Premium
Spatial variability in relative survival from female breast cancer
Author(s) -
Saez Marc,
Barceló Maria Antònia,
Martos Carmen,
Saurina Carme,
MarcosGragera Rafael,
Renart Gemma,
OcañaRiola Ricardo,
Feja Cristina,
Alcalá Tomás
Publication year - 2012
Publication title -
journal of the royal statistical society: series a (statistics in society)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.103
H-Index - 84
eISSN - 1467-985X
pISSN - 0964-1998
DOI - 10.1111/j.1467-985x.2011.00720.x
Subject(s) - relative survival , relative risk , breast cancer , statistics , bayesian hierarchical modeling , estimation , survival analysis , bayesian probability , efficiency , demography , cancer , population , econometrics , geography , medicine , cancer registry , bayes' theorem , mathematics , environmental health , confidence interval , sociology , management , estimator , economics
Summary. Relative survival is a measure of survival corrected for the effect of independent causes of death other than the cancer of interest. It is provided by population‐based cancer registries and constitutes an indicator of the quality of cancer patient management. Geographical variations in relative survival from cancer could reflect differences in the effectiveness of healthcare. Studies comparing geographical variations use estimates of relative survival obtained independently in each of the corresponding geographical units. When the units havesmall populations, the statistical stability of survival estimates could be seriously compromised.Our main objective is to assess the geographical variation in relative survival ratios from femalebreast cancer in the Girona Health Region (corresponding, virtually, to the province of Girona inCatalonia, in north‐eastern Spain). Firstly, we propose smoothing relative survival estimates bymeans of a (full) Bayesian hierarchical model. Secondly, we investigate which geographicalunit will give more stable estimates and could therefore be used for the analysis of cancersurvival as a healthcare performance indicator. We find that the model with more success in controlling for extra variability in the estimation of relative survival from female breast cancer in theGirona Health Region is a (full) Bayesian hierarchical model that incorporates both heterogeneity and spatial random effects. The greatest stability of these relative survival estimates wasachieved when basic health areas were taken as geographical units.