Spatial Distribution of Cancer Mortality in Andalusia, Spain (2002-2013)
Marco Breschi, University of Sassari
Juan Antonio Córdoba, Consejería de Salud de la Junta de Andalucía. Delegación Territorial de Cádiz
Antonio Escolar, Consejería de Salud de la Junta de Andalucía. Delegación Territorial de Cádiz
Lucia Pozzi, University of Sassari
Rebeca Ramis, Instituto de Salud Carlos III
Vanessa Santos, University of Sassari
Francisco Viciana, Junta de Andalucia
Several studies show the association between socioeconomic inequalities and cancer, increasing mortality in groups with lower socioeconomic levels.
The aim is to study the spatial pattern of all-cause cancer mortality in Andalusia and its association with the deprivation index of the census tract between 2002-2013.
An ecological study of small area is carried out for the 5,381 Andalusian census tracts.
Mortality data and person-years have been obtained from the Longitudinal Statistics of Survival and Longevity in Andalusia (Institute of Statistics and Cartography of Andalusia). We use deprivation index created by Escolar, classified in five levels.
We calculate standardized mortality ratios and smoothed relative risks (RR) by census tract with their corresponding 95% credibility intervals. Posterior probabilties of what smoothed RR were greater than 1 are provided.
For map plotting purposes, smoothed RR were calculated using the conditional autoregressive model proposed by Besag, York and Mollie.
Deprivation index has been included in the models as covariate.
Integrated nested Laplace approximations (INLA) were used as a tool for Bayesian inference through the R-INLA package.
In both sexes, mortality maps show a high-risk pattern in the provinces of Cadiz, Huelva and Seville for some causes, among them, lung in men and breast in women.
A social gradient is observed for men in oral cavity and pharynx, colorectal, esophagus, stomach, liver, larynx, lung and kidney, and for women in stomach, breast, lung and gallbladder. Most models adjusts better including covariate, according to the deviance information criterion (DIC).
For many causes, the existing historical mortality pattern is maintained, as well as the association between mortality and deprivation index.
Presented in Session 47: Spatial Analysis of Mortality