STRENGTHS AND LIMITATIONS OF THIS STUDY
-
8-year data linkage study providing analysis of a large number of patients spanning 180 public hospital emergency departments (EDs) across all New South Wales.
-
Criteria used to define violence against women aimed to capture all forms of violence.
-
The rate of violence against women was age-standardised to increase the accuracy of the trends and patterns.
-
Potentially underestimated incidence of violence against women, given many women did not disclose, nor were asked about family and domestic violence in ED settings.
-
Grouping statistical areas beyond metropolitan and non-metropolitan areas may reveal other valuable geographical factors impacting the rate of violence against women.
Introduction
Violence against women (VAW) is a major public health concern worldwide, including Australia,1–3 and constitutes one of the most common types of violence.4 For the purposes of this study, we adopt VAW as an umbrella term that encompasses domestic violence (DV), intimate partner violence (IPV) and family and DV (FDV). This framing is consistent with the broader concept of family, domestic and sexual violence used by the Australian Institute of Health and Welfare (AIHW), which includes physical, sexual, economic and psychological abuse perpetrated by a current or former partner or family member. All of these terms reflect the use of force to harm, threaten, coerce or control women. Commensurate with all these definitions is the manifestation of force to harm, threaten, coerce, oppress and control women.4 In 2018, the WHO reported nearly one in three women globally, aged ≥15 years experienced either physical and/or sexual violence at least once in their lifetime.1 In Australia, a third of adult assaults requiring hospitalisation are due to FDV.5 6 Additionally, the AIHW has reported that one woman is killed by an intimate partner every 11 days in Australia and at least one in six women has experienced IPV since the age of 15 years.5 7 8 Unfortunately, VAW has also been known to intensify in times of crisis, as evidenced during the COVID-19 pandemic or in armed-conflict zones.9 Some studies described worsening incidences of VAW during the COVID-19 pandemic as ‘stay at home orders’ further isolated women in abusive relationships and left them at greater risk for harm.10–12
VAW has devastating health consequences, causing serious short-term and long-term problems for women including chronic disease, stress, anxiety, miscarriages, pregnancy terminations, sexually transmitted infections, depression and increased suicidality.1 2 11 This form of violence also significantly impacts children and has been associated with higher rates of infant and child mortality and morbidity.1 In addition to health outcomes, VAW is known to have a high social and economic cost for women, their families and societies.1
Emergency Departments (EDs) may be the first and only point of access for women experiencing violence.2 Australian research has shown a gap in emergency healthcare workers’ responses to IPV, particularly the underreporting of this form of violence in regional hospitals.2 Women experiencing violence may not disclose their injury cause when presenting to EDs out of fear or shame. ED clinicians do not routinely or systematically screen for VAW in their triage assessment. This research is in keeping with international studies suggesting that IPV is often a missed opportunity for screening in EDs.4 13 Although some controversy exists regarding the appropriateness of routine screening for VAW, some studies have recommended prioritising screening among groups with higher observed prevalence, such as younger women, women with lower socioeconomic status, Indigenous women, and those living in rural or remote areas. However, it is recognised that relying solely on demographic or social characteristics risks missing women at risk, and therefore targeted screening should complement, rather than replace, broader clinical awareness and appropriate training for all ED clinicians.13 14 There are few reported studies examining trends and characteristics of ED presentations for this vulnerable patient population in Australia and this gap in knowledge hampers the ability to improve screening strategies at the point of care.15
The present study aimed to describe features and direct age-standardised rates of ED presentations identified as related to VAW and confirmed cases of FDV inpatient admissions and compare any differences in geographic locations across New South Wales (NSW) Australia.
