injury patients are not a well-studied population; therefore we cannot assess the impact of population on the results of each screening tool.  determined, using a known sample of abused versus non-abused women, that each question on the WAST (including the direct questions) correlated significantly with the Abuse Risk Inventory (ARI), and that there are significant differences between the answers of abused and non-abused women on the direct physical and emotional abuse questions. The direct sexual abuse question was not studied. To our knowledge, no study has compared prevalence of IPV using direct questions alone to prevalence of IPV using the PVS.

Limitations
A limitation of the current study is a small sample size of women who screened positive for IPV using our composite of the three measures. The questionnaire was self-reported, and did not require verification of the answers. Since the research coordinator who conducted the interviews did not look at the completed questionnaires before the patients left (for privacy reasons), information that was missing or incorrect could not be corrected. The questionnaire 189 was anonymous and the results were not shared with the treating surgeon or any other health care provider, so answers could not be verified verbally with the patient or by consulting medical records. Some women may not feel comfortable with direct questions if they do not trust the medical system. They may fear being judged or not believed. Another limitation is that the study excludes non-English speaking women and previous studies have shown that people of many races and ethnicities are affected by IPV (McCloskey et al, 2005, Azziz-Baumgartner et al, 2010, El-Bassel et al, 2007). Each participant was asked the PVS questions, followed by the WAST questions, including the direct questions last. This may mean that the participants were primed to see abuse after answering questions about various abusive behaviours. This may have lead to an overestimation of prevalence of IPV on the direct questions. Another potential limitation is that participants may have misread the instruction to disclose IPV in the past year only, which could have inflated the one year prevalence rate.

Future Considerations
It is important to consider the definition of IPV that one wishes to use before screening for IPV in a clinic setting or in IPV research initiatives. Since the WAST, PVS and many other tools screen for different types of violence, selecting the most appropriate screening tool for the task is essential. Using a combination of screening tools may broaden the types of IPV victims identified (i.e. physical, emotional, sexual abuse victims), and a direct questioning approach can broaden the definition even further. Researchers and clinicians could consider asking directly about physical, emotional, and sexual abuse when screening for IPV as a means of broadening the definition of IPV. However, validation studies are required before this approach should be widely implemented. Future research should focus on using one reference standard to evaluate existing screening tools. The direct approach could be used as the reference standard for other screening tools because of its broad definition, ease of administration, and simple to understand questions.

Conclusions
Despite being tested for validity and being developed and used for the same purpose, the WAST and the PVS largely identified different women as victims of IPV, as there is very little overlap in the women who screened positive for IPV on the WAST and the PVS. This difference in identification is concerning for investigators and clinicians who use only one of the WAST or PVS to detect IPV in their study or practice. Some women who are victims of IPV may remain undetected when using only one screening instrument. In the current study, the direct approach identified more women as victims of IPV than the WAST and PVS, and missed fewer victims compared to the WAST and PVS. We suggest that the direct questioning approach should be used as a preliminary assessment for IPV, and then the clinician can initiate a more in-depth discussion of the patient's situation and determine which actions are appropriate for her situation. By ssing this approach clinicians can err on the side of detecting more victims. All three measures are quick to administer in a clinical setting.

Background
Intimate partner violence (IPV) is a serious global health issue and is a large source of preventable morbidity and mortality among women 1 . It is characterized by a pattern of assaultive and coercive behaviours, including physical, sexual, and psychological attacks as well as economic coercion committed by both men and women against their partners 2 . Consequences of IPV include a host of physical and mental health problems documented in numerous studies 3 . The economic costs of IPV against women are substantial. Max and colleagues reported that IPV against women costs $5.8 billion, considering expenditures for increased medical care, mental health services, and lost productivity stemming from injury and premature death 4 . These large health consequences and costs to the health care system have led to widespread research efforts in the area of IPV over the past several decades, primarily focusing on obstetrics, emergency medicine, and primary care settings 5 .
Until recently, there has been a paucity of research focusing on IPV in the realm of orthopaedic surgery. Renewed interest in the underlying risk of IPV among women with musculoskeletal injuries has fueled several important studies 6,7,8,9,10,11,12,13 to determine the nature and scope of this issue in orthopaedic surgery. Because of the high prevalence of IPV in orthopaedics 11 , the evidence that patients rarely have their IPV documented by emergency department staff 14 , and because orthopaedic surgeons see patients several times for follow up appointments, often when they are not in acute pain (as opposed to the emergency department), orthopaedic surgeons are in an ideal position to help victims of IPV 15 . This information has led several organizations to recommend that orthopaedic surgeons take a lead role in IPV identification and offering assistance to victims of IPV including the Canadian Orthopaedic Association and the American Academy of Orthopaedic Surgeons 15,16,17 .
We present an overview of the current knowledge in the field of IPV and orthopaedics as well as where the knowledge gaps lie and we propose a method of developing a support program for IPV victims in orthopaedic fracture clinics based on the best available evidence.

