zbMATH — the first resource for mathematics

Ecological regression with errors in covariates: An application. (English) Zbl 0965.62093
Lawson, Andrew (ed.) et al., Disease mapping and risk assessment for public health. Chichester: Wiley. 329-348 (1999).
From the introduction: We describe a Bayesian hierarchical-spatial model for ecological regression aimed at investigating the relationship between insulin dependent diabetes mellitus (IDDM) incidence, the proportion of glucose-6-phosphate-dehydrogenase (G6PD) deficient individuals and past prevalence of malaria in Sardinia. Our model is composed of two regression submodels. The first model allows us to estimate the effect of malaria on IDDM risk, while the second model allows us to estimate the effect of malaria on G6PD deficiency. Both models allow also for a common unknown underlying non-malaria factor which we suppose to be responsible for the association between IDDM and G6PD deficiency. Smoothing spatial priors are posited to reduce random geographical variability in the estimates. Measurement error in the ecological covariates is also accounted for.
For the entire collection see [Zbl 0942.00010].
62P10 Applications of statistics to biology and medical sciences; meta analysis
62M30 Inference from spatial processes
62F15 Bayesian inference
62P12 Applications of statistics to environmental and related topics