Estimates of effect of price discounts (n=513) versus no price discounts (n=515) on mean change from baseline in nutrient and food purchases by ethnicity†
| Māori | Pacific | European/other | All participants | Interaction with ethnicity‡ | |||||
| Difference in means (SE) | 95% CI | Difference in means (SE) | 95% CI | Difference in means (SE) | 95% CI | Difference in means (SE) | 95% CI | p Value | |
| Saturated fat (% energy) | |||||||||
| Intervention (6 months) | 0.35 (0.52) | −0.68 to 1.39 | 1.09 (0.74) | −0.38 to 2.56 | −0.34 (0.21) | −0.75 to 0.06 | −0.02 (0.19) | −0.40 to 0.36 | 0.02 |
| Follow-up (12 months) | −0.48 (0.57) | −1.60 to 0.65 | −1.04 (0.81) | −2.64 to 0.55 | −0.06 (0.21) | −0.47 to 0.36 | −0.12 (0.20) | −0.51 to 0.27 | 0.30 |
| Food purchases (kg/week) | |||||||||
| All foods | |||||||||
| Intervention (6 months) | −0.13 (0.94) | −1.98 to 1.71 | 1.51 (1.16) | −0.78 to 3.80 | 1.13** (0.32) | 0.50 to 1.76 | 0.90* (0.31) | 0.29 to 1.52 | 0.24 |
| Follow-up (12 months) | −0.19 (0.99) | −2.14 to 1.75 | 2.40 (1.25) | −0.06 to 4.87 | 0.34 (0.32) | −0.29 to 0.98 | 0.37 (0.32) | −0.26 to 1.00 | 0.24 |
| All healthier discounted foods | |||||||||
| Intervention (6 months) | −0.15 (0.48) | −1.10 to 0.80 | 1.20* (0.57) | 0.06 to 2.34 | 1.02** (0.21) | 0.60 to 1.43 | 0.79** (0.19) | 0.43 to 1.16 | 0.04 |
| Follow-up (12 months) | −0.12 (0.51) | −1.13 to 0.89 | 1.07 (0.62) | −0.15 to 2.29 | 0.47* (0.21) | 0.05 to 0.89 | 0.38* (0.19) | 0.01 to 0.76 | 0.43 |
| Healthier discounted fruit and vegetables | |||||||||
| Intervention (6 months) | −0.05 (0.37) | −0.78 to 0.67 | 0.86* (0.42) | 0.03 to 1.69 | 0.59** (0.16) | 0.28 to 0.89 | 0.48** (0.14) | 0.21 to 0.75 | 0.13 |
| Follow-up (12 months) | 0.01 (0.39) | −0.75 to 0.78 | 0.65 (0.46) | −0.25 to 1.56 | 0.33* (0.16) | 0.02 to 0.64 | 0.28* (0.14) | 0.00 to 0.56 | 0.63 |
*p<0.05 for intervention effect.
**p<0.01 for interaction between intervention and period.
***p<0.001 for intervention effect.
† Change from baseline with intervention versus change from baseline without intervention calculated by using repeated-measures mixed model regression analysis. The model for the study population as a whole was adjusted for baseline nutrient and food purchases, ethnicity, income, age and sex.
‡ Significance of interaction with ethnicity was performed using analysis of covariance regression (ANCOVA) under a similar regression model as described above (excluding ethnicity), with separate models for ‘intervention’ and ‘follow-up’ outcomes.