What do we know about intimate partner violence and musculoskeletal injury?
The field of orthopaedics has historically been closely involved in the identification and management of cases of child abuse involving physical injuries. Orthopaedic research in the area of child abuse and injuries began as early as 1974 with a description of battered child syndrome 18 . However, descriptions of IPV and musculoskeletal injuries did not appear in the orthopaedic literature until 1993 19 . Varvaro and Lasko described the most common physical injuries from IPV as contusions, abrasions/lacerations, and fractures/strains/sprains with most injuries occurring on the face, neck, head, extremities, or in multiple locations 19 . Bhandari and colleagues evaluated 263 women referred to a domestic violence therapy and advocacy center and found that the most prevalent forms of abuse were emotional (84%), psychological (68%), physical (43%), sexual (41%), and financial (38%) 6 . Among those women who reported physical abuse, 46% sought medical attention. The authors identified 144 injuries in the 218 women who experienced physical abuse. Head and neck injuries were the most common, followed by musculoskeletal injuries, which included sprains, fractures and dislocations, and foot injuries 6 . The authors concluded that recognizing musculoskeletal injuries in women as a potential result of IPV is warranted 6 . 195 A recent systematic review and meta-analysis examined the pattern of physical injury associated with IPV in women presenting to emergency rooms 7 . The association between head, neck, and facial injuries and IPV was higher among studies that excluded women with verifiable injuries such as witnessed falls or motor vehicle accidents. Thoracic, abdominal, or pelvic injuries were generally non-specific for IPV whereas upper extremities were suggestive of non-IPV. The authors cautioned that the quality of evidence pertaining to thoracic, abdominal, and extremity injuries may be limited by methodological issues and lack of data to a greater extent than findings that pertained to head, neck, and facial injuries 7 .
A recent meta-analysis examined the prevalence of IPV across different medical specialities 20 . The authors included 37 articles in their study and found that most studies took place in family medicine clinics (15/37, 40.5%) and emergency departments (12/37, 32.4%). Pooled prevalence was reported for emergency (lifetime: 38%; one year: 19.9%) and family medicine (lifetime: 40%, one year: 19.5%). Data from other specialty areas suggested that from 43% (obstetrics and gynecology) to 73% (addiction recovery) of women experienced IPV during their lifetime of which 4% (pediatric emergency) to 21% (obstetrics and gynecology) of women experience physical abuse. This review did not find any data examining the prevalence of IPV in orthopaedic fracture clinics and identified this as a gap in the literature.
A cross-sectional study of women reporting to two trauma centres in Ontario found that one third of the respondents had experienced IPV in the last twelve months 11 . Emotional abuse was the most prevalent form of abuse (30.5%), followed by physical abuse (8.5%), and sexual abuse (3.3%) 11 . Seven women (2.5%) presented to the orthopaedic fracture clinic as a direct result of their abuse 11 . This study also reported that none of the women included in this study were asked about IPV by their attending orthopaedic surgeon. This is the first study that attempted to ascertain the number of women reporting to orthopaedic fracture clinics who were victims of IPV. The estimate is similar to previously reported one-year rates in other medical specialties, including internal medicine, pediatrics, and obstetrics and gynecology 20 .

Knowledge Gaps Among Orthopaedic Surgeons
Knowledge and comfort with IPV identification has been assessed in large scale surveys and smaller qualitative studies. Of 186 surgeon members of the Canadian Orthopaedic Association surveyed, 148 (80%) believed that IPV was exceedingly rare among the women they treated, affecting less than one percent of their patients 8 . Furthermore, one in two surgeons expressed that they lacked knowledge of the appropriate resources available to IPV victims (53%). Many surgeons held a number of misperceptions about IPV including 1) victims must be getting something out of the abusive relationships (14%); 2) some women have personalities that cause the abuse (20%); and 3) the battering would stop if the batterer quit abusing alcohol (43%) 8 . These findings were supported by a similar evaluation of US surgeon members of the Orthopaedic Trauma Association which found that among 153 surgeon members of the Orthopaedic Trauma Association, several misconceptions were evident: 1) victims must be getting something out of the abusive relationships (16%); 2) some women have personalities that cause the abuse (20%); and 3) the battering would stop if the batterer quit abusing alcohol (40%). In the past year, only 4% of respondents currently screened for IPV among female patients with injuries 9 .
Whether the knowledge gaps resulted from a lack of reinforcing early education during medical school and residency training into later practice or instead, a systemic lack of education in the system is unclear. However, a recent online survey demonstrated knowledge gaps and 196 discomfort in both surgical residents and medical students 10 . Respondents reported feelings that physicians should not interfere with a couple's conflicts (21%), that patient's personalities caused them to be abused (41%), and the majority (84% of medical students and 60% of surgical residents) felt that their training on IPV was inadequate [10]. Over 90% of both residents and medical students estimated the prevalence of IPV in their intended practice is less than 10% 10 .

