Table 6 Multivariable weighted stepwise linear regression models of subject‐specific means of log transformed styrene and SO levels in blood (weights are the frequencies of subject‐specific repeated measurements of styrene and SO in blood, respectively).
| Response variable | Overall r2 | Predictor variables | Parameter estimates | Δr2 | 95% CI |
|---|---|---|---|---|---|
| ln(Styrene in blood, mg/l, n = 292) | 0.72 | Intercept | −5.27 | −5.52 to −5.02 | |
| ln(Styrene in air, ppm) | 0.994 | 0.68 | 0.899 to 1.09 | ||
| Truck manufacture | 0.807 | 0.02 | 0.525 to 1.09 | ||
| Current alcohol use | 0.358 | 0.01 | 0.135 to 0.581 | ||
| Pipe and tank production | 0.535 | 0.01 | 0.204 to 0.866 | ||
| RV production | 0.540 | 0.01 | 0.168 to 0.912 | ||
| ln(SO in blood, μg/l, n = 211) | 0.34 | Intercept | −3.09 | −3.26 to −2.92 | |
| RV production | 0.750 | 0.30 | 0.517 to 0.982 | ||
| ln(Styrene in air, ppm) | 0.140 | 0.04 | 0.065 to 0.214 |