Developing the atlas of cancer in Queensland: methodological issues

ACEMID cohort study: protocol of a prospective cohort study using 3D total body photography for melanoma imaging and diagnosis

INTRODUCTION

Three-dimensional (3D) total body photography may improve early detection of melanoma and facilitate surveillance, leading to better prognosis and lower healthcare costs. The Australian Centre of Excellence in Melanoma Imaging and Diagnosis (ACEMID) cohort study will assess long-term outcomes from delivery of a precision strategy of monitoring skin lesions using skin surface imaging technology embedded into health services across Australia.

METHODS AND ANALYSIS

A prospective cohort study will enrol 15 000 participants aged 18 years and above, across 15 Australian sites. Participants will attend study visits according to their melanoma risk category: very high risk, high risk or low/average risk, every 6, 12 and 24 months, respectively, over 3 years. Participants will undergo 3D total body photography and dermoscopy imaging at study visits. A baseline questionnaire will be administered to collect sociodemographic, phenotypic, quality of life and sun behaviour data. A follow-up questionnaire will be administered every 12 months to obtain changes in sun behaviour and quality of life. A saliva sample will be collected at the baseline visit from a subsample.

ETHICS AND DISSEMINATION

The ACEMID cohort study was approved by the Metro South Health Human Research Ethics Committee (approval number: HREC/2019/QMS/57206) and the University of Queensland Human Research Ethics Committee (approval number: 2019003077). The findings will be reported through peer-reviewed and lay publications and presentations at conferences.

TRIAL REGISTRATION NUMBER

ACTRN12619001706167.

Mapping the Morbidity Risk Associated with Coal Mining in Queensland, Australia

The populations in the vicinity of surface coal mining activities have a higher risk of morbidity due to diseases, such as cardiovascular, respiratory and hypertensive diseases, as well as cancer and diabetes mellitus. Despite the large and historical volume of coal production in Queensland, the main Australian coal mining state, there is little research on the association of coal mining exposures with morbidity in non-occupational populations in this region. This study explored the association of coal production (Gross Raw Output—GRO) with hospitalisations due to six disease groups in Queensland using a Bayesian spatial hierarchical analysis and considering the spatial distribution of the Local Government Areas (LGAs). There is a positive association of GRO with hospitalisations due to circulatory diseases (1.022, 99% CI: 1.002–1.043) and respiratory diseases (1.031, 95% CI: 1.001–1.062) for the whole of Queensland. A higher risk of circulatory, respiratory and chronic lower respiratory diseases is found in LGAs in northwest and central Queensland; and a higher risk of hypertensive diseases, diabetes mellitus and lung cancer is found in LGAs in north, west, and north and southeast Queensland, respectively. These findings can be used to support public health strategies to protect communities at risk. Further research is needed to identify the causal links between coal mining and morbidity in non-occupational populations in Queensland.

Assessing the effect of environmental and socio-economic factors on skin melanoma incidence: an island-wide spatial study in Gran Canaria (Spain), 2007-2018

INTRODUCTION

Skin melanoma incidence has risen in the last decades becoming a major public health problem in many regions of the world. Geographic variation of rates is not well understood.

PURPOSE

To assess the spatial distribution of skin melanoma in Gran Canaria Island (Canary Islands, Spain) and to evaluate the role of environmental, socio-economic, and demographic factors in this distribution.

METHODS

We performed a small-area study with disease mapping at the census-tract level (CT) in Gran Canaria between 2007 and 2018. After testing for spatial autocorrelation, we integrated individual-level health data with census-based demographic and socio-economic indicators, and satellite-based environmental data. Finally, we assessed the role of demographic, socio-economic and environmental factors on skin melanoma incidence using a Bayesian analytical framework, with options for non-spatial and spatial random effects.

RESULTS

1058 patients were diagnosed with invasive skin melanoma in the study period and geolocated to a CT (number of CT in Gran Canaria = 565). We found evidence of global spatial autocorrelation in skin melanoma incidence (Moran's I = 0.09, pseudo p-value = 0.001). A few hotspots were detected, fundamentally in urban northern tracts. A radial pattern of high values was also observed in selected ravines with historical isolation. Multivariable conditional autoregressive models identified urbanicity, percent of females, and a high socio-economic status as risk factors for disease. Solar radiation did not show a significant role.

CONCLUSION

Urbanicity and a high socio-economic status were identified as the main risk factors for skin melanoma. These associations might reflect differential melanoma susceptibilities or be explained by health inequalities in detection. This study also uncovered high-risk areas in particular ravines. Future targeted research in these regions might help better understand the role of genetic and toxic factors in melanoma pathogenesis.

Environmental and sociodemographic risk factors associated with environmentally transmitted zoonoses hospitalisations in Queensland, Australia

Zoonoses impart a significant public health burden in Australia particularly in Queensland, a state with increasing environmental stress due to extreme weather events and rapid expansion of agriculture and urban developments. Depending on the organism and the environment, a proportion of zoonotic pathogens may survive from hours to years outside the animal host and contaminate the air, water, food, or inanimate objects facilitating their transmission through the environment (i.e. environmentally transmitted). Although most of these zoonotic infections are asymptomatic, severe cases that require hospitalisation are an important indicator of zoonotic infection risk. To date, no studies have investigated the risk of hospitalisation due to environmentally transmitted zoonotic diseases and its association with proxies of sociodemographic and environmental stress. In this study we analysed hospitalisation data for a group of environmentally transmitted zoonoses during a 15-year period using a Bayesian spatial hierarchical model. The analysis incorporated the longest intercensal-year period of consistent Local Government Area (LGA) boundaries in Queensland (1996-2010). Our results showed an increased risk of environmentally transmitted zoonoses hospitalisation in people in occupations such as animal farming, and hunting and trapping animals in natural habitats. This risk was higher in females, compared to the general population. Spatially, the higher risk was in a discrete set of north-eastern, central and southern LGAs of the state, and a probability of 1.5-fold or more risk was identified in two separate LGA clusters in the northeast and south of the state. The increased risk of environmentally transmitted zoonoses hospitalisations in some LGAs indicates that the morbidity due these diseases can be partly attributed to spatial variations in sociodemographic and occupational risk factors in Queensland. The identified high-risk areas can be prioritised for health support and zoonosis control strategies in Queensland.

Modelling the Relationship between Rainfall and Mental Health Using Different Spatial and Temporal Units

Drought is thought to impact upon the mental health of agricultural communities, but studies of this relationship have reported inconsistent results. A source of inconsistency could be the aggregation of data by a single spatiotemporal unit of analysis, which induces the modifiable areal and temporal unit problems. To investigate this, mental health-related emergency department (MHED) presentations among residents of the Wheat Belt region of Western Australia, between 2002 and 2017, were examined. Average daily rainfall was used as a measure of drought. Associations between MHED presentations and rainfall were estimated based on various spatial aggregations of underlying data, at multiple temporal windows. Wide variation amongst results was observed. Despite this, two key features were found: Associations between MHED presentations and rainfall were generally positive when rainfall was measured in summer months (rate ratios up to 1.05 per 0.5 mm of daily rainfall) and generally negative when rainfall was measured in winter months (rate ratios as low as 0.96 per 0.5 mm of daily rainfall). These results demonstrate that the association between drought and mental health is quantifiable; however, the effect size is small and varies depending on the spatial and temporal arrangement of the underlying data. To improve understanding of this association, more studies should be undertaken with longer time spans and examining specific mental health outcomes, using a wide variety of spatiotemporal units.

Augmenting disease maps: a Bayesian meta-analysis approach

Analysis of spatial patterns of disease is a significant field of research. However, access to unit-level disease data can be difficult for privacy and other reasons. As a consequence, estimates of interest are often published at the small area level as disease maps. This motivates the development of methods for analysis of these ecological estimates directly. Such analyses can widen the scope of research by drawing more insights from published disease maps or atlases. The present study proposes a hierarchical Bayesian meta-analysis model that analyses the point and interval estimates from an online atlas. The proposed model is illustrated by modelling the published cancer incidence estimates available as part of the online Australian Cancer Atlas (ACA). The proposed model aims to reveal patterns of cancer incidence for the 20 cancers included in ACA in major cities, regional and remote areas. The model results are validated using the observed areal data created from unit-level data on cancer incidence in each of 2148 small areas. It is found that the meta-analysis models can generate similar patterns of cancer incidence based on urban/rural status of small areas compared with those already known or revealed by the analysis of observed data. The proposed approach can be generalized to other online disease maps and atlases.

Bayesian disease mapping and the 'High-Risk' oral cancer population in Hong Kong

BACKGROUND

Preventive and early diagnostic methods such as health promotion and disease screening are increasingly advocated to improve detection and survival rates for oral cancer. These strategies are most effective when targeted at "high-risk" individuals and populations. Bayesian disease-mapping modelling is a statistical method to quantify and explain spatial and temporal patterns for risk and covariate factor influence, thereby identifying "high-risk" sub-regions or "case clustering" for targeted intervention. Rarely applied to oral cancer epidemiology, this paper highlights the efficacy of disease mapping for the Hong Kong population.

