Aberrant deaths: a conundrum in domestic homicide reviews

Abstract

The Family Violence Death Review Committee (FVDRC) have historically classified those deaths that do not apparently fit within a pattern of violence as ‘aberrant’. The article aims to determine if the FVDRC classification system has been incorrectly identifying deaths as aberrant when there is the potential to prevent such deaths.

Within this investigation, we reviewed the classification of apparently aberrant deaths within the Aotearoa New Zealand Family Violence Death Review system. We also reviewed the literature that describes the nexus between mental disorder and family violence homicide to inform future classifications. Most homicides, including those associated with mental disorder, do not occur ‘out of the blue’ and are, as such, incorrectly classified as aberrant. Some of those incorrectly classified describe non-compliance with medication in combination with drug/alcohol abuse and/or a history of violence or trauma in the family.

Engagement with the psychiatric community will increase awareness of psycho-bio-social risks across the life course of families and individuals and allow preventative interventions and solutions to be identified. This provides hope for families who are supporting people with severe mental disorders, highlighting the importance of effective, comprehensive support in the community. There are a small number of cases that are truly aberrant. In these, the homicide occurred during the sudden onset of an acute first episode psychosis after a very short prodromal phase, often identified in retrospect and atypical in the life of the offender.

Background

Domestic homicide reviews (DHRs) are undertaken to prevent such deaths occurring in the future. Their impact has primarily been in understanding policy and practice issues associated with the prevention of intimate partner homicide (IPH) (1).

In Aotearoa New Zealand (hereafter Aotearoa), the Family Violence Death Review Committee (FVDRC) was established in 2009 with the purpose of:

Reviewing and reporting to the Minister on family violence deaths, with a view to reducing the numbers of family violence deaths, and to maintain continuous quality improvement through the promotion of ongoing quality assurance programmes. (2)

In 2022, there was a restructure of the Aotearoa mortality review system (3), and the FVDRC members became Family Violence Death Review Subject Matter Experts (FVDR SMEs), functioning on behalf of the National Mortality Review Committee, with the same purpose. The FVDR SMEs’ reviews are intimate partner violence (IPV), child abuse and neglect (CAN) and intrafamilial violence (IFV) homicides, which includes violence between cousins, adult children on their parents, siblings, and other familial relationships. IPV deaths as well as other forms of family violence homicides are understood as the culmination of a pattern of behaviour, and a lack of responsiveness from services tasked with supporting people’s health and wellness.

However, the death review process has also identified homicide events that appear aberrant: ‘someone or something regarded as atypical’ (4). Contrary to dictionary definitions suggesting these deaths are ‘able to be ignored or discounted’, those captured by the death review process, while atypical and a minority of cases, may help establish further opportunities for prevention.

Of the 72 intrafamilial violence deaths recorded in the FV SMEs data that occurred between 1 January 2009 and 31 December 2020, 15 (21%) were classified as ‘aberrant’ where the offender was recorded as having either schizophrenia or psychosis at the time of the death event and there was a nexus between mental illness and the use of violence.

The FVDR SMEs have been debating the nature and classification of these deaths for at least five years. The term aberrant implies a death that is so atypical that the death is inevitable or unable to be prevented. However, when reviewing such deaths there has been a growing sense that opportunities for prevention are evident in the lives of the offender, and so they cannot truly be considered aberrant. Given the FVDR SME’s role in making recommendations designed to prevent such deaths from occurring, transparency around the classification of these deaths is essential, as is the identification of change in practice to reduce the likelihood of death events.

The objective of the research presented is to provide a critical review of the FVDR SME death classification of aberrant against a review of the literature about the nature of family violence in the context of serious mental disorder. This manuscript provides an exploration of the methods used to classify ‘aberrant’ cases and the strengths and limitations of the data on which these are based. Through this we recommend a new approach to understanding those deaths currently considered aberrant, and make suggestions for future practice.

Materials and methods

This investigation uses two approaches to determining an appropriate approach to understanding those deaths currently classified as aberrant. The first is a review of the literature that describes the nexus between serious mental disorder and family violence homicide. The second is a critical review of the classification category ‘aberrant’.

Literature review

We sought to explore if there was evidence of an association between particular mental disorder diagnoses and the different categories of family violence homicide perpetrated by people living with serious mental disorder.

A detailed literature review was conducted and is provided in Supplemental File 1. The PsychInfo, Scopus and Google Scholar databases were searched on the following search terms (including subcategories): ‘Domestic Violence/or Spouse Abuse/or Intimate Partner Violence/’ or ‘homicide/’ or ‘violence and (domestic or famil* or intrafamilial or dating or partner* or spouse or wife or wives or husband* or ‘intimate partner’ or child*)’ or ‘abuse and (domestic or famil* or intrafamilial or dating or partner* or spouse or wife or wives or husband* or ‘intimate partner’ or child*)’ and ‘Mental Disorders/’ or ‘Mental Health/’ or (‘mental disorder*’ or ‘mental health disorder*’ or psychiatric or ‘mental illness*’ or ‘behaviour disorder*’ or ‘behaviour disorder*’ or ‘mental disorder*’ or schizophreni* or bipolar or psychosis or psychotic) or ‘(filicide or matricide or parricide or patricide or uxoricide or femicide or ‘domestic homicide*’)’ and perpetrator and ‘(mental or psychiat*)’.

Articles included were published within the period 2014–2023. Those considered relevant were reviewed, along with their reference lists. Where additional relevant articles were identified through the reference list, those were also included in the review. A total of 84 articles were reviewed.

Critical review of classification

We describe the development of the aberrant category as a process of understanding patterns of violence leading to a death event. We then review the strengths and limitations of data available to determine if a death is aberrant. Finally, we draw on findings from in-depth reviews of family violence death events to identify potential points of intervention.

Results

Re-examining the evidence base for aberrant homicide and mental disorder

A review of the literature describing the associations between various forms of family violence and mental disorder is presented in Supplemental File 1. It is important to note at the outset that most perpetrators do not have a diagnosable mental health disorder. While there is an association between a wide variety of mental health disorders, perpetration and victimisation from violence, from the literature review, it is not possible to identify a causal relationship between specific disorders and family violence homicide. Instead, it appears that mental disorders are part of a constellation of risk factors that enhance vulnerability to violence perpetration and victimisation.

Indeed, there may be differences in the nature of the association between mental disorder and the type of family violence perpetrated. As such, the literature review has been structured according to type of family violence.

A summary of the findings from the review are presented in Table 1.

Table 1.

Summary of literature review findings.

Family violence type	Summary of findings from literature review
IPV	Associations between mental disorder and both victimisation and perpetration of violence. Mental health variables do not consistently predict or moderate intervention outcomes, although there is a higher prevalence of mental disorders amongst male perpetrators of IPV. To prevent IPV, there is a need to address both violence perpetration and mental health needs of victims and perpetrators.
Intimate partner homicide (IPH)	A minority of IPH offenders present with psychotic symptoms at the time of the homicide. Where there may be relationships with homicide-suicide events, those who perpetrate the homicides are often older, have threatened or attempted suicide previously and are more likely to have previously used violence.
IFV	The presence of persecutory delusions and auditory command hallucinations may be elevated in IFV homicides. Schizophrenia may be a key factor explaining fatal violence from adult-aged children towards parents. Missed appointments, poor compliance with medication and drug or alcohol misuse may exacerbate risk. Mothers may be at heightened risk due to their proximity as carers.
CAN	The perinatal period is a time of increased vulnerability to the onset or exacerbation of mental disorder. While the onset of post-partum psychosis is rapid and requires urgent psychiatric intervention, when mothers present with post-partum mood disorders, interventions to prevent filicide should be possible. The known factors associated with maternal filicide seem to be non-specific and few factors are consistent across populations. In cases of filicide-suicide, there are frequently confounding risk factors, including relationship conflict, chronic health issues for the child and recent separation.
Homicide-suicide	Offenders are often older, male, and have previously reached out to medical professionals for support prior to the death event. In middle-aged groups there is a strong relationship with relationship separation. For older age groups, there is some suggestion of an overlap with chronic health conditions and caring duties. While there is evidence of mental disorder for some offenders, previous suicidal ideation, financial issues and alcohol or drug abuse were also identified as risk factors.
Familicide	Familicide is rare worldwide with an incidence rate of approximately one to two per 10 million persons. Familicides are almost exclusively committed by men, usually in their 30s or 40s. Mental health problems, relationship problems and financial difficulties are prevalent.

