Indigenous women and intimate partner homicide in Australia: confronting the impunity of policing failures

ABSTRACT In Australia, Indigenous women experience disproportionate levels of violence compared to others in our society. Recent, horrific examples of Indigenous women's deaths have come to light through coronial courts' public inquests. This article examined 151 Australian coronial court investigations and inquests over a 20-year period (2000-2020), analysing them thematically. The analysis highlighted the vulnerability of Indigenous women to intimate partner homicides, noting that the women's deaths, in most instances, were entirely preventable. This article explores specifically the actions of police, given they are often the first responders to these situations. The findings of this study concur with a coroner who labelled the actions and inactions of police as being akin to systemic racism, or at the very least, lazy policing. The profound system failings at the intersections of law, policy and practice have arguably in some of these cases contributed to these women having lost their lives. This article explores firstly, the nuances of the coronial findings, secondly, it interrogates policing practices and processes. It concludes critically reflecting on ‘systemic racism', coroners’ recommendations for addressing this problem and the significance of this issue in the broader context of a recent Queensland inquiry into policing that affirms similar findings.


Introduction
Intimate partner violence in Australian Indigenous communities is unfortunately a common experience. Indigenous women are more likely to experience this type of violence than any other population group (Australian Institute of Health and Welfare, 2018, p. 83). Available national data tells us that despite comprising less than 3% of the population, Indigenous women have consistently experienced higher rates of homicide compared to non-Indigenous women since [2005][2006], with the average rate being eight times higher than non-Indigenous women (AIC, 2022). Further, national data on unsolved missing persons cases highlights that Indigenous women represent up to 10% of cases, with many being presumed dead (Collard, 2019). Moreover, when compared to non-Indigenous women, Indigenous women are 32 times more likely to be hospitalised for assault-related injuries (Australian Institute of Health and Welfare, 2018, p. 83). However, it should be noted that in some jurisdictions the rates of violence and homicide are far greater. For example, in the Northern Territory, the Indigenous victimisation rate is approximately 18 times higher than in any other jurisdiction, and homicide rates are also disproportionately higher, given the isolation that comes with being in a state largely defined by remoteness (Kerr, Whyte, & Strang, 2017;Lloyd, 2014;Lackey, 2018).
Whilst the statistics above paint a grim picture, they provide little insight into the private nature of most intimate partner relationships. This means that rarely do we see, hear or appreciate the full extent of the violence experienced, and its impact upon women, children and families. Their stories, unfortunately, become muted in the numbers. In contrast, coronial inquests provide an opportunity for the stories of victims to be heard, albeit in the aftermath of their deaths. Importantly, they can also provide closure to the victims' families. The inquests seek out and record as many of the facts concerning a death as the public interest requires. The focus is on discovering what happened, not on ascribing guilt, attributing blame or apportioning liability. They are a vehicle through which understanding can be achieved, and change can be advocated, which may in turn prevent future intimate partner deaths.
Indeed, over the past decade, an increasing number of coronial inquests have attracted media attention in response to horrific examples of intimate partner homicides in multiple Australian jurisdictions. These have been supplemented by Royal Commissions including the Victorian Royal Commission into Family Violence in 2015 and independent inquiries, including the current Independent Commission of Inquiry into Queensland Police Service responses to domestic and family violence examining structural failures and ways to ameliorate them (Queensland Government, 2022;Maddocks, 2016). But despite these efforts, in June 2022, the Northern Territory Acting Coroner Elisabeth Armitage reported in the media, after handing down her finding in the death from a fatal assault of Indigenous woman Roberta, that police 'did nothing to help her'. In fact, the fatal assault was the seventh time that her partner had abused her in less than two weeks, after a long history of previous assaults and a period in jail. It was five days after Roberta had been told by police to 'stop calling us'. Acting Coroner Elizabeth Armitage summed up this case as one in which police failed to follow any of their procedures concerning domestic violence complaints, and that their manner towards Roberta was rude and dismissive. But these actions and failures were not confined to the actions of police alone; the triple zero call taker incorrectly classified Roberta's calls for help, and the parole officer tasked with supervising Roberta's partner was oblivious to his breaches of parole conditions. The breakdown in communication across these services and the lack of integrated support contributed to Roberta's death (Hynnes, 2022).
As can be demonstrated from the above case, coroners' findings in these instances highlight that women's deaths are rarely without warning and are generally preceded by increasing violent or abusive incidents indicating a heightened risk of future harm. It is because of these indicators that these types of deaths are considered some of the most preventable. However, prevention is only possible if the systems designed to protect women are capable of doing so and heed the warnings through practices and processes focused on addressing women's safety.