Methods
Design and setting
This was a retrospective analysis of linked data using routinely collected public hospital data in NSW, Australia. NSW has a population of 8.4 million people with approximately 70% of the population residing in metropolitan areas along the eastern seaboard.16 There are approximately 180 designated public hospital EDs in NSW ranging from 12 level six major tertiary referral centres to 32 level one rural multipurpose centres.17
Study population
Study participants were all individuals with female sex recorded and aged ≥15 years, who presented to any NSW ED between 2015 and 2022, with one or more of the predetermined criteria indicating VAW. In this study, VAW was operationalised based on diagnostic codes and presenting problem text available in NSW ED and inpatient datasets. As such, our definition captures violence, assault, abuse and neglect presenting to hospital settings, but does not extend to all recognised forms of VAW (eg, female genital mutilation, reproductive coercion, workplace sexual harassment or other forms that may not be identifiable in routine health data).18 ED separation diagnosis or diagnostic codes where the patient was admitted to an inpatient unit were used. ED triage presenting problems were based on text fields and diagnoses were based on International Coding Diagnoses Version 10 Australian Modification (ICD-10-AM), ICD Version 9 Clinical Modification (ICD-9-CM) or Systematized Nomenclature for Medical terms Concept Terms (SNOMED-CT) (table 1). We included records with any inpatient diagnosis in the following ICD-10-AM code ranges T73.8-T74.9, W50.00-W51.99, X85.00-Y09.99, Y87.1, R45.6, Z61.4-Z61.6 and Z62.4–62.5. These codes were interpreted to include all forms of violence, assault, abuse and neglect against women, in any location, during any activity and by any perpetrator. Confirmed FDV was defined as a subgroup of the above, who had ICD codes that identified the perpetrator as spouse, domestic partner or other family members. Information about the perpetrator themselves was beyond the scope of this study. Additional codes from SNOMED-CT and ICD-9-CM and descriptors can be seen in table 1. Patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research.
SNOMED-CT codes, ICD codes and examples of presenting problems
Data collection and variables
Patient-level health data were obtained from NSW Centre for Health Record Linkage, which was linked probabilistically based on date and time of admission and master linkage keys to generate a single linked record across the ED Data Collection (EDDC), Admitted Patient Data Collection (APDC) and Registry of Births Deaths and Marriages for a given patient presentation.
The EDDC includes routinely collected health information on all ED episodes of care across NSW. Collected variables include age, gender, triage categories (Australasian Triage Scale, ATS), presenting problem, ED diagnosis, mode of arrival (ambulance, self-presents, police/correctional, other) and Indigenous Status (identified as Aboriginal and/or Torres Strait Islander origin in any record). Location of residence in EDDC was coded using Statistical Area Level 2 (SA2 2016 code) and converted to SA Level 4 units for the purposes of this study and only the location of residence documented in the incident presentation for a given patient was used. The EDDC data dictionary is publicly available.
EDDC records were linked to APDC records if individuals had an inpatient admission within 3 days of an ED presentation. The APDC collects data on inpatient episodes of care at public hospitals in NSW, including inpatient diagnosis codes, principal diagnosis and length of stay (days). Physical injuries and body regions injured were based on inpatient diagnosis codes and Injury Severity Scores (ISS) estimated using a previously described injury code mapping tool.19 During the study period 2015–2022 inclusive, there were approximately 24 million ED episodes of care from all public hospitals except six small rural multipurpose units representing less than 0.1% of total presentations. Estimated residential female population (aged ≥15 years) by SA2 by year was obtained from Australian Bureau of Statistics and metropolitan and non-metropolitan areas of residence defined by Australian Statistical Geography Standard.20
Outcomes
The primary outcome was the age-standardised rate of VAW presentations to ED per year and confirmed cases of FDV inpatient admissions in both metropolitan and non-metropolitan geographical locations in NSW Australia.
Data analyses
Overall population rates of VAW presentations to ED were calculated using age-standardised populations using 2016 as reference and trends over time summarised using the annual rate of change. Rates were presented as median and IQR. Descriptive statistics were used to compare metropolitan and non-metropolitan ED presentations with respect to clinical and demographic characteristics. Cases were only counted once per calendar year to estimate incidence and geospatial correlation. χ2 tests were used to compare proportions and Mann Whitney U tests were used to compare normally distributed continuous variables. We also compared inpatient characteristics and ED presenting problem related to confirmed FDV inpatient admissions. All statistical analyses were performed using SAS Enterprise Guide V.6.3.
Patient and public involvement
Patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research.
Results
Baseline characteristics for VAW presentations are presented in table 2. The proportion of VAW presentations was highest in the 25–44 year age group (n=9705, 45.7%) and 60% of presentations occurred after hours (18:00–08:00). Almost 20% of VAW presentations were women who identified as Aboriginal and/or Torres Strait Islander (19.6%, n=4153). The proportion of Indigenous presentations was higher in non-metropolitan regions (30.9%, n=2675), while presentations in older women aged ≥65 years were greater in the metropolitan area (3.5%, n=437), compared with non-metropolitan areas (2.3%, n=197).
