Table 2.

Deviance information criteria and empirical variances of the Bayesian models (with their 95 % credible intervals, CI)

	Deviance information criteria				Empirical variances (95 % CI)
Cancer	α(+βχι)	α + Vi(+βχι)	α + Ui(+βχι)	α + Ui + Vi(+βχι)	s2u	s2v
Without covariate
Lung	2172.8	2071.2	2082.8	2068.3	0.031 (0.004; 0.498)	0.057 (0.031; 0.085)
Prostate	2765.3	2631.4	2585.4	2589.2	0.008 (0.003; 0.019)	0.617 (0.310; 0.998)
Colon-rectum	1929.0	1931.0	1930.3	1932.5	0.0091 (0.002; 0.022)	0.0354 (0.007; 0.180)
Bladder	1374.6	1373.2	1371.0	1371.4	0.018 (0.005; 0.052)	0.241 (0.025; 1.015)
With covariate
Lung	2016.5	2007.2	2009.9	2007.9	0.066 (0.013; 0.374)	0.011 (0.002; 0.030)
Prostate	2736.4	2630.2	2585.7	2588.5	0.007 (0.017; 0.002)	0.577 (0.261; 0.981)
Colon-rectum	1930.7	1933.1	1932.0	1933.3	0.007 (0.002; 0.024)	0.006 (0.002; 0.143)
Bladder	1368.3	1371.0	1367.4	1370.3	0.017 (0.0035; 0.056)	0.195 (0.030; 2.529)

(+βχ _ι) in case of additional covariate - $\mathcal{U}$ : Autocorrelation - $\mathcal{V}$ : heterogeneity - s ²_u: Autocorrelation empirical variance - s ²_v: Heterogeneity empirical variance – $\mathcal{X}$ Townsend index components