To date, research into intimate partner homicide in Australian Indigenous contexts has been limited (McGlade & Tarrant, 2021) and focused on single specific jurisdictions (Lloyd, 2014;Mouzos, 2001). This research differs in that it utilised coronial data available through the National Coronial Information System (NCIS) to analyse 151 Indigenous women's intimate partner deaths examined by coroners over a 20-year period (2000-2020). In analysing this data, this article firstly presents the patterns in victimisation, offending and the detail of intimate partner homicides over this period to appreciate the context of these deaths. Secondly, the article focuses on systems designed to protect women, specifically policing, and what coroners have said within this sample about the practices and processes of policing in preventing Indigenous women's deaths. The third and final section of the article critically reflects on 'systemic racism', coroners' recommendations for addressing this problem and the significance of this issue in the broader context of the recent Queensland state government inquiry into policing that affirms similar findings (Queensland Government, 2022). This article is a timely contribution to a national conversation occurring through a Senate Inquiry into Missing and Murdered First Nations Women and Children that is giving rise to more thoughtful consideration on what we can learn from coroners, coronial processes and the critical review of women's deaths in the aftermath of intimate partner violence.

Method
Coroners' courts operate in all Australian jurisdictions. They are typically governed by the specific jurisdictional coroners act. These acts define the functions and powers of coroners, obligations in reporting deaths, the process and duties placed on coroners for investigating deaths and the manner and form in which an inquest is undertaken. Typically, unexpected, unnatural or violent deaths must be reported to the police and/or the state coroner. In broad terms, the coroner is then responsible for enquiring into the circumstances surrounding such deaths. They determine the identity of the deceased, the time, place, cause and manner of death. They also, when relevant, make comments on matters pertaining to public health, safety and/or the administration of justice (Coroners Act, 2017 (NT), Queensland Courts, 2013).
Whilst it is noted that all reportable deaths must be investigated, most do not result in the coroner conducting an inquest. The coroner decides whether an inquest is necessary. They are typically held when there is reasonable doubt about the cause or circumstances of the death. They are also conducted when it is in the public interest to do so, for example, when addressing systemic defects or risks to assist in maintaining public confidence in the administration of justice, health services or other public agencies, or to prevent other similar deaths (Coroners Court of Victoria, 2018). The proceedings are formal judicial hearings into the causes and circumstances of death, in which the parties involved present evidence, cross-examine witnesses and are often represented by lawyers (Coroners Act, 2017 (NT), Queensland Courts, 2013).
The investigative process is essential to informing the decisions a coroner makes in respect of an inquest. A coroner has to be careful, particularly in cases where homicide is suspected, as their decisions need to be made so as not to jeopardise the fairness of a criminal trial. Given this, the coroners act in all Australian jurisdictions prohibits a coroner indicating or suggesting that an offence has been committed by a specific individual. It is important to note that the coronial inquest will be halted if, during the process, a person of interest is charged with an offence relating to the death being investigated. This is done to protect the fairness of criminal processes. The coroner has the discretion to restart the inquest once the criminal process has concluded. However, this rarely happens, as the criminal process will have identified the deceased person, the time and date of their death and the manner and cause of their death, making the coronial inquest redundant. Lloyd confirms this practice in her examination of intimate partner homicides and missing women that took place in the Australian central desert area between 2000 and 2008; none of these proceeded to inquest for the reasons noted above (Lloyd, 2014, p. 102). In this study, only 12.6% of 151 deaths in a 20-year period proceeded to full inquest. Whilst on face value, a coroner's decision not to proceed may seem appropriate given the limited resources of coronial offices, it also presents as a missed opportunity to reflect on the broader preventive issues relevant in specific cases, for example, system failures that require reform to prevent future deaths (Coroners Act, 2017 (NT), Queensland Courts, 2013).
To some extent this is being rectified by State and Territory Domestic and Family Violence Death Review Teams (DFVDRT), which have progressively come online since 2011. However, two jurisdictions are still without such systems in place-Tasmania and the Australian Capital Territory (AFVDR & ANROWS, 2022). These DFVDRT teams have available to them additional information outside of coronial sources including briefs of evidence, police reports, media reporting, sentencing remarks and agency records (ADFVDR & ANROWS, 2022). This has impacts for providing a complete jurisdictional but also national snapshot of Australian Domestic and Family Violence Death Reviews, specifically findings and recommendations across jurisdictions in order to improve the response system and thereby prevent future deaths. To build on the work of these teams and the network more broadly requires input from Indigenous people experienced in the violence against women field, including researchers able to contextualise the complexities of the findings and to reflect on the accountability of failures within systems.

Locating cases
Typically, inquest findings are published on the office of the state or territory's coronial website. This is consistent with the public nature of the inquest process and a requirement under governing legislation (eg, Coroners Act, 2017 (NT)). Instances in which inquests are not made public include when it may prejudice a person's fair trial, it would be contrary to the administration of justice, national security or personal security, or involve the disclosure of details of sensitive personal matters that may not be in the public interest as determined by the coroner. In these circumstances, inquest findings may be anonymized or not made publicly available at all (Queensland Courts, 2013). For this research, application was made to the National Coronial Information System (NCIS) to access a more complete dataset. This does involve several layers of ethical clearance and agreement to respect the anonymity of victims in the reporting of data (ethics clearance approval obtained: AIATSIS EO177-25052020; AH&MRC 14425873; JHREC CF/19/30712; NCIS NRC M0440; UNSW HC190655; WACEC EC08/2021). NCIS holds data for all Australian states and territories, from 1 July 2000, with the exception of Queensland, whose data came online 1 January 2001, and for New Zealand from 1 July 2007 (NCIS, 2022).