From the literature review we note that, where an identified relationship between mental disorder and homicide exists, missed appointments, non-compliance with medication and drug and alcohol abuse are also a feature. Further, there are frequently recorded histories of conflict, IPV or relationship breakdown.

There appear to be some high-risk groups identified within the literature. In particular, parents who care for adult children with severe mental illness (in particular, mothers – see ‘Matricide and Mental Disorder’), and mothers who present with post-partum mood disorders (see ‘Filicide and Mental Disorder’).

The review revealed prevention opportunities for family violence homicides. In particular:

• Psychiatrists, psychologists and court assessors need to undertake rigorous examination of psychotic symptoms and violent thoughts, fantasies and behaviours. The homicide potential in depressed patients and early warning signs such as threats of suicide and death need to be taken seriously.

• A specific focus is required on the peri- and post-partum period.

• There is a need for effective illness management, including treatment and medication compliance, to prevent both perpetration and victimisation.

• Substance use disorders in those experiencing mental disorders exacerbate the risks of violence.

• There appears the potential for homicide-suicide events to occur with men who present as depressed, where there is a pattern of violence in a relationship, and/or where they have taken on a caregiving role for an elderly partner.

Further, some symptoms attributed to mental disorder may be due to an underlying medical condition. Attributing violent symptoms to a diagnosis of a mental disorder involves the exclusion of significant medical pathology that may be either causative or contributory. Tardiff (5) lists the more common of these as head trauma, stroke, delirium, dementias, postictal states, electrolyte imbalances, hypoxia, thyroid disorders, and other abnormal metabolic states, which must always be ruled out (5). Space occupying lesions such as brain tumours may present with psychosis, personality change and aggression. While some can be of slow, insidious onset, others may be fast-growing and present with a sudden onset of symptoms. Klein and Hirachin note that if present, medical disorders must be treated alongside any co-occurring mental disorder or substance use (6).

Critical review of classification

Family violence as a pattern of behaviour

‘A pattern of behaviour’ as related to family violence is primarily drawn from conceptualisations of IPV (7) from feminist literature (8). As well as physical harm, other forms of violence including psychological and financial abuse contribute to this pattern (9).

Describing violence as a pattern of behaviour places a responsibility on support services to be responsive to that pattern (10). However, by describing violence as a pattern of behaviour, the FVDR SMEs may have contributed to a blinkered approach, whereby recorded pattern is necessary for family violence to be considered.

Dutton and Kerry underscore the difficulty of predicting IPH solely on the basis that previous violence has occurred in a relationship (11). Previous criminal charges are not predictive of IPH (12). In a review of previous investigations of male-perpetrated IPH, Kivisto identified that between 22% and 77% of male IPH perpetrators had a previous recorded history of IPV (13). Lower estimates were drawn from official statistics, while higher estimates were drawn from proxy informants rather than police-reported crimes.

Legal definitions of ‘aberrant’

Definitions of the term ‘aberrant’ in relation to homicide are based on understandings that (for example) the ‘aberration of mind must have been absolute in degree, and so complete at the time of the deed, as to have deprived the individual of a knowledge of right and wrong, of knowing friend from foe’ (14). This is encapsulated in the insanity defence of England and Wales, as set out in the M’Naghten Rules (1843), on which the Aotearoa defence of insanity is broadly based (15). The rule arose after Daniel M’Naghten shot Edward Drummond, private secretary to British Prime Minister Sir Robert Peel, in mistake for Peel, in the context of a paranoid delusional state. The trial set a precedent in English law and related jurisdictions.

Box 1.

. M’Naghten Rules (1843)

It must be clearly proved that, at the time of the committing of the act, the party accused was labouring under such a defect of reason, from disease of the mind, as not to know the nature and quality of the act he was doing; or, if he did know it, that he did not know he was doing what was wrong.

R v. M’Naghten (1843) 10 CI & F 200, p.210

The M’Naghten Rules uses the unqualified term ‘wrong’, as did the New Zealand Codes of 1893 and 1908. Aotearoa moved from the position of ‘wrong’ as simply meaning ‘contrary to the law’, to a qualified position of ‘moral wrongfulness’ (‘having regard to the commonly accepted standards of right and wrong’) as enacted by the Crimes Act 1961 (s23 (2)(b) (16)) (Box 2). This is an important legislative difference. Also of note, while understanding and knowledge are a matter of degree, section 23(2)(a) Crimes Act 1961 uses the term ‘understanding’, rather than ‘knowledge’ as expressed in the M’Naghten Rules.

Box 2.

Section 23(2) Crimes Act (1961)

23 Insanity

…

(2) No person shall be convicted of an offence by reason of an act done or omitted by him or her when labouring under natural imbecility or disease of the mind to such an extent as to render him or her incapable—

(a) of understanding the nature and quality of the act or omission; or

(b) of knowing that the act or omission was morally wrong, having regard to the commonly accepted standards of right and wrong.

When raised as a defence, such definitions associate the offending with serious mental disorder or ‘disease of the mind’. This is a legal rather than clinical definition, which the courts generally interpret as psychotic and major mood disorders. It is captured by the modern day finding of ‘act proven but not criminally responsible on account of insanity’ (s20 (17)). Importantly, this does not extend to the range of existential experiences normally associated with being human, such as states of anger or misery.

Further, while the death event may have occurred at the nexus between a serious mental disorder and a violent event, this does not preclude intervention opportunities to prevent this event from occurring. For the purposes of FVDR classification, it also doesn’t establish the event as aberrant, as there may have been a pattern of violence that previously existed or, as highlighted in the literature review, compounding risk factors that exacerbated the likelihood of violence occurring.

Developing an ‘aberrational’ category in the Aotearoa death review classifications

The term aberrational was first used to describe a family violence death event in the Fourth Annual Report (Box 3). The term was considered necessary to reduce the likelihood of including death events where a pattern of behaviour was not evident, resulting in spurious findings or recommendations (8). ‘Aberrant’ deaths were excluded from further analysis, while being included in the overall tally of family violence homicides.

Box 3.

‘Aberrational’ description

Some cases have aberrational features. Whilst there may have been an intimate relationship between the offender and the deceased, the killing does not appear to be an act of family violence, for example, cases in which the offender in the death event … has killed the deceased for material gain. The Committee has labelled these cases as aberrational. (8)

As the FVDRC’s understanding of IPV, as well as other forms of family violence, has developed, killing for ‘material gain’ is now considered an extreme form of financial abuse.

In preparation of the Fifth Report Data, court reports written pursuant to Section 38 (1)(a) and (2)(b)(c) of the Criminal Procedure (Mentally Impaired Persons) Act 2003 (CP(MIP)A) were collated for IFV death events. Section 38 reports can provide evidence of whether the defendant is fit to stand trial or whether they could have been impaired at the time of the offending. They are also sought to assist the court with sentencing matters. s38 CP(MIP)A outlines the process for obtaining reports for defendants in respect of fitness to stand trial, and insanity within the meaning of s23 of the Crimes Act 1961 (Box 4).

Box 4.

Section 38, Criminal Procedures (Mentally Impaired Persons) Act 2003

Power of court to require assessment report

(1) When a person is in custody at any stage of a proceeding against the person, whether before or during the hearing or trial, or while awaiting sentence or the determination of an appeal, a court may, on the application of the prosecution or the defence or on its own initiative, order that a health assessor prepare an assessment report on the person for the purpose of assisting the court to determine 1 or more of the following matters:

      (a) whether the person is unfit to stand trial:

       (b) whether the person is insane within the meaning of section 23 of the Crimes Act 1961:

       (c) the type and length of sentence that might be imposed on the person:

       (d) the nature of a requirement that the court may impose on the person as part of, or as a condition of, a sentence or order.