How Can We Help IPV Victims in the Orthopaedic Fracture Clinic Setting
The varied research exploring the relationship between IPV and orthopaedics allows for greater understanding of the issue at hand and helps to generate possible responses to this significant public health issue. The demonstrated high prevalence of IPV among patients presenting to orthopaedic fracture clinics 11 , suggests that fracture clinics are an opportune setting for offering support to IPV victims. There is reason to believe that patients in orthopaedic clinics may experience more severe IPV than other types of patients because their fractures and other injuries warrant an orthopaedic surgeon's care. Furthermore, IPV victims may not be effectively assisted by other health care professionals (HCPs), so asking about IPV at the level of the fracture clinic may be a window of opportunity. For example, a study of family medicine records showed that fewer than 15% of abused women had their abuse documented in a medical chart 21 .
There are multiple initiatives that could be implemented within the orthopaedic fracture clinic setting to help IPV victims and provide them with the appropriate social support. Prior to the widespread implementation of any new initiative or program, it should be appropriately evaluated, following an evidence-based approach. Several of these programs are currently underway and are described in the next section. There are also simple steps that orthopaedic surgeons can take immediately and they are outlined in Table 1.

Practical Approaches to Identification and Support of IPV Victims in a Clinic Setting
Multiple system-level factors must be considered to enable orthopaedic surgeons to effectively identify and help patients who have experienced IPV, including case-finding protocols and support plans. Multiple items need to be carefully considered when developing an IPV case-finding protocol in the fracture clinic setting, which are described below and summarized in Figure 1. An evidence-based approach should be followed when possible. Abuse Screening tool (WAST) 41 [41], Partner Violence Screen (PVS) 42 , and Abuse Assessment Screen (AAS) 43 are among the most widely used and validated short screening tools 38 . In orthopaedic trauma populations, previous research has found that the WAST and PVS have very good specificity, but the sensitivity is relatively low 12 . It is important that we maximize sensitivity with IPV screening tools to avoid missing the opportunity to assist victims. Although good specificity is fairly important, having a good set of questions to begin with is essential from which HCPs can initiate a more in-depth conversation with the patient 12 . Although these tools are frequently used in other settings, there is a subset of the WAST questions that have increased sensitivity in the orthopaedic trauma population 12 (Figure 2). A simple set of three questions is easy to remember, does not take a lot of time, and the results can be determined easily, not with a complicated scoring procedure. In addition, the American Medical Association and the Canadian Orthopaedic Association recommend using direct questions because they are easy to understand and tend to elicit direct responses 15,44 .
Another simple measure that surgeons and other HCPs should consider is the mobile phone app called "R3" made by Harbour House shelter in Florida that guides HCPs step-by-step through the process of asking about IPV in a medical setting using the well-studied HITS screening tool 45 . The app also provides some local, state, and national (USA) referral resources for victims of IPV and it also provides some specific instructions such as ensuring the victim is connected with an advocate before leaving the appointment (if available), using the patient's own words to document abuse, and tips on how to assess safety and follow up appropriately. Future endeavours could include expanding upon available apps to include resources for other countries or developing new apps for specific purposes such as screening in orthopaedics specifically.

Item 4: Ensure Confidentiality and Patient Safety in the Clinic Setting
When discussing IPV, it is critical that HCPs minimize the victims' risk of harm and maximize potential benefits 17 . Discussing IPV could have negative consequences for IPV victims including anxiety, shame, fear, and physical harm 46 . These risks can be minimized through careful consideration of the type of environment in which screening takes place. Maintaining the confidentiality of IPV victims is of paramount importance because of the sensitive nature of the topic 47 . The subject should not be broached with the patient's partner, friends, or family in the area, to respect the patient's safety and confidentiality 17 . The vast majority of patients agree that it is important to ask about IPV in a private location 30 . Creating an environment conducive to maintaining confidentiality may be challenging; one of the most commonly reported barriers to screening for IPV is that clinics are not private enough 31 . Many fracture clinics (especially in older hospitals) are not constructed with having a personal and delicate conversation in mind. However, in even the most open clinics, there is often a room or small space with a bit more privacy which an IPV coordinator could use to speak to patients about IPV, thereby reducing the likelihood of harm. Victims should be treated in a manner that will minimize their anxiety, shame, and fear, assuring them that their abuse is not their fault.

Item 5: Develop Social Support Plans
When assisting with a complex issue such as IPV, the victim's choices must be respected 17 . If a patient does not wish to disclose about an IPV experience, the HCP should understand that it is the patient's choice and not to force the conversation. Similarly, if the patient chooses to disclose but not take action (leave the relationship, accept referrals, seek counselling etc.), then the patient's wishes should be respected. IPV is a very complex issue that involves social, financial, physical, and psychological considerations and consequences. Following disclosure, it may be difficult for an IPV victim to accept help for fear of losing their children, losing their partner, being financially worse off, or experiencing retaliation from their partner 48 . IPV management must be individualized, effectively addressing each victim's particular needs 17 . The IPV Coordinator must be sufficiently experienced with assisting victims of IPV and must be aware of the complex social, psychological, financial, and familial issues that surround IPV to help to provide IPV victims with the appropriate social support.Health care guidelines for implementing IPV identification and support programs include the following components to assist women who have experienced IPV: 1) Listen to her concerns in a nonjudgmental way; use phrases like "I am concerned for your safety" or "The abuse is not your fault", 2) Provide information that helps to reduce misconceptions and alleviate fear and anxiety, 3) Create an immediate safety plan with the patient; make sure that she is safe to go home and offer immediate assistance such as a social worker or legal/police assistance if required and if she agrees, 4) Offer longer-term resources such as health-care or community support system access, women's shelters, other specialized local services, etc 17 .