Inclusion criteria
Inquest and investigation findings were included subject to the following criteria: firstly, that the victim could be identified as Aboriginal and/or Torres Strait Islander; secondly, the victim was female; and thirdly, the death was attributable to intimate partner violence. Intimate partner violence was defined as acts of violence (physical, sexual, emotional, psychological, financial, spiritual) that occur between people who have, or have had, an intimate relationship.

Study sample
151 cases were identified as meeting the inclusion criteria. These cases represented six Australian jurisdictions-Northern Territory, Western Australia, Queensland, South Australia, New South Wales and Victoria, and covered reported deaths from 2000-2018 subject to coronial investigations or inquests in the period 2000-2020. The sample is smaller than anticipated. This can be attributed to the known underreporting of Aboriginal and Torres Strait Islander status in death records, and to the conflicts that can arise between the reporting of identity on death records, and birth, death and marriage records that are influenced by historical and contemporary issues (Cripps, 2008;NCIS, 2022). As Indigenous researchers, we were able to verify the accuracy of the identity records through additional screening steps including notations on coronial files indicating Indigenous organisational involvement with victims.

Analysis
As an Indigenous woman, I come to this research project already appreciating that much of the research about Indigenous violence has been written by, as Dodson has said, 'colonising cultures [who] have had a preoccupation with observing, analysing, studying, classifying and labelling [us]' as 'a problem to be solved ' (2003, p. 27), without really appreciating what the problem is about. Indeed, Nakata (2007) explains that Indigenous knowledges and how they are represented is a result of a 'complex set of interests and contestations ' (2007, p. 190) that has already been 'ordered and organised by others and their sets of interests ' (p. 191). Certainly, in examining coronial records, this is true. The stories of victims, offenders, witnesses, but also state representatives such as medical examiners, police and others, are constructed and reported through the voice of a coroner and their proceedings. In this article, however, I prioritise elevating the women's experiences primarily by appreciating their story of violence and of death. I am reminded of an Aboriginal elder who has now passed, Aunty Rosalie Kunoth-Monks, who said 'I am not the problem' (ABC, 2014). Indeed, these women whose stories lie at the heart of this article are 'not the problem'. This article works to reframe the narrative constructed of these women's experiences of violence by giving a fuller, more contextualised and nuanced appreciation of the events that led to their deaths (Tuhiwai-Smith, 1999, pp. 153-154); albeit through records that are constructed through an institutional lens. It is in the critical examination of these records, positioned as we are as Indigenous researchers, that we are able to give space to honouring these women's lives, and their stories, by learning as much as we can to prevent future deaths. It is also in this process that we create space for the women's lives to be appropriately 'grieved'.
I ask readers to respect the stories shared, feel the emotion and appreciate the impact for Indigenous women and their families. As an Aboriginal woman, I take the responsibility to hold these stories and to communicate them respectfully and sensitively, seriously. I urge the reader 'witnessing' these for the first time to not turn away or remove themselves for a 'more comfortable reflexivity' (Fickin et al., 2022, p. 56). This is because it is in the uncomfortable spaces that we grieve for lives lost, we engage our humanity and we act with urgency through advocacy to address the injustices experienced by Indigenous women and girls to prevent future harm.
To analyse the data available through the NCIS, the researchers 1 used a grounded theory approach. This is an iterative approach that involves going back and forth over the data ensuring that the analysis accurately reflects context, capturing actions and processes, as detailed in the data (Charmaz, 2006;Bainbridge, Whiteside, & McCalman, 2013). In practice, this involves a constant reflection and revision of coding instruments to ensure that we are accurately capturing 'what is being told, how it is being told and the conditions of its being told, but also the conditions surrounding what is being told' (Glaser, 2001, p. 145). Reinforcing the cultural strength and appropriateness of this method enables intersectionality to be examined at multiple levels, including in the research process itself. This type and depth of engagement with Indigenous intimate partner homicide has been lacking in the existing literature. Its significance, then, is the insight and detail that it provides to our understanding of the occurrence of violence, how institutions are interacting with it and how violence may be prevented. This is the first in what will be a series examining this dataset, speaking, in this instance, to the practices and processes of policing in Indigenous intimate partner violence, and the lessons that can be gleaned from the data. In practical terms, a combination of computer-assisted software such as Excel and NVIVO has been utilised to compile the results communicated in this article.

Part 1: patterns in victimisation, offending and the context of Indigenous intimate partner homicides 2000-2018
In this sample of 151 cases, the average age of Indigenous women subjected to intimate partner homicide was 35 years. The youngest victim was in her teens, and the oldest was in her sixties at the time of their deaths. For offenders, their age was mostly unknown, as this information is not routinely recorded in the NCIS database.
The most common type of relationship between the parties was de facto/husband (72.2%); boyfriend (15.9%); ex-partner (5.9%) and other intimate partner (6%), characterising relationships that were not clearly defined in the coronial records. The duration of the relationships was not recorded in 57% of the case files. Of the case files that did record relationship duration, that being only 43% of the sample, we were able to discern that the relationship duration ranged from less than a year (25.4%) to 10 plus years (22.4%), with a significant proportion (17.9%) of relationships lasting 1-3 years, and the rest between 4-9 years duration. Irrespective of the duration of the relationship, the relationships themselves were characterised by patterns of persistent violence, escalating in severity over time, which is consistent with recent research in this field (see, for example, ADFVDR & ANROWS, 2022;Boxall, Doherty, Lawler, Franks, & Bricknell, 2022).
