Table 6.
Joint model fits of the data from cystic fibrosis patients†
| Parameter | Results for the following models: | |||
|---|---|---|---|---|
| SGP | Lagged SGP | IS | SGP + IS | |
| Longitudinal | ||||
| Intercept | 74.899 (4.868) | 74.904 (4.948) | 76.683 (5.069) | 76.669 (4.406) |
| Time | −1.502 (0.074) | −1.502 (0.074) | −1.762 (0.082) | −1.652 (0.080) |
| Age at t0 | −0.454 (0.250) | −0.454 (0.256) | −0.577 (0.262) | −0.568 (0.218) |
| Sex (males) | 0.786 (1.415) