Towards development of functional climate-driven early warning systems for climate-sensitive infectious diseases: Statistical models and recommendations.

Climate, weather and environmental change have significantly influenced patterns of infectious disease transmission, necessitating the development of early warning systems to anticipate potential impacts and respond in a timely and effective way. Statistical modelling plays a pivotal role in understanding the intricate relationships between climatic factors and infectious disease transmission. For example, time series regression modelling and spatial cluster analysis have been employed to identify risk factors and predict spatial and temporal patterns of infectious diseases. Recently advanced spatio-temporal models and machine learning offer an increasingly robust framework for modelling uncertainty, which is essential in climate-driven disease surveillance due to the dynamic and multifaceted nature of the data. Moreover, Artificial Intelligence (AI) techniques, including deep learning and neural networks, excel in capturing intricate patterns and hidden relationships within climate and environmental data sets. Web-based data has emerged as a powerful complement to other datasets encompassing climate variables and disease occurrences. However, given the complexity and non-linearity of climate-disease interactions, advanced techniques are required to integrate and analyse these diverse data to obtain more accurate predictions of impending outbreaks, epidemics or pandemics. This article presents an overview of an approach to creating climate-driven early warning systems with a focus on statistical model suitability and selection, along with recommendations for utilizing spatio-temporal and machine learning techniques. By addressing the limitations and embracing the recommendations for future research, we could enhance preparedness and response strategies, ultimately contributing to the safeguarding of public health in the face of evolving climate challenges.

Impacts of heatwaves on type 2 diabetes mortality in China: a comparative analysis between coastal and inland cities.

The impacts of extreme temperatures on diabetes have been explored in previous studies. However, it is unknown whether the impacts of heatwaves appear variations between inland and coastal regions. This study aims to quantify the associations between heat exposure and type 2 diabetes mellitus (T2DM) deaths in two cities with different climate features in Shandong Province, China. We used a case-crossover design by quasi-Poisson generalized additive regression with a distributed lag model with lag 2 weeks, controlling for relative humidity, the concentration of air pollution particles with a diameter of 2.5 µm or less (PM2.5), and seasonality. The wet- bulb temperature (Tw) was used to measure the heat stress of the heatwaves. A significant association between heatwaves and T2DM deaths was only found in the coastal city (Qingdao) at the lag of 2 weeks at the lowest Tw = 14℃ (relative risk (RR) = 1.49, 95% confidence interval (CI): 1.11-2.02; women: RR = 1.51, 95% CI: 1.02-2.24; elderly: RR = 1.50, 95% CI: 1.08-2.09). The lag-specific effects were significant associated with Tw at lag of 1 week at the lowest Tw = 14℃ (RR = 1.14, 95% CI: 1.03-1.26; women: RR = 1.15, 95% CI: 1.01-1.31; elderly: RR = 1.15, 95% CI: 1.03-1.28). However, no significant association was found in Jian city. The research suggested that Tw was significantly associated with T2DM mortality in the coastal city during heatwaves on T2DM mortality. Future strategies should be implemented with considering socio-environmental contexts in regions.

Geographical and spatial disparities in the incidence and survival of rare cancers in Australia.

Rare cancers collectively account for around a quarter of cancer diagnoses and deaths. However, epidemiological studies are sparse. We describe spatial and geographical patterns in incidence and survival of rare cancers across Australia using a population-based cancer registry cohort of rare cancer cases diagnosed among Australians aged at least 15 years, 2007 to 2016. Rare cancers were defined using site- and histology-based categories from the European RARECARE study, as individual cancer types having crude annual incidence rates of less than 6/100 000. Incidence and survival patterns were modelled with generalised linear and Bayesian spatial Leroux models. Spatial heterogeneity was tested using the maximised excess events test. Rare cancers (n = 268 070) collectively comprised 22% of all invasive cancer diagnoses and accounted for 27% of all cancer-related deaths in Australia, 2007 to 2016 with an overall 5-year relative survival of around 53%. Males and those living in more remote or more disadvantaged areas had higher incidence but lower survival. There was substantial evidence for spatial variation in both incidence and survival for rare cancers between small geographical areas across Australia, with similar patterns so that those areas with higher incidence tended to have lower survival. Rare cancers are a substantial health burden in Australia. Our study has highlighted the need to better understand the higher burden of these cancers in rural and disadvantaged regions where the logistical challenges in their diagnosis, treatment and support are magnified.

Being Bayesian in the 2020s: opportunities and challenges in the practice of modern applied Bayesian statistics.