(2) If a court orders that an assessment report on a person be prepared under subsection (1), the court may—

       (a) make it a condition of a grant of bail that the person go to a place approved by the court for the purpose of the assessment; or

       (b) order that the person be detained in a prison for the purpose of the assessment for any period not exceeding 14 days as the court thinks fit; or

       (c) order that the person be detained in a hospital or secure facility for the purpose of the assessment for any period not exceeding 14 days as the court thinks fit, if—

              (i) a remand to a prison for that purpose would be inappropriate for any reason; and

              (ii) a health assessor has expressed the opinion, in a certificate or in evidence, that it would be desirable if an assessment, or a further assessment, take place in a hospital or in a secure facility.

For s38 reports to be ordered, they must be considered relevant by judicial officers. Of the 53 defendants included in the IFV data set at the time, s38 reports were available for 10. Eight of these defendants were European, while 58% of offenders were European and 47% were of Māori ethnicity.

The analysis resulted in a refinement in the definition of aberrant deaths, and separate definitions were documented for IPV and IFV homicides (Box 5).

Box 5.

Aberrant deaths in IPV and IFV

Intimate partner violence: Some death events have aberrational features. While there may have been an intimate relationship between the offender and the deceased, the killing does not appear to be an act of family violence. Examples are death events in which the offender appears to have killed the deceased for material gain, or where the offender had mental health issues but had not shown any previous coercive or controlling behaviours towards the deceased. (18).

Intrafamilial violence: Some cases involve IFV death events that fall within the Committee’s terms of reference, because the death was an episode of family violence, but they have aberrational features as there is no identifiable history of family violence. In these IFV death events there is an identifiable mental health history (but no known acts of violence against family members outside of the mental health disorder) and the homicide is usually committed by a family member who was mentally unwell at the time. (18)

In these situations, if the nexus between the use of violence and the mental disorder was such that the violence only occurred during a psychotic episode (for example), it was attributed to the mental disorder.

The FVDR SMEs are not constrained by legal findings in their classification processes. A finding of insanity based on a legal construct of mental disorder as a ‘disease of the mind’ may fail to capture the nuances and contextualisation of the death event and cannot be considered a proxy for classification of an aberrational family violence death (19). While s38 report-writers may be expert in matters of mental disorder, analysis of these reports as part of the death review process indicates varying levels of understanding and expertise in family violence (20).

Strengths and limitations of the approach

Quality of the information on which the assessment is based

Much like DHRs across the world, the FVDR SMEs are primarily dependent on the use of data collected through service delivery. A strategic approach is taken, whereby two to four in-depth reviews are conducted of the approximately 30–50 homicide events that occur per year. For the remaining, the following (tier 1) information is collected:

• Electronic health information.

• Police family harm notifications.

• Sentencing decisions (High Court outcomes).

• Pre-sentencing reports.

• Coronial reports.

This information allows a clear description of the pattern of violence within a relationship as described through police interactions and recorded through judicial processes (8).

Where in-depth reviews are conducted, this information is supplemented with case records and clinical notes from health providers, non-governmental organisations, the statutory child protection agency, statutory welfare providers, housing providers, prison records and interviews with friends and surviving family members (tier 2 data). A life-course approach is undertaken.

The detailed analysis of these records provides examples of where family have reported the existence of violence or control which hasn’t been captured in agency records. For example, where police have responded to the primary victim and noted:

Self-harmed during domestic with partner. Removed from house for mental health assessment and medical check-up.

In such cases, his role in the event is obfuscated and her response is primarily considered an emotional reaction to a situation. No criminal offences become evident. However, while such notes are routinely accessed for in-depth reviews, they are not available for the tier 1 data set. Only the outcome of such reporting is recorded, and this will be a recording of wellbeing rather than an offence. Where such situations occur, it is possible that as well as obfuscating his role, there is a reduced ability to determine a nexus between violence and mental disorder that may exist.

Given the strong relationship between trauma and violence experience and the subsequent development of mental disorders (21), the collation of lifetime trauma experience is necessary in DHRs. Mental health clinical records and medico-legal reports may capture trauma histories throughout the life course, including experiences of childhood abuse and neglect. However, as highlighted, there is a lack of consistency in the collection of these reports and the expertise of those charged with writing them.