METHODS

Following ethical approval, anonymized individual-level data for oral cancer diagnoses were obtained retrospectively from the Clinical Data Analysis and Reporting System (CDARS) of the Hong Kong Hospital Authority (HA) database for a 7-year period (January 2013 to December 2019). Data facilitated disease mapping and estimation of relative risks of oral cancer incidence and mortality.

RESULTS

A total of 3,341 new oral cancer cases and 1,506 oral cancer-related deaths were recorded during the 7-year study period. Five districts, located in Hong Kong Island and Kowloon, exhibited considerably higher relative incidence risks with 1 significant "case cluster" hotspot. Six districts displayed higher mortality risks than expected from territory-wide values, with highest risk identified for two districts of Hong Kong Island.

CONCLUSION

Bayesian disease mapping is successful in identifying and characterizing "high-risk" areas for oral cancer incidence and mortality within a community. This should facilitate targeted preventive and interventional strategies. Further work is encouraged to enhance global-level data and comprehensive mapping of oral cancer incidence, mortality and survival.

Development of the Australian Cancer Atlas: spatial modelling, visualisation, and reporting of estimates

BACKGROUND

It is well known that the burden caused by cancer can vary geographically, which may relate to differences in health, economics or lifestyle. However, to date, there was no comprehensive picture of how the cancer burden, measured by cancer incidence and survival, varied by small geographical area across Australia.

METHODS

The Atlas consists of 2148 Statistical Areas level 2 across Australia defined by the Australian Statistical Geography Standard which provide the best compromise between small population and small area. Cancer burden was estimated for males, females, and persons separately, with 50 unique sex-specific (males, females, all persons) cancer types analysed. Incidence and relative survival were modelled with Bayesian spatial models using the Leroux prior which was carefully selected to provide adequate spatial smoothing while reflecting genuine geographic variation. Markov Chain Monte Carlo estimation was used because it facilitates quantifying the uncertainty of the posterior estimates numerically and visually.

RESULTS

The results of the statistical model and visualisation development were published through the release of the Australian Cancer Atlas ( https://atlas.cancer.org.au ) in September, 2018. The Australian Cancer Atlas provides the first freely available, digital, interactive picture of cancer incidence and survival at the small geographical level across Australia with a focus on incorporating uncertainty, while also providing the tools necessary for accurate estimation and appropriate interpretation and decision making.

CONCLUSIONS

The success of the Atlas will be measured by how widely it is used by key stakeholders to guide research and inform decision making. It is hoped that the Atlas and the methodology behind it motivates new research opportunities that lead to improvements in our understanding of the geographical patterns of cancer burden, possible causes or risk factors, and the reasons for differences in variation between cancer types, both within Australia and globally. Future versions of the Atlas are planned to include new data sources to include indicators such as cancer screening and treatment, and extensions to the statistical methods to incorporate changes in geographical patterns over time.

Area-Level Walkability and the Geographic Distribution of High Body Mass in Sydney, Australia: A Spatial Analysis Using the 45 and Up Study

Improving the walkability of built environments to promote healthy lifestyles and reduce high body mass is increasingly considered in regional development plans. Walkability indexes have the potential to inform, benchmark and monitor these plans if they are associated with variation in body mass outcomes at spatial scales used for health and urban planning. We assessed relationships between area-level walkability and prevalence and geographic variation in overweight and obesity using an Australian population-based cohort comprising 92,157 Sydney respondents to the 45 and Up Study baseline survey between January 2006 and April 2009. Individual-level data on overweight and obesity were aggregated to 2006 Australian postal areas and analysed as a function of area-level Sydney Walkability Index quartiles using conditional auto regression spatial models adjusted for demographic, social, economic, health and socioeconomic factors. Both overweight and obesity were highly clustered with higher-than-expected prevalence concentrated in the urban sprawl region of western Sydney, and lower-than-expected prevalence in central and eastern Sydney. In fully adjusted spatial models, prevalence of overweight and obesity was 6% and 11% lower in medium-high versus low, and 10% and 15% lower in high versus low walkability postcodes, respectively. Postal area walkability explained approximately 20% and 9% of the excess spatial variation in overweight and obesity that remained after accounting for other individual- and area-level factors. These findings provide support for the potential of area-level walkability indexes to inform, benchmark and monitor regional plans aimed at targeted approaches to reducing population-levels of high body mass through environmental interventions. Future research should consider potential confounding due to neighbourhood self-selection on area-level walkability relations.