Building on a strong foundation of philosophy, theory, methods and computation over the past three decades, Bayesian approaches are now an integral part of the toolkit for most statisticians and data scientists. Whether they are dedicated Bayesians or opportunistic users, applied professionals can now reap many of the benefits afforded by the Bayesian paradigm. In this paper, we touch on six modern opportunities and challenges in applied Bayesian statistics: intelligent data collection, new data sources, federated analysis, inference for implicit models, model transfer and purposeful software products. This article is part of the theme issue 'Bayesian inference: challenges, perspectives, and prospects'.

Mapping the prevalence of cancer risk factors at the small area level in Australia.

BACKGROUND: Cancer is a significant health issue globally and it is well known that cancer risk varies geographically. However in many countries there are no small area-level data on cancer risk factors with high resolution and complete reach, which hinders the development of targeted prevention strategies. METHODS: Using Australia as a case study, the 2017-2018 National Health Survey was used to generate prevalence estimates for 2221 small areas across Australia for eight cancer risk factor measures covering smoking, alcohol, physical activity, diet and weight. Utilising a recently developed Bayesian two-stage small area estimation methodology, the model incorporated survey-only covariates, spatial smoothing and hierarchical modelling techniques, along with a vast array of small area-level auxiliary data, including census, remoteness, and socioeconomic data. The models borrowed strength from previously published cancer risk estimates provided by the Social Health Atlases of Australia. Estimates were internally and externally validated. RESULTS: We illustrated that in 2017-2018 health behaviours across Australia exhibited more spatial disparities than previously realised by improving the reach and resolution of formerly published cancer risk factors. The derived estimates revealed higher prevalence of unhealthy behaviours in more remote areas, and areas of lower socioeconomic status; a trend that aligned well with previous work. CONCLUSIONS: Our study addresses the gaps in small area level cancer risk factor estimates in Australia. The new estimates provide improved spatial resolution and reach and will enable more targeted cancer prevention strategies at the small area level. Furthermore, by including the results in the next release of the Australian Cancer Atlas, which currently provides small area level estimates of cancer incidence and relative survival, this work will help to provide a more comprehensive picture of cancer in Australia by supporting policy makers, researchers, and the general public in understanding the spatial distribution of cancer risk factors. The methodology applied in this work is generalisable to other small area estimation applications and has been shown to perform well when the survey data are sparse.

Modeling total predation to avoid perverse outcomes from cat control in a data-poor island ecosystem.

Data hungry, complex ecosystem models are often used to predict the consequences of threatened species management, including perverse outcomes. Unfortunately, this approach is impractical in many systems, which have insufficient data to parameterize ecosystem interactions or reliably calibrate or validate such models. Here we demonstrate a different approach, using a minimum realistic model to guide decisions in data- and resource-scarce systems. We illustrate our approach with a case-study in an invaded ecosystem from Christmas Island, Australia, where there are concerns that cat eradication to protect native species, including the red-tailed tropicbird, could release meso-predation by invasive rats. We use biophysical constraints (metabolic demand) and observable parameters (e.g. prey preferences) to assess the combined cat and rat abundances which would threaten the tropicbird population. We find that the population of tropicbirds cannot be sustained if predated by 1607 rats (95% credible interval (CI) [103, 5910]) in the absence of cats, or 21 cats (95% CI [2, 82]) in the absence of rats. For every cat removed from the island, the bird's net population growth rate improves, provided that the rats do not increase by more than 77 individuals (95% CI [30, 174]). Thus, in this context, one cat is equivalent to 30-174 rats. Our methods are especially useful for on-the-ground predator control in the absence of knowledge of predator-predator interactions, to assess whether 1) the current abundance of predators threatens the prey population of interest, 2) managing one predator species alone is sufficient to protect the prey species given potential release of another predator, and 3) control of multiple predator species is needed to meet the conservation goal. Our approach demonstrates how to use limited information for maximum value in data-poor systems, by shifting the focus from predicting future trajectories, to identifying conditions which threaten the conservation goal. This article is protected by copyright. All rights reserved.

Spatial statistical machine learning models to assess the relationship between development vulnerabilities and educational factors in children in Queensland, Australia.