Baseline characteristics of emergency department presentations related to violence against women (VAW) and comparison between metropolitan and non-metropolitan geographical locations in New South Wales, Australia
More than half of women presenting with VAW self-presented (52.5%, n=11 142) while almost 40% (39.6%, n=8394) arrived by ambulance. The majority were allocated an ATS Triage category of 3 (41.0%, n=8687), followed by category 4 (36.9%, n=7829) (p<0.0001). Almost a quarter of all VAW presentations were for sexual assault (24.4%, n=5181), with significantly more sexual assault presentations in metropolitan areas (n=3626, 28.9%) compared with non-metropolitan areas (n=1555, 17.9%) (p<0.0001). 8% of VAW presentations (n=1696) had a confirmed inpatient diagnosis of FDV and 14% of VAW presentations (n=3018) were admitted to hospital. 6.5% did not wait rate and the death rate was less than 1% although this accounted for 20 deaths (table 1). The median rate of VAW presentations was 6.12 (IQR 3.97–9.68) cases per 1000 population per SA2 location in non-metropolitan areas and 3.91 (2.53–5.48) cases in metropolitan areas.
For women with confirmed FDV inpatient admissions (table 3), the majority were in the 25–44 year age group and less than half (46.1%, n=782) were widowed. The most common injured body regions were the skin (52.0%), head and neck (38.3%) followed by facial injuries (23.7%), and severe injuries occurred in 6% of admissions based on the ISS (table 3). Confirmed FDV inpatient admissions were most likely to present with alleged assault, multiple trauma and head injury, whereas other VAW admissions were more likely to present with alleged assault, care review and sexual assault (table 4).
General characteristics of confirmed family and domestic violence inpatient admissions
Emergency department presenting problems as listed by triage for confirmed family and domestic violence (FDV) inpatient admissions and other violence against women (VAW) presentations
Rates of VAW presentations increased remarkably in the 65–84 year age group, increasing from 45 per year in 2015 to 79 presentations per year in 2022 with an annual rate of change of 7.3%. (figure 1).
Age-standardised incidence of violence against women (VAW) emergency department (ED) presentations. yo, year-old.
Figure 2 provides heatmaps that illustrate the presenters and presentations per 10 000 female population in 2015 versus 2022 and the raw presentation totals, across the greater Sydney, outside Sydney and the rest of NSW. A schematic of how the three regions were defined is presented in figure 3. Areas where the populations are below 50 or are not usual places of residence are excluded from the heatmaps.
Heatmaps for presenters and presentations per 10 000 female population in 2015 versus 2022 and raw presentation totals. Missing (grey) areas due to low or no residence recorded by Australian Bureau of Statistics (ABS) at time of Census. NSW, New South Wales; VAW, violence against women.
Schematic of region definitions. CC, Central Coast; NSW, New South Wales.
Discussion
The present study aimed to identify the clinical and demographic characteristics for VAW ED presentations and confirmed FDV inpatient admissions, and to compare these between metropolitan and non-metropolitan geographical locations in NSW.
Our study demonstrated significantly different patterns and rates of VAW ED presentations by geographic area. We found that the median rate of VAW ED presentations per SA2 per 1000 population was 6.12 (IQR 3.97–9.68) cases per 1000 population per SA2 location for non-metropolitan areas and 3.91 (IQR 2.53–5.48) cases per 1000 population per SA2 location in metropolitan areas, consistent with previous studies showing higher overall rates of VAW in rural areas when compared with major cities.7 Additionally, previous investigations have reported the impact of socioeconomic factors such as unemployment and lower educational levels on the severity of VAW.21 The socioeconomic and demographic factors resulting in the higher rate of VAW in rural areas of Australia will have to be further investigated. Indigenous women were over-represented in non-metropolitan area VAW presentations compared with metropolitan VAW presentations. Overall, with almost 20% of all VAW ED presentations being from Indigenous women, this alarmingly is almost five times greater than the Indigenous female population in NSW (4.2%). The National Plan to End VAW and Children 2022–20323 supports measures designed to achieve Closing the Gap Target 13 By 2031, aiming to reduce family violence and abuse against Aboriginal and Torres Strait Islander people by 50%, as a target for the future standalone First Nations National Plan. Data from this study provide evidence against which such progress can be assessed over time.
A specific set of characteristics was noted for VAW ED presentations and confirmed FDV inpatient admissions. For VAW ED presentations, we discovered a higher proportion of women in the 25–44 age group, a majority of after-hours presentations, a significant proportion identifying as Aboriginal and/or Torres Strait Islander, and many presentations being allocated a triage category of 3. Additionally, sexual assault made a high proportion of VAW (25%), and 8% had a confirmed inpatient diagnosis of FDV with 14% having been admitted in hospital. For patients with confirmed FDV inpatient admissions, we again identified the predominance of admissions in the 25–44 age category. Common injury patterns mainly affected the following respectively: skin, head and neck and facial injuries. Less than 10% of these patients had severe injuries based on their ISS.
