Table 4. Results of linear regression for complete cases and imputed data using May-adjusted vitamin D as the outcome.
| Multivariable models | ||||
|---|---|---|---|---|
| Model 1a (n = 628) | Model 2b (n = 1949) | |||
| Predictors | Est (95% CI) | p-value | Est (95% CI) | p-value |
| Intercept | 32.4 (18.5, 46.2) | 5.25x10-6 | 30.3 (22.2, 38.3) | 2.2x10-13 |
| Body mass index (kg/m2) | -0.75 (-1.02, -0.47) | 1.95x10-7 | -0.54 (-0.70, -0.38) | 7.5x10-11 |
| Holidays outside the UK | ||||
| < once a year | -0.93 (-4.38, 2.52) | 0.59 | 0.75 (-1.34, 2.85) | 0.48 |
| Once year | 5.03 (0.20, 9.86) | 0.041 | 6.47 (3.47, 9.47) | 0.000024 |
| > once a year | 18.7 (11.3, 26.2) | 1.04x10-6 | 13.5 (9.07, 18.0) | 3.4x10-9 |
| Age at venepuncture | 0.24 (0.11, 0.36) | 0.0003 | 0.14 (0.07, 0.22) | 0.00030 |
| Physical activity | 1.66 (0.58, 2.75) | 0.003 | 1.42 (0.65, 2.19) | 0.00032 |
| Socio-economic status 3 (“non-traditional”) | -2.10 (-3.57, -0.64) | 0.005 | -1.74 (-2.71, -0.78) | 0.00043 |
| Summer minutes outside | 0.0046 (-0.00398, 0.013) | 0.29 | 0.006 (-0.00028, 0.012) | 0.062 |
| Socio-economic status 2 | 0.32 (-1.30, 1.94) | 0.70 | 0.69 (-0.34, 1.72) | 0.19 |
| Vitamin D intake (μg) | 0.201 (-0.18, 0.60) | 0.30 | 0.14 (-0.17, 0.46) | 0.37 |
| Working (not retired) | -2.07 (-6.61, 2.47) | 0.37 | -0.18 (-2.45, 2.09) | 0.88 |
| Sex (male) | 1.06 (-1.55, 3.67) | 0.43 | -0.12 (-1.77, 1.52) | 0.88 |
| Socio-economic status 1 | -0.013 (-1.85, 1.82) | 0.99 | -0.075 (-1.17, 1.02) | 0.89 |
a Model 1 constructed using complete cases in the original dataset, R2 = 0.204.
b Model 2 constructed using 68 datasets with missing data completed by imputation (100 cycles), R2 = 0.111 (People missing outcome data were excluded from the imputation model).
Socio-economic status was derived from principal components analysis.