While gambling has expanded rapidly globally, systems to monitor harms are underdeveloped. Operators have been shown to use predatory practices in pursuit of profits and maximise ‘revenue per available customer’ (revpac). The capacity of operators to exploit customers is further increased by digital technologies that track customer activity and push marketing that encourages gambling. This study was designed to consider ways to improve the identification of gambling suicides to inform prevention efforts.
We conducted 32 in-depth interviews in Australia. Participants were people who had survived a gambling-related suicide attempt, professionals working in death investigation, data custodians, researchers and lived experience advocates. Transcripts were analysed using reflexive thematic analysis and informed by our experiences reviewing Australian mortality datasets.
Identifying a gambling-related suicide relies on manual review of evidence that might spontaneously emerge during the death investigation process. Unlike drugs, alcohol or relationship difficulties, very few systems prompt death investigators or coders to capture whether gambling was a factor in a suicide. Social and cultural norms may discourage death investigators, clinicians and witnesses from considering gambling as a relevant factor in a suicide death. Sources of financial distress rarely required to be identified, compounding the underestimation of gambling harms. In cases where gambling is noted, relevant detail that could inform prevention efforts - such as the gambling product that was the main source of harm - is not consistently recorded.
Urgent efforts are needed to improve monitoring and surveillance of gambling suicides. Gambling operators hold transaction data that should be centrally accessible to regulators and death investigators to enable them to review account data in a centralized ‘data vault’. Account based gambling would provide people who gamble the ability to set loss limits. Yet vested interests often obstruct efforts by policy makers to introduce these measures.

Climate change is increasingly recognised as a major public health threat, with the physical health impacts of high temperatures well established. However, national evidence on temperature-related effects on mental health outcomes in Australia remains limited. This study examined the association of both higher and lower outdoor temperatures with mental health-related hospitalisations across Australia and estimated the heat-attributable health and hospital cost burden.
We used weekly Statistical Area Level 4 mental health-related hospitalisation data from AIHW, linked with population-weighted SA4-level weekly mean temperature. A two-stage time-series framework using distributed lag non-linear models was applied to estimate temperature-related risks and attributable burden; hospital costs were estimated by combining attributable numbers with 2022–2023 mental health hospitalisation episode benchmark costs.
Over 2006–2019, there were 8.6 million mental health-related hospitalisations in Australia. In the standardised national cumulative exposure–response curve, higher weekly mean temperatures were associated with increased mental health hospitalisation risk, while lower temperatures were not associated with increased risk relative to the national median reference temperature. Relative risks were 1.05 (95% CI: 1.03, 1.08) and 1.06 (95% CI: 1.02, 1.10) at the 90th and 99th percentiles of weekly mean temperature, respectively. Heat exposure above the 90th percentile was associated with approximately 2,438 heat-attributable mental health hospitalisations per year, corresponding to an estimated hospital cost burden of AUD 51.1 million annually. Higher heat-attributable mental health hospitalisation fractions were observed in the Northern Territory, northern Western Australia and inland Queensland, suggesting greater heat vulnerability in hotter and more remote settings.
These findings indicate that heat, rather than cold, is an important and geographically uneven contributor to mental health hospitalisations in Australia. Mental health should be more explicitly incorporated into climate adaptation planning, including heat-health warning systems, health service preparedness and place-based strategies for communities facing increasing heat exposure under climate change.

AI-supported mental health tools are moving rapidly from promise to deployment. An emerging model is digital twin-supported care: AI systems configured around a therapist’s approach to extend continuity of care between sessions through exercises, check-ins, progress tracking, and conversational support – without replacing the therapist. This model may address persistent challenges in psychological injury pathways, including treatment discontinuity, low homework engagement, clinician workload, and delayed escalation. Yet it also raises questions about safety, accountability, boundaries, privacy, trust, and governance.
This presentation draws on our research on compassionate AI and emotionally responsive chatbots to identify lessons for public health. Across this work, three lessons have become clear. First, AI systems that simulate care must be governed not only for technical accuracy, but also for emotional safety, relational boundaries, and human dignity. Second, AI support is most defensible when it augments rather than substitutes human care, keeping professional judgement, escalation, and accountability with humans. Third, implementation matters: the same system can produce different consequences depending on workflow integration, incentives, data governance, and institutional setting.
We apply these lessons to Ecko Health, an Australian platform that provides clinicians with a “Clinical Double” – a digital twin-inspired system that captures session insights, supports continuity between sessions, and operates under the therapist’s clinical governance. Rather than presenting Ecko as a proven clinical intervention, we examine it as a live implementation challenge: under what conditions can digital twin-supported care be feasibly, acceptably, and responsibly introduced into human-led psychological care?
The expected contribution is an implementation-oriented public health framework for responsible human–AI co-therapy. The presentation clarifies what must be known before scale-up: how clinicians and clients experience digital twin-supported care; how therapeutic continuity, homework engagement, workload, trust, and perceived safety are affected; and what governance arrangements are required before clinical outcomes or cost-effectiveness trials are justified.

Abstract
Background: Maternal depression, gestational diabetes mellitus (GDM), and hypertensive disorders of pregnancy (HDP) are each associated with increased risks of infections and allergic diseases in offspring. However, their combined effects, common in clinical practice, remain poorly characterised. We examined whether maternal depressive symptoms co-occurring with GDM or HDP are associated with increased risks of infections and allergic diseases in offspring.
Methods: In this population-based cohort study, we analysed linked administrative data from the Born in Queensland Study, including all births between 2015 and 2021. The primary exposure was maternal depressive symptoms comorbid with GDM or HDP. Depressive symptoms were assessed using the Edinburgh Postnatal Depression Scale (EPDS), and GDM and HDP were identified using ICD-10-AM codes. Outcomes were neonatal and early childhood infections and allergic diseases, also ascertained using ICD-10-AM codes. Adjusted risk ratios (aRRs) were estimated using generalised linear models with a binomial distribution and log link, adjusting for sociodemographic, maternal, and infant characteristics.
Results: A total of 149,025 mother–child pairs were included in the analysis. Offspring of mothers with comorbid depressive symptoms and HDP had a higher risk of neonatal infections (aRR 1·22, 95% CI 1·07–1·39), as did those exposed to comorbid depressive symptoms and GDM (1·30, 1·19–1·42). Risks of early childhood infections were also ele[GY1.1]vated for both comorbid exposures (HDP: 1·15, 1·10–1·21; GDM: 1·15, 1·12–1·19). For allergic diseases, comorbid depressive symptoms and GDM were associated with increased risk (1·20, 1·06–1·37), whereas the association with HDP was not statistically significant (1·17, 0·95–1·44). Effect estimates for comorbid exposures were consistently greater than those observed for each condition alone.
Conclusion: Maternal depressive symptoms co-occurring with GDM or HDP are associated with increased risks of infections and allergic diseases in offspring, with stronger effects than individual conditions alone. These findings highlight the importance of integrated antenatal care that simultaneously addresses maternal mental health and cardiometabolic conditions to improve early-life health outcomes.
Keywords: Antenatal depression; Gestational diabetes mellitus; Hypertensive disorders of pregnancy; Pregnancy comorbidity, Infections; Allergic diseases