A multilevel analysis on the relationship between neighbourhood poverty and public hospital utilization: is the high Indigenous morbidity avoidable?

Table 1 Multilevel logistic models for number of total and potentially avoidable hospital patients per population

	(1) Disadvantage model				(2) Advantage and disadvantage model
Coefficient	β	(SE)	OR	(95%CI)	β	(SE)	OR	(95%CI)
Intercept	-1.963	(0.008)			-1.761	(0.008)
Indigenous status	0.598	(0.011)	1.818	(1.777-1.860)	0.342	(0.012)	1.408	(1.375-1.441)
Sex	0.308	(0.008)	1.361	(1.339-1.383)	0.311	(0.008)	1.365	(1.343-1.388)
Aged	0.144	(0.012)	1.155	(1.127-1.184)	0.138	(0.013)	1.148	(1.118-1.179)
SEIFA 1	-0.041	(0.004)	0.960	(0.953-0.967)
SEIFA 2					-0.079	(0.004)	0.924	(0.917-0.931)
-Log likelihood	185142				185246
Variance of random effect
Intercept	0.503	(0.007)			0.476	(0.007)
Indigenous status	0.393	(0.012)			0.345	(0.012)
Aged	0.082	(0.006)			0.076	(0.006)
	(3) Income and education model
	a. Total hospital morbidity				b. Avoidable hospital morbidity
Intercept	-1.811	(0.008)			-2.378	(0.012)
Indigenous status	0.260	(0.013)	1.297	(1.265-1.329)	0.529	(0.015)	1.697	(1.646-1.748)
Sex	0.312	(0.008)	1.366	(1.344-1.389)	0.270	(0.010)	1.310	(1.284-1.336)
Aged	0.202	(0.013)	1.223	(1.192-1.256)	0.003	(0.000)	1.003	(1.002-1.003)
SEIFA 3	-0.077	(0.004)	0.926	(0.918-0.934)	-0.045	(0.006)	0.956	(0.945-0.967)
SEIFA 4	-0.087	(0.004)	0.917	(0.909-0.925)	-0.139	(0.005)	0.870	(0.861-0.879)
-Log likelihood	185076				138337
Variance of random effect
Intercept	0.556	(0.009)			0.336	(0.009)
Indigenous status	0.399	(0.013)			0.200	(0.009)
Aged	0.042	(0.003)			0.000	(0.000)

Note: β= coefficient estimate; CI = confidence interval; OR = Odds Ratio; SE = Standard error; SEIFA = Socio-Economic Indexes for Areas; SEIFA1 measures relative socio-economic disadvantage; SEIFA2 represents relative socio-economic advantage and disadvantage; SEIFA3 indicates household income; SEIFA4 is index of education; SLA = Statistical Local Area.

ISSN: 1471-2458