Table 2.
Perceived access to healthy foods, grocery shopping practices, diet and BMI stratified by residence in a food desert among primary food shoppers in the USA, 2015
| Descriptive statistics |
Regression results |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Does not live in a food desert (n 3294) |
Lives in a food desert (n 411) |
Unweighted bivariate model |
Inverse probability of treatment-weighted model |
|||||||
| Mean or % | sd or n | Mean or % | sd or n | OR or b | 95 % CI | P | OR or b | 95 % CI | P | |
|
| ||||||||||
| Perceived access to healthy foods (mean and sd) | 22·2 | 6·16 | 21·2 | 6·83 | −1·02 | −1·66, −0·38 | 0·002 | −0·62 | −1·39, 0·15 | 0·12 |
| Primary store type (% and n) | ||||||||||
| Large grocery store | 54·1 | 1781 | 48·7 | 200 | Ref. | – | – | Ref. | – | – |
| Superstore or supercentre | 26·4 | 870 | 34·1 | 140 | 1·26 | 0·91, 1·74 | 0·17 | 1·32 | 1·02, 1·71 | 0·03 |
| Smaller grocery store | 11·2 | 369 | 12·7 | 52 | 1·43 | 1·14, 1·81 | 0·002 | 1·18 | 0·82, 1·70 | 0·37 |
| Warehouse club | 4·5 | 149 | 2·9 | 12 | 0·72 | 0·39, 1·32 | 0·28 | 0·76 | 0·37, 1·54 | 0·44 |
| Farmers’ market, co-op or specialty store | 2·6 | 84 | 1·2 | 5 | † | – | – | † | – | – |
| Dollar store or convenience store | 1·3 | 41 | 0·5 | 2 | † | – | – | † | – | – |
| Frequency of grocery shopping (% and n) | ||||||||||
| >4 times/month | 33·8 | 1112 | 28·5 | 117 | Ref. | Ref. | ||||
| 4 times/month | 32·2 | 1062 | 26·8 | 110 | 0·98 | 0·75, 1·29 | 0·91 | 0·96 | 0·71, 1·30 | 0·79 |
| <4 times/month | 34·0 | 1120 | 44·8 | 184 | 1·56 | 1·22, 2·00 | 0·0004 | 1·28 | 0·97, 1·69 | 0·08 |
| Distance from home to primary store (miles*; mean and sd) | 5·4 | 7·46 | 5·8 | 7·95 | 0·08 | −0·06, 0·22 | 0·27 | 0·04 | −0·10, 0·17 | 0·58 |
| Dollar amount spent per shopping trip (% and n) | ||||||||||
| ≥$US 200 | 5·2 | 172 | 6·1 | 25 | Ref. | Ref. | ||||
| $US 150–$199·99 | 7·4 | 245 | 7·5 | 31 | 0·87 | 0·50, 1·53 | 0·63 | 1·01 | 0·54, 1·90 | 0·98 |
| $US 100–$149·99 | 18·9 | 622 | 22·1 | 91 | 1·01 | 0·63, 1·62 | 0·98 | 1·12 | 0·66, 1·90 | 0·68 |
| $US 50–$99·99 | 37·4 | 1233 | 34·6 | 142 | 0·79 | 0·50, 1·25 | 0·32 | 0·82 | 0·49, 1·37 | 0·45 |
| $US 25–$49·99 | 24·4 | 803 | 22·9 | 94 | 0·81 | 0·50, 1·29 | 0·37 | 0·74 | 0·44, 1·26 | 0·27 |
| <$US 25 | 6·7 | 219 | 6·8 | 28 | 0·88 | 0·50, 1·56 | 0·66 | 0·73 | 0·39, 1·37 | 0·33 |
| Fruit and vegetable intake (servings/d; mean and sd) | 2·5 | 2·03 | 2·3 | 1·96 | −0·14 | −0·35, 0·06 | 0·18 | −0·06 | −0·28, 0·16 | 0·58 |
| Dietary fat intake (% of total daily energy intake; mean and sd) | 34·7 | 4·72 | 34·4 | 3·94 | −0·29 | −0·77, 0·18 | 0·23 | −0·45 | −0·93, 0·03 | 0·06 |
| BMI (kg/m2; mean and sd) | 27·6 | 6·69 | 29·4 | 7·75 | 1·77 | 1·07, 2·47 | <0·0001 | 1·14 | 0·36, 1·93 | 0·004 |
Ref., reference category.
P values were generated by performing the regression of each dependent variable v. an indicator variable for residence in a food desert or not. Linear regression was used for the following outcomes: perceived access to healthy foods, fruit and vegetable intake, dietary fat intake and BMI. Poisson regression was used for distance from home to primary store. Multinomial logistic regression was used for primary store type, frequency of grocery shopping and dollar amount spent per shopping trip. Robust standard errors were used for the weighted regression analyses to account for the inverse probability of treatment weights.
1 mile ≈ 1·61 km.
The categories for farmers’ market, co-op or specialty store and dollar store or convenience store were omitted from the models due to sparse data.