Table 2.

Negative binomial spatial regression analyses of incident Tuberculosis case rates

Variable/Indicator	Null model	Null spatial model	Bivariate models		Multivariable model
Log Population Densitya			0.95	[0.85–1.06]
Log Mean Monthly Incomea			0.66	[0.57–0.77]
Literacy rate			0.40	[0.22–0.72]	0.59	[0.44–0.80]
Mean household size			1.94	[1.25–3.03]
Electricity connectivity			0.05	[0.01–0.17]
Water supply			0.92	[0.80–1.07]
Garbage collection			0.65	[0.50–0.85]
Sewerage coverage			1.00	[0.97–1.04]	1.05	[1.02–1.08]
AIDS rateb			1.01	[0.93–1.10]
Homicide rateb			1.16	[1.07–1.25]	1.08	[1.01–1.16]
Proportion of population white			0.87	[0.80–0.94]
Proportion of population black			1.20	[1.12–1.30]	1.12	[1.03–1.21]
Proportion of population yellow			1.04	[0.96–1.12]
Proportion of population brown			1.13	[1.04–1.23]
Proportion of population indigenous			1.05	[0.97–1.13]
District I			-	-
District II			0.78	[0.51–1.18]
District III			0.81	[0.58–1.13]
District IV			0.72	[0.50–1.08]
District V			0.72	[0.48–1.11]
District VI			0.77	[0.51–1.24]
Deviance Information Criterion	1137.3	1109.5			1065.5
Spatial variance		0.117			0.031
Non-spatial variance	0.025	0.015			0.012
Spatial variance ratio		0.88			0.73
Global Moran’s I	0.33	0.11			0.11
Global Moran’s I p-value	0.00	0.03			0.03

Values are incidence rate ratios and 95% confidence intervals associated with a 10 percentage point change in the value of each covariate, with the exceptions of log values (^a: one log change), incidence rates (^b: one standard deviation change) or indicators for District (difference relative to District I). Model diagnostics not shown for the 16 bivariate models; bivariate model for District contained five indicator variables