Effects of density† of fast food restaurant and various types of food stores‡,§ on body mass index as an outcome variable, independent of individual-level factors and neighbourhood socioeconomic characteristics
| Independent variables | Regression coefficient (SE) with BMI (kg/m2) as the outcome variable | ||||
|---|---|---|---|---|---|
| Fast food restaurant | Convenience store | Small grocery store | Ethnic market | Supermarket | |
| BMI, body mass index; NS, not significant; SES, socioeconomic status. | |||||
| *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001 | |||||
| †Store density was estimated by dividing the number of stores in an area by the number of stores located in an area comprising the neighbourhood and a 0.5 mile buffer zone (expressed as number of stores per square mile). | |||||
| ‡The table shows regression coefficient (SE) derived from multilevel modelling procedures constructed for each store type with BMI as the outcome variable; city and survey year were also included in the regression equations. | |||||
| §Definitions of store types (table 2). | |||||
| ¶A composite variable derived from annual household income and educational attainment. | |||||
| **Smoking was defined by a categorical variable (current smoker, past smoker, never smoker). | |||||
| ††Physical activity was assessed by asking participants to rate their level of physical activity relative to others of their age on a scale of 1–7 (higher scores indicate higher levels of activity). | |||||
| ‡‡A dichotomous variable derived from 17 questions was developed to assess nutrition knowledge. The median score, 4/17, was the cut-off point used to create this dichotomous variable. | |||||
| §§Neighbourhood SES index was derived from census information on median family income, median housing value, percentage of blue collar workers, percentage of unemployed and percentage having less than high-school education and categorised into survey-specific tertiles, compared with high tertile. | |||||
| ¶¶Interactions between sex and food store characteristics of the neighbourhood were observed; therefore, the regression coefficients are presented separately for men and women. Interactions between neighbourhood SES and store density were not observed for any store type. | |||||
| Individual factors | ' | ||||
| Age (year) | 0.061 (0.004)**** | 0.061 (0.004)**** | 0.061 (0.004)**** | 0.061 (0.004)**** | 0.061 (0.004)**** |
| Hispanic ethnicity | 1.386 (0.192)**** | 1.393 (0.192)**** | 1.377 (0.192)**** | 1.386 (0.192)**** | 1.388 (0.192)**** |
| Individual SES¶ | −0.530 (0.072)**** | −0.531 (0.072)**** | −0.528 (0.072)**** | −0.530 (0.072)**** | −0.531 (0.072)**** |
| Smoking** | |||||
| Current | −1.022 (0.133)**** | −1.021 (0.133)**** | −1.015 (0.133)**** | −1.022 (0.133)**** | −1.017 (0.133)**** |
| Past | NS | NS | NS | NS | NS |
| Physical activity†† | 0.569 (0.034)**** | −0.571 (0.034)**** | −0.569 (0.034)**** | −0.569 (0.034)**** | −0.571 (0.034)**** |
| Nutrition knowledge‡‡ | 0.231 (0.113)* | −0.230 (0.113)* | −0.233 (0.113)* | −0.231 (0.113)* | −0.232 (0.113)* |
| Neighbourhood SES§§ | |||||
| Low tertile | 0.631 (0.200)* | 0.624 (0.198)* | 0.574 (0.206)* | 0.631 (0.200)* | 0.619 (0.199)* |
| Middle tertile | NS | NS | NS | NS | NS |
| Geographic accessibility to store | |||||
| Store density†, ¶¶ | |||||
| Men | NS | NS | NS | NS | NS |
| Women | NS | NS | 0.053 (0.023)* | NS | NS |