Item 6: Evaluate the Program
The next step in advancing IPV advocacy in orthopaedic fracture clinics is to pilot and evaluate a universal screening program within the orthopaedic fracture clinic setting. This program would include having a trained IPV Coordinator to ask three screening questions (Figure 2) to all patients who present to the orthopaedic fracture clinic. With patients who screen positive, the IPV coordinator would assist the patient with providing the appropriate social support as described above. This is a challenging and costly initiative to both implement and 203 evaluate. The cluster randomized controlled trial study design may be the most appropriate because the intervention would be meant to be applied at the clinic level as opposed to the individual level. Cluster randomized designs are best for when the unit of randomization is a clinic, school, family, or other easily defined group 49 .

Future Directions -Bridging the Gaps in Knowledge and Research
The current literature has identified numerous gaps in both IPV knowledge and research in the field of orthopaedic surgery. Multiple research initiatives are currently underway and are being developed to further advance this important field which include assessing the prevalence of IPV, evaluating an IPV fracture clinic tool kit, and developing and evaluating an IPV casefinding program in the orthopaedic setting. Each of these is described in detail below.

Assessing the Prevalence of IPV
A larger multi-centre IPV prevalence study is currently being conducted at eleven sites in North America, Europe, and Asia 13 . Approximately 3000 women will be included in this initiative. This study will provide a more accurate estimate of the prevalence of IPV in the orthopaedic fracture clinic setting and will demonstrate the differences in prevalence rates between nations 13 . The results of this study will further inform the need for IPV screening and offering victim's support programs within the orthopaedic fracture clinic setting.

Providing an IPV Fracture Clinic Tool Kit
Another way that fracture clinic personnel may be able to help IPV victims is through an IPV "toolkit" that includes IPV awareness posters, buttons, and pamphlets to post throughout the fracture clinic setting. The Family Violence Prevention Project found that the implementation of pamphlets, resource cards and examination room posters increased the number of clinician referrals and patient self-referrals to an on-site domestic violence evaluator more than twofold 50 , indicating that such an intervention may have a significant impact on a patient's willingness to discuss IPV in the clinic. HCPs can also be provided with a simple set of steps to follow in case of disclosure. Previous research has shown that orthopaedic surgeons, medical students, and surgical residents are largely unsure of what to do if a patient discloses 8,9,10 . A pre-and postinterventional study is currently under way that aims to evaluate whether the presence of posters, buttons, and pamphlets on IPV changes patients perceptions about IPV and their comfort level with discussing IPV within theorthopedic setting.

Developing and Evaluating an IPV Identification Program -The Initial Steps
Identifying victims of IPV within the fracture clinic setting through a case-finding program may be another method of assisting individuals who are being abused. Multiple researchers have suggested that screening for IPV in a medical setting is not supported by evidence 51,52 and "universal screening" for IPV within medical settings remains highly controversial, as is exemplified in the debate between Wathen and MacMillan, and Taket 22 . MacMillan and Wathen hold the opinion that routinely screening all women for IPV in any setting (universal screening) is not appropriate and has potential harms 22 . They support targeted "case-finding" for women who present with certain signs and symptoms of IPV 22 . Conversely, Taket holds the opinion that universal screening should be applied because it "contributes to changing social attitudes to domestic abuse" among other benefits such as decreasing 204 stigmatization, possible increased safety compared to selective screening, and avoiding incorrect stereotypes of IPV among HCPs 22 .
A recent randomized controlled trial on identification of IPV in primary care clinics concluded that screening all women for IPV did not increase quality of life 53 . This trial had three intervention groups: a group that was screened using a computer and received a list of local resources, a group that received a list of local resources only, and a group that received no intervention. This study, along with MacMillan and colleagues' randomized trial shows that passive interventions are not effective 51 . This evidence suggests there is a need for a high quality trial evaluating "active" screening or case finding programs that include both an identification component and a component where women who disclose can get the support they require.

Conclusions
Based on the available evidence in surgery and other medical fields, surgeons should recognize that IPV is a serious public health issue that affects a large proportion of orthopaedic patients. Surgeons and other HCPs should be aware of the various issues and complexities surrounding the problem of IPV. We propose a stepwise, structured approach to developing a support program that includes the following: Item 1) Decide who to include in IPV identification programs; Item 2) Determine who should ask about IPV; Item 3) Determine the method of identification; Item 4) Ensure confidentiality and patient safety in the clinic setting; Item 5) Develop social support programs; Item 6) Evaluate the program. Additional research is currently underway to inform the development of an IPV screening program within the orthopaedic fracture clinic setting.

