Table 5.

Associations between fast food count^* and school disadvantaged/advantaged socio-economic status (SES), adjusting for both food retail kernel density (K-density) and retail zoning within the road network buffers around schools, metropolitan Adelaide, South Australia, 2013

				95 % CI for RR
Poisson regression	β coefficient	P value	RR & OR=exp(β)	Lower	Upper
SES disadvantaged 1000 m	0·403	<0·0001	1·495	1·391	1·608
SES disadvantaged 1500 m	0·325	<0·0001	1·385	1·318	1·456
Logit (fast food=0):
Retail zoning 1000 m	−1·578	<0·0001	0·206	0·125	0·340
Logit (fast food=0):
Retail zoning 1500 m	−2·770	<0·0001	0·063	0·026	0·150
K-density 1000 m (mid)	0·961	<0·0001	2·615	2·123	3·214
K-density 1000 m (high)	2·772	<0·0001	15·984	13·184	19·379
SES disadvantaged	0·571	<0·0001	1·770	1·645	1·901
Logit (fast food=0):
Retail zoning 1000 m	−1·930	<0·0001	0·145	0·063	0·337
K-density 1500 m (mid)	0·997	<0·0001	2·710	2·282	3·217
K-density 1500 m (high)	2·727	<0·0001	15·281	12·984	17·984
SES disadvantaged	0·439	<0·0001	1·550	1·475	1·629
Logit (fast food=0):
Retail zoning 1500 m	−4·260	<0·0001	0·014	0·002	0·125
SES disadvantaged	0·564	<0·0001	1·757	1·634	1·890
K-density 1000 m (mid)	0·945	<0·0001	2·573	2·100	3·153
K-density 1000 m (high)	2·692	<0·0001	14·766	12·207	17·861
Retail zoning 1000 m	0·707	<0·0001	2·028	1·770	2·323
Logit (fast food=0):
Retail zoning 1000 m	−1·039	0·0519	0·354	0·124	1·009
SES disadvantaged	0·431	<0·0001	1·538	1·464	1·616
K-density 1500 m (mid)	0·948	<0·0001	2·579	2·173	3·061
K-density 1000 m (high)	2·657	<0·0001	14·251	12·099	16·784
Retail zoning 1500 m	0·362	<0·0001	1·436	1·223	1·686
Logit (fast food=0):
Retail zoning 1500 m	−3·682	0·0029	0·025	0·002	0·285

RR, relative risk.

^*Poisson regression model for the fast-food count and logit regression model for the absence of fast food (fast food=0) to predict the membership to the ‘certain zero’ group; SES disadvantaged=referent; K-density low=referent; retail zoning no=referent; significance was set at 0·05.