Table III.
Log odds ratios of preterm neonatal death for six-point increase in deprivation score.
| Method | log OR | SE | 95% CI | p | |
|---|---|---|---|---|---|
| IEE | 0.600 | 0.178 | 0.251 | 0.949 | 0.001 |
| ML estimation of GLMM | 0.546 | 0.189 | 0.175 | 0.917 | 0.004 |
| Conditional ML | 0.460 | 0.195 | 0.076 | 0.843 | 0.019 |
| Poor Man's method | |||||
| Within-cluster | 0.470 | 0.198 | 0.083 | 0.858 | 0.017 |
| Between-cluster | 1.262 | 0.641 | 0.005 | 2.519 | 0.049 |
| Model random intercept | |||||
| Within-cluster | 0.464 | 0.197 | 0.079 | 0.850 | 0.018 |
| Between-cluster | 0.987 | 0.630 | −0.248 | 2.222 | 0.117 |
| Cluster size (×1000) | 0.007 | 0.003 | 0.000 | 0.013 | 0.035 |
Methods are independence estimating equations, maximum likelihood estimate from random-intercept logistic regression model, conditional ML estimate from the same model, poor man's method, and modelling expectation of random intercept as linear function of mean deprivation in cluster and cluster size.