Skip to main content

Table 3 Predictors of childhood undernutrition: results from two-level logistic regression analysis

From: A survey of undernutrition in children under three years of age in rural Western China

Characteristics Null model Random intercept model with individual variables
Fixed effects (OR, 95% CI)   
Individual level variables   
Province   
Gansu   1.18 (0.70, 1.99)
Guangxi   1.25 (0.90, 1.74)
Guizhou   2.15 (1.50, 3.08)
Jiangxi   0.99 (0.71, 1.38)
Inner Mongolia   0.41 (0.28, 0.60)
Ningxia   0.49 (0.34, 0.72)
Qinghai   0.82 (0.59, 1.14)
Sichuan   1.77 (1.28, 2.45)
Xinjiang   0.79 (0.58, 1.06)
Chongqing   1.40 (0.97, 2.02)
Children’ age (months)   
0 ~ 6   1.00 (Reference)
7 ~ 12   1.27 (1.07, 1.50)
13 ~ 18   1.84 (1.55, 2.18)
19 ~ 24   2.41 (2.03, 2.86)
25 ~ 30   2.47 (2.06, 2.96)
31 ~ 36   2.86 (2.39, 3.41)
Parity   1.05 (0.97, 1.13)
Gender   
girl   1.00 (Reference)
boy   1.27 (1.16, 1.39)
Currently breastfeeding   
no   1.00 (Reference)
yes   0.94 (0.76, 1.17)
Ethnicity   
Minority   1.00 (Reference)
Han   0.77 (0.65, 0.90)
Wealth index   
poor   1.00 (Reference)
middle   0.86 (0.76, 0.97)
rich   0.79 (0.69, 0.90)
Mothers’ educations years   
no education   1.00 (Reference)
1 ~ 5   0.89 (0.76, 1.05)
6 ~ 9   0.75 (0.64, 0.89)
10 ~ 12   0.47 (0.36, 0.61)
>12   0.52 (0.34, 0.79)
Mothers’ age (years)   
<20   1.00 (Reference)
20 ~ 29   0.88 (0.66, 1.17)
30 ~ 39   0.81 (0.60, 1.10)
40~   0.65 (0.39, 1.07)
Random effects   
County level variance (SE) 0.67 (0.08) 0.36 (0.05)
VPC (latent variable method) 17.0% 9.8%
PCV Reference 46.3%
MOR (95% CI) 2.18 (2.01,2.39) 1.76 (1.65, 1.90)
  1. VPC, Variance partition coefficient; PCV, Proportional change in variance; MOR, Median odds ratio; SE, Standard error.