Internationally, previous studies have looked at the characteristics of VAW ED presentations.4 10 13 21 These support some of our above findings, notably regarding injury patterns4 13 and the approximate age of women affected.4 21 As previously mentioned, some of these studies have also made the link between VAW demographic and socioeconomic characteristics and the severity of injuries, where unemployed women and women with primary educations only experienced more severe violence.21 In Australia, there are only a few reported studies15 examining trends and characteristics of ED presentations for this vulnerable patient population. Future investigations22 exploring risk factors of VAW can supplement our findings in order to better detect and prevent VAW. Highlighted by Gharfounia et al,15 the use of screening tools has allowed the improvement of VAW patient identification worldwide, resulting in a higher rate of VAW than previously thought. The NSW Health Strategy for Preventing and Responding to Domestic and Family Violence 2021–2026 recognised FDV as an urgent public health issue, which predominantly impacts women. Public health efforts in NSW are currently aimed at addressing these issues; however, routine screening for FDV has not yet been implemented across all NSW EDs.15 By identifying characteristics of VAW ED presentations, screening efforts in our EDs can be improved. Educating staff and creating tailored screening tools should help identify patients that may otherwise be missed as potential VAW patients. Furthermore, the differences seen between metropolitan versus non-metropolitan areas are significant. Overall, in non-metropolitan areas, we see a higher rate of VAW ED presentations and a larger proportion of Indigenous patients. This difference noted between metropolitan and non-metropolitan can support the allocation of healthcare resources to areas of need in NSW. Our research has echoed previous findings both in Australia23 and overseas24 showing that Indigenous women are at greater risk of violence and over-represented as victims of VAW. The context in which Indigenous women and people experience this form of violence involves the effects of colonial and institutional violence leading to intergenerational trauma that has ongoing devastating and long-term impacts on the first peoples of Australia.23 Given NSW has the largest population of Aboriginal people in Australia, a tailored focus on this patient population needs to be addressed when implementing screening efforts and should involve Aboriginal healthcare workers. Routine FDV screening has been piloted in three EDs across NSW; however, further trials are required to enact permanent clinical practices.23 We hope that our research can aid supplement efforts to improve screening tools and provide evidence that tailored screening tools can help address this important public health issue.
Limitations
The most notable limitation of this paper is the potentially underrepresented ‘true incidence’ of VAW given many women do not disclose, nor are asked about FDV in ED settings. For example, patients presenting with facial injuries and being discharged home with facial injuries with no documentation of confirmed FDV would inevitably be missed from the dataset. Additionally, we acknowledge the limitation of using only the first episode of ED presentation to derive geographic residential location; a key stratification variable over a study period of 8 years. It is plausible that individuals were counted in one geographic group (ie, non-metropolitan vs metropolitan), but over time may have relocated to another geographic group. It is not anticipated that this would vary the results substantially; however, an exact estimate has not been calculated.
Second, these data do not provide information on perpetrators of VAW. While FDV and IPV often relate to partners being perpetrators, the high number of widowed patients in our cohort of confirmed FDV inpatient admissions might indicate violence perpetrated by other people in these women’s lives, although there is insufficient information in the administrative data about personal relationship. Lastly, although this study included data collected over 7 years which encompassed the COVID-19 pandemic period, the role of COVID-19 on VAW was out of the scope of our study.
Conclusion
Our study showed some significant differences between ED presentations of VAW in metropolitan and non-metropolitan areas. These findings may contribute to further discussions regarding routine FDV screening in EDs and inform public health efforts and allocation of health resources and referral pathways within NSW.
Data availability statement
Data may be obtained from a third party and are not publicly available. Data subject to third party restrictions. The data that support the findings of this study are not publicly available due to privacy or ethical restrictions.
Ethics statements
Patient consent for publication
Ethics approval
The study was undertaken as part of the previous ethics approved Sydney Triage to Admission Risk Tool project which aimed to determine predictors of various clinical outcomes and disposition in ED. Ethics approval was obtained from the NSW Population and Health Services Research Ethics Committee (2019/ETH01600). A waiver of consent was granted.
Acknowledgments
We acknowledge the assistance of staff from the NSW Centre for Health Record Linkage for their assistance and provision of linked data used in this study. We acknowledge and thank the staff in the Family, Domestic and Sexual Violence Unit of the Australian Institute of Health and Welfare for their advice regarding coding selections. We also acknowledge the lived experience of the women whose experience of violence these data represent.
