Maps of regional morbidity and mortality rates are useful tools in det
ermining spatial patterns of disease. Combined with sociodemographic c
ensus information, they also permit assessment of environmental justic
e; that is, whether certain subgroups suffer disproportionately from c
ertain diseases or other adverse effects of harmful environmental expo
sures. Bayes and empirical Bayes methods have proven useful in smoothi
ng crude maps of disease risk, eliminating the instability of estimate
s in low-population areas while maintaining geographic resolution. In
this article we extend existing hierarchical spatial models to account
for temporal effects and spatio-temporal interactions. Fitting the re
sulting highly parameterized models requires careful implementation of
Markov chain Monte Carlo (MCMC) methods, as well as novel techniques
for model evaluation and selection. We illustrate our approach using a
dataset of county-specific lung cancer rates in the state of Ohio dur
ing the period 1968-1988.