Spatial modeling for correlated cancers using bivariate directed graphs
Year: 2020
Type: article
Abstract: Background: Disease maps are an important tool in cancer epidemiology used for the analysis of geographical variations in disease rates and the investigation of environmental risk factors underlying spatial patterns. Cancer maps help epidemiologists highlight geographic areas with high and low prevalence, incidence, or mortality rates of cancers, and the variability of such rates over a spatial domain. They can also be used to detect “hot-spots” or spatial clusters which may arise due to common environmental, demographic, or cultural effects shared by neighboring regions. Statistical methods for spatial data formulate models to capture spatial autocorrelation and produce cancer maps to better detect clustering and hotspots. When more than one cancer is of interest, the models must also capture the inherent or endemic association between the diseases in addition to the spatial association. This article develops interpretable and easily implementable spatial autocorrelation models for two or more cancers. Methods: The article builds upon recent developments in univariate disease mapping that have shown the use of mathematical structures such as directed acyclic graphs (DAGs) to capture spatial association for a single cancer. The advantage of using DAGs over other existing models is the easier interpretation of spatial association. The current manuscript extends this family of directed acyclic graphical models to estimate inherent or endemic association for two cancers in addition to the association over space (clustering) for each of the cancers. The method builds a Bayesian hierarchical model where the spatial effects are introduced as latent random effects for each cancer. A valid joint probability model is constructed by first modeling the marginal distribution of one disease followed by the second disease conditional on the first. This approach ensures easier interpretation of model parameters and helps to separate the spatial autocorrelation for each cancer from the association between the two cancers. Results: We analyze the relationship between esophagus and lung cancer extracted from the Surveillance, Epidemiology, and End Results (SEER) Program for their incidence rates in the years 2012–2016 across 58 counties in California. Our analysis shows statistically significant association between the county-wide incidence rates of lung and esophagus cancer across California. After accounting for explanatory variables (smoking, age, education, employment, sex, race, health insurance and poverty), esophagus cancer rates exhibit weaker spatial association than lung cancer rates for data counties in California. Conclusions: The bivariate directed acyclic graphical model performs better than competing bivariate spatial models in the existing literature. This improvement is seen both in terms of the model’s fit to the data and complexity of the model.
Source: Annals of Cancer Epidemiology
DOI: 10.21037/ace-19-41
Authors: Leiwen Gao, Sudipto Banerjee, Abhirup Datta
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W2990199412