[The cartographic depiction of regional variation in morbidity : Data analysis options using the example of the small-scale cancer atlas for Schleswig-Holstein]

The cancer registry in Germany collects area-wide small-area data that can be presented in themes (disease mapping). Because of the occurrence of random extreme values of rates, mapping without prior spatial-statistical data analysis is problematic from a methodological and risk-communicative viewpoint - the extreme values easily mislead the card reader and obscure actual spatial patterns.The problem of data instability can generally be met by aggregation or by smoothing. The cancer atlas for Schleswig-Holstein is based on data from 1142 municipalities (median population: 721) for the diagnostic years 2001-2010. Maps for incidence (as a standardized incidence ratio), mortality (as a standardized mortality ratio), and relative survival (as a relative excess risk) were smoothed by using a Bayesian method (BYM model). The maps show that spatial differences can be made visible by smoothing.Data aggregation is the methodically simpler way, but means a loss of information. The atlas shows that small-scale mapping is possible while preserving the entire spatial information. The method of smoothing is complex, but useful for generating hypotheses. The spatial patterns found are complex, difficult to interpret, and require the collaboration of specialists from different professions, because of the diverse influencing factors (data collection, lifestyle factors, early detection, risk factors, etc.). The effort required to explain the methodology in a language easy to understand should not be underestimated.

Cancer mortality rates and spillover effects among different areas: A case study in Campania (southern Italy)

The present study analyses the spatial distribution of cancer mortality rates in Campania (an Italian region with the highest population density), in which residents in several areas are exposed to major environmental health hazards. The paper has the methodological aims of verifying the existence, or otherwise, of a spatial correlation between mortality from different types of cancer and the occurrence of some specific area characteristics, using both Bayesian statistics and spatial econometrics. We show that the use of the Spatial Empirical Bayes Smoothed Rate, instead of the more commonly used Raw Rate, allows a more comprehensive analysis of the mortality rate, highlighting the existence of different cluster sizes throughout the region, according to the type of cancer mortality rate analysed. By using a Spatial Durbin model we verify that cancer mortality rates are related to the environmental characteristics of specific areas with spatial spillover effects. Our results validate the hypothesis that living along the coast by Mt Vesuvius and, to a lesser extent, along the Domitio-Flegreo coast NW of Naples and in more urbanised municipalities, increases the risk of dying of cancer. By contrast, living in less urbanised municipalities, with the presence of natural and historical attractions, has a positive effect on the residents' health, reducing their risk of disease. In both cases significant spillover effects (negative and positive) are found in municipalities close to the areas in question. Despite a number of reasonable limitations, our findings may provide useful information support for policy makers to foster knowledge, awareness and informed participation of citizens.

Does Walkability Contribute to Geographic Variation in Psychosocial Distress? A Spatial Analysis of 91,142 Members of the 45 and Up Study in Sydney, Australia

Walkability describes the capacity of the built environment to promote walking, and has been proposed as a potential focus for community-level mental health planning. We evaluated this possibility by examining the contribution of area-level walkability to variation in psychosocial distress in a population cohort at spatial scales comparable to those used for regional planning in Sydney, Australia. Data on psychosocial distress were analysed for 91,142 respondents to the 45 and Up Study baseline survey between January 2006 and April 2009. We fit conditional auto regression models at the postal area level to obtain smoothed “disease maps” for psychosocial distress, and assess its association with area-level walkability after adjusting for individual- and area-level factors. Prevalence of psychosocial distress was 7.8%; similar for low (7.9%), low-medium (7.9%), medium-high (8.0%), and high (7.4%) walkability areas; and decreased with reducing postal area socioeconomic disadvantage: 12.2% (most), 9.3%, 7.5%, 5.9%, and 4.7% (least). Unadjusted disease maps indicated strong geographic clustering of psychosocial distress with 99.0% of excess prevalence due to unobserved and spatially structured factors, which was reduced to 55.3% in fully adjusted maps. Spatial and unstructured variance decreased by 97.3% and 39.8% after adjusting for individual-level factors, and another 2.3% and 4.2% with the inclusions of area-level factors. Excess prevalence of psychosocial distress in postal areas was attenuated in adjusted models but remained spatially structured. Postal area prevalence of high psychosocial distress is geographically clustered in Sydney, but is unrelated to postal area walkability. Area-level socioeconomic disadvantage makes a small contribution to this spatial structure; however, community-level mental health planning will likely deliver greatest benefits by focusing on individual-level contributors to disease burden and inequality associated with psychosocial distress.