The potential for bias in the data

Like many colonised societies, Aotearoa has a history of discriminatory and biased practices towards Māori, the indigenous population. The impact of this has been the over-surveillance of Māori, and the under-surveillance of Pākehā, those of primarily European descent (22). In combination with their marginalisation and forced removal from ancestral lands, this produces an over-representation of Māori in Police, Justice and statutory child protection data sets. Where the corollary exists for Pākehā, and there is an under-representation in statutory data sets, there is the potential for the relationship between childhood trauma and the development of psychoses to be missed.

As highlighted above, our review of s 38 reports also showed that New Zealand Europeans were more likely to access s 38 reports in defence of their crime compared with Māori.

While it could be considered that these two experiences of inequity would cancel each other out, criminal court judges refer to s 38 reports in their sentencing decisions. This establishes the nexus between mental disorders and violence behaviours for the FVDR SMEs to consider and therefore has the potential to influence how these homicides are classified.

Disentangling the complex relationship between mental disorder and family violence

There are risks associated with capturing an aberrant homicide as those solely related to ‘mental disorders’. Using such a broad, poorly defined term has the potential to increase stigma (23). As highlighted in the literature review, other factors are implicated in the relationship between mental illness and violence, including previous use of or victimisation from violence, substance misuse, caregiving responsibilities, financial stress and impending or recent separation.

Oram et al. (24) found differences in the time course, risk factor profile and consequences of violence against partners compared to stranger violence. While stranger violence tends to be sporadic, partner violence is usually a recurrent pattern of coercive, controlling behaviour (24), impacting on dependent children. Risk factors include childhood abuse, witnessing parental domestic violence, social deprivation, substance misuse, and violent victimisation (24). These risk factors are also associated with having a mental disorder, ranging from common disorders such as anxiety and mild/moderate depression, to serious mental illnesses such as psychoses, bipolar disorder and severe recurrent depression (24). This raises the possibility of an indirect association between mental illness and IPV through risk factors that are evident where mental disorders do not exist.

For only three of the 20 IFV death events categorised as aberrant between January 2009 and 31 December 2020 (15%) did the FVDR SMEs have access to information concerning childhood trauma for the offender. For six of the IFV aberrant death events (30%) there was an identified relationship between drug and alcohol abuse and the illness. In the absence of information about other risk factors, it is possible that additional weight was placed on the involvement of the mental disorder.

Where relationships have been established between schizophrenia and other psychoses with respect to increased risk of homicide, some of this excess risk appears to be mediated by substance abuse comorbidity (21,25,26). Identification of such comorbidities provides an opportunity to intervene to reduce homicide.

People experiencing psychosis are two to eight times more likely to commit violence compared with non-affected members of the general population and four to ten times more likely to commit homicide (27). In contrast, there appear to be higher rates of affective disorders in the specific subset of homicides against intimate partners (28). This may suggest that homicide associated with psychosis represents a particular manifestation of symptom-driven behaviour (27).

Further, despite the growing body of evidence of a relationship between mental illness and interpersonal violence, as highlighted in our literature review, most people who experience mental illness are more likely to be the victims rather than the perpetrators (29,30). People with schizophrenia are nearly twice as likely to be victims than perpetrators of homicide (31,32) and non-lethal forms of violence (33,34).

Those with serious mental illness are more often assaulted by people they know, including intimate partners and family members (35,36). Concurrently, family members are the most common victims of violence by people with mental disorders (37). A Canadian study by Penney et al. (2023) of people admitted to forensic mental health care for a homicide act between 2011 and 2021, found that victims were commonly known to the individual (35.7%) and were often family members in care-providing roles (55.4%) (38). Limit-setting practices, including management of the family member’s finances, regardless of intention, may be perceived as coercive and lead to an increased risk of conflict and escalating violence (37), as may an insistence on medication compliance.

Where there is non-compliance with medication in combination with drug/alcohol abuse and/or a history of violence or trauma in the family, it could be considered that there will be increased risk of violence being used (39). Further, where there is a pattern of violence that forms a nexus with poorly controlled psychosis or schizophrenia and families have been actively seeking the support of mental health services, there is the potential for an intervention to be put in place. While mental disorders may have been a central vulnerability for which the family was seeking help in these death events, they are not truly aberrant. A pattern existed so prevention was possible.

