Table 3.

Associations of neighbourhood-level variables and regular fruit and vegetable (FV) consumption (≥5 d/week) adjusted for individual characteristics, São Paulo, Brazil, 2011

	FV consumption
			Adjusted models
	Model 1		Model 2: Density		Model 3: Proximity
Variable	PR	95 % CI	PR	95 % CI	PR	95 % CI
Community food environment
Supermarket and fresh produce market density (per 10 000 residents) (log-transformed)
Q1	0·80	0·64, 0·99	0·79	0·63, 0·99
Q2	0·83	0·63, 1·08	0·82	0·63, 1·08
Q3	0·84	0·65, 1·08	0·84	0·62, 1·15
Q4 (reference)	1·00	–	1·00	–
Supermarket and fresh produce market distance (m)
Q1	1·29	1·12, 1·49			1·25	1·08, 1·46
Q2	1·18	1·02, 1·37			1·15	0·98, 1·35
Q3	1·32	1·16, 1·50			1·27	1·12, 1·45
Q4 (reference)	1·00	–			1·00	–
Consumer food environment
Availability*	1·41	1·19, 1·67
Quality
No stores with fresh FV	2·20	0·96, 5·05	2·07	0·99, 4·32	2·11	0·91, 4·91
No store in upper quartile	2·33	1·08, 5·01	2·25	1·03, 4·89	2·33	1·08, 5·03
Store present in upper quartile (reference)	1·00	–	1·00	–	1·00	–
Price (Z-score) (R$)†	0·96	0·89, 1·04	0·99	0·88, 1·11	0·97	0·89, 1·05

PR, prevalence ratio; SSB, sugar-sweetened beverages.

Model 1: each consumer or community food environment was associated with FV and SSB consumption separately and adjusted for individual-level variables (age, sex, education and income).

Model 2: adjusted for individual-level variables (age, sex, education and income)+neighbourhood-level income + community (density of supermarkets and fresh produce markets) and consumer food environment variables.

Model 3: adjusted for individual-level variables (age, sex, education and income)+neighbourhood-level income + community (proximity to supermarkets and fresh produce markets) and consumer food environment variables.

^*Due to high collinearity between variables, FV availability was not included in the fully adjusted models.

^†As of March 2011, $US 1·00=R$ 1·66.