Background
Intimate partner violence (IPV) or domestic violence is a common and serious public health problem around the globe. Victims of IPV frequently present to health care practitioners including orthopaedic surgeons. Substantial research has been conducted on IPV over the past few decades, but very little research has focused on IPV in the field of orthopaedic surgery. Orthopaedic surgeons may be well positioned to help women who are experiencing IPV and position statements from both the American Academy of Orthopaedic Surgeons and the Canadian Orthopaedic Association exist and provide guidance on the topic (1,2). This thesis originated from the lack of understanding of IPV in orthopaedic patients as well as the desire to develop a program for orthopaedic surgeons to assist IPV victims presenting to orthopaedic fracture clinics.

Aims of this Thesis
The overarching purpose of this thesis was to conduct research to understand the opportunities and challenges facing orthopaedic surgeons in assisting IPV victims in their orthopaedic fracture clinics. The specific aims of this thesis were: 1) to investigate orthopaedic surgeons', surgical trainees', and medical students' perceptions about IPV, 2) to determine the prevalence of IPV in orthopaedic fracture clinic patients, 3) to assess the barriers to and facilitators for screening for IPV in orthopaedic settings, and 4) to discuss the development of a screening program for IPV in orthopaedic fracture clinics.

Section I: Orthopaedic Surgeons and Surgical Trainees Perceptions about Intimate Partner Violence
To inform the current situation, we assessed the perceptions, attitudes, and knowledge of orthopaedic surgeons, surgical trainees, and medical students on the topic of IPV. This section reported on the findings of two surveys. We began by surveying the Canadian Orthopaedic Association membership, which is comprised of over 700 orthopaedic surgeons practicing in Canada (Chapter 2). A qualitative survey was used to collect demographic data about the surgeons and their practices, as well as to measure their attitudes towards IPV by identifying their beliefs regarding victims, batterers, and their own responsibilities as healthcare practitioners. One-hundred-and-eighty-six orthopaedic surgeon members of the Canadian Orthopaedic Association completed the survey, which equates to a response rate of 51%. The majority of the respondents (95 percent) estimated that less than 10 percent of their patients were victims of IPV and most respondents believed that IPV was rare (a prevalence of less than one percent). Multiple misperceptions were identified including: 1) the belief that asking about IPV is an invasion of the victims' privacy, 2) investigating IPV is not the surgeon's duty, 3) IPV victims choose to be victims, and 4) IPV victims play a role in causing their abuse. In summary, this survey found that discomfort about IPV and a lack of education on IPV have led to many misconceptions among Canadian orthopaedic surgeons.
Misperceptions about IPV were also evident in a survey of Ontario medical students and surgical trainees (Chapter 3). Misperceptions about IPV among respondents included the following: 1) victims must get something from the abusive relationships, 2) physicians should not interfere with a couple's conflicts, 3) asking about IPV risks offending patients, 4) victims choose to be victims, 5) it usually takes 'two to tango', and 6) some patients' personalities cause them to 211 be abused. The majority of medical students and surgical residents estimated the IPV prevalence in their intended practice to be 10 percent or less. However, the majority of respondents believed identifying IPV was very relevant to clinical practice. Most of the medical students and surgical residents felt that their level of training on IPV was inadequate and over three quarters of respondents expressed a desire to receive additional education and training on IPV.
The surveys of the Canadian Orthopaedic Association (Chapter 2) and of Ontario medical students and surgical trainees (Chapter 3) are strengthened by the use of a survey that has previously been well described. In our survey of the Canadian Orthopaedic Association we obtained a response rate of 51 percent. Our survey was limited to Canadian orthopaedic surgeons who are practicing members of the Canadian Orthopaedic Association, which represents approximately 80 percent of orthopaedic surgeons in Canada. This may limit the generalizability of our study results.
Although our survey of medical students and surgical trainees had an adequate sample size, our response rate was 29 percent (Chapter 3), which may be a potential source of bias. In addition, this survey was restricted to medical students and surgical trainees at McMaster University who could be contacted via email. It is unclear as to whether our findings are generalizable to other academic institutions and jurisdictions.
In summary, this section found that orthopaedic surgeons, surgical trainees, and medical students have multiple misperceptions about IPV. Our surveys also found that respondents believed that the prevalence of IPV was very low in their patient population. These surveys also identified the need for further education of orthopaedic surgeons, surgical trainees, and medical students on the topic of IPV. Through education and awareness, the misperceptions must be corrected before we can develop IPV screening programs and IPV support programs within orthopaedic fracture clinics.