Violence can emerge because of psychotic symptoms such as command hallucinations, where ‘voices’ instruct an individual to harm someone, and persecutory delusions, where an individual falsely believes that someone intends to harm them (40). Humphreys et al., found that 20% of 253 first-episode psychosis patients acted violently preceding their first psychiatric admission (41). The most common symptoms reported during the prodromal phase (prior to the onset of acute symptoms and treatment commencing, often identified in retrospect) were depression, anxiety, irritability, religiosity and withdrawal (42). These symptoms may not come to the awareness of family or health services if the prodromal phase is very short, there is no prior history of violence or aggression and there is an acute onset of symptoms associated with homicidal violence. A family violence homicide event in such circumstances would be uncommon and atypical and for the purposes of classification, could be considered aberrant. In only one case of those considered for this analysis was the death event the first reported experience of acute psychosis.

Discussion

Within this investigation, we have drawn on the growing body of international literature that seeks to understand the relationship between serious mental disorder and family violence to critically review the FVDR SMEs’ classification of aberrant deaths. Feminist understandings of IPV have shaped the FVDR SMEs’ approach to understanding violence within families. This has directed the Committee to look to patterns of violence in the relationship. However, there is also a need to understand the pattern of violence in the life of the family. Where there exists a nexus in the use of violence and identified, sub-optimally treated mental disorder, there exists the potential to prevent future violence. The data presented in this study has highlighted that appropriate classification of death events requires an understanding of the constellation of risk factors as they exist within the lives of the perpetrator, the victim and their family. Focusing on only one risk factor (in this case serious mental disorder) may lead to inappropriate classifications and reduce the opportunities for intervention.

The FVDR SME classification system does not disaggregate by type of serious mental disorder. Similarly, many studies within the academic literature reviewed refer broadly to mental disorder and fail to disaggregate data by diagnosis. The schizophrenia spectrum/psychotic disorders are the most studied diagnostic group, followed by the major affective disorders, but these are not homogenous groups and cannot be assumed to carry the same risk.

Diagnoses alone limit our understanding of the relationship between mental disorder and family violence. The challenge is to identify symptoms within a diagnosis that confer a higher risk of violence, the likelihood of that risk being reduced if they are absent, and to then identify whether those symptoms are shared across diagnoses (26). Examination of specific symptoms or symptom clusters may provide a more nuanced understanding (43) and improved identification of risk factors for victimisation or perpetration (44). This is particularly important to reduce stigma for those people living with serious mental illness, the majority of whom remain at greater risk of victimisation rather than perpetration. Most perpetrators of family violence are not mentally disordered.

Implications for classification

The Merriam Webster dictionary identifies aberrations as ‘someone or something regarded as atypical and therefore able to be ignored or discounted’ (45). However, with consistent data collection, patterns emerge in apparent ‘aberrant’ family violence homicides. Those that may have once been considered ‘aberrant’ may now be illustrative of a different pattern of behaviour when the full complement of risk factors are understood.

In the case of family violence homicide committed in the context of serious mental disorder, the pattern of violence may reflect a relapse of psychotic or severe depressive symptoms, such as the presence of persecutory thinking or preoccupation with thoughts of death and suicide; paranoid delusional beliefs or delusional beliefs of guilt or blame; or command auditory or derogatory hallucinations. However, while present, these symptoms may be unrelated to the homicide, which may have been committed for the same reasons as those who do not have a serious mental disorder.

Science does not stand still, instead it is a ‘supremely social undertaking’ (46). When considered with the benefit of hindsight, there will be improvements available for classification definitions that lead to more nuanced approaches to prevention (46). The work of the FVDR SMEs is reflective of this. The classification of death events occurs at a classification day where, after Committee members have had the opportunity to classify deaths independently, those inconsistently classified are discussed and debated to ensure consensus. It is those that are potentially aberrant that result in the most discussion and debate. In part, this results from the need to take a reductionist approach to the complexity of the lives of individuals and their families which may not be adequately captured through accessed information. Equally, this results from an increased understanding of emergent patterns of violence and help-seeking in the lives of the victim and perpetrator.