In the context of where the intimate partner homicides took place, 47.7% of the cases were within significant city limits. This is an important finding, noting that, given their location, this could have meant, provided the circumstances facilitated it, the possibility of access to emergency and police services that may have prevented the death. In the other 53.3% of cases this involved more regional and remote settings where access to such services is more limited.
The coronial case files also provide graphic detail pertaining to the means by which these women died. In 61.6% of these cases, the women died from blunt force traumaassaults that went on for hours. In these assaults, the offenders used their bodies to inflict injuries, hitting, kicking, punching and stomping on the women. They also used whatever else was at their disposal-rocks, sticks, pieces of concrete, furniture, saucepans, power cords, tyre rims. This was in stark contrast to non-Indigenous offenders who, Whittle and Hall (2018) found, 'rarely used bodily force to kill their victims … the weapon used to kill their partners was predominantly a knife, a gun or strangulation device' (p. 927). The experiences endured by Indigenous women in this sample were indeed brutal attacks, as explained by coroners in this sample, but also by judges in criminal matters, in that these were circumstances that resulted in overkill. The perpetrators in many instances failed to stop, even when the victim was no longer conscious (Whittle & Hall, 2018). Yet, their crimes were often seen as less serious by the courts as they are not aggravated crimes involving weapons such as guns or knives, a finding also consistent with Whittle and Hall's findings. Within the case files we examined, a further 31.1% of Indigenous women lost their lives as a result of being stabbed, shot, strangled or drowned (4%); and 3.3% of the deaths were unable to be classified.
The case files indicate that alcohol was involved in 69.5% of the deaths-in most instances the offender being affected, but in others the victim also. Drugs and other substances were also noted in 21.8% of case files. These findings are unsurprising, given reports and inquiries into Indigenous domestic and family violence that have long associated alcohol and drugs with violence (see, for example, Bolger, 1991;Robertson, 1999;Wild & Anderson, 2007). These findings are also consistent with findings in the broader Australian population that 'over 60 per cent of male homicide offenders who killed a female intimate partner engaged in problematic drug and/or alcohol use in the lead-up to or at the time the homicide' (AFVDR & ANROWS, 2022 p. 33). These results do need to be read with caution and will require further analysis to understand the additional other factors that are likely to be associated with the homicides analysed in this article.
In respect of the 151 Indigenous women who lost their lives, the justice system ultimately dealt with 106 offenders to a sentencing outcome. These offenders are both Indigenous and non-Indigenous people. The exact proportions are still to be determined, as the NCIS classifications are focused on those who have died, and thus the capture of Indigenous identity for offenders is not a core component of the data set. A further project is required to data match these records with sentencing and criminal justice records to capture this detail, subject to these systems having recorded Indigenous identity accurately. Certainly, the methodology used by Buxton-Naminsyk (2021) would suggest that increased access to a broader linked dataset would likely confirm what is suspected-a high proportion of these deaths are committed by Indigenous partners as opposed to non-Indigenous partners. She found in her study of 68 Indigenous women's deaths from 2006-2016 that 57 were killed by First Nations partners (Buxton-Naminsyk, 2021, p. 5).
The ripple effects of these women's deaths are also acknowledged in this dataset-41.7% of the Indigenous women killed were identified from the records as mothers. Seven women were also pregnant at the time of their death, with one in the late stages of pregnancy. It is notable that 50.33% of the records make no mention of the parental status of the women, so the number of mothers could be much higher. Of these women, and of the records that included details pertaining to their children, we were able to ascertain that approximately 154 children would have experienced the loss of their mother, and likely their father, had he been the offending parent subsequently dealt with by the justice system. Thirty-nine children are recorded as having witnessed the violence, and possibly the homicide, demonstrating the importance for these children of sustained access to trauma-informed care.
Part 2: patterns in policing Indigenous intimate partner violence As mentioned earlier, the study design included both coronial investigations (87.4%) and inquests (12.6%). Both provide insight into intimate partner deaths, but it should be noted that coronial investigations provide less detail reflecting on systems. Their focus is more specifically on replicating the information needed to record a victim's death, largely derived from the conclusion of matters before the criminal justice system. This is a limitation of this study. This next section of the article reflects on police involvement and actions in cases ultimately involving the death of Indigenous women. These reflections are important, given that 37.7% of the sample had a documented history of engagement with police with respect to intimate partner violence. It is likely that victims' engagement with police could well be higher, and that further data linkage of these coronial files to criminal justice files or, in later years, to domestic and family violence death reviews, may certainly yield a higher result. Of the cases that had a documented history of police engagement, this section draws on 14.6% of coronial investigations and inquests from the sample that made a more detailed specific mention of the role and practices of police in responding to intimate partner violence. It is important to reflect on the significance of these mentions, given that police are often first responders, and that a safe and supportive response is the desired outcome in what is inevitably a complex situation. Certainly, the goal of coroners in shining a light on police practice is to give credit where it is due, should there be good practice, but equally to recommend improvements in practice to prevent future deaths.