Section II: Prevalence of Intimate Partner Violence in Orthopaedic Fracture Clinic Patients
The aim of this section was to determine the prevalence of IPV in patients presenting to orthopaedic fracture clinics. We began by conducting a systematic review and meta-analysis of the published literature to examine the best estimates of IPV prevalence across different medical subspecialties (Chapter 4). This meta-analysis found that the best estimates of lifetime prevalence of any type of IPV were 38 percent in family medicine and 40 percent in emergency medicine. We did not identify any studies assessing the prevalence of IPV in orthopaedic patients, revealing a gap in the present published literature. This review and meta-analysis (Chapter 4) is strengthened by our comprehensive search of the literature. We screened articles for inclusion in duplicate and two independent reviewers abstracted data from each included study, ensuring accuracy. Despite these strengths, this study is limited by a large degree of heterogeneity across the included studies, which limited our ability to directly compare studies. Another limitation is that only articles published in English were included. We also did not include grey literature in our review.
To measure the prevalence of IPV in orthopaedic fracture clinic patients, we developed the protocol for the PRevlance of Abuse and Intimate Partner Surgical Evaluation (P.R.A.I.S.E.) study (Chapter 5). This is a multi-centre cross-sectional study in which female patients presenting to fracture clinics completed two validated self-reported questionnaires (Woman Abuse Screening Tool (WAST) and the Partner Violence Screen (PVS)), as well as a direct screening approach, to determine the prevalence of IPV in the past 12 months and in their lifetime. Prior to completing the large cross-sectional study, we completed a pilot P.R.A.I.S.E. study at two level I trauma centres in Ontario (Chapter 6). In our pilot study, we found that the 212 overall prevalence of IPV (defined as emotional, physical, and/or sexual abuse) within the last 12 months was 32 percent. This study suggested a high prevalence of IPV among female patients with injuries who presented to fracture clinics within trauma centres in Ontario. The pilot study helped to inform the protocol for the definitive P.R.A.I.S.E. study and the high prevalence that we found in the pilot study provides a strong rationale to complete the larger definitive study.
The definitive P.R.A.I.S.E. study included 2,945 women from 12 clinical sites in Canada, the United States, the Netherlands, Demark, and India. This study found that the 12-month prevalence of IPV is 16.0 percent and the lifetime prevalence is 34.6 percent (Chapter 7). This study also found that the rates of IPV differed across the different jurisdictions.
Our Prevalence of Abuse and Intimate Partner Surgical Evaluation (P.R.A.I.S.E.) study (Chapters 5, 6, and 7), was strengthened by the use of previously developed screening questionnaires and direct questions, broad eligibility criteria, the use of female study coordinators in order to maximize enrolment, the completion of self-administered questionnaires in a private location, and the assurance of anonymity of the respondents. Using multiple screening tools may help to identify patients who have experienced forms of abuse that may not be detected with only utilizing one tool. However, there is a risk that broadening the definition of IPV created a high rate of false-positives. The P.R.A.I.S.E. pilot study (Chapter 6) is limited by 113 women declining to participate in the study. It is plausible that the non-participants differed from the participants in terms of prevalence of IPV. In addition, the generalizability of the findings may be limited to fracture clinics with similar referral and triage practices as the two included in this pilot study.
Our cross-sectional multi-centre study assessing the prevalence of IPV in the fracture clinic setting was strengthened by including multiple international centres (Chapter 7). This study was limited by the use of a self-completed questionnaire, which resulted in some missing data. This study was also limited by including only patients who present to the clinic for treatment or follow-up of an injury, which decreases the generalizability of the findings. In addition, the exclusion of non-English speaking patients further decreases the generalizability of the findings. Finally, this study only included female patients and it consequently does not provide an estimate of the prevalence of IPV in male orthopaedic fracture clinic patients.
This research from this section provides suggests that the prevalence of IPV is as high as other medical subspecialties (3 -24) and provides a strong rationale to consider implementing a screening program to assist IPV victims in orthopaedic fracture clinics.