Future work in this field could involve a careful analysis by psychiatric and psychological subject matter experts of those death events where serious mental disorder is a factor. Such analysis could include findings of insanity and unfitness to stand trial, identification of those psychiatric symptoms and psychological characteristics known to be associated with an increased risk of family violence and identification of any causative or contributory medical disorders.

The role of the FVDR SMEs is to understand factors leading to a homicide event to prevent such deaths occurring in the future. Terms such as ‘aberrant’ and ‘mental health homicide’ suggest an element of inevitability about the death event and the inability to prevent such homicides from occurring. They also provide a hierarchy of risk factors, where the serious mental disorder is considered at the expense of other risk factors that are evident. The classification of a death event as a ‘mental health homicide’ has the potential to be stigmatising and create fear, conflating mental disorder with dangerousness and obfuscating the role of other risk factors as well as support services.

In a review of approaches to reduce antisocial and aggressive behaviours amongst people with schizophrenia, Hodgins has noted that a combination of approaches is required – to address the illness, to identify ‘aggressive environments’ which heighten the likelihood of both victimisation and perpetration, and to address aggressive behaviours (39). Focusing on symptom specificity can reduce stigma by recognising that most people living with schizophrenia (for example), for whom risky symptoms may be either absent or well-controlled, are not violent. Given the strong association with childhood trauma and substance abuse, it is essential that these are also components of long-term interventions.

As has been highlighted in this article, the collation of information about apparently aberrant events over a significant period can result in an understanding of such events to the point where the term aberrant is no longer applicable. At over 20% of IFV death events, those associated with severe mental illness can no longer be considered atypical or aberrant events. Despite this, there will be a small number of cases where the event is truly aberrant – where it is atypical in the life of the offender. Although homicide is a rare complication of acute psychosis, it is disproportionately common in the first episode of a psychotic illness and strongly associated with when it occurs in the life course (42,47).

Going forward, it is important for the FVDR SMEs to note the nexus between particular symptoms present in some serious mental disorders and the use of violence. It is also important to note that the risk of violence, where present, is dynamic and multifactorial. Current violence risk assessment tools to predict the likelihood of community violence in psychiatric patients, have been shown to have little utility in individuals with schizophrenia and there is large variation in item content between instruments (48). Further research is necessary to determine whether the inclusion of alternative factors could improve risk assessment.

Conclusion

Homicide associated with a sudden onset of an acute first episode psychosis after a very short prodromal phase, often identified in retrospect, where associated psycho-bio-social risks were atypical in the life of the offender, are uncommon and in our present state of knowledge, are properly classified as aberrant.

Most homicides do not occur ‘out of the blue’, including those in which there is an association with symptoms of mental disorder. Many individuals admitted to forensic care for homicide have had inadequate mental health care leading up to this point. Effective responses to reduce and manage risk should encompass services that proactively address illness-related (e.g. earlier access and better maintenance in care) and criminogenic (e.g. substance use treatment, employment and psychosocial supports) domains (38).

Identifying the nexus between those symptoms present in some serious mental disorders and the use of violence enables recommendations to be drawn. To do so requires engagement with the psychiatric community to improve risk assessments, increase awareness of psycho-bio-social risks that weave themselves across the life course of families and individuals, and for preventative interventions and solutions to be identified. This would provide hope for families who are supporting people with severe mental disorders and highlights the importance of effective, comprehensive support in the community.

Funding Statement

The activities of the national mortality review programme within Aotearoa New Zealand are supported by the Health Quality & Safety Commission.

Ethical standards

Declaration of conflicts of interest

Jacqueline Short has declared no conflicts of interest.

Pauline Gulliver is an employee at the Health Quality & Safety Commission.

Fiona Cram has declared no conflicts of interest.

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.

CRediT author statement

Jackie Short: Conceptualisation, Methodology, Formal Analysis, Data Curation and Writing; Pauline Gulliver: Conceptualisation, Methodology, Formal Analysis, Data Curation and Writing; Fiona Cram: Conceptualisation, Methodology, Data Curation and Writing.