Initial calls to 000
Several cases spotlighted the significance of time lapses between initial calls to 000 and police units arriving at the scene. These time lapses in part related to systems being unable to account for the precarious situations that victims found themselves in, where they may be unable to speak due to safety concerns, or because of the status of their injuries as a consequence of the violence. This was demonstrated in one example, where the victim called on the evening before her death but was unable to speak. However, the audio recording was able to capture the violence and the victim's distress. The automatic 000 system disconnected the call with no follow up. In the early hours of the following morning, the same residence called again, asking to be connected to police, and police could hear moaning and whimpering, but no one engaging with them, until a male asked for an ambulance. The officer attempted to get the required information, but the information provided by the male was inconsistent with information they had on the residence the call was coming from. The officer tried calling back, but was unable to reestablish contact, after which police did not follow up the call further. Approximately an hour later police received multiple calls from neighbors that a man had alerted them to an issue with his partner. Police arrived within 10-15 minutes of these calls, but the victim was unresponsive, CPR was commenced, an ambulance arrived and confirmed that the victim had died. The coroner's question in this case was: Had earlier calls been adequately addressed by sending police earlier the night before, or more specifically at the first call that morning, would the victim have survived?
Whilst the above case speaks to a system that failed to log these details in their call system and to adequately follow up, another case spoke to categorizing what was clearly reported as a domestic violence case as a priority 3 welfare check that ultimately delayed police arriving at the residence by over an hour. The coroner in this matter also asked whether the victim's life might have been saved had the police arrived earlier. But equally in this sample there were cases that were prioritized higher but that also experienced significant delays. These delays were due to emergency service policies that dictated that, for example, in town camps, under a 'safety first' principle, police and ambulance personnel are to be escorted. Whilst the delay in this scenario may well have contributed to the loss of a woman's life, the coroner was heartened that the inquest process had prompted the policy to be reviewed.

Initial assessment of situation and decision making of police on scene
When police arrive on scene their engagement with victims and witnesses is critical. Diemer, Ross, Humphreys, and Healey (2017) have highlighted that police action in these moments is primarily focused on keeping victims safe, assessing risk and the most effective means by which to produce perpetrator accountability (2017, p. 340). Indeed, police general orders and/or operating guidelines with respect to domestic violence give specific and detailed directions to officers on how cases are to be handled at every step, from initial reporting, on scene, through to investigations. Therefore, there should be a consistency in approach; though police do exercise discretion. It is in this area that coroners asked serious questions about whether police were exercising discretion or operating clearly outside of their institutional orders/guidelines. In one case, police officers attending decided initially to take the victim into protective custody due to what they reported was 'her intoxication'. But upon further questioning from the coroner, the assessment of intoxication was questioned on the basis that no conversation was had with the victim or other witnesses, including those who had phoned 000. The assessment was made based on her appearance-she was sitting on the curb when they arrived in just a pair of knickers, her partner having moments before their arrival removed her clothes as she fled from him. The only conversations the police officers had were with the alleged offender, who, during the time they were there, 'was helpful and appeared sober'. In one police officer's words, his perception of the victim was that she was 'just a drunk Aboriginal female'. It was this perception that led to the victim being placed into protective custody, however, when a higher priority call came across the police radio, the officers requested the perpetrator remove the victim from the back of the police vehicle. The victim was promptly swung over the perpetrator's shoulder and taken inside their home. When another police vehicle attended 15-20 minutes later, police found the victim alive, lying on the floor in the doorway of a bedroom. Witnesses inside the house reported that there had been further violence since the other police unit had left. Police requested an ambulance, however, by the time the ambulance arrived, she had stopped breathing and resuscitation efforts were necessary. Ambulance and hospital staff were unable to save her. The coroner in this case referred to the actions of the first officers on scene as 'abysmal'. The Police Commissioner's office did not disagree, and they proffered an unqualified apology to the victim's family and friends. This is not an isolated example of police practices being non-compliant with police general orders and/or operating guidelines with respect to domestic violence. There were many examples of police having had extensive engagements with victims and their partners over a long period of time. The files documented frustration that Indigenous women were clearly calling the police to diffuse violent situations with their partners, but would often not want to proceed with police holding their partners accountable through the criminal justice system. This created, as Cripps and Habibis (2019) have described, a 'degree of complacency and resignation among service providers [police], creating a sense that nothing they do is going to change the situation' and the acceptance of violence in Indigenous families and communities as acceptable or something that does not require action or intervention 'because they keep going back to the situation' (2019, p. 38). This logic blames the victim and fails to appreciate the complex intersectional drivers that underpin it, and which consequently-and as demonstrated in this sample-condemns victims to deaths that were likely wholly preventable.
In some of the cases there was an acknowledgement by police and the victim that she had had 'a little bit' of trouble before, but without any real appreciation of what that actually meant. Cripps (2008) has written previously about how Indigenous victims will use a language of minimisation to protect themselves and that it requires an astute professional to take the time to develop rapport with the victim to unpack the meaning of such statements. In this victim's situation she had been repeatedly hit and stabbed with a stick, she had previous lacerations, and she had multiple soft-tissue injuries in addition to infections from her wounds. In her communications with police though, she was clear: she wanted a domestic violence order. She wanted the violence to stop, and she thought, as so many victims do, that the order would give her partner that message and provide an extra level of protection. However, police in response to this request asked 'if she wanted her boyfriend to go to court for the assault', and she did not answer. The police took no further action, and no domestic violence order was issued. This questioning by police in this instance reflected their complacency with this woman's situation.
