Association between child's ethnicity and known risk factors for later obesity (with adjustment for sociodemographic characteristics)
| European N=493 | Māori or Pacific N=115 | Other N=79 | |
|---|---|---|---|
| Categorical risk factors; ORs and 95% CIs | |||
| Maternal pre-pregnancy BMI, kg/m2 | |||
| <25 | 1 (ref) | 1 (ref) | 1 (ref) |
| 25 to <30 | 1 (ref) | 1.47 (0.80 to 2.68) p=0.21 | 1.03 (0.58 to 1.82) p=0.92 |
| ≥30 | 1 (ref) | 2.27 (1.21 to 4.26) p=0.01 | 0.44 (0.15 to 1.29) p=0.14 |
| Mother did physical activity pre-pregnancy | 1 (ref) | 1.79 (0.92 to 3.48) p=0.09 | 1.30 (0.60 to 2.81) p=0.50 |
| Hypertension in pregnancy | 1 (ref) | 1.50 (0.82 to 2.75) p=0.19 | 0.97 (0.45 to 2.07) p=0.93 |
| Diabetes in pregnancy | 1 (ref) | 0.74 (0.20 to 2.73) p=0.65 | 1.70 (0.59 to 4.91) p=0.32 |
| Exclusive breastfeeding in first 3 days | 1 (ref) | 0.72 (0.43 to 1.23) p=0.23 | 0.95 (0.52 to 1.74) p=0.87 |
| Exclusive breastfeeding at 3 months | 1 (ref) | 0.98 (0.63 to 1.52) p=0.92 | 1.24 (0.75 to 2.03) p=0.40 |
| Solid foods introduced by 3 months | 1 (ref) | 1.24 (0.77 to 1.98) p=0.38 | 0.80 (0.45 to 1.40) p=0.43 |
| Continuous risk factors; mean differences and 95% CIs | |||
| Mother's diet in pregnancy* | |||
| Processed foods (SD) | 0 (ref) | 0.07 (−0.18 to 0.33) p=0.57 | −0.11 (−0.37 to 0.15) p=0.41 |
| Snacks (SD) | 0 (ref) | 0.37 (0.12 to 0.62) p=0.004 | −0.27 (−0.53 to −0.01) p=0.05 |
| Healthy foods (SD) | 0 (ref) | −0.11 (−0.37 to 0.15) p=0.40 | −0.02 (−0.30 to 0.26) p=0.90 |
| Sweet foods (SD) | 0 (ref) | −0.20 (−0.46 to 0.06) p=0.12 | −0.09 (−0.36 to 0.17) p=0.50 |
| Starchy foods (SD) | 0 (ref) | 0.23 (−0.00 to 0.47) p=0.05 | 0.30 (0.02 to 0.58) p=0.03 |
| Child's weight in first week (kg) | 0 (ref) | 0.08 (−0.04 to 0.21) p=0.20 | −0.05 (−0.19 to 0.10) p=0.53 |
| Child's weight at 3 months (kg) | 0 (ref) | 0.16 (−0.04 to 0.36) p=0.11 | 0.07 (−0.15 to 0.29) p=0.65 |
N=687 using multiply imputed data.
Values are ORs or mean differences (95% CIs) comparing Māori or Pacific children and other ethnicity children to European children. Multinomial logistic regression was used for maternal pre-pregnancy BMI, with <25 kg/m2 as the reference category. Logistic regression was used for binary risk factors, and linear regression for continuous risk factors.
*Dietary pattern scores are standardised to have a mean of zero and a variance of one; values in this table can therefore be interpreted as SDs from the overall population mean.
BMI, body mass index.