Abstract
Background: Unhealthy diet is a major risk factor for CRC. This study assessed diet-related CRC burden from 1990 to 2021 in Organization for Economic Co-operation and Development (OECD) nations and estimated the burden until 2050
Methods: Data for OECD countries on diet-related CRC disability-adjusted life years (DALYs) and deaths were obtained from the Global Burden of Disease (GBD), 2021. The estimated annual percent change (EAPC) was calculated to analyse the CRC burden attributable to dietary factors. A generalised additive model with negative binomial distribution was used to predict the future burden of CRC attributable to dietary factors from 2021 to 2050.
Results: In 2021, the age-standardised percentage of diet-related CRC DALYs and deaths were 39.1% (95% uncertainty interval (UI): 9.3, 61.3) and 39.0% (95% UI: 9.7, 60.9), respectively, in the OECD countries. Between 1990 and 2021, the age-standardised DALYs decreased from 185 to 129 per 100,000, and deaths decreased from 8 to 6 per 100,000 population for OECD countries. Similarly, the EAPC of rates showed a downward trend (EAPCdeaths = -1.26, and EAPCDALYs = -1.20).
The estimated diet-related CRC DALYs and deaths are projected to increase to 4.1 million DALYs and 0.2 million deaths by 2050. There has been a downward trend in CRC deaths (EAPC = 1.33 for both sexes) and in DALYs (-0.90 for males and -1.0 for females) from 1990 to 2050.
Conclusion: Diet-related CRC burden remains significant. Implementation of nutrition intervention programmes is necessary to promote access to affordable and nutritious foods and to raise awareness about the importance of a healthy diet in reducing CRC risk.

Introduction: Digital health records have facilitated large-scale clinical studies to improve patient outcomes, but a large amount of health data remains underutilised in research due to its unstructured format. Unstructured data are often made up of free text in the form of clinicians’ notes, discharge summaries and medical imaging reports, which can be valuable in clinical and epidemiological research. One barrier to using unstructured data is the presence of high volumes of personal identifiable information present therein. However, this can be addressed using de-identification tools.
Methods: We scoped recent literature and tested several de-identification tools on clinical unstructured data, including samples of admission/discharge summaries, pathology reports, consultation notes and progress notes. We assessed the tools based on their precision, recall and F1 scores, along with several parameters assessing practicality.
Results: We tested Microsoft Presidio, Philter (beta) and National Library of Medicine Scrubber tool to de-identify a sample of unstructured health data. Out of the three tools, we found that Microsoft Presidio, when customised using patterns and context words relevant to the format of the data, had the best performance in de-identifying unstructured health data.
Conclusion: There is a small number of tools currently available that can de-identify personal identifiable information. These tools yield high precision but low recall, which reduces the integrity of the data. However, customisation of the tools, where available, can lead to a high level of de-identification. Incorporation of further customisation options relevant to medical records would make future de-identification tools more effective for use on unstructured health data.

Background: The Good Life with OsteoArthritis:Denmark (GLA:D®) program is an evidence-based education and exercise program for hip and knee osteoarthritis (OA). The program largely operates in private settings around the world. This study evaluated the implementation and delivery of GLA:D® in public outpatient settings in Tasmania, Australia.
Methods: A process and outcome evaluation was conducted to evaluate the implementation and delivery of the GLA:D® program within the Tasmanian Health Service. The RE-AIM QuEST Framework was used to assess Reach, Effectiveness, Adoption, Implementation, Maintenance, and Qualitative Evaluation Systematic Translation using health system, service-level metrics, patient-level data and fidelity. Semi-structured focus groups and one-on-one qualitative interviews were conducted with staff and patients who were involved in the program. Interviews were audio recorded, transcribed, coded and thematically analysed.
Results: Implementation and Adoption: The GLA:D® program was implemented at three public outpatient centres in Southern Tasmania between May 2022 and February 2024. Successful implementation factors included GLA:D® being a recognisable evidence-based program that was straightforward to deliver, THS executive support and buy-in, local physiotherapist champions who developed processes and promoted the program, and the establishment of an eReferral pathway which improved efficiency. Reach: 12 physiotherapists were trained in GLA:D®. 89 patients enrolled in the program (21 cohorts), of which 63 had complete GLA:D® registry data at baseline. Effectiveness: In interviews, most patients noted improvements in symptoms; however, some patients with co-morbidities or complex pain did not benefit. Staff and patients raised concerns about the accessibility and long length of the GLA:D® registry questionnaires. Maintenance. All three centres have elected to continue the program.
Conclusions: GLA:D® can be effectively, efficiently and sustainably implemented in public outpatient settings in Tasmania. Operational and administrative considerations have been identified and provide a foundation for broader implementation of the program in public settings around the world.

Background: Poor cardiovascular health (CVH) is a risk factor for current and future adverse health outcomes. Identifying CVH trajectories is essential for early intervention. However, the trajectories of composite CVH scores across the life course from childhood are poorly understood. This study modelled the trajectories of composite CVH scores from childhood to mid-adulthood and identified its childhood sociodemographic determinants.
Methods: Data were analysed from the Childhood Determinants of Adult Health study, which tracked participants from 1985 to 2019. CVH factors (physical activity, smoking, body mass index, sleep, and diet) were assessed in childhood (ages 7-15, n=1801) and at least once in young adulthood (ages 26-36 years, n=1415), adulthood (ages 31-41 years, n=666), or mid-adulthood (ages 36-50, n=923). Composite CVH scores were calculated according to the American Heart Association's Life's Essential 8 criteria. Group-based trajectory modeling was applied using the Stata traj command to identify CVH trajectory groups. Multinomial logistic regression was used to identify sociodemographic determinants of CVH trajectory groups.
Results: The mean follow-up duration was 29.1 (±4.8) years. Four discrete life course CVH trajectory groups were identified: declining from moderate to low (n=143, 9.5%), declining from high moderate to moderate (n=660, 34.8%), improving from moderate to high (n=156, 12%), and consistently high (n=842, 43.8%). Males, older baseline age, and low academic attainment were associated with a higher likelihood of CVH decline, while higher school engagement index related to a lower chance of CVH decline. School-only parental education and poor self-rated health were associated with both an increased chance of declining and a higher likelihood of improving CVH, suggesting complex, bidirectional associations with CVH changes over time.
Conclusion: While over half of the participants had moderate CVH in childhood, only 12% improved to ideal CVH in adulthood. Sociodemographic variables play a significant role in both worsening and improving CVH trajectories.

Background: Pharmaceutical claims data are often used as the primary information source to define medicine exposure periods in pharmacoepidemiological studies. However, often critical information on directions for use and the intended duration of medicine supply are not available. In the absence of this information, alternative approaches are needed to support the assignment of exposure periods.

Methods: This study summarises the key methods commonly used to estimate medicine exposure periods and dose from pharmaceutical claims data; and describes a method using individualised dispensing patterns (IDP) to define time-dependent estimates of medicine exposure and dose.

Results: The IDP method extends on important features of existing methods and also accounts for recent changes in an individual's medicine use. Specifically, this method constructs medicine exposure periods and estimates the dose used by considering characteristics from an individual's prior dispensings, accounting for the time between prior dispensings and the amount supplied at prior dispensings. Guidance on the practical applications of this method is also provided. Although developed primarily for application to databases, which do not contain duration of supply or dose information, use of this method may also facilitate investigations when such information is available and there is a need to consider individualised and/or changing dosing regimens.

Conclusion: By shifting the reliance on prescribed duration and dose to determine exposure and dose estimates, individualised dispensing information is used to estimate patterns of exposure and dose for an individual. Reflecting real-world individualised use of medicines with complex and variable dosing regimens, this method offers a pragmatic approach that can be applied to all medicine classes.

Background: The increased availability of data through cohort studies and large population-linked datasets has improved the capacity of epidemiologists to examine complex questions related to health equity for marginalised populations. Although data science methodologies enable reproducible and complex quantitative analyses for epidemiological projects, using these approaches without sufficient contextualisation risks erroneous findings or interpretation, resulting in ineffective or potentially damaging public health recommendations that do not meet the needs of the population. Incorporating lived experience into the research process is one way to improve the quality and relevance of the research and its findings. Here we present a case study of a secondary data analysis informed by lived experience expertise. Methods: This research used the Household, Income and Labour Dynamics in Australia (HILDA) Survey, a population-based national longitudinal study established in 2001 of 7682 Australian households, to investigate the associations between loneliness, social isolation, and mental health in young people (aged 15-24 years) with disability. Four lived experience research assistants were employed to contribute experiential expertise as young people with disability. Results: We incorporated lived experience into most steps of the research process, including refining the research questions and analytic strategy, choosing relevant confounders and mediators by refining a DAG, interpreting and contextualising the results, and determining the next steps for research and policy recommendations. Lived experience brought in critical expertise that altered the project aims, how we defined and selected variables and commonly used indices, and interpreted the research findings, making the research more focussed on the needs of young people with disability and potentially reducing the inherent statistical biases that large observational datasets are prone to. Conclusion: Our case study demonstrates a new way for undertaking health equity research when using secondary data that promotes the voice of the population of interest.

Background: Many studies in the field of physical activity epidemiology are observational. Failing to adequately consider and account for confounders in observational studies can lead to biased estimates due to uncontrolled confounding and/or collider bias, and may result in misleading or imprecise estimates of the effect of physical activity-related exposures (such as physical activity, sedentary behaviour and muscle-strengthening exercises) on health outcomes. There are many different confounder selection approaches and it is considered best practice, and recommended by reporting guidelines (such as STROBE), that the choice of confounders is described and justified. We aimed to describe current practices in confounder selection in studies using observational data to estimate the effect of physical activity-related exposures on health outcomes.

Methods: A structured search strategy was used to identify studies estimating the effect of physical activity-related exposure on a health-related outcome using individual-level observational data, published in 2022-23. A total of 1353 articles were identified in the screening process, 300 of which were randomly selected for full-text screening and data extraction. Data extracted included the approach(es) used to select confounders, and study characteristics.

Results: Of the 300 articles, 279 were eligible. Most (61%) did not describe their approach to choosing confounders. Around 27% of studies used a confounder selection approach based on prior knowledge or theory, and a further 3% used a directed acyclic graph. The remaining studies used a data-driven approach.

Conclusion: Most published studies aiming to estimate the effect of physical activity-related exposures on health outcomes do not report the method they used to choose confounders. This goes against reporting guidelines, and negatively impacts transparency and reproducibility. These results suggest more training and awareness is required among people conducting observational studies in physical activity epidemiology, and/or that greater adherence to reporting guidelines is required by study authors, reviewers and journal editors.

Background: Over the past 10 years there has been increasing recognition of the importance of respectful maternity care as an essential component of good-quality care, including in limited-resource settings. There is currently no consensus approach on the best way to measure women’s experience of care at birth and limited data on women’s experiences of care from the Asia-Pacific region.

Methods: The Global Maternal and Newborn Health Platform (GMP) is a multi-country, prospective, facility-based, observational study. We implemented GMP in 10 Asia-Pacific countries; using multi-stage sampling, we recruited 74 facilities (>=2000 births per year) to collect data on clinical practices and birth outcomes for 3 to 6 months. At each facility, 400 women are randomly sampled and invited to complete an audio computer-assisted self-interview (ACASI) via digital tablet and headphones that asks about their experiences of care. With ACASI, multiple-choice questions are presented using text and pictures, with the questions and answer options spoken in the woman’s language of choice.

Results: Thus far, 31,526 women were eligible for the ACASI survey, of which we have data from 23,975 women (76%). In total, 14% of women described some form of mistreatment during hospital admission (such as physical, verbal abuse or stigmatisation); 13% stated that health workers did not obtain their permission prior to vaginal examinations. Most women (75%) who experienced labour wanted a birth companion, though amongst this group, 26% were unable to have one.

Conclusions: We found a high prevalence of mistreatment and examinations being conducted without consent. We also demonstrated a considerable unmet need for companionship during labour. These findings highlight that further work is needed to deliver the goal of providing respectful maternity care to all women. The ACASI technology is flexible and enables women across languages and literacy levels to confidentially report on their experiences of maternity care.

Background: Perinatal pharmacoepidemiology is often limited by power due to low prevalence of exposure to medicines during pregnancy and rare outcomes. To overcome this, we aimed to transform an existing Australian linked administrative dataset into the Nordic Pregnancy Drug Safety Studies (NorPreSS) common data model (CDM) to facilitate distributed analyses of pregnancy and childhood outcomes following medicine exposure using common analytical code.
Methods: We linked health, social and education data from the Australian Early Life Course Platform, a population-based cohort of 1.7 million children born between 2001 and 2019 in NSW. We then mapped common elements from the Australian Perinatal Data Collection (PDC), the Pharmaceutical Benefits Scheme (PBS) and from the National Assessment Program – Literacy and Numeracy (NAPLAN) onto equivalent elements in the NorPreSS CDM. Mapping to several event datasets that identify clinical events such as hospitalisations, procedures and death is still ongoing.
Results: Of the information from the NSW Perinatal Data Collection, 34 (81%) items could be mapped directly to the pregnancy and 19 (86%) to the child study population datasets. Most information (6 out of 11 items) missing from PDC, regard the health of the pregnant mother and exist in other datasets such as hospitalisations and Medicare billing records which are yet to be mapped. Australian PBS medicine use and child academic performance (NAPLAN) directly mapped to elements of the NorPreSS CDM.
Conclusion: Previous collaboration with Nordic countries relied on a common protocol and involved significant work by country-based analysts to a run a comparable analysis in their respective datasets. By harmonising NSW linked data into the NorPreSS common data model, we anticipate a more streamlined analytical approach to facilitate international collaboration. Similar data holdings in different states in Australia and potentially other countries could benefit from this approach.

Background: While the immediate physical health consequences of hypertensive disorders during pregnancy (HDP) are well-documented, their potential impact on children's mental health and cognitive outcomes remains relatively under-investigated. This study aimed to examine the associations between HDP and intelligence quotient (IQ) in children at the ages of 8 and 16 years.

Methods: Our study sample comprised participants in the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort, an ongoing population-based longitudinal birth cohort in Bristol, United Kingdom. Children’s IQ was measured using the Wechsler Intelligence Scale for Children (WISC-III). This study included over 4900 and 3300 mother-child pairs at ages 8 and 16, respectively. Binary and multinomial logistic regression analyses were used to estimate odds ratios for the associations.

Results: HDP (gestational hypertension and/or pre-eclampsia) were not found to be associated with lower IQ scores in children at ages 8 and 16. In the multinomial logistic model, we found children born to mothers with gestational hypertension but not pre-eclampsia were more likely to have above-average IQ at age 16 compared to average IQ children born to mothers without gestational hypertension (OR = 1.42; 95% CI: 1.03 – 1.94). However, this association did not persist when children with below-average IQ were used as the reference category and no such associations were also observed at the age of 8 years.

Conclusions: Our findings revealed no evidence of associations between HDP and lower IQ scores in children ages 8 and 16. The observed association between gestational hypertension and higher odds of having an above-average IQ at age 16 warrants further investigation.

Background: The association between paternal smoking and attention deficit hyperactivity disorder (ADHD) in offspring is not well established. We summarised the available evidence to determine the effect of paternal smoking on the risk of ADHD in children.
Methods: This systematic review and meta-analysis was performed following the Preferred Reporting Items for Systematic Review and Meta-analysis guidelines. The review protocol was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO). Pooled odds ratios with 95% confidence intervals (CIs) were calculated using a random-effects meta-analysis. The heterogeneity among studies was assessed using the I2 test. Publication bias and the presence of small study effects were evaluated using funnel plots and Egger's test. Sensitivity and subgroup analyses were also performed.
Results: Twenty observational studies involving over 294,236 study participants from 16 different countries were included. We found that paternal smoking was associated with a 22% increased risk of ADHD in children (RR=1.22, 95% CI: 1.12, 1.33). Our subgroup analysis revealed that this association was not evident among studies that accounted for maternal smoking (OR=1.23, 95% CI: 1.10, 1.38) and paternal alcohol use (OR=1.02, 95% CI: 0.79, 1.33) and education (OR=1.14, 95% CI: 0.98, 1.34).
Conclusion: The observed association between paternal smoking and the increased risk of ADHD in children may be confounded by maternal smoking, paternal education and other confounders. Future studies should focus on maternal and paternal comparisons to disentangle the independent and combined effects of parental smoking on ADHD risk in children while also better controlling for potential confounders.

Background. In anatomy higher education, face-to-face learning is preferred to online learning, however its benefits have not been examined under a causal structure. This study aimed to estimate the causal effect of face-to-face learning, compared to online learning, on student performance in anatomy by analysing examination marks under a causal structure.
Methods. We specified a causal graph to indicate how the mode of learning affected student performance. We purposively sampled observational data to obtain examination marks of undergraduate and postgraduate students who learned using face-to-face (pre-COVID, year 2019) or online modes (post-COVID, year 2020). Under the causal graph, the total effect of the mode of learning on student performance can be assessed by comparing the difference in the end-semester marks between cohorts that learned by face-to-face versus online modes. Marks were compared using linear regression, and sensitivity analyses were performed to assess if causal effects were robust to unmeasured confounding.
Results. On average, compared to online learning, face-to-face learning improved undergraduate student performance in the end-semester examination (gain of mean 8.3%, 95% CI 3.3 to 13.4%; E-value 2.77, lower limit of 95% CI 1.80) but lowered performance in postgraduate students (loss of 8.1%, 95% CI 3.6 to 12.6%; E-value 2.89, lower limit of 95% CI 1.88).
Conclusion. Under the causal graph, we found that compared to online learning, face-to-face learning improved student performance in the end-semester examination in undergraduate students, but lowered student performance in postgraduate students. These findings suggest that different modes of learning may suit different types of students. Importantly, this is the first attempt to investigate causal effects of the mode of learning on student performance in higher education under a causal structure. This approach makes investigator assumptions transparent, informs data analysis, and is recommended when using observational data to make causal inferences.

Abstract
Background
Stunting still represents a major public health problem in Egypt, impacting around one-third of children under five and leading to a 2-3% drop in Gross domestic product (GDP). Therefore, this study examined the trends in the prevalence of stunting and its associated factors among children aged 0-23 months, 24-59 months and 0-59 months in Egypt.

Methods
The study used combined data from the 2005, 2008, and 2014 Egypt Demographic and Health Surveys (EDHS), with a total sample of 39,857 children aged 0-59 months. A multilevel logistic regression that adjusts for sampling weights and clustering was performed to analyse the factors associated with stunting among children in Egypt.

Results
The prevalence of stunting was 24.8% in children aged 0-23 months, 24.3% in children aged 24-59 months, and 24.5 % in children aged 0-59 months. After controlling for potential confounders, the common factors associated with stunting in three age groups were mothers with primary education and mothers who did not have an antenatal care (ANC) visit during pregnancy. Maternal short stature (height < 155 cm) was associated with an increased risk of stunting in three age groups compared to mothers taller than 160 cm.

Conclusion
Addressing the causes of stunting in Egypt necessitates sectoral collaboration, including health, water, sanitation and hygiene, education, and social protection, with the implementation of targeted interventions focusing on children born to short mothers. Interventions must be accessible for families with low socioeconomic status, focusing on improved healthcare access, parent education, and infant and young child feeding practices. A systematic nutrition monitoring framework that enables routine data collection on nutritional outcomes to monitor and improve child health is recommended.

Keywords
Undernutrition, childhood, 0-59 months, Growth defect, North Africa

Background
Despite significant gains in gender equality in Australia, inequalities persist. Examining the way that gender equality varies geographically across Australia, and the way this is related to health outcomes is important in identifying and supporting locally implemented strategies to improve gender equality and population health. In this analysis, we aimed to examine spatial variations in gender equality across Australia and its intersections with some key determinants of health.

Methods
We used the newly created Australian Gender Equality Index (AGEI), a composite measure of 12 dimensions of structural gender equality operationalised at the Statistical Area-3 (SA3) level across Australia. All analyses were conducted in R. The AGEI was quintiled then mapped using ggplot2. The Viridis colour scale was applied to enable visualisation of variations in gender equality. Associations between gender equality and area level socioeconomic position, cultural and linguistic diversity, and family composition were examined using bivariate maps.

Results
Spatial analysis showed consistent gender inequality across Australia. Area-level gender inequality was linked to measures of advantage, with greater area advantage associated with higher gender inequality. Gender inequality was also more pronounced in areas with a high proportion of families with children. Across major cities, spatial variation in gender inequality was associated with factors such as household structure, area variations in age distribution, and wealth.

Discussion
While gender inequality was consistently apparent across Australia, there was heterogeneity in its patterning. Some variation in the spatial patterning of gender inequality across cities appears to be related to some key social determinants of health, and future analysis will examine the way that the patterning of gender inequality is related to health outcomes for different population groups across Australia.

Funding: This study was supported by an Australian Research Council Linkage Grant (LP180100035) and an Australian Research Council Discovery Project (DP250101371).

Background: Medicine dispensing claims, including Australian Pharmaceutical Benefits Scheme (PBS) data, are used to generate real world evidence about the use, utilisation, effectiveness and safety of medicines. Critical to these studies is defining medicine exposure and a necessary first step is estimating dispensing intervals, i.e. how long a person is considered to be on treatment after the medicine is dispensed.
Methods: The parametric waiting time distribution (WTD) enables a data-driven approach yielding unbiased estimates of dispensing intervals. It is a frequency distribution of the time to the last dispensing for each person within a fixed time window.
Results: We have developed a freely available R package implementing WTD methods including approaches to address variability between and within people, stockpiling/seasonality, and starting and stopping treatment. We have applied WTD analysis to PBS data and demonstrated some advantages over other approaches including a simple method to incorporate pack size.
Conclusion: Availability of WTD methods in an open-source R package enables a variety of applications in estimating dispensing intervals; a tutorial is forthcoming. The method is mature with over a dozen published papers describing refinements and applications.

Background
Diabetes in pregnancy increases the risk of adverse outcomes for mothers and babies. More than 30 years ago, the St Vincent Declaration set targets to achieve similar pregnancy outcomes between women with and without diabetes. It is unclear whether targets are being met in Australia. This study compared trends in adverse outcomes among women with and without diabetes.

Methods
Population-based data linkage study including all women with a singleton birth ≥20 weeks gestation in New South Wales (NSW) during July 2006 to December 2019. Exposure groups: Type 1 (T1DM), Type 2 (T2DM), Gestational (GDM) and no diabetes. Outcomes: stillbirth, preterm birth, small- and large-for-gestational age (SGA/LGA) and planned birth. Comparisons were made using Modified Poisson regression models adjusted for demographic and pregnancy factors. Analyses for GDM were restricted to women with births ≥30 weeks.

Results
Overall, women with T1DM and T2DM had increased risk of stillbirth, preterm birth and LGA but decreased risk of SGA. There were increasing trends over time in the risk of preterm birth (aRR₂₀₀₆-₂₀₀₉ 5.29, 4.76-5.87 versus aRR₂₀₁₅-₂₀₁₉ 6.55, 6.12-7.01) and LGA (aRR₂₀₀₆-₂₀₀₉ 3.30, 3.09-3.53 versus aRR₂₀₁₅-₂₀₁₉ 4.25, 4.08-4.44) but decreasing risk of planned birth (aRR₂₀₀₆-₂₀₀₉ 1.88, 1.81-1.95 versus aRR₂₀₁₅-₂₀₁₉ 1.65, 1.58-1.71) for women with T1DM. There was an increasing trend in the risk of planned birth (aRR₂₀₀₆-₂₀₀₉ 1.43, 1.34-1.52 versus aRR₂₀₁₀-₂₀₁₄ 1.59, 1.54-1.64) for women with T2DM. Overall, women with GDM had increased risk of preterm birth, LGA and planned birth. There were decreases over time in the risk of stillbirth (aRR₂₀₁₀-₂₀₁₄ 1.37, 1.14-1.66 versus aRR₂₀₁₅-₂₀₁₉ 0.91, 0.75-1.10) and planned birth (aRR₂₀₁₀-₂₀₁₄ 1.46, 1.45-1.47 versus aRR₂₀₁₅-₂₀₁₉ 1.37, 1.36-1.37).

Conclusion
Women with diabetes in pregnancy in New South Wales remain at increased risk of adverse outcomes compared to those without diabetes. However, there is evidence of decreasing stillbirth risk for women with gestational diabetes.

Background
In Australia, the prevalence of gestational diabetes (GDM) and obesity have been increasing. There is uncertainty about whether GDM increases the risk of stillbirth, especially with increased adoption of ‘stricter’ diagnostic criteria. This study examined the risk of stillbirth and neonatal death among women with GDM, with and without obesity.

Methods
Population-based data linkage study involving all women with a singleton pregnancy ≥30 weeks gestation in New South Wales, Australia during 2016-2020. Linked perinatal, hospital, mortality and pathology data were used. Women with pre-existing diabetes were excluded. Study outcomes of stillbirth and neonatal death were examined by GDM management (diet, insulin, oral hypoglycaemics) and body mass index (BMI) categories. Outcomes were also examined by gestational age groups (30-31, 32-33, 34-36, 37-38, 39-41, ≥42 weeks). Cox Proportional hazard models were used to estimate adjusted hazard ratios for GDM relative to no diabetes groups.

Results
Among 426,878 women, 66,113 (16%) were diagnosed with GDM [management: diet 52%, insulin 34%, oral hypoglycaemics alone 6%] and 19,233 (29%) had BMI ≥30kg/m². Stillbirth and neonatal death rates were 1.9/1000 births and 0.86/1000 livebirths (GDM), 2.0/1000 and 1.09/1000 (GDM with obesity), 2.2/1000 and 1.02/1000 (no diabetes with obesity) and 2.0/1000 and 0.76/1000 (no diabetes). Compared to no diabetes, we found no evidence of increased risk of perinatal death for GDM (stillbirth: aHR 0.99, 95%CI 0.81-1.21, neonatal: aHR 1.21, 95%CI 0.89-1.64) or GDM with obesity (stillbirth: aHR 1.02, 95%CI 0.70-1.47, neonatal: aHR 1.21, 95%CI 0.73-2.02). There was increased risk of neonatal death at full term for GDM with obesity (aHR range 2.44-5.58].

Conclusion
There was no increased risk of perinatal mortality for women with GDM alone but increased risk of neonatal death among full term infants for women with combined GDM and obesity.

Introduction
Oral health literacy is an important aspect of understanding oral health as it has an impact in attaining a healthy lifestyle. It helps in identifying ways to access oral health services and improving oral hygiene habits. This study aimed to measure the oral health literacy level of children and find the factors influencing it.
Methods
A cross-sectional survey was conducted from May- June 2024 in four Islamic boarding schools in Lahore, Pakistan. The children, aged 12-17 years, were able to read and write in the Urdu language, and were asked to respond to Health Literacy in Dentistry short-form (HeLD-14) to estimate their oral health literacy. Frequency distribution, mean-standard deviation, and linear regression were used to analyse the data.
Results
A total of 534 adolescents participated in the study with a mean age of 15.37±1.63 and boys 269 (50.4%) were almost evenly distributed against the girls 265 (49.6%). Cumulative oral health literacy score was calculated as 17.55 ± 5.14 mean and standard deviation (range 0 - 28). Visiting the dentist in last 12 months [Adjusted Odds Ratio 1.89 (95% Confidence Intervals 0.89–2.89)], higher age-group of 15-17 years [1.29 (0.30–2.29)], regular brushing [2.2 (0.09–4.31)], being able to pay a dentist [3.47 (2.46–4.48)], and higher sugar intake [1.86 (0.98–2.75)] were strongly associated with the oral health literacy score of the children.
Conclusion
These results suggest that a significant number of children do not fall in the category of having high level of oral health literacy. Health promotion campaigns at school level are needed to educate children at an early age about the importance of oral health, teaching them about the ways to seek dental treatment, and improving oral hygiene and eating habits among them.

Introduction
Oral diseases are common worldwide, particularly among children, who should attend regular checkups and avoid negative oral habits such as tobacco use, to prevent oral diseases at early age. Raising oral health awareness and improving oral habits can improve the oral health amongst children.
Methods
A cross-sectional survey was conducted from May- June 2024 in four Islamic boarding schools in Lahore, Pakistan. The children, aged 12-17 years, were able to read and write in the Urdu language, and were asked to respond to a self-reported questionnaire about their oral health and use of different tobacco products. Frequency distribution, chi-square and binary logistic regression were performed to analyse the data.
Results
Out of 534 adolescents who participated in the study, 155 (29%) reported that they have used tobacco products. Among them, 109 (70.3%) had used one product and 46 (26.7%) had consumed 2-4 products. Betel leaf (n 45) and betel nuts (n 143) were the most used tobacco products. The number of boys (n 105, 67.3%) consuming tobacco products was more than twice higher than girls (n 51, 32.6%). Teeth health [Adjusted Odds Ratio 0.64 (95% Confidence Intervals 0.41-1.00)], higher prevalence of toothaches [0.60 (0.37–0.96)] and healthy gums [0.68 (0.43–1.07)] were significantly associated with the use of tobacco products.
Conclusion
These results suggest that a high number of children are exposed to tobacco products, and it can become a burgeoning crisis if neglected by health authorities. Health promotion and oral health awareness campaigns at school level are needed to educate children at an early age about the importance of oral health and the adverse effects of tobacco products.

Introduction
Oral health-related quality of life OHRQoL is a patient-centred approach in assessing perceived physical / mental health. The aim of this study is to assess the association between OHRQoL and different aspects of self-reported oral health status and oral health-seeking behaviour.
Methods
A cross-sectional survey was conducted from May- June 2024 in four Islamic boarding schools in Lahore, Pakistan. The children, aged 12-17 years, were able to read and write in the Urdu language, and were asked to respond to Child Oral Health Impact Profile (COHIP) questionnaire to assess their Oral Health-Related Quality of Life OHRQoL. Frequency distribution, mean-standard deviation and linear regression were used to analyse the data.
Results
The questionnaire was answered by 534 participants. The mean ± SD scores were calculated as OHRQoL 99.4 ±23.44, oral health and wellbeing 24.78±7.08, functional wellbeing 19.32±4.76, social and emotional wellbeing 24.64±7.77, school environment 14.32±2.61, self-image 14.46±5.91, treatment expectancies 8.51±2.30 and overall health 2.64±1.17. Healthy teeth [Adjusted Odds Ratio 0.12 (95% Confidence Intervals 0.10–0.14)], healthy gums [0.11 (0.09–0.13)], no gum bleeding [0.09 (0.07–0.11)], no toothache in last 12 months [0.08 (0.06–0.11)], planning to seek dental treatment [0.06 (0.04–0.08)], and girls [0.19 (0.02–0.07)] were strongly associated with OHRQoL score.
Conclusion
The results suggest that OHRQoL can impart a strong effect on a children’s oral health and health-seeking behaviour. If the children are visiting the dentist for regular checkups and have better oral health, then their overall quality of life will be improved as well. To achieve that parents, teachers and government should play their role in providing education and easy access to dental services to these children.

Background

Culturally and linguistically diverse (CALD) populations experience persistent health disparities, yet epidemiological studies often rely on inconsistent and overly simplistic definitions of CALD status. These methodological choices are not neutral, they reflect broader assumptions about identity, belonging, migration, and inequities. Additionally, measurement bias can arise when study variables like CALD status are inaccurately measured or classified, potentially skewing results and interpretations. This study examines the impact of different CALD definitions and analytical approaches across three datasets to assess how these choices shape epidemiological findings and their implications for public health.

Methods
Three population-based datasets were analysed, each examining different health-related outcomes. CALD status was operationalised using multiple thresholds: (1) born overseas, (2) born in a non-English-speaking country, (3) region of birth, (4) born in a racially marginalised country, and (5) ancestry (where available). CALD was incorporated into epidemiological models in four ways: (1) as a descriptive variable, (2) as a stratification factor, (3) as a confounder, and (4) as an exposure variable. The impact of these different approaches was assessed across the three datasets.

Results
Findings show that estimates of health disparities varied depending on how CALD status was defined and analysed. Broad definitions resulted in smaller observed disparities, while more specific operationalisations identified greater differences in health outcomes.

Descriptive analysis masked subgroup variations, while stratification exposed disparities within CALD groups. Confounder adjustments altered key associations, and exposure modelling showed CALD as an independent risk factor, with stronger effects using precise definitions.

Conclusion
The operationalisation and analysis of CALD status significantly influences epidemiological estimates and reflect ideological positions that shape how health disparities are framed and understood. The findings advocate for a reflexive epidemiology that interrogates the assumptions in methodological choices, ensuring research on health disparities exposes rather than obscures public health inequities.

Background
Interpersonal trust- the belief that other people can be trusted is a key but often overlooked factor in public health compliance. While previous research has highlighted the role of institutional trust, less is known about how interpersonal trust influenced compliance to public health orders during the pandemic and how this has changed in the post pandemic period. This study examines the relationship between interpersonal trust and compliance with public health measures and assesses longitudinal trends in interpersonal trust over the past decade.

Methods
This study conducted secondary analysis of the Household, Income and Labour dynamics in Australia (HILDA) survey, an annual household based panel study. The analysis used 4 waves of data (pre pandemic, year 1 and year 2 of pandemic and post pandemic) with 15, 450 adult respondents. The data were analysed using dynamic structural equation modelling.

Results
The SEM analysis suggests that interpersonal trust, age, ethnicity, area diversity and income explained 21% of the variance in Covid-19 compliance behaviours. Higher interpersonal trust was significantly associated with some COVID-19 behaviours (vaccination and wearing a face mask) but not others (staying at home and social distancing). However, despite remaining stable for a decade, interpersonal trust declined significantly in the post pandemic period.

Conclusion
Interpersonal trust influence some but not all COVID-19 compliance behaviours, suggesting that its role in public health varies by intervention type. The post pandemic decline in trust poses challenges for future crisis response efforts, emphasising the need for strategies that rebuild trust and strengthen social cohesion at the community level.

Background
Global pandemics and seasonal epidemics have significant economic and health impacts on the aviation industry, travelling public, and to public health bodies from the translocation of disease. Despite the impact, the aviation industry currently lacks evidence-based knowledge of the mechanistic dynamics of infectious disease transmission in air travel and how multiple control measures fit together to prevent disease transmission outside of specific individual events. Here we develop a systematic and holistic framework approach to examine disease transmission risk that is transferable to multiple pathogens and settings.
Methods
The framework provides a method to integrate a number of empirically driven mathematical models, including screening and testing, computational fluid dynamics, pathogen transmission; agent-based human movement and behaviour; and epidemiological models within the operational environment of commercial aircraft and airports. The framework defines the functionality required in each sub-model, and the required interfaces between models for data exchange. The integrated framework is then run using Monte Carlo methods to evaluate transmission risk and to optimise control strategies for a given pathogen.
Results
We show that the framework can be used to simulate a variety of infectious agents and characterise control measures used in risk reduction and mitigation. The framework is able to be run in a systematic way that allows data to be transferred between the different sub-models and for each sub-model to be run independently and at the numerical fidelity required.
Conclusion
This framework was developed in response to the U.S. Department of Transportation National Aviation Preparedness Plan. It provides a quantitative evidence-based disease transmission risk framework and suite of modelling tools needed to improve understanding of disease transmission throughout the global air travel ecosystem and to develop a risk management tool that is scalable to changing complex environments and adaptable to multiple infectious pathogens.

Background
The transmission of pathogens from surfaces is a significant concern for many industries, healthcare, and transportation settings, further highlighted during the COVID-19 pandemic. Surface bound antimicrobials play an important role in mitigating disease spread from fomites, however they can be slow acting, only affect specific organisms, require reapplication, and degrade the material they are applied on. Here we present a novel persistent antimicrobial coating that targets a wide range of pathogens and can be tailored to specific agents. Further, the polymer has unique capability to respond to environmental cues such as temperature, fluids and pH for specific anti-microbial targeting. We discuss its effectiveness and use in transportation and other settings.
Methods
Here we report on the antimicrobial efficacy from several studies that include live bacterial and viral pathogen testing trails on a variety of high-touch surface materials, and a six-month flight efficacy and durability test on board the international space stations.
Results
Our study demonstrates the efficacy of a novel water-borne polymer coating in inactivating SARS-CoV-2 variants, Influenza A and several bacterial species. We further show that the antimicrobial remains persistent on multiple surface types when being physically touched and in unique environments over extended periods without the need for reapplication.
Conclusion
The development of this polymer coating represents a fast and effective means at controlling the transmission fomite transmitted pathogens. By effectively inactivating microbial particles on high-touch surfaces, this technology could play a crucial role in reducing the spread of current and emerging infections disease threats. Its ease of application, fast mode of action, persistence, and human safety make it uniquely suited to spaces that see high human traffic and areas vulnerable to greater disease transmission risk and on personal protective equipment.

Background
A comprehensive assessment of the morbidity and mortality burden of cancer and their evolving trends is crucial for making informed policy decisions and effective allocating resources. We aimed to examine the burden of cancer in Taiwan from 2000 to 2021 using a national population-based database.
Methods
Linking data from the Taiwan Cancer Registry and National Death Registry, we calculated years lived with disability (YLDs), years of life lost (YLLs), and disability-adjusted life years (DALYs) for 30 specific cancer groups. Our methodology aligns with the Global Burden of Disease Study.
Results
From 2000 to 2021, the age-standardized cancer mortality rate decreased by 15.8%, while the morbidity rate increased by 76.4%. In 2021, the age-standardized DALYs for total cancer were 3317.6 per 100,000 population. For men, lung, liver, and colorectal cancers were the leading contributors to cancer-related DALYs, whereas for women, breast, lung, and colorectal cancers were predominant. Decomposition analysis revealed distinct life expectancy patterns by sex, with significant gains for specific cancers from 2000 to 2021: cervical, stomach, and liver cancers in women (0.20, 0.13, and 0.12 years, respectively) and liver, lung, and stomach cancers in men (0.37, 0.17, and 0.16 years, respectively).
Conclusion
Our study provides a comprehensive overview of the intricate landscape of cancer burden, mortality, and their effects on life expectancy over time in Taiwan. These findings offer critical insights for policy making, resource allocation, and future research.

Background
Evaluating long-term health outcomes and the economic burden of smoking is critical for assessing the effectiveness of tobacco control policies. Our aim was to estimate the lifetime productivity loss attributable to smoking in adults in Taiwan.
Methods
We analyzed data from the MJ cohort, linking medical and demographic information with national insurance and vital records to estimate participants' monthly earnings and confirm vital status. Lifetime productivity loss due to smoking was assessed through the lifetime earnings and survival functions estimated with the rolling extrapolation algorithm, incorporating the inverse probability of treatment weighting approach to address confounders.
Results
Our analysis of 215,852 participants aged 30–64 years revealed that ever-smokers exhibited relatively low baseline employment rates and median monthly incomes. Smoking reduced life expectancy by an average of 2.00 years (95% CI: 1.54 – 2.35) for men and 1.37 years (95% CI: −0.47 – 2.12) for women and reduced average employment time by 0.70 years (95% CI: 0.26 – 1.30) for men and 1.82 years (95% CI: 0.97 – 2.44) for women. Smoking led to significant lifetime productivity loss, with an average loss of US$ 45,572 for men and US$ 61,552 for women. The total cost of productivity loss per annum attributed to smoking in Taiwan was US$ 4,563 million (95% CI: 3,602 million – 5,574 million), which was equivalent to 0.6% of Taiwan's GDP in 2022.
Conclusion
Our study demonstrates that expanding tobacco control measures may yield substantial economic benefits, achieve significant productivity gains, and alleviate the economic burden of smoking.

Background
Influenza surveillance is critical for supporting a timely public health response to influenza outbreaks and seasonal peaks. However, existing systems are often not optimised for predictive modelling. Given the growing role of Artificial Intelligence (AI) and Machine Learning (ML) in modelling future short-term disease scenarios, surveillance systems need to be selected and adapted to effectively support these applications. This study proposes a framework, based on key characteristics required for short-term forecasts, to evaluate and then enhance surveillance systems for AI/ML-driven forecasting, using New Zealand as a case study.

Methods
This framework focuses on eight key attributes—timeliness, sensitivity, specificity, representativeness, coverage, robustness, completeness, and historical data—selected from 16 data quality-related attributes, in a broader pool of 31. Attributes were chosen based on expert input, AI/ML requirements, and established evaluation frameworks. A comprehensive review of government reports, official data, and literature was undertaken to characterise New Zealand’s influenza surveillance systems, with weighted scoring used to evaluate their suitability for model training and short-term forecasting.

Results
In New Zealand, one sentinel community-based surveillance system (SHIVERS) and one hospital-based system (SARI) emerged as the most suitable surveillance systems for AI/ML-based predictive models in both training and forecasting. While the national hospital discharge system (NMDS) and national mortality database offer strong potential for training, their delayed reporting limits short-term forecasting. Laboratory-based virological surveillance plays a pivotal role in bridging community and hospital data, enhancing model accuracy by integrating virological confirmation with broader surveillance insights.

Conclusions
This study highlights gaps where surveillance systems could be improved for better predictive modelling. Enhancing real-time data collection and leveraging multiple data sources could improve forecasting accuracy, helping to make timely, more informed decisions for managing influenza outbreaks and seasonal peaks in healthcare demand.

Keywords: Influenza, Surveillance, Artificial Intelligence, Machine Learning, Short-term Forecasting.

Background: Older adults have been disproportionately impacted by the COVID-19 pandemic. In Australia, measures to mitigate and prevent the spread of SARS-CoV-2 within older adults focused primarily on those living in residential long-term care. Consequently, there are currently no published national-level epidemiological data describing how the pandemic affected older adults receiving aged care services within their home. Using the Australian Institute of Health and Welfare’s COVID-19 Register, this analysis will describe COVID-19 case notification rates, hospitalisation and mortality rates, and vaccination status in community-dwelling older adults receiving Home Care Packages, from 2020-2022.
Methods: Drawn from the Medicare Consumer directory, the study population includes all adults aged 65 years or older at the start of the study period (50+ for First Nations), and those who turned 65 during the study period (50+ for First Nations), retaining person time after cohort entry only. Rates over time will be calculated for each of the three cohorts: i) receiving aged care within their own home (Home Care Packages, by level of care), ii) receiving aged care within a residential facility or iii) not receiving aged care services (including older people receiving Commonwealth Home Support Program services).
Results: Analyses will include an examination of temporal trends based on the two main circulating SARS-CoV-2 variants in Australia. Key outcomes include a comparison by cohort of COVID-19 notification rates; hospitalisations and emergency department presentations with a COVID-19 diagnosis; deaths where COVID-19 is a cause; and vaccination status. Primary analyses will be at the national level only.
Conclusion: Policy that strengthens and supports sustainable models of care for community-dwelling older adults needs to be evidence based. Robust epidemiological data is required to inform planning and preparedness strategies to increase resilience to future waves of COVID-19, other respiratory pandemics and current endemic respiratory viruses, within the home care sector.

Background: Multiple sclerosis (MS) risk has been attributed to a complex interplay between multiple environmental and genetic factors. Here, we systematically reviewed the literature for observational studies of gene-environment interaction (GxE) of low vitamin D and/or low sun exposure associations with MS risk.
Methods: Relevant data from each study were extracted and four-level additive interaction measures of association (E+/G-, E-/G+, E+/G-, E+/G+), from which interaction parameters were estimated, including synergy index (S), relative excess risk due to interaction (RERI), and attributable proportion due to interaction (AP). Studies were assessed for quality and risk of bias and publication bias. In the subsets of comparable articles, meta-analysis was performed to estimate pooled measures of association and interactions.
Results: We included 11 eligible studies on GxE of low vitamin D and/or sun associations with MS risk. For vitamin D, most studies (5/7) measured serum 25-hydroxyvitamin D (25(OH)D). For sun exposure (n=8), measures used were more varied - only four studies used the same approach (questionnaire-based days per week in sun). Across both vitamin D and sun GxE analyses, HLA-DRB1*15 was the most common genotype evaluated, along with those for genes involved in vitamin D metabolism (e.g., GC) and signalling (e.g., VDR), and skin colour regulation (MC1R).
Using meta-analysis to pool comparable studies, significant additive interactions were observed between low vitamin D and HLA-DRB1*15 risk genotype with MS risk (S=1.65; AP=0.39; RERI=2.5) and between low sun exposure and HLA-DRB1*15 (S=1.49; AP=0.28; RERI=1.42). Additionally, those with low sun exposure and HLA-DRB1*15 positivity were five-fold were more frequent among MS cases (OR=5.17, 95%CI=4.39-6.17) compared to those with neither risk factor.
Conclusion: These results substantiate the dynamic interplay between low sun and low vitamin D and genetic risk factors for MS - particularly HLA-DRB1*15:01 and suggest potential utility of these factors in screening for MS.

Background: White matter hyperintensities (WMHs) and hippocampal atrophy are early imaging markers of dementia, that can be detected decades before clinical dementia symptoms. However, there is little evidence showing that factors affecting hippocampal atrophy and WMHs also impact the risk of Alzheimer's disease (AD).
Methods: We examined more than 100 exposures for the consistency of their prospective associations with hippocampal volume (n=42,152), volume of WMHs (n=44,053), and AD (n=502,340) using the UK Biobank. In further two-sample Mendelian randomization (2SMR) analyses on AD, exposures were instrumented by genetic variants associated with the exposure at genome-wide significance (p<5 x 10-8), with outcome associations assessed using data from the International Genomics of Alzheimer's Project (21,982 cases and 41,944 controls). Primary analyses were conducted using inverse-variance weighted MR, with sensitivity analyses using pleiotropy robust methods.
Results: Higher adiposity was associated with adverse differences in brain morphometry but with a consistently lower risk of AD (e.g. body fat percentage, MRIVW per standard deviation (SD) OR= 0.83, 95%CI: 0.71-0.97), with similar contrast seen for basal metabolic rate (MRIVW per SD OR= 0.77, 95%CI: 0.68-0.87). Higher coffee consumption (> 6 cups/day) was observationally associated with adverse brain morphometry and AD, with inconclusive evidence from MR (MRIVW per SD OR=1.13, 95%CI: 0.89-1.43). Higher 25-hydroxyvitamin D [25(OH)D], physical activity, and measures of better lung function, were associated with favourable brain outcomes and lower AD risk, with limited support from MR analyses. Multiple sclerosis was consistently related to adverse brain morphometry and AD risk. Several blood biomarkers associated with brain morphometry with directionally consistent evidence for and urate and IGF-1
Conclusion: Some risk factors associated with adverse brain outcomes are linked to AD, but observed associations are not always consistent, suggesting a need for a nuanced evaluation and a better understanding of the drivers of selection and competing risks.

Background: Ultra-Processed Food (UPF) consumption is associated with depression in longitudinal cohort studies. However, the effect of UPF on depressive and mental health symptoms in older adults has not been determined.
Objective: This study investigated the effect of UPF on depression and mental health in community-dwelling older adults.
Methods: A pragmatic preventive target trial was designed and emulated using the ASPirin in Reducing Events in the Elderly longitudinal data. Participants were community-dwelling older adults (>70 years) in Australia. We specified and emulated the protocol of a two-arm randomised pragmatic clinical trial using level of UPF consumption as the intervention. Dietary consumption was assessed using a diet screening questionnaire and the level of food processing was classified based on the NOVA classification. The outcomes of the study were depression, defined a score of ≥ 8 on the Center for Epidemiological Studies Depression 10-item scale, and general mental health, defined by the mental component summary score of the Short Form-12. We applied the inverse probability treatment weighting to balance covariates. Marginal structural models were employed to estimate the population-level average effect of intervention using generalized estimated equations.
Result: A total of 11,192 participants (3,415 intervention and 7,777 control) were eligible for the emulation. High UPF consumption at time zero was associated with an increased risk of depression at follow-ups (RR: 1.10; CI: 1.04-1.18). The finding was consistent with sensitivity analyses, after excluding participants on antidepressants at time zero, the risk of depression in the intervention group was increased by 11% compared to control (RR: 1.11; 95% CI: (1.04-1.20)). Consumption of UPF adversely affected the mental component quality of life (β: -0.40; CI: -0.65- -0.15).
Conclusion: A higher level of UPF consumption was associated with a higher risk of depression and adversely affected mental health among older adults.

BACKGROUND: Common respiratory viruses have substantial impact on morbidity and mortality. SnotWatch was established in 2019 as an ecological data platform for real-time population-level analyses of spatiotemporal associations between viral activity and important disease outcomes in Victoria. We aimed to analyse respiratory virus detections and describe the epidemiology of the baseline viral exposure dataset.

METHODS: SnotWatch has an ongoing data feed of results from laboratory polymerase chain reaction (PCR) tests from major pathology services in Victoria performed from 2010 onwards. We analysed all available data as of January 2025. Variables included age, sex, specimen collection date, location of patient residence mapped to Statistical Area 2 (SA2) and pathogen-specific result.

RESULTS: From January 2010-September 2024, data were available for 2,272,587 respiratory samples from six public laboratories for analysis. Baseline demographics were available for 1,917,273 adults and 337,596 children; the median age at sampling was 37 years (interquartile range 26-56) and predominantly from females (53%). Of 369,407 positive tests, the most common virus detected was rhinovirus/enterovirus (122,658 [33%]), followed by SARS-CoV-2 (COVID-19) (75,184 [20%]), respiratory syncytial virus (RSV) (52,806 [14%]), influenza A (40,196 [11%]) and adenovirus (16,090 [4%]). Virus detections varied by season, age groups and over time; rhinovirus/enterovirus was identified from continued circulation among young children despite the COVID-19 pandemic. Spatial analyses of infections across Victoria will be presented. If available, updated laboratory data will be presented.

CONCLUSION: SnotWatch provides population-level, comprehensive coverage of respiratory viral activity in Victoria. Analysis of seasonal patterns may be impacted by inconsistencies across different time periods among laboratories, due to electronic record data availability and technological infrastructure. As the platform matures, continued collaboration with laboratory services will facilitate improved data on viral activity and, in turn, the modelling of important health conditions.

Background
Respiratory syncytial virus (RSV) is the leading cause of acute lower respiratory tract infections (LRTI) in young children. In Australia, children under five years account for most RSV-related hospitalisations, with the highest rates in infants under six months of age. Nirsevimab is a long-acting monoclonal antibody which protects infants and young children against RSV infection. Clinical trials and real-world effectiveness data from the Northern Hemisphere’s 2023-24 winter season showed good protection against RSV infection and a reduction in hospitalisations and intensive care admissions. In Australia, state/territory-based programs started in March 2024; broad population-based programs were introduced in NSW, QLD and WA, while the ACT, NT and TAS introduced nirsevimab for a very small cohort of high-risk infants and children.

Methods and Results
This ecological study aims to provide a national overview of the early impact of nirsevimab introduction on the epidemiology of RSV cases and hospitalisations in children under five years of age in Australia. We hypothesise that RSV cases and hospitalisations in this age group decreased nationally post-introduction in jurisdictions that introduced broad population-based programs. We will use population-level data from the National Notifiable Diseases Surveillance System to compare jurisdictions with and without broad population-based RSV prevention programs in the pre-introduction (January 2022 – March 2024) and post-introduction (April – December 2024) periods. We will present age-stratified trends on RSV positive cases, RSV-related LRTI hospitalisations and RSV-related LRTI intensive care admissions. Where available, we will describe national and jurisdictional nirsevimab coverage rates.

Conclusions
Characterising the impact of nirsevimab introduction will add to the evidence for immunisation policy considerations, which could improve equitable protection from RSV nationwide in 2025.

Background:
Poor healthcare access among Cambodian women poses a worse health outcome as it leads to unwanted pregnancies and maternal death. Therefore, this study aims to identify the problems in accessing general health care and the associated factors among women age 15-49 years in Cambodia.
Method:
This study used the existing data of the 2021-2022 Cambodia Demographic and Health Survey by including 19,496 reproductive age women from 15-49 years old. Descriptive statistics were used to describe continuous variables and weighted frequency and percentage were performed to report categorical variables. Multivariate logistic regression was employed to assess the strength and magnitude of the association between factors and problems in accessing health care services and to control for confounding. Weighted analysis was applied in this study.
Results:
The prevalence of problems in accessing healthcare among women age 15-49 years old was 60%. Women age group of 25-34 (AOR=0.79, 95%CI: 0.69-0.89) and age group of 35-49 (AOR=0.78, 95%CI: 0.68-0.90), married women or living together with partner (AOR= 0.83, 95%CI: 0.69-0.99), women with no education (AOR=2.23, 95%CI: 1.74-2.85), primary school (AOR=1.82, 95%CI: 1.49-2.22) and secondary school (AOR=1.60, 95%CI: 1.34-1.91), rural area women (AOR=1.22, 95%CI: 1.08-1.38), poorest (AOR=3.32, 95%CI: 2.72-4.06), poorer (AOR=2.15, 95%CI: 1.80-2.56), middle (AOR=1.88, 95%CI: 1.59-2.21) and richer wealth status (AOR=1.61, 95%CI: 1.37-1.89) and distance from health care facilities with more than 2 hours (AOR=3.73, 95%CI: 1.31-10.56), 60-119 minutes (AOR=2.49, 95%CI: 1.70-3.66) and 30-59 minutes (AOR=1.71, 95%CI: 1.44-2.03) were associated with problems in healthcare access.
Conclusion:
The findings of this study are very important to suggest and recommend health policy makers, civil society organizations and other stakeholders to pay attention and strengthen health care services for women living in rural areas and far from health care facilities and those from low socioeconomic group in Cambodia.

Background
Campylobacteriosis is the most commonly reported bacterial gastrointestinal infection in New Zealand, with environmental factors playing a crucial role in its transmission. Various aspects such as extreme climates, water supply characteristics, water quality, land use, and rurality may influence disease incidence. Understanding the interplay between these factors is essential for effective public health interventions. This study applies machine learning techniques to examine the complex relationships between environmental risk factors and Campylobacteriosis notification rates across New Zealand. We hypothesize that water quality indicators, such as microbial contamination levels, along with land use patterns and rurality, are strongly associated with Campylobacteriosis rates.

Methods
We are conducting a cross-sectional analysis using national Campylobacteriosis notification data combined with environmental, social, and demographic datasets. Machine learning models, including random forests and gradient boosting algorithms, are being applied to identify key environmental and social predictors of disease occurrence. Model performance is assessed using accuracy metrics and feature importance scores to determine the most influential risk factors.
Results
We expect to show that microbial contamination in untreated or minimally treated water supplies is a key predictor of increased Campylobacteriosis notification rates. high livestock density shows higher disease incidence, due to waterborne contamination. Rural areas, particularly those relying on private water supplies, also experience elevated rates. Final results, including predictive model outputs and variable importance rankings, will be presented at the meeting.

Conclusion
This study demonstrates the utility of machine learning in uncovering complex associations between environmental factors and Campylobacteriosis risk. Identifying key predictors can inform targeted interventions, such as improving water treatment infrastructure and land use management strategies.

Background
Survivors of COVID-19 are susceptible to diminished health-related quality of life (HRQoL) and adverse psychological health, which may be exacerbated by their experiences of fear and the impact of the pandemic itself. This study aims to identify distinct fear and impact patterns related to the COVID-19 pandemic among survivors through latent profile analysis (LPA) and examine the associations of fear and impact patterns with post-traumatic stress symptoms (PTSS) and HRQoL.
Methods
A total of 5,890 Vietnamese COVID-19 survivors completed the COVID-19 Impact Battery- Disability Scale (CIB-D), the Fear of COVID-19 Scale (FCoV-19 S), the Impact of Event Scale-Revised for PTSS, and the 36-Item Short Form Survey (SF-36) for HRQoL.
Results
Four distinct groups of fear and impact were identified: "Fearful and highly impacted" (26.8%), "moderately impacted yet not fearful" (22.9%), "less impacted and less fearful" (18.6%), and "mildly impacted and neutral" (31.7%). Survivors who were "less impacted and less fearful" exhibited significantly higher HRQoL scores (regression coefficient, B: 10.9; 95% confidence interval (CI): 10.0 - 11.7), both in terms of physical (B: 12.0; 95%CI: 11.1 - 12.9) and mental health (B: 19.4; 95%CI: 9.6 - 11.1), and lower PTSS levels (B: -24.5; 95%CI: -25.8 - -23.3) compared to those who were "highly impacted and fearful".
Conclusion
It is imperative to acknowledge the intricate association between fear, impact, and mental health to comprehensively address the diverse needs of this distinct population post-COVID-19. These findings provide insights for designing interventions and support mechanisms for COVID-19 survivors.

Introduction/Background
Patient-reported outcomes (PROs) can identify and address unmet patient needs, and can also be used as an indicator of health care quality when collected in clinical quality registries (CQRs). Digital PRO collection can reduce costs and improve sustainability, but may also result in equity due to gaps in digital access. We sought to examine PRO collection through a prostate cancer CQR and compare collection in regional and non-regional areas to understand digital uptake.

Methods
PRO data collection methods and completion rates from 2015 – 2022 were compared between Tasmania, a fully regional jurisdiction within the binational Prostate Cancer Outcomes Registry – Australia and New Zealand (PCOR-ANZ), and seven participating jurisdictions (NZ plus all states/territories save WA). PROs were collected via phone, email, or mail.

Results
From 2015 to 2022, 57,910 surveys were collected in PCOR-ANZ (1,928 in Tasmania). In 2015, the proportion of PROs collected across all ANZ by phone, email and mail changed from 48%, 9% and 43% in 2015 to 9%, 21% and 70% in 2022, respectively. Across all ANZ, email collection was highest for younger patients (27% of age <60, vs 15% for age >74); for Tasmania, this was 25% and 9%, respectively. Mail collection was used by 66% of Tasmanian patients <60 years old, compared to 54% of patients in all ANZ jurisdictions. Tasmanian PROM completion rates were generally high: in 2020, 75.9% of Tasmanian patients completed PROMs, compared to a completion rate of 52.5% across all ANZ jurisdictions.

Conclusions
Digital collection of PROs in CQRs is possible and appears to be increasing over time. Data collection via mail may improve completion rates. Working to identify and address barriers to digital access may enhance completion. Relying solely on digital methods may be inequitable and unrepresentative at a population-based level, particularly in regional areas.

Background:
ACT Health uses Computer Assisted Telephone Interviews (CATI) to conduct the ACT General Health Survey (ACTGHS) to monitor population health trends. The ACTGHS has been conducted annually since 2007. In 2024, the Integrated Phone Number Database (IPND) was used as the sampling frame for the first time; listed numbers were used prior to 2024. A major benefit of the IPND is that it contains almost every assigned telephone number in Australia, along with the postcode of each number. This eliminates the need to screen the mobile phone sample which has considerable cost implications for CATI surveys.

The aim of this presentation is to compare sample coverage and investigate results bias between the IPND and listed mobile numbers using data obtained through the 2020–2024 ACTGHS.

Method:
Gender, age and geographic location as well as self-rated health statuses were analysed to compare sample coverage and to identify any bias in survey results obtained from the IPND and listed mobile numbers sampling frames. The analyses are based on 8,739 ACT residents aged five years and over who participated in the ACTGHS between 2020 and 2024.