It was, however, in contravention of their police general orders and was further compounded by their failure to check or log details about the victim or alleged offender in the police data system. Had they done this basic task, they would have known of the alleged offender's domestic violence history with this victim and others. It is no coincidence that she died just over a week later from further violence, with the medical examiner documenting 37 recent injuries contributing to her death. The Constable and Acting Sergeant overseeing this case were both subject to an internal police inquiry. It was determined that they had not acted appropriately, disciplinary proceedings were instigated, and both were formally cautioned in writing and received remedial training on domestic violence.
Further examples of how the initial assessment of situations encountered by police have gone awry are reflected in several coronial investigations where coroners note with frustration that police focused their entire investigations (or lack thereof) on alleged offenders' statements that the victim injured herself, without critically testing the facts. This was often the case in circumstances where the victim was said to have stabbed herself. Indeed, one coroner reflected on available research evidence at the time of presentations of stab wounds to a regional hospital; 53% were young women, and only 6.8% of cases were for self-inflicted injuries, with the large majority of injuries inflicted by others. Additionally, the types of stab-related injuries seen and labelled 'self harm' raised suspicion, given their presentation as singular wounds, most notably in the thigh region, representing a pattern not seen in other populations suggestive of intentional harm. Coroners and medical examiners had alerted police and prosecutors to this pattern. However, despite this, in their experience when such cases continued to occur, when it was suggested on scene that it was 'self harm' by witnesses, police were often likely to follow that line of inquiry. The investigations that followed were consequently less rigorous, there was less attention given to the proper collection and examination of evidence, and these cases were subject to extensive time delays. In one case the delay was nearly four years before the woman's death was declared a major crime. The likelihood of achieving a successful prosecution in this matter given this delay is slim. The coroner in this case commented that the fact that nobody has been convicted of this awful crime is an affront to the victim's family, compounded by continued police inaction and delay.

Breaches of orders
This study, like Buxton-Naminsyk (2021), found that there was a consistent practice by police to not enforce domestic violence orders (DVOs) designed to protect Indigenous women. Many Indigenous women requested DVOs, but equally many women also had DVOs that they had not requested, and therefore had not consented to, and which police had instigated for their protection. Of this sub-sample of cases, 54.5% had DVOs in place at the time of their deaths, and 18.2% of women had partners whose parole conditions specifically prohibited them from contacting them. Irrespective of the circumstances related to the issuance of a DVO, Indigenous women are in the aftermath navigating the consequences of the conditions of those orders, for example, their impact on a father's access to children and whether they can safely remain in their housing-one victim had mentioned to service providers that, given the housing she shared with her partner was allocated with his employment, she felt he had all the control and could kick her out of the house whenever he wanted to.
There are also financial implications, managing caring responsibilities, especially if there are known health issues in the family unit, and pressures from family and community not to cause trouble. Faced with these consequences, Indigenous women will often contemplate their safety and that of others and their choices with respect to breaches of those orders. These are not at all straightforward or easy decisions, and certainly the strength and courage of women to pursue a report of a breach to police deserves to be respected and taken seriously. Yet, too often in the sample for this research, the women's reports of breaches were not taken seriously, especially in instances where there was a clear escalation of repeated and more serious actions by the offender. There were also significant time delays in investigations of breaches. For example, in one case one offender in contravention of a DVO repeatedly breached conditions of the order over several days, and these were reported to police, yet the case remained unallocated for investigation until three months later. Police then attended the victim's address in relation to the matter but had not realized that the offender was in prison on remand in relation to another breach of the DVO.
In another case, a report of a breach was made to police where an ex-partner had broken into the victim's home, leaving behind a letter and a machete. Detectives received the file the next day for investigation. Had they followed up immediately, they would have known the offender was on parole for previous threats to kill the victim, and that she was at a heightened risk of a fatality or serious injury. This victim's case had not been allocated to any office for investigation of the DVO breach at the time of the victim's murder four days later.
Coroners reflected that there appeared to be a practice occurring in some states and territories where police would delay investigations and actioning the breaches until the victim had time to reflect whether this was the action she wanted to proceed withnot dissimilar to the point made above, asking victims whether they really want to proceed with charges, so as to discourage them from doing so or make them secondguess themselves. Police would then follow up a few days or a week later, to find that, in most instances, the victim no longer wished to pursue the matter. In one case police did advise a victim that 'she had the option of telling the Magistrate that she did not want to give evidence against her partner due to her fear that he may do something to her'. It was the only case that demonstrated that police understood the bind that women were in and were willing to find alternative solutions to enable the breach to proceed through the courts to secure the safety of the victim.
