Augmenting disease maps: a Bayesian meta-analysis approach
Open Access
- 5 August 2020
- journal article
- research article
- Published by The Royal Society in Royal Society Open Science
- Vol. 7 (8)
Abstract
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.Keywords
Funding Information
- Decision Making in the Big Data Era (FL150100150)
This publication has 44 references indexed in Scilit:
- Disease Mapping and Regression with Count Data in the Presence of Overdispersion and Spatial Autocorrelation: A Bayesian Model Averaging ApproachInternational Journal of Environmental Research and Public Health, 2014
- Spatial Autocorrelation of Cancer Incidence in Saudi ArabiaInternational Journal of Environmental Research and Public Health, 2013
- Geographic remoteness and risk of advanced colorectal cancer at diagnosis in Queensland: a multilevel studyBritish Journal of Cancer, 2011
- Identification of area-level influences on regions of high cancer incidence in Queensland, Australia: a classification tree approachBMC Cancer, 2011
- The Simpson's paradox unraveledInternational Journal of Epidemiology, 2011
- Developing the atlas of cancer in Queensland: methodological issuesInternational Journal of Health Geographics, 2011
- A basic introduction to fixed-effect and random-effects models for meta-analysisResearch Synthesis Methods, 2010
- Meta-analysis of standardized incidence and mortality rates of childhood leukaemia in proximity to nuclear facilitiesEuropean Journal of Cancer Care, 2007
- Inference from Iterative Simulation Using Multiple SequencesStatistical Science, 1992
- On Simpson's Paradox and the Sure-Thing PrincipleJournal of the American Statistical Association, 1972