Correlates of fatality risk of vulnerable road users in Delhi

Pedestrians, cyclists, and users of motorised two-wheelers account for more than 85% of all the road fatality victims in Delhi. The three categories are often referred to as vulnerable road users (VRUs). Using Bayesian hierarchical approach with a Poisson-lognormal regression model, we present spatial analysis of road fatalities of VRUs with wards as areal units. The model accounts for spatially uncorrelated as well as correlated error. The explanatory variables include demographic factors, traffic characteristics, as well as built environment features. We found that fatality risk has a negative association with socio-economic status (literacy rate), population density, and number of roundabouts, and has a positive association with percentage of population as workers, number of bus stops, number of flyovers (grade separators), and vehicle kilometers travelled. The negative effect of roundabouts, though statistically insignificant, is in accordance with their speed calming effects for which they have been used to replace signalised junctions in various parts of the world. Fatality risk is 80% higher at the density of 50 persons per hectare (pph) than at overall city-wide density of 250 pph. The presence of a flyover increases the relative risk by 15% compared to no flyover. Future studies should investigate the causal mechanism through which denser neighborhoods become safer. Given the risk posed by flyovers, their use as congestion mitigation measure should be discontinued within urban areas.

The contribution of area-level walkability to geographic variation in physical activity: a spatial analysis of 95,837 participants from the 45 and Up Study living in Sydney, Australia

BACKGROUND

Individual-level studies support a positive relation between walkable built environments and participation in moderate-intensity walking. However, the utility of this evidence for population-level planning is less clear as it is derived at much finer spatial scales than those used for regional programming. The aims of this study were to: evaluate if individual-level relations between walkability and walking to improve health manifest at population-level spatial scales; assess the specificity of area-level walkability for walking relative to other moderate and vigorous physical activity (MVPA); describe geographic variation in walking and other MVPA; and quantify the contribution of walkability to this variation.

METHODS

Data on sufficient walking, sufficient MVPA, and high MVPA to improve health were analyzed for 95,837 Sydney respondents to the baseline survey of the 45 and Up Study between January 2006 and April 2010. We used conditional autoregressive models to create smoothed MVPA "disease maps" and assess relations between sufficient MVPA to improve health and area-level walkability adjusted for individual-level demographic, socioeconomic, and health factors, and area-level relative socioeconomic disadvantage.

RESULTS

Within-cohort prevalence of meeting recommendations for sufficient walking, sufficient MVPA, and high MVPA were 31.7 (95% CI 31.4-32.0), 69.4 (95% CI 69.1-69.7), and 56.1 (95% CI 55.8-56.4) percent. Prevalence of sufficient walking was increased by 1.20 (95% CrI 1.12-1.29) and 1.07 (95% CrI 1.01-1.13) for high and medium-high versus low walkability postal areas, and for sufficient MVPA by 1.05 (95% CrI 1.01-1.08) for high versus low walkability postal areas. Walkability was not related to high MVPA. Postal area walkability explained 65.8 and 47.4 percent of residual geographic variation in sufficient walking and sufficient MVPA not attributable to individual-level factors.

CONCLUSIONS

Walkability is associated with area-level prevalence and geographic variation in sufficient walking and sufficient MVPA to improve health in Sydney, Australia. Our study supports the use of walkability indexes at multiple spatial scales for informing population-level action to increase physical activity and the utility of spatial analysis for walkability research and planning.

A flexible parametric approach to examining spatial variation in relative survival

Most of the few published models used to obtain small-area estimates of relative survival are based on a generalized linear model with piecewise constant hazards under a Bayesian formulation. Limitations of these models include the need to artificially split the time scale, restricted ability to include continuous covariates, and limited predictive capacity. Here, an alternative Bayesian approach is proposed: a spatial flexible parametric relative survival model. This overcomes previous limitations by combining the benefits of flexible parametric models: the smooth, well-fitting baseline hazard functions and predictive ability, with the Bayesian benefits of robust and reliable small-area estimates. Both spatially structured and unstructured frailty components are included. Spatial smoothing is conducted using the intrinsic conditional autoregressive prior. The model was applied to breast, colorectal, and lung cancer data from the Queensland Cancer Registry across 478 geographical areas. Advantages of this approach include the ease of including more realistic complexity, the feasibility of using individual-level input data, and the capacity to conduct overall, cause-specific, and relative survival analysis within the same framework. Spatial flexible parametric survival models have great potential for exploring small-area survival inequalities, and we hope to stimulate further use of these models within wider contexts. Copyright © 2016 John Wiley & Sons, Ltd.