BACKGROUND: The health and development of children during their first year of full time school is known to impact their social, emotional, and academic capabilities throughout and beyond early education. Physical health, motor development, social and emotional well-being, learning styles, language and communication, cognitive skills, and general knowledge are all considered to be important aspects of a child's health and development. It is important for many organisations and governmental agencies to continually improve their understanding of the factors which determine or influence development vulnerabilities among children. This article studies the relationships between development vulnerabilities and educational factors among children in Queensland, Australia. METHODS: Spatial statistical machine learning models are reviewed and compared in the context of a study of geographic variation in the association between development vulnerabilities and attendance at preschool among children in Queensland, Australia. A new spatial random forest (SRF) model is suggested that can explain more of the spatial variation in data than other approaches. RESULTS: In the case study, spatial models were shown to provide a better fit compared to models that ignored the spatial variation in the data. The SRF model was shown to be the only model which can explain all of the spatial variation in each of the development vulnerabilities considered in the case study. The spatial analysis revealed that the attendance at preschool factor has a strong influence on the physical health domain vulnerability and emotional maturity vulnerability among children in their first year of school. CONCLUSION: This study confirmed that it is important to take into account the spatial nature of data when fitting statistical machine learning models. A new spatial random forest model was introduced and was shown to explain more of the spatial variation and provide a better model fit in the case study of development vulnerabilities among children in Queensland. At small-area population level, increased attendance at preschool was strongly associated with reduced physical and emotional development vulnerabilities among children in their first year of school.

Ensuring Prevention Science Research is Synthesis-Ready for Immediate and Lasting Scientific Impact.

When seeking to inform and improve prevention efforts and policy, it is important to be able to robustly synthesize all available evidence. But evidence sources are often large and heterogeneous, so understanding what works, for whom, and in what contexts can only be achieved through a systematic and comprehensive synthesis of evidence. Many barriers impede comprehensive evidence synthesis, which leads to uncertainty about the generalizability of intervention effectiveness, including inaccurate titles/abstracts/keywords terminology (hampering literature search efforts), ambiguous reporting of study methods (resulting in inaccurate assessments of study rigor), and poorly reported participant characteristics, outcomes, and key variables (obstructing the calculation of an overall effect or the examination of effect modifiers). To address these issues and improve the reach of primary studies through their inclusion in evidence syntheses, we provide a set of practical guidelines to help prevention scientists prepare synthesis-ready research. We use a recent mindfulness trial as an empirical example to ground the discussion and demonstrate ways to ensure the following: (1) primary studies are discoverable; (2) the types of data needed for synthesis are present; and (3) these data are readily synthesizable. We highlight several tools and practices that can aid authors in these efforts, such as using a data-driven approach for crafting titles, abstracts, and keywords or by creating a repository for each project to host all study-related data files. We also provide step-by-step guidance and software suggestions for standardizing data design and public archiving to facilitate synthesis-ready research.

Bayesian prediction of winning times for elite swimming events.

To develop a statistical model of winning times for international swimming events with the aim of predicting winning time distributions and the probability of winning for the 2020 and 2024 Olympic Games. The data set included first and third place times from all individual swimming events from the Olympics and World Championships from 1990 to 2019. We compared different model formulations fitted with Bayesian inference to obtain predictive distributions; comparisons were based on mean percentage error in out-of-sample predictions of Olympics and World Championships winning swim times from 2011 to 2019. The Bayesian time series regression model, comprising auto-regressive and moving average terms and other predictors, had the smallest mean prediction error of 0.57% (CI 0.46-0.74%). For context, using the respective previous Olympics or World Championships winning time resulted in a mean prediction error of 0.70% (CI 0.59-0.82%). The Olympics were on average 0.5% (CI 0.3-0.7%) faster than World Championships over the study period. The model computes the posterior predictive distribution, which allows coaches and athletes to evaluate the probability of winning given an individual's swim time, and the probability of being faster or slower than the previous winning time or even the world record.

Estimating a novel stochastic model for within-field disease dynamics of banana bunchy top virus via approximate Bayesian computation.

The Banana Bunchy Top Virus (BBTV) is one of the most economically important vector-borne banana diseases throughout the Asia-Pacific Basin and presents a significant challenge to the agricultural sector. Current models of BBTV are largely deterministic, limited by an incomplete understanding of interactions in complex natural systems, and the appropriate identification of parameters. A stochastic network-based Susceptible-Infected-Susceptible model has been created which simulates the spread of BBTV across the subsections of a banana plantation, parameterising nodal recovery, neighbouring and distant infectivity across summer and winter. Findings from posterior results achieved through Markov Chain Monte Carlo approach to approximate Bayesian computation suggest seasonality in all parameters, which are influenced by correlated changes in inspection accuracy, temperatures and aphid activity. This paper demonstrates how the model may be used for monitoring and forecasting of various disease management strategies to support policy-level decision making.