Section III: Barriers to and Facilitators for Screening for Intimate Partner Violence in Orthopaedic Fracture Clinic Patients
The results of the first section of this thesis demonstrated that orthopaedic surgeons and surgical trainees had multiple misperceptions and a lack of education about IPV, which is a significant barrier to implementing a screening program within a fracture clinic. This section further investigated some of the barriers to screening for IPV in an orthopaedic fracture clinic setting and also suggests recommendations on how to overcome some of these barriers.
We conducted a systematic review which examined health care providers' perceived barriers to screening for IPV (Chapter 8) and identified and explored five categories of IPV screening barriers: personal barriers, resource barriers, perceptions and attitudes, fears, and patient-related barriers. The most frequently reported barriers included personal discomfort with IPV, lack of knowledge about IPV, and time constraints within a clinical setting. Health care provider barriers were reported more frequently than patient-related barriers. This systematic review was strengthened by a thorough and systematic search of the literature. We reviewed all 213 potentially eligible articles and independently abstracted all data in duplicate. In addition, this systematic review is limited by the small sample sizes and the low response rates in some of the included studies. In addition, there was heterogeneity across the included studies as a result of variations in study design, methodology, provider characteristics, and medical specialties. Consequently, this heterogeneity limited the direct comparisons of the included studies. Another limitation was the high number of low-and moderate-quality studies included in the review. In addition, we included only articles that were published in English. Finally, we did not include grey literature in this review.
To further explore the barriers to and facilitators for IPV screening in fracture clinics, we conducted a series of focus groups with orthopaedic surgeons and orthopaedic surgical trainees (Chapter 9). To provide additional context to the focus group findings, we also interviewed two opinion leaders in the field of orthopaedics. Similar to the previous published literature, the focus groups of orthopaedic surgeons and surgical trainees identified multiple perceived fracture clinic barriers, barriers related to patients, barriers specific to surgical trainees, and barriers specific to surgeons. The participants of the focus groups also discussed and identified facilitators for IPV screening including system-level facilitators, fracture clinic processes, and personnel resources. The interviews with the opinion leaders identified several facilitators for the implementation of policies for IPV screening including the need for champions, the need to increase awareness both locally and nationally, local policy implementation, and the need for research to inform policy. This qualitative study was strengthened by utilizing an experienced facilitator and interviewer who was knowledgeable about IPV. In addition, we worded questions within our focus group and interview guide to ensure that they were neutral and did not lead the participants. Participants in our focus groups were from a single academic institution, limiting the generalizability of the results. However, we did interview two opinion leaders from different academic institutions, which may improve the generalizability of our findings.
We then evaluated whether it was acceptable to fracture clinic patients for health care practitioners to screen for IPV in the orthopaedic fracture clinic setting. Seven-hundred and fifty patients at five different hospitals in Canada and the Netherlands completed a survey that addressed this question (Chapter 10). This study found that the majority of patients agreed that the fracture clinic was an appropriate setting for health care practitioners to ask about IPV. This survey was strengthened by including multiple centres and the use of broad inclusion criteria. This study was limited by only including English (Canada) and Dutch (Netherlands) speaking patients, which may mask the effects of cultural differences. In addition, the use of a selfcompleted questionnaire resulted in some missing data.
In summary, this section identified multiple barriers to screening for IPV, which are similar to those identified in other specialties (25)(26)(27). Fortunately, many of which are possible to overcome through championship, education and additional resources. In addition, orthopaedic surgeons provided multiple facilitators for implementing a screening program within the facture clinic and patients were acceptable to the idea of a universal screening program for IPV being implemented in the fracture clinic setting.

Section IV: Moving Towards Developing a Screening Program for Intimate Partner Violence in the Orthopaedic Fracture Clinic
There are a number of potential options for screening for IPV in health care settings and the optimal method of screening for IPV remains highly controversial (28,29). This section provides a discussion on several different screening methods and the challenges for implementing a screening program within an orthopaedic fracture clinic setting. 214 Multiple screening instruments have been developed and implemented in various health care settings (30)(31)(32)(33)(34)(35)(36)(37). For the P.R.A.I.S.E. study, we utilized two previously developed instruments (the WAST and the PVS), and we also directly asked patients whether they had experienced physical, emotional, and/or sexual abuse. We found that the prevalence rates varied across these instruments (Chapter 11). Specifically, the prevalence was 30.5 percent with the direct questioning approach, 12.4 percent using the WAST, and 9.2 percent using the PVS. This study suggests that previously developed screening tools may not be broad enough and are consequently under-estimating the prevalence of IPV. A more direct approach to screening for IPV may be more effective. This study is limited by the small sample size of women who screened positive using the composite of the three measures. Please refer to the previous discussion about the strengths and limitations for Chapter 6.
Multiple factors need to be considered prior to implementing an IPV program in an orthopaedic fracture clinic setting. This chapter (Chapter 12) provides a summary of the research conducted to date on IPV in the field of orthopaedic surgery This chapter also sets the stage for future research initiatives which include providing orthopaedic fracture clinics with an IPV tool kit and determining its effectiveness. We also discussed several items that need to be considered when developing a program to identify cases of IPV within the orthopaedic fracture clinic setting.
Recommendations for future research are also discussed. This review is strengthened by a thorough review of the literature and using an evidence-based approach in making recommendations for future research.
This section has discussed several items that need to be considered when implementing an initiative to help identify patients who are experiencing IPV within the orthopaedic fracture clinic setting. Screening for IPV remains controversial and recent randomized controlled trials evaluating different screening programs have not shown positive results (38,39). Future research needs to rigorously evaluate active screening programs using the appropriate study designs and outcome measures.