Spatially Enabling the Health Sector

Spatial information describes the physical location of either people or objects, and the measured relationships between them. In this article, we offer the view that greater utilization of spatial information and its related technology, as part of a broader redesign of the architecture of health information at local and national levels, could assist and speed up the process of health reform, which is taking place across the globe in richer and poorer countries alike. In making this point, we describe the impetus for health sector reform, recent developments in spatial information and analytics, and current Australasian spatial health research. We highlight examples of uptake of spatial information by the health sector, as well as missed opportunities. Our recommendations to spatially enable the health sector are applicable to high- and low-resource settings.

Spatial variations in cervical cancer prevention in Colombia: Geographical differences and associated socio-demographic factors

We examined spatial variations in the frequencies of women who had not heard of human papillomavirus vaccination (NHrd-Vac) and who had not had Pap testing (NHd-Pap) among Colombian administrative divisions (departments), before and after considering differences in socio-demographic factors. Following global and local tests for clustering, Bayesian Poisson hierarchical models identified department factors associated with NHrd-Vac and NHd-Pap, as well as the extent of the spatially structured and unstructured heterogeneity. Models of spatial variations for both outcomes included the department percentage of women with subsidised health insurance. The relative risks of NHrd-Vac and NHd-Pap were highest in several departments adjacent to the Colombian border. Our finding that the risk of not having adequate access to cervical cancer (CC) prevention programmes in Colombia was location-dependent, could be used to focus resources for CC prevention programmes.

Maps and atlases of cancer mortality: a review of a useful tool to trigger new questions

In this review we illustrate our view on the epidemiological relevance of geographically mapping cancer mortality. In the first part of this work, after delineating the history of cancer mapping with a view on interpretation of Cancer Mortality Atlases, we briefly illustrate the 'art' of cancer mapping. Later we summarise in a non-mathematical way basic methods of spatial statistics. In the second part of this paper, we employ the 'Atlas of Cancer Mortality in the European Union and the European Economic Area 1993-1997' in order to illustrate spatial aspects of cancer mortality in Europe. In particular, we focus on the cancer related to tobacco and alcohol epidemics and on breast cancer which is of particular interest in cancer mapping. Here we suggest and reiterate two key concepts. The first is that a cancer atlas is not only a visual tool, but it also contain appropriate spatial statistical analyses that quantify the qualitative visual impressions to the readers even though at times revealing fallacy. The second is that a cancer atlas is by no means a book where answers to questions can be found. On the contrary, it ought to be considered as a tool to trigger new questions.

Spatially Varying Coefficient Inequalities: Evaluating How the Impact of Patient Characteristics on Breast Cancer Survival Varies by Location

An increasing number of studies have identified spatial differences in breast cancer survival. However little is known about whether the structure and dynamics of this spatial inequality are consistent across a region. This study aims to evaluate the spatially varying nature of predictors of spatial inequality in relative survival for women diagnosed with breast cancer across Queensland, Australia. All Queensland women aged less than 90 years diagnosed with invasive breast cancer from 1997 to 2007 and followed up to the end of 2008 were extracted from linked Queensland Cancer Registry and BreastScreen Queensland data. Bayesian relative survival models were fitted using various model structures (a spatial regression model, a varying coefficient model and a finite mixture of regressions model) to evaluate the relative excess risk of breast cancer, with the use of Markov chain Monte Carlo computation. The spatially varying coefficient models revealed that some covariate effects may not be constant across the geographic regions of the study. The overall spatial patterns showed lower survival among women living in more remote areas, and higher survival among the urbanised south-east corner. Notwithstanding this, the spatial survival pattern for younger women contrasted with that for older women as well as single women. This complex spatial interplay may be indicative of different factors impacting on survival patterns for these women.

Spatio-Temporal Pattern of Breast Cancer - Case Study of Southern Karnataka, India

INTRODUCTION

Spatio-ecological study of disease provides a framework to study the interaction of genetic, environmental, social, cultural and behavioural factors on people's health. The occurrence and interaction of these factors are different in different places, giving rise to distinct geographic or spatial variation. Diseases like breast cancer have variation both spatially and temporally. Public health practitioners can use Geographic Information System (GIS) as a visualization tool to effectively present geographic phenomenon and depict it in maps that might remain otherwise undiscovered in tabular form.

AIM

To demonstrate how GIS can be used to understand and communicate breast cancer data through spatial visualization techniques.

OBJECTIVES

(i) To visualize the Spatial Distribution of Breast cancer incidences by a point map. (ii) To visualize the Temporal distribution of breast cancer incidences by thematic maps for the study period of 2007 -2011.