Bayesian meta-analysis models for cross cancer genomic investigation of pleiotropic effects using group structure.

An increasing number of genome-wide association studies (GWAS) summary statistics is made available to the scientific community. Exploiting these results from multiple phenotypes would permit identification of novel pleiotropic associations. In addition, incorporating prior biological information in GWAS such as group structure information (gene or pathway) has shown some success in classical GWAS approaches. However, this has not been widely explored in the context of pleiotropy. We propose a Bayesian meta-analysis approach (termed GCPBayes) that uses summary-level GWAS data across multiple phenotypes to detect pleiotropy at both group-level (gene or pathway) and within group (eg, at the SNP level). We consider both continuous and Dirac spike and slab priors for group selection. We also use a Bayesian sparse group selection approach with hierarchical spike and slab priors that enables us to select important variables both at the group level and within group. GCPBayes uses a Bayesian statistical framework based on Markov chain Monte Carlo (MCMC) Gibbs sampling. It can be applied to multiple types of phenotypes for studies with overlapping or nonoverlapping subjects, and takes into account heterogeneity in the effect size and allows for the opposite direction of the genetic effects across traits. Simulations show that the proposed methods outperform benchmark approaches such as ASSET and CPBayes in the ability to retrieve pleiotropic associations at both SNP and gene-levels. To illustrate the GCPBayes method, we investigate the shared genetic effects between thyroid cancer and breast cancer in candidate pathways.

A call for better understanding of social media in surveillance and management of noncommunicable diseases.

Using social media for health purposes has attracted much attention over the past decade. Given the challenges of population ageing and changes in national health profile and disease patterns following the epidemiologic transition, researchers and policy-makers should pay attention to the potential of social media in chronic disease surveillance, management and support. This commentary overviews the evidence base for this inquiry and outlines the key challenges to research laying ahead. The authors provide concrete suggestions and recommendations for developing a research agenda to guide future investigation and action on this topic.

Using internet-based query and climate data to predict climate-sensitive infectious disease risks: a systematic review of epidemiological evidence.

The use of internet-based query data offers a novel approach to improve disease surveillance and provides timely disease information. This paper systematically reviewed the literature on infectious disease predictions using internet-based query data and climate factors, discussed the current research progress and challenges, and provided some recommendations for future studies. We searched the relevant articles in the PubMed, Scopus, and Web of Science databases between January 2000 and December 2019. We initially included studies that used internet-based query data to predict infectious disease epidemics, then we further filtered and appraised the studies that used both internet-based query data and climate factors. In total, 129 relevant papers were included in the review. The results showed that most studies used a simple descriptive approach (n=80; 62%) to detect epidemics of influenza (including influenza-like illness (ILI)) (n=88; 68%) and dengue (n=9; 7%). Most studies (n=61; 47%) purely used internet search metrics to predict the epidemics of infectious diseases, while only 3 out of the 129 papers included both climate variables and internet-based query data. Our research shows that including internet-based query data and climate variables could better predict climate-sensitive infectious disease epidemics; however, this method has not been widely used to date. Moreover, previous studies did not sufficiently consider the spatiotemporal uncertainty of infectious diseases. Our review suggests that further research should use both internet-based query and climate data to develop predictive models for climate-sensitive infectious diseases based on spatiotemporal models.

Informing management decisions for ecological networks, using dynamic models calibrated to noisy time-series data.

Well-intentioned environmental management can backfire, causing unforeseen damage. To avoid this, managers and ecologists seek accurate predictions of the ecosystem-wide impacts of interventions, given small and imprecise datasets, which is an incredibly difficult task. We generated and analysed thousands of ecosystem population time series to investigate whether fitted models can aid decision-makers to select interventions. Using these time-series data (sparse and noisy datasets drawn from deterministic Lotka-Volterra systems with two to nine species, of known network structure), dynamic model forecasts of whether a species' future population will be positively or negatively affected by rapid eradication of another species were correct > 70% of the time. Although 70% correct classifications is only slightly better than an uninformative prediction (50%), this classification accuracy can be feasibly improved by increasing monitoring accuracy and frequency. Our findings suggest that models may not need to produce well-constrained predictions before they can inform decisions that improve environmental outcomes.

Global, regional, and national burden of lung cancer and its attributable risk factors, 1990 to 2017.

