Could the Bondi Junction tragedy have been prevented? Lessons for forensic-informed community care

Australia’s mental health system needs urgent reform

The tragic event at the Bondi Junction Westfield shopping centre on Saturday, 13 April 2024, in which Joel Cauchi fatally stabbed six people and injured many others, has sparked intense scrutiny of Australia’s mental health system. The New South Wales (NSW) coronial inquest into the deaths identified major failings in the mental health systems in both NSW and Queensland, with experts calling them ‘broken’ and in urgent need of reform.

What is apparent is that this incident highlights a deeper, systemic issue in how Australia responds to mental health – particularly regarding those with serious mental illness who have contact with the criminal justice system.

What does the research tell us?

Psychosis is characterised by a distortion in thinking and perception of objective reality. Schizophrenia, one of the most common serious psychotic disorders, is characterised by at least 6 months of continuous psychotic symptoms, alongside the presence of negative symptoms (such as amotivation) and cognitive and functional impairment. Over the past decade, researchers at the University of NSW conducted a series of studies based on whole-of-population data of those diagnosed with psychosis which showed an association between psychosis and criminal conviction, including convictions for violent offences (Chowdhury et al., 2021). While this relationship and mechanisms involved are complex and multifactorial, the research confirmed that contact with mental health services significantly reduced the risk of re-offending in those with psychosis, and those receiving earlier and more frequent treatment following offending had the lowest risk of re-offending (Adily et al., 2023).

Inequitable access to mental health diversion

The rate of serious mental illness, including psychosis, is disproportionately higher in prison settings than in the non-incarcerated population worldwide. In response to this, jurisdictions across Australia have introduced court diversion programmes allowing courts, using powers vested in respective Mental Health Acts, to dismiss charges and divert eligible mentally ill defendants, who meet criteria for harm to self or others, away from the criminal justice system and into mental health treatment (MacDonald and Ellis, 2025). Those with serious mental illness diverted into treatment have lower rates of re-offending compared with those who receive a punitive sanction (Albalawi et al., 2019). Despite this, only one in four defendants diagnosed with a psychotic disorder in NSW was diverted, suggesting poor uptake of this option (Albalawi et al., 2019).

What the experts think an optimal model of care should look like

The above-mentioned research culminated in a Delphi study (a structured process to gather expert opinions and arrive at a group consensus) regarding an ‘optimal’ model of care for those with psychosis embroiled in the justice system post-arrest/detention. 25 experts were selected for their knowledge, insight, and experience in the mental health and criminal justice systems. Among them were psychologists, psychiatrists, academics, police, stakeholder organisations (prisons, prison medical services, and non-government organisations), policymakers, and consumers (Simpson et al., 2021).

Following several rounds of deliberation, four key principles were determined to underpin an optimal model of care: integration and coordination of multidisciplinary services, person-centred care, evidence-based practice, and government support – alongside 32 service attributes spanning diversion from detention, pre-release planning prior to release from detention, integration with state-wide community mental health services (including drug and alcohol services), and ongoing evaluation and follow-up. These themes and attributes embrace elements of the Assertive Community Treatment (ACT) approach, an intensive multidisciplinary case management model to aid community re-integration for those with severe mental illness (Simpson et al., 2021).

Both models emphasise a person-centred, trauma-informed approach to care, as well as multi-agency communication and collaboration. Distinct from the ACT model, however, the expert panel favoured an approach based on a specialised community health service team, independent of both government health and justice agencies, to deliver certain treatments and which equips individuals with access to external supports and services. Adaptations of the ACT model have been implemented overseas, including for forensic patients through, for example, the Forensic Assertive Community Treatment (FACT) model. This adaptation responds to the specific needs of patients with a history of offending, moving beyond a ‘one-size-fits-all’ approach (see Simpson et al., 2021).