MATERIALS AND METHODS

Total 1090 breast cancer case records collected for the year 2007-2012 were segregated taluk wise for the 29 taluks and geocoded using the address of the patient, creating a point map. ArcGIS 10.2 software was used to prepare thematic map of breast cancer cases. The taluk wise aggregated breast cancer incidence from the year 2007 to 2011 was then attributed into polygon map representing taluks (Base Map). Natural break data classification technique was used to classify the breast cancer incidence data and breast cancer incidences were classified as low, moderate, high and very high.

RESULTS

Spatial distribution of breast cancer incidences using thematic mapping methods high incidences were reported in MY_ T24 (Hunsur), MY_ T25 (KR Nagar), MY_27 (Nanjangud), CH_T1 (Chamrajnagar) and CH-T2 (Gundlupet). Temporal maps prepared for the study from 2007 to 2011 showed that Mysore Taluk had very high Incidence level and the same was observed throughout the study period. The taluks which have high and moderate intensities seem to be fluctuating. However, 25 taluks do not fall into very high category during the study period. Taluks such Gundlupet (CH_T2), K R Nagar (MY_T25), Kollegal (CH_T3) have been observed to enter high intensity category during the year 2011 from moderate intensity. It is also observed that Nanjangud (MY_T27) is in high intensity category throughout the study period which might be due to its proximity to Mysore urban.

CONCLUSION

Analysis of Breast Cancer in southern Karnataka using GIS has revealed that urban areas of Mysore has the highest risk of breast cancer and the temporal trends reveal that even rural areas with moderate risk are moving towards high risk areas.

Missing in space: an evaluation of imputation methods for missing data in spatial analysis of risk factors for type II diabetes

Background

Spatial analysis is increasingly important for identifying modifiable geographic risk factors for disease. However, spatial health data from surveys are often incomplete, ranging from missing data for only a few variables, to missing data for many variables. For spatial analyses of health outcomes, selection of an appropriate imputation method is critical in order to produce the most accurate inferences.

Methods

We present a cross-validation approach to select between three imputation methods for health survey data with correlated lifestyle covariates, using as a case study, type II diabetes mellitus (DM II) risk across 71 Queensland Local Government Areas (LGAs). We compare the accuracy of mean imputation to imputation using multivariate normal and conditional autoregressive prior distributions.

Results

Choice of imputation method depends upon the application and is not necessarily the most complex method. Mean imputation was selected as the most accurate method in this application.

Conclusions

Selecting an appropriate imputation method for health survey data, after accounting for spatial correlation and correlation between covariates, allows more complete analysis of geographic risk factors for disease with more confidence in the results to inform public policy decision-making.

Electronic supplementary material

The online version of this article (doi:10.1186/1476-072X-13-47) contains supplementary material, which is available to authorized users.

Comparing multilevel and Bayesian spatial random effects survival models to assess geographical inequalities in colorectal cancer survival: a case study

BACKGROUND

Multilevel and spatial models are being increasingly used to obtain substantive information on area-level inequalities in cancer survival. Multilevel models assume independent geographical areas, whereas spatial models explicitly incorporate geographical correlation, often via a conditional autoregressive prior. However the relative merits of these methods for large population-based studies have not been explored. Using a case-study approach, we report on the implications of using multilevel and spatial survival models to study geographical inequalities in all-cause survival.

METHODS

Multilevel discrete-time and Bayesian spatial survival models were used to study geographical inequalities in all-cause survival for a population-based colorectal cancer cohort of 22,727 cases aged 20-84 years diagnosed during 1997-2007 from Queensland, Australia.

RESULTS

Both approaches were viable on this large dataset, and produced similar estimates of the fixed effects. After adding area-level covariates, the between-area variability in survival using multilevel discrete-time models was no longer significant. Spatial inequalities in survival were also markedly reduced after adjusting for aggregated area-level covariates. Only the multilevel approach however, provided an estimation of the contribution of geographical variation to the total variation in survival between individual patients.

CONCLUSIONS

With little difference observed between the two approaches in the estimation of fixed effects, multilevel models should be favored if there is a clear hierarchical data structure and measuring the independent impact of individual- and area-level effects on survival differences is of primary interest. Bayesian spatial analyses may be preferred if spatial correlation between areas is important and if the priority is to assess small-area variations in survival and map spatial patterns. Both approaches can be readily fitted to geographically enabled survival data from international settings.