Results:
Sample coverage by gender, age group and geography remained stable over 2020–2024. There was no break in trend over survey years observed for the weighted estimates of self-rated health or self-rated mental health.

Conclusion:
Any changes to the survey methodology of ongoing surveys such as the ACTGHS need to be carefully considered to ensure results can be reliably compared over time. Transitioning to the IPND sampling frame from the current listed mobile numbers was shown to produce similar sample coverage and trend estimates for self-rated health statuses.

Background: There are limited studies on the link between maternal depression and ASD, and those that exist have significant limitations, including insufficient control for confounders, reliance on self-reported data, small sample sizes, and a lack of investigation into potential mediating factors. This study aims to fill these gaps by examining both the direct relationship and the potential mediating effects of preterm birth, low birth weight, and low APGAR scores on this association.
Methods: We analysed linked administrative health data involving 223,068 mother-offspring pairs in New South Wales (NSW), Australia. Maternal perinatal depressive disorders and offspring ASD were assessed using the International Classification of Diseases (ICD-10). A generalised linear model with a binomial distribution and a log link function was employed to examine the association. The mediation effects of preterm birth, low birth weight, and low APGAR scores were assessed through mediation analysis.
Results: After adjusting for a range of potential confounders, offspring of mothers with antenatal, postnatal, and overall perinatal depressive disorders had a 61% (RR=1.60, 95% CI=1.12–2.32), 89% (RR=1.89, 95% CI=1.22–2.94), and 80% (RR=1.80, 95% CI=1.33–2.43) higher risk of ASD, respectively. Only about 1.24% and 1.27% of the effect of maternal antenatal depressive disorders on offspring ASD was mediated by preterm birth and low APGAR scores, respectively, which is approximately 61 times smaller than the direct effect. Low birth weight had no significant mediating effect on the association.
Conclusion: Maternal perinatal depressive disorders are associated with an increased risk of offspring ASD. Preterm birth and low APGAR scores were weak mediators of this association. Early intervention strategies that aim to enhance maternal mental health and mitigate the risk of exposed offspring are needed.

Background
Streptococcus pneumoniae is an important cause of pneumonia in older adults, however, serotyping and indirect impact information from low-middle-income countries is lacking. Mongolia has a childhood 13-valent pneumococcal conjugate vaccine (PCV13) program, but no adult pneumococcal vaccination program. We describe pneumococcal carriage rates, disease and serotype distribution among adult pneumonia cases, and explore changes over the COVID-19 pandemic period.

Methods
Adults (≥18 years) hospitalised with clinical pneumonia were enrolled over 3 years (2019-2022) into a prospective pneumonia surveillance program. Nasopharyngeal swabs were tested to detect pneumococci using lytA qPCR and molecular serotyping by DNA microarray and metagenomics. Pneumococcal pneumonia was identified using serotype-specific urinary antigen detection and BinaxNOW® assays. Pneumococcal carriage and pneumonia prevalence were assessed over the COVID-19 period with log-binomial regression used to estimate prevalence and adjusted prevalence ratios (pre- versus early- and late-COVID periods).

Results
Overall 3442 pneumonia cases were enrolled with 3178 (92.3%) meeting the clinical pneumonia case definition. Of these 1370 (43.1%) were male and 1928 (60.8%) had ≥1 underlying medical condition. There was a higher percentage of adults ≥65 years (23.2% vs 30.3%) and patients with multiple medical conditions (22.4% vs 33.6%) in the late- versus pre-COVID-19 period.
Pneumococcal carriage prevalence was 12.1% (333/2,759); in those able to be serotyped, PCV13 and non-PCV13 serotype carriage prevalence was 3.1% (82/2,663) and 5.7% (152/2,663), respectively. Pneumococcal pneumonia prevalence was 8.6% (253/2,925).
In the late-COVID-19 period, pneumococcal carriage prevalence was reduced by 66% (aPR 0.34, 95%CI 0.25–0.46) and pneumococcal pneumonia by 82% (aPR 0.18, 95%CI 0.12–0.27) compared with pre-COVID-19 period.

Conclusion
The COVID pandemic impacted both pneumococcal carriage and disease. Despite paediatric vaccination with high coverage, we identified some residual PCV13 serotypes with predominance of non-PCV13 serotypes carried and causing disease in adults. Direct adult vaccination will potentially reduce disease in adults in Mongolia.

Background: A long-standing debate in the field of medically assisted reproduction (MAR) is whether the medicines used in these treatments are carcinogenic. While biologically plausible, evidence for the link between MAR medicines and cancer is mixed. Assessing the true causal relationship between MAR medicines and cancer is difficult, as studies rely on real-world data (RWD). A major challenge with causal inference from RWD is dealing with bias (systematic deviation of the observed estimate from the true causal relationship). The current work outlines two methods used to estimate the impact of bias (quantitative bias analysis, and E-values), and shows how these methods can be applied to the study of MAR and cancer.
Methods: We reviewed both recent literature into the relationship between MAR and cancer, and literature into bias in observational studies, including quantitative bias analysis and E-values. We constructed a directed acyclic graph to outline potential sources of bias in studies of MAR and cancer, based on current literature. Leveraging this graph, we identified potential sources of bias that may have affected four published studies on MAR and cancer. Finally, we estimated E-values for these studies to provide a demonstration of how this method are used to measure the impact of bias.
Results: Our E-value analysis suggested one study may have incorrectly concluded that a causal relationship existed between MAR and thyroid cancer, as this observed relationship may have been caused by failing to control for endometriosis as a confounder.
Conclusions: Our analysis shows that estimating and reporting the impact of bias is important for studies on MAR and cancer. More broadly, we argue that explicitly estimating bias is crucial when estimating causal effects using RWD, especially for health policy and practice implications, and epidemiologists should be familiar with common methods for doing so.

Background: Early life is a critical period for the biological embedding of the social environment, with disease risk accumulating across the life course. Systemic inflammation is a key mechanism in social-to-biological transitions. We investigated the extent to which neighbourhood, household and individual-level socioeconomic exposures are associated with inflammation in childhood and adolescence and assessed the extent to which this relationship is mediated by body mass index (BMI).
Methods: We conducted cohort-specific analysis in four longitudinal studies: the Barwon Infant Study (BIS, Australia, n=708, 4 years); Born in Bradford (BiB, UK, n=4576, 7–11 years); the Longitudinal Study of Australian Children’s Child Health CheckPoint (LSAC-CP, n=1874, 12 years); and the Northern Finland Birth Cohort 1986 (NFBC 1986, n=9467, 15–16 years). Exposures, measured in pregnancy/infancy, were neighbourhood disadvantage, household income, and maternal education. Outcomes were log-transformed high-sensitivity C-reactive protein (hsCRP) and glycoprotein acetyls (GlycA). We conducted confounder-adjusted multiple linear regression and explored the role of BMI through causal mediation analysis based on interventional effects.
Results: Greater neighbourhood disadvantage and lower maternal education were consistently associated with higher inflammation in BiB, LSAC-CP and NFBC 1986 (e.g., difference in mean log GlycA between: i. most and least disadvantaged neighbourhoods in BiB: 0.14 standard deviations [95% CI 0.03–0.24]; ii. most and least educated mothers in LSAC-CP: 0.23 [0.00–0.46]). Fewer effects were observed for household income and minimal effect sizes between SEP and inflammation were found in BIS at 4 years. BMI was an important mediator in some models (e.g., intervening to reduce BMI in BiB children from disadvantaged neighbourhoods could reduce absolute GlycA differences by 45%).
Conclusion: Socioeconomic inequalities in systemic inflammation are evident in mid-childhood and adolescence across Australia, Finland and the UK, partly mediated by BMI. Interventions should address upstream socioeconomic conditions as well as downstream targets such as obesity.

In New Zealand, 7% of deaths are related to respiratory diseases, with Pacific people at higher risk. Our work investigates the causal effects of early-life risks and resilience factors on early-adulthood lung function amongst Pacific Islands Families Study (PIFS) cohort members (n=1,398). 466 from the cohort participated in the respiratory study. Primary outcome was forced expiratory volume in 1 second (FEV1) z-score at age 18 years. FEV1 and healthy lung function (HLF), defined as the z-score being larger than -1.64, were secondary outcomes. A previous study had evaluated the effects of early-life nutrition factors on the respiratory health of Pacific youth. The results suggested a positive impact of consuming more fruit and vegetables during childhood on respiratory health later in life. The follow-up study will continue to explore the effects of factors from relevant domains based on the PIFS cohort, where a new integrated model will be applied. A simulation will be conducted to determine this model.