Cauchi’s experience: a breakdown in continuity of care

Cauchi exhibited multiple criminogenic risk factors in the years prior to the Bondi Junction event (O’Sullivan, 2026). Criminogenic risk factors (e.g. a history of prior arrests and convictions, early age of onset of offending, prior supervision failures, antisocial personality traits, propensity to impulsive behaviour, antisocial justifications for offending behaviour, antisocial associates, substance use, poor educational attainment, unemployment, and poor engagement in follow-up) are strong predictors of violence than clinical symptomology alone (e.g. active symptoms of psychotic or mood disorders, affective instability and agitation/irritability, reduced inhibition, poor insight, and cognitive impairment) (Chowdhury et al., 2021). While Cauchi may not have been a typical ‘forensic patient’ by legal definition, in that he was never criminally convicted or incarcerated, he most certainly was when considering his longitudinal history and risk profile.

Indeed, Cauchi was well known to the Queensland police and mental health systems regarding concerns with ongoing psychotic symptoms, on a background of treatment-resistant schizophrenia (TRS). For Cauchi, this was characterised by a deterioration in functioning and self-care, the presence of intrusive thoughts and auditory hallucinations, psychotic logic behind his possession of sharp knives and interest in owning a gun, and threats to harm others.

In addition, Cauchi’s experience reflects a pattern familiar to those involved in the mental health sector: intermittent engagement with services and medication non-adherence secondary to a subjective sense of well-being (Adily et al., 2023). Perhaps most significantly, his experience reflects a breakdown in communication between service providers. This fragmentation between Cauchi’s family, the police and both public and private mental health services was heightened when he relocated from Queensland to NSW. Overall, Cauchi’s experience highlights a myriad of missed opportunities which underscore the need for a national, coordinated approach in the provision of mental health care for high-risk populations.

There is often a failure by general mental health services to consider a person’s longitudinal history and criminogenic factors. General mental health services often focus on cross-sectional assessments, without taking into consideration all information that is potentially available to them about the individual, because of the very high demands on providing services, the limited resources, and inadequate transfer and communication of historical information between various service providers and clinicians. General mental health services are often unaware of the kinds of risk factors to evaluate and manage, focusing predominantly on active current clinical symptoms and not on historical criminogenic factors. And rightly so, general mental health services do not regard themselves as forensic specialists. The focus tends to be on whether a person is at risk in the moment and meets the criteria for the necessary Mental Health Act, often overlooking the longitudinal and long-term risks.

What does this mean?

The care for mentally ill offender populations should be tailored to address the kinds of behaviours associated with offending and have a clear forensic mental health level of expertise, rather than legal status alone. Those with psychosis, including those with TRS, with high loading of risk factors for offending (i.e. criminogenic risk factors combined with complex clinical and social factors), would benefit from adaptations of the Delphi model (Chowdhury et al., 2021; Simpson et al., 2021), regardless of prior convictions.

Within the community, the lack of an assertive approach to manage people with mental illness who cause concern for future offending or recidivist offending can result in a repetitive cycle of police contact and symptomatic relapse without intervention to stop the cycle, as in Cauchi’s case.

As such, we propose that the optimal model of care outlined in the Delphi study report should not be restricted to post-prison populations. Rather, core components from the model could be expanded to inform a coordinated, specialised, and preventive ‘primary care’ model for non-forensic patients within the community who raise significant concern for offending, based on their longitudinal history, psychosocial circumstances, and clinical presentation.

The recommendations from the inquest into the deaths at Westfield Bondi Junction (5 February 2026) by NSW State Coroner Magistrate Teresa O’Sullivan echo the outcomes of the Delphi study. Both underscore that fragmented, agency-based responses are insufficient. Rather, governments must provide additional resourcing, clearer lines of responsibility, and greater forensic mental health awareness in the acute and long-term management of people with mental illness, including those with TRS, who raise concern for offending. Central to this is the capacity to divert individuals with high loading of risk factors for offending or who cause significant concern based on their historical behaviour, into a specialised, coordinated service or team established specifically to manage such complexity. Whether through independent multidisciplinary teams, dedicated mental health intervention coordinators or co-responder models, integrated mechanisms are essential to ensure continuity of care across police, health, and community services.

Had such a service been operational, Cauchi’s repeated contacts with police and mental health services may have functioned as a ‘forensic gateway’ to intensive rehabilitation years earlier. Whether the outcomes from the Delphi model and the coroner’s recommendation will be acted on remains to be seen.