|
Cox regression analysis for models exploring the relationship between exposure to the occupation and CVD in later life (B). |
||||
| Covariate (referent category) |
Hazard Ratio |
95% confidence intervals |
p value |
|
|
|
||||
| Unadjusted |
Exposure to occupation (unexposed) |
2.88 |
1.78 – 4.66 |
0.01 |
| Unstratified, adjusted for potential confounders1 |
Exposure to occupation (unexposed) |
2.52 |
1.54 – 4.13 |
0.01 |
| Preterm gestational age (term) |
0.28 |
0.04 – 2.05 |
0.21 |
|
| Birth weight (per kg increase) |
1.15 |
0.73 – 1.83 |
0.55 |
|
| Manual paternal occupation (non-manual) |
1.41 |
0.87 – 2.29 |
0.16 |
|
| Rural parish of birth (urban) |
1.61 |
0.97 – 2.69 |
0.07 |
|
| Stratified by parish, and adjusted for potential confounders1 |
Exposure to occupation in urban parishes (unexposed) |
2.75 |
1.41 – 5.35 |
0.012 |
| Exposure to occupation in rural parishes (unexposed) |
2.28 |
1.08 – 4.81 |
||
|
1 Gestational age at birth, birth weight, paternal occupation at birth and parish of residence at birth were all considered potential confounders. It is important to note that the interpretation of the estimated hazard ratios for potential confounders is not straightforward since the main exposure variable (exposure to the occupation) lies on the pathway between these potential confounders and the outcome variable (CVD in later life). 2 p value for interaction between birth weight and sex. | ||||
Head et al. BMC Public Health 2008 8:303 doi:10.1186/1471-2458-8-303 |
||||