BACKGROUND: China's lung cancer (LC) burden plays a pivotal role in the global cancer epidemic. Comparing LC burden and population attributable fractions (PAFs) of risk factors between China and other countries/regions is essential to inform effective intervention. The Global Burden of Disease (GBD) study provides a unique opportunity for such comparisons. METHODS: We extracted the number of LC deaths, age-standardized death rates (ASDRs), age-standardized disability-adjusted life-year (DALY) rates, and PAFs of risk factors for LC deaths between 1990 and 2017 from GBD 2017. The annual percentage change (APC) was used to quantify the trends of LC ASDRs and age-standardized DALY rates. The relationship between the APC of LC ASDR and Socio-demographic Index was assessed among China and other countries. RESULTS: Globally, the ASDR for LC decreased in men (APC, -0.66% [95% CI, -0.69 to -0.62]) but increased in women (APC, 0.31% [95% CI, 0.26 to 0.36]) from 1990 to 2017. The ASDRs in China increased both for men (APC, 1.12% [95% CI, 1.03 to 1.20]) and women (APC, 0.80% [95% CI, 0.70 to 0.89]). The increased LC death numbers among men (312,798) and women (139,115) in China accounted for 59.39% and 43.01% of global increases. LC years of life lost accounted for the majority of LC DALYs globally and in China. The risk factors with the highest PAFs of LC death in China were smoking and ambient particulate matter. The ASDRs for LC associated with ambient particulate matter in China ranked second globally. CONCLUSIONS: The trends of LC ASDRs and age-standardized DALY rates and the PAFs of risk factors vary markedly by region, indicating a need for tailored measures to reduce LC burden and improve health equality. China's LC ASDRs are among the highest in the world, and the primary intervention priorities in China should be control of ambient particulate matter and tobacco usage.

Forecasting intensifying disturbance effects on coral reefs.

Anticipating future changes of an ecosystem's dynamics requires knowledge of how its key communities respond to current environmental regimes. The Great Barrier Reef (GBR) is under threat, with rapid changes of its reef-building hard coral (HC) community structure already evident across broad spatial scales. While several underlying relationships between HC and multiple disturbances have been documented, responses of other benthic communities to disturbances are not well understood. Here we used statistical modelling to explore the effects of broad-scale climate-related disturbances on benthic communities to predict their structure under scenarios of increasing disturbance frequency. We parameterized a multivariate model using the composition of benthic communities estimated by 145,000 observations from the northern GBR between 2012 and 2017. During this time, surveyed reefs were variously impacted by two tropical cyclones and two heat stress events that resulted in extensive HC mortality. This unprecedented sequence of disturbances was used to estimate the effects of discrete versus interacting disturbances on the compositional structure of HC, soft corals (SC) and algae. Discrete disturbances increased the prevalence of algae relative to HC while the interaction between cyclones and heat stress was the main driver of the increase in SC relative to algae and HC. Predictions from disturbance scenarios included relative increases in algae versus SC that varied by the frequency and types of disturbance interactions. However, high uncertainty of compositional changes in the presence of several disturbances shows that responses of algae and SC to the decline in HC needs further research. Better understanding of the effects of multiple disturbances on benthic communities as a whole is essential for predicting the future status of coral reefs and managing them in the light of new environmental regimes. The approach we develop here opens new opportunities for reaching this goal.

Relative rate of change in cognitive score network dynamics via Bayesian hierarchical models reveal spatial patterns of neurodegeneration.

The degeneration of the human brain is a complex process, which often affects certain brain regions due to healthy aging or disease. This degeneration can be evaluated on regions of interest (ROI) in the brain through probabilistic networks and morphological estimates. Current approaches for finding such networks are limited to analyses at discrete neuropsychological stages, which cannot appropriately account for connectivity dynamics over the onset of cognitive deterioration, and morphological changes are seldom unified with connectivity networks, despite known dependencies. To overcome these limitations, a probabilistic wombling model is proposed to simultaneously estimate ROI cortical thickness and covariance networks contingent on rates of change in cognitive decline. This proposed model was applied to analyze longitudinal data from healthy control (HC) and Alzheimer's disease (AD) groups and found connection differences pertaining to regions, which play a crucial role in lasting cognitive impairment, such as the entorhinal area and temporal regions. Moreover, HC cortical thickness estimates were significantly higher than those in the AD group across all ROIs. The analyses presented in this work will help practitioners jointly analyze brain tissue atrophy at the ROI-level conditional on neuropsychological networks, which could potentially allow for more targeted therapeutic interventions.

Detecting Technical Anomalies in High-Frequency Water-Quality Data Using Artificial Neural Networks.