A brief review of spatial analysis concepts and tools used for mapping, containment and risk modelling of infectious diseases and other illnesses

Fast response and decision making about containment, management, eradication and prevention of diseases, are increasingly important aspects of the work of public health officers and medical providers. Diseases and the agents causing them are spatially and temporally distributed, and effective countermeasures rely on methods that can timely locate the foci of infection, predict the distribution of illnesses and their causes, and evaluate the likelihood of epidemics. These methods require the use of large datasets from ecology, microbiology, health and environmental geography. Geodatabases integrating data from multiple sets of information are managed within the frame of geographic information systems (GIS). Many GIS software packages can be used with minimal training to query, map, analyse and interpret the data. In combination with other statistical or modelling software, predictive and spatio-temporal modelling can be carried out. This paper reviews some of the concepts and tools used in epidemiology and parasitology. The purpose of this review is to provide public health officers with the critical tools to decide about spatial analysis resources and the architecture for the prevention and surveillance systems best suited to their situations.

How to interpret and choose a Bayesian spatial model and a Poisson regression model in the context of describing small area cancer risks variations

BACKGROUND

The statistical Bayesian approach is widely used in disease mapping and Poisson regression. Results differ depending on the underlying hypothesis. Our objective is to give a comprehensive presentation of the tools that can be used to interpret results and choose between the different hypotheses. Data from the Isere cancer registry (France) illustrate this presentation.

METHOD

We consider, first, Bayesian models for disease mapping. Classic heterogeneity (Potthoff-Whithinghill statistic) and spatial autocorrelation tests (Moran statistic) of the SIRs, the DIC criteria of the different Bayesian models and finally the comparison of the empirical variance of the unstructured and structured heterogeneity components of the BYM model are considered. The last two criteria are considered for Bayesian Poisson regression including a covariate. Mapping the components of the BYM model with a covariate is also considered.

RESULTS

Four cancer sites (prostate, lung, colon-rectum and bladder) in men diagnosed during the 1998-2007 period are used to illustrate our presentation. We show that the different criteria used to interpret and to choose a model give coherent results.

CONCLUSION

A relevant interpretation of results is a necessary step in choosing the best-adapted Bayesian model. This choice is easy to make with criteria such as the DIC. The comparison of the empirical variance of the unstructured and structured heterogeneity components of the BYM model is also informative.

The Spatial Distribution of Cancer Incidence in Fars Province: A GIS-Based Analysis of Cancer Registry Data

BACKGROUND

Cancer is a major health problem in the developing countries. Variations of its incidence rate among geographical areas are due to various contributing factors. This study was performed to assess the spatial patterns of cancer incidence in the Fars Province, based on cancer registry data and to determine geographical clusters.

METHODS

In this cross sectional study, the new cases of cancer were recorded from 2001 to 2009. Crude incidence rate was estimated based on age groups and sex in the counties of the Fars Province. Age-standardized incidence rates (ASR) per 100,000 was calculated in each year. Spatial autocorrelation analysis was performed in measuring the geographic patterns and clusters using geographic information system (GIS). Also, comparisons were made between ASRs in each county.

RESULTS

A total of 28,411 new cases were diagnosed with cancer during 2001-2009 in the Fars Province, 55.5% of which were men. The average age was 61.6 ± 0.5 years. The highest ASR was observed in Shiraz, which is the largest county in Fars. The Moran's Index of cancer was significantly clustered in 2004, 2005, and 2006 in total, men, and women. The type of spatial clustering was high-high cluster, that to indicate from north-west to south-east of Fars Province.

CONCLUSIONS

Analysis of the spatial distribution of cancer shows significant differences from year to year and between different areas. However, a clear spatial autocorrelation is observed, which can be of great interest and importance to researchers for future epidemiological studies, and to policymakers for applying preventive measures.

Identification of area-level influences on regions of high cancer incidence in Queensland, Australia: a classification tree approach

Background

Strategies for cancer reduction and management are targeted at both individual and area levels. Area-level strategies require careful understanding of geographic differences in cancer incidence, in particular the association with factors such as socioeconomic status, ethnicity and accessibility. This study aimed to identify the complex interplay of area-level factors associated with high area-specific incidence of Australian priority cancers using a classification and regression tree (CART) approach.

Methods

Area-specific smoothed standardised incidence ratios were estimated for priority-area cancers across 478 statistical local areas in Queensland, Australia (1998-2007, n = 186,075). For those cancers with significant spatial variation, CART models were used to identify whether area-level accessibility, socioeconomic status and ethnicity were associated with high area-specific incidence.

Results

The accessibility of a person's residence had the most consistent association with the risk of cancer diagnosis across the specific cancers. Many cancers were likely to have high incidence in more urban areas, although male lung cancer and cervical cancer tended to have high incidence in more remote areas. The impact of socioeconomic status and ethnicity on these associations differed by type of cancer.

Conclusions

These results highlight the complex interactions between accessibility, socioeconomic status and ethnicity in determining cancer incidence risk.