Implications of this Thesis
Unlike child abuse, it is not mandatory to screen for and report IPV in adult patients in most jurisdictions. As a consequence, most hospitals and clinics have procedures and protocols in place for screening for and managing cases of child abuse, but not for IPV. In addition, most physicians receive education on how to identify and manage children who are being abused and very little education on how to manage IPV in a clinical setting. While the mandatory reporting of IPV in the adult population is not appropriate, we can learn from the establishment of detailed protocols and processes and educational initiatives that are in place to guide health care professionals in the screening and management of child abuse.
Traditionally, general practitioners have been well-positioned to screen for IPV and assist IPV victims due to their gatekeeper role within the health care system. Many patients have close relationships with the family physicians which leads one to believe that general practitioners are appropriately helping IPV victims within their practice. Unfortunately, not everyone has a general practitioner and not all general practitioners actively ask patients about IPV. In addition, patients may be hesitant to disclose IPV to someone that they and their families are very close to. Given these limitations, it is vital that other health care practitioners look to identify cases of IPV.
Many of the women who present to emergency departments with injuries are referred to orthopaedic surgeons for treatment of their injuries. Although these patients may be screened for IPV in the emergency room, the emergency department is not the perfect location to manage IPV 215 (2). The orthopaedic fracture clinic may provide an ideal opportunity for a second chance to screen for IPV and subsequently offer appropriate assistance for patients with orthopaedic injuries (2). At the initiation of this thesis, very little research had been conducted on IPV in orthopaedic trauma and this thesis helps us to move towards an evidence-based approach for addressing how orthopaedic surgeons can help IPV victims.
We hypothesized that orthopaedic surgeons are well positioned to help victims of IPV; however, as is evident by the research conducted in this thesis, multiple barriers must be overcome before large-scale IPV programs can be implemented in the orthopedic fracture clinic. One of the key barriers identified in this thesis is the lack of education on IPV for medical students, surgical trainees, and orthopaedic surgeons. Broad education initiatives need to be developed, evaluated, and then implemented to ensure that orthopaedic surgeons are comfortable with assisting IPV victims. Another key barrier identified included was the lack of privacy within the orthopedic fracture clinic setting. As fracture clinics are renovated and redesigned, patient privacy should be taken into consideration. We need to move from a model of openconcept, to private exam rooms. Orthopaedic surgeons and the surgical trainees also identified that the lack of time during the patient's fracture clinic visit was a critical barrier. This suggests that it may be more efficient to have a nurse or social worker screen for IPV and provide support to patients who are victims of IPV.
Our surveys found that orthopaedic surgeons believed that the prevalence of IPV was very low in the patients presenting to their fracture clinics. This is a misperception, as the crosssectional prevalence studies conducted as part of this thesis found a higher than anticipated rate of IPV in patients presenting to fracture clinics with injuries. This finding further supports our belief that screening for IPV in the orthopaedic fracture clinic presents an opportunity to help many women who are victims of IPV.
The last section of this thesis touched on some of the items to consider when developing a program to identify IPV victims within the fracture clinic setting. Our research found that a direct approach when asking about IPV and computer based screening may be more effective than face-to-face screening and self-administered written screening. There are numerous factors to take under consideration when developing, evaluating, and implementing a new IPV initiative and it is vital that a multi-disciplinary approach be utilized. Future research needs to be conducted to evaluate different screening options. In summary, through education, restructuring, forward thinking, and additional research, the barriers to identifying and managing IPV patients in orthopaedic fracture clinics needs to be urgently overcome.

Future Research
Future research needs to begin with a formal evaluation of educational initiatives in identifying and managing IPV for medical students, surgical trainees, and orthopaedic surgeons. Educational opportunities regarding IPV may include an addition to the core curriculum at the medical school level. Individual surgical trainee programs also need to incorporate information about the management of IPV into their programs to ensure that surgical trainees are comfortable with screening and managing IPV victims. Workshops and educational rounds may also provide an excellent opportunity for creating awareness and improving the level of comfort about IPV amongst orthopaedic surgeons and surgical trainees.
This thesis identified several different facilitators to screening for IPV, along with multiple methods for creating awareness and managing IPV in the fracture clinic setting. Despite the initial research on IPV screening in the fracture clinic setting conducted as part of this thesis, 216 numerous questions remain unanswered. Some of these questions include: what is the optimal method of screening for IPV, is passive or active screening more effective, who should conduct the screening, and what are the most effective means of helping fracture clinic patients who are identified as IPV victims. These initiatives need to be formally developed and appropriately evaluated using an evidence-based approach before they are widely implemented to ensure that they are an effective and safe means of helping victims of IPV. In addition, very little research has been conducted on IPV in male patients. Future research should estimate the prevalence of IPV in male patients as well as design, implement, and evaluate screening and management programs aimed at male IPV victims.

Conclusions
We are moving towards an evidence-based approach to assisting IPV victims in orthopaedic fracture clinics. We found that multiple barriers to screening exist, and that orthopaedic surgeons, surgical trainees, and medical students have many misperceptions about IPV. Educational programs are urgently needed to address these misperceptions and to provide factual knowledge about IPV. We also identified a higher than anticipated prevalence of IPV (physical, emotional, and sexual abuse) in orthopaedic patients, which suggests that the orthopaedic fracture clinic may provide an opportunity to identify and subsequently help many women who are victims of IPV. Our focus groups with orthopaedic surgeons and surgical trainees and our interviews with opinion leaders also identified multiple facilitators for screening for IPV in orthopaedic fracture clinics. Most focus group and interview participants are in favour of IPV screening programs and would be supportive of their implementation. Multiple factors need to be considered when designing an IPV screening and assistance program and future research needs to focus on the evaluation of educational programs as well as the effectiveness of different IPV screening and assistance programs.