Anomaly detection (AD) in high-volume environmental data requires one to tackle a series of challenges associated with the typical low frequency of anomalous events, the broad-range of possible anomaly types, and local nonstationary environmental conditions, suggesting the need for flexible statistical methods that are able to cope with unbalanced high-volume data problems. Here, we aimed to detect anomalies caused by technical errors in water-quality (turbidity and conductivity) data collected by automated in situ sensors deployed in contrasting riverine and estuarine environments. We first applied a range of artificial neural networks that differed in both learning method and hyperparameter values, then calibrated models using a Bayesian multiobjective optimization procedure, and selected and evaluated the "best" model for each water-quality variable, environment, and anomaly type. We found that semi-supervised classification was better able to detect sudden spikes, sudden shifts, and small sudden spikes, whereas supervised classification had higher accuracy for predicting long-term anomalies associated with drifts and periods of otherwise unexplained high variability.

Evaluating the impact of a small number of areas on spatial estimation.

BACKGROUND: There is an expanding literature on different representations of spatial random effects for different types of spatial correlation structure within the conditional autoregressive class of priors for Bayesian spatial models. However, little is known about the impact of these different priors when the number of areas is small. This paper aimed to investigate this problem both in the context of a case study of spatial analysis of dengue fever and more generally through a simulation study. METHODS: Both the simulation study and the case study considered count data aggregated to a small area level in a region. Five different conditional autoregressive priors for a simple Bayesian Poisson model were considered: independent, Besag-York-Mollié, Leroux, and two variants of a localised clustering model. Data were simulated with eight different sizes of areal grids, ranging from 4 to 2500 areas, and two different levels of both spatial autocorrelation and disease counts. Model goodness-of-fit measures and model estimates were compared. A case study involving dengue fever cases in 14 local areas in Makassar, Indonesia, was also considered. RESULTS: The simulation study showed that model performance varied under different scenarios. When areas had low autocorrelation and high counts, and the number of areas was at most 25, the BYM, Leroux and localised [Formula: see text] models performed similarly and better than the independent and localised [Formula: see text] models. However, when the number of areas were at least 100, all models performed differently, and the Leroux model performed the best. Overall, the Leroux model performed the best for every scenario especially when there were at least 16 areas. Based on the case study, the comparative performance of spatial models may also vary for a small number of areas, especially when the data have a relatively large mean and variance over areas. In this case, the localised model with G = 3 was a better choice. CONCLUSION: Detecting spatial patterns can be difficult when there are very few areas. Understanding the characteristics of the data and the relative influence of alternative conditional autoregressive priors is essential in selecting an appropriate Bayesian spatial model.

Multivariate Bayesian meta-analysis: joint modelling of multiple cancer types using summary statistics.

BACKGROUND: Cancer atlases often provide estimates of cancer incidence, mortality or survival across small areas of a region or country. A recent example of a cancer atlas is the Australian cancer atlas (ACA), that provides interactive maps to visualise spatially smoothed estimates of cancer incidence and survival for 20 different cancer types over 2148 small areas across Australia. METHODS: The present study proposes a multivariate Bayesian meta-analysis model, which can model multiple cancers jointly using summary measures without requiring access to the unit record data. This new approach is illustrated by modelling the publicly available spatially smoothed standardised incidence ratios for multiple cancers in the ACA divided into three groups: common, rare/less common and smoking-related. The multivariate Bayesian meta-analysis models are fitted to each group in order to explore any possible association between the cancers in three remoteness regions: major cities, regional and remote areas across Australia. The correlation between the pairs of cancers included in each multivariate model for a group was examined by computing the posterior correlation matrix for each cancer group in each region. The posterior correlation matrices in different remoteness regions were compared using Jennrich's test of equality of correlation matrices (Jennrich in J Am Stat Assoc. 1970;65(330):904-12. https://doi.org/10.1080/01621459.1970.10481133 ). RESULTS: Substantive correlation was observed among some cancer types. There was evidence that the magnitude of this correlation varied according to remoteness of a region. For example, there has been significant negative correlation between prostate and lung cancer in major cities, but zero correlation found in regional and remote areas for the same pair of cancer types. High risk areas for specific combinations of cancer types were identified and visualised from the proposed model. CONCLUSIONS: Publicly available spatially smoothed disease estimates can be used to explore additional research questions by modelling multiple cancer types jointly. These proposed multivariate meta-analysis models could be useful when unit record data are unavailable because of privacy and confidentiality requirements.