Coroners, however, seemed to have more insight than police who are on the front line, with one coroner commenting, '[i]n my view, the decision not to properly pursue DVO infringements on the basis of information and direction given by complainants is not a good idea. In many cases, complainants are disadvantaged indigenous women, unsophisticated, "battered and bruised", and still scared of their partners, drunk or with diminished faculties caused by alcoholism. Such persons need to be protected from themselves and no doubt, this is why police general orders, in this regard, are prescriptive in nature'. General orders may well be prescriptive that breaches should be followed through but police practice in many instances was often to the contrary, as was demonstrated in this sample.
Another point to make about breaches of domestic violence orders concerns the clear frustration that appears in the handling of some cases by police, such that they perceive the only way to secure safety is to breach both the victim and the offender, as this quote demonstrates: The main issue here is that both the father and mother are constantly breaching orders by living together … it would be best to breach both of them, put the matter before the court, and let the Magistrate make an informed decision.
If the mother returns with the child … we will have to look at breaches of orders and getting Family and Community Services involved to care for the kids whilst the parents are in our custody.
The difficulty with this approach is that it is not appreciating the drivers of violence. Why is it that the victim and perpetrator continue to live together? Has alternative housing been offered? What support has been offered to enable the father to have ongoing access to his children? These are only a few of the questions that need to be considered, in addition to engagement with services and supports to enable an order to work effectively. Incarcerating a victim and offender is not going to resolve this problem. It only creates further problems, in particular, the removal of Indigenous children from their families who may not find their way home. The latter is a stark reality, given state/territory requirements on parents to achieve parenting standards within set timeframes. However, the finality of death, as these cases have resulted in, has meant that the children involved in the above case have lost both parents, irrespective of a police response to a domestic violence order. We clearly need to do better in responding and preventing violence to avoid this outcome.

Police accountability
Many of the cases (45.5%) were subject to a police internal review. A significant number highlighted discrepancies in compliance with police general orders with respect to domestic violence, if not at this stage, then certainly at the conclusion of the coronial investigation/inquest stage. Over the course of twenty years, increasingly it became practice for police to conduct internal reviews. Many of these reviews would often reflect that the investigation did not 'highlight any poor practice or inadequacy on the part of police or any other agency, however it does highlight the difficulties in providing effective protection for victims who find themselves in this type of abusive relationship'. The framing of this statement is problematic on several levels, as it puts the victims in the frame of responsibility for the abuse they are experiencing, rather than reflecting on the broader dynamics of attitudinal, institutional and societal change needed to support victims in these situations.
In terms of accountability, an examination of these files discovered two cases where police commissioners offered formal apologies to families with respect to the handling of these women's cases. Eight officers of various ranks were disciplined, which ranged from formal cautions, counselling and one 'no confidence notice', with this person ultimately retiring. Two officers in this group were also referred for remedial domestic violence training. A further two officers resigned, with their actions referred by the coroner back to the police commissioner and the director of public prosecutions for consideration with respect to charges of criminal negligence. According to the coroner, these charges 'cast a wide net, so as to cover all acts or omissions endangering the life, health or safety of any member of the public where the risk ought to have been clearly foreseen and the act or omission avoided' (Baumer v The Queen (1988) 166 CLR 51 at [55]). It also imposes a general liability when the omission to act seriously endangers others. The coroner felt in this matter, having had regard to all of the evidence, that the actions taken by the police officers concerned demonstrated a 'risk to the deceased [that] ought to have been clearly foreseen'. Despite an extensive search of legal and case databases, it does not appear at this stage that these officers have been dealt with by the courts.

Part 3: 'Systemic racism' in policing Indigenous intimate partner violence and coroners' recommendations for addressing this problem
Given the consistent finding in this sample of investigations and inquests noting that police practices deviated from police operational practices and guidelines on domestic violence, it poses the question: Why? One coroner, who had conducted 17.9% of the inquests and investigations in this sample, reflected in one report that he felt the issue was institutional racism. He said it was a more subtle concept than overt racism, which influenced police service delivery through 'a systematic tendency that could unconsciously influence police performance'. He referenced the police investigation of the murder of Stephen Lawrence in London in 1993 that defined this problem: The only explanation or excuse offered to us for the failures and mistakes in this case are that they are the result of incompetence and misjudgement … incompetence does not without more, become discrimination … [but] … mere incompetence cannot of itself account for the whole catalogue of failures, mistakes, misjudgements, and lack of direction and control which bedevilled the … investigation.
He argued that his investigation of three Aboriginal women's deaths spoke to a system that was broken and that, if it was not impacted by institutional racism, then '[p]erhaps, it is lazy policing. Just trying to find the easiest way to wind up an investigation. Or perhaps, it is cultural ignorance'. These are profound statements, and given the findings presented in this article, they are reinforced by the practices highlighted. But what hope is there for change? Certainly, the Stephen Lawrence case and the subsequent Macpherson report (1999) brought sweeping changes in law and police practices in the United Kingdom, but can we expect the same here? The current Senate Inquiry into missing and murdered First Nations women and children may bring pressure for system change, but it is hard to tell at this stage, with its report not due until the end of 2023. However, at the start of this article, I noted that prevention is possible if systems designed to protect women are heeding the warnings of their system failures and are responding accordingly. There was evidence of this in the sample, with coroners scrutinising specific failures and police hierarchy, but also other justice-related systems such as Department of Corrections and Parole Services responding that policy and practice guidelines were changing as a consequence. They collectively stressed the necessity to communicate the importance of compliance with these policies and practice guidelines, at every level, in every way possible, as many cases demonstrate that they can save women's lives.
In addition, coroners and police hierarchy need to send strong messages that failure to abide by these policies and procedures carries consequences. In this sample, there was evidence that this was the case to some degree. However, the overriding message of bias, stigmatisation, stereotyping and devaluing of Indigenous women as victims of violence constitute, as the UN Special Rapporteur on Violence Reem Alsalem has noted, 'serious barriers within the justice system at large, resulting in discrimination and revictimization, reinforcing the already existing and deep-seated fear and distrust of the justice system' (2022, p. 14). Indigenous women's own testimony, advocacy and writing in this area also affirm these findings (see, for example, Brennan, 2022;Longbottom, 2022;McGlade & Perera, 2021a;McGlade & Tarrant, 2021b;McQuire, 2022).
Coroners have made consistent findings to use these cases of demonstrated failures for the education and training of police to improve their collective responses to domestic violence. But this on its own does not guarantee the institutional change needed to address the problems identified in this article. Coroners also referred back to the evidence encouraging government agencies and services to engage with, and implement, the recommendations of Domestic and Family Violence Inquiries.
The latter is significant in light of the most recent Inquiry into Queensland Police Service responses to Domestic and Family Violence published in November 2022. This inquiry affirmed many of the findings and comments made by coroners referenced in this article. The inquiry found in Queensland-though I would argue that there are similarities in these findings in other jurisdictions-that there is a lack of understanding of the dynamics of, and power imbalances within, domestic violence relationships. That there is a significant under-resourcing in this area, which leads to reactive, and at times short-lived, reform, and on the frontline, confusion as to expectations in police practice (2022, p. 10). But more damning is when this collides with a 'culture where attitudes of misogyny, sexism and racism are allowed to be expressed, at times acted upon, largely unchecked', and that this impacts directly on the response of police to victim-survivors (2022, p. 11).
It is the destructiveness of this culture that the Inquiry into Queensland Police Service responses to Domestic and Family Violence found existed institutionally, and which urgently needs to be addressed to prevent the loss of lives noted in this article. The Inquiry recommended a range of initiatives not dissimilar to those being recommended by coroners to address this, including training, improved guidelines and systems that can be followed consistently, and new resources such as specialist positions in domestic and family violence. In the Indigenous space, and in recognition of the historical legacy of harmful police practice, there were recommendations for a reimagining of the relationship, including through a tailored, local and specific coresponding model. These are not quick fixes. They require commitment to doing this work differently to save lives at all levels of policing practice. It requires, too, a commitment to accountability. The work of coroners will be influential in producing the checks and balances to achieve this.

Conclusion
This article has highlighted the utility of examining the information produced by coroners in investigations and inquiries of intimate partner deaths of Indigenous women. It has provided a snapshot, if you will, into the lives of women who are too often unseen in public discourse. Whose stories are often untold, in preference to those that are characterised as 'ideal victims' or who are deemed to be more 'grievable'. The narratives produced through this process enable checks to be done on attitudes that have been allowed to permeate systems; systems that, as coroners in this sample have attested, and that the recent Queensland state inquiry into police practice has affirmed, have arguably devalued Indigenous women's lives through their actions and inactions.
Whilst the sample in this article is not without limitations, reflecting on Indigenous women's deaths over time and across jurisdictions has enabled an analysis of how systems have changed, but also remained the same. The reflection on systemic racism so poignantly raised by one coroner named the culture of policing as significant, and this finding would not have held on an evidential basis without a substantial sample of cases. This finding, amongst the many others presented in this article, is also timely in light of the current Senate Inquiry into Missing and Murdered Indigenous Women in Australia. We must learn from these women's deaths. Their lives mattered to all of us but especially to their families. We honor their memories by doing all we can to improve the systems, the institutional cultures and societal attitudes that failed to hear their cries for help. We must heed what we know is international best practice and complete Domestic and Family Violence Death Reviews on each and every death that involves Indigenous women and girls. Six Australian jurisdictions are already doing this, which has been a long time coming. Regular checks and balances on systems and services engaging with Indigenous women and girls are needed to produce accountability. Critically, these reviews need to involve Indigenous community members and researchers experienced in this sector. This is important both for oversight purposes, but also for knowledge exchange to ensure that the learnings from these deaths can help prevent future deaths and the occurrence of intimate partner violence more broadly.
However, this on its own is still not enough, and so, this article also urges Australia to take part in the UN Global Femicide Watch initiative. Its aim is to promote evidencebased policies and strategies for the prevention of femicide through the collection of comparable data on femicide rates at the national, regional and global level. Australia's participation in this initiative with the provision of disaggregated data on Indigenous women would enable us to learn about efforts to combat this problem on a worldwide scale, bringing additional scrutiny and potentially new practices and opportunities to improve systems and services for the better. Indigenous women and girls who have experienced violence deserve to be treated with humility, respect and dignity. Working with them-and for them-to achieve safety must always be at the centre of the work we do. The stories presented in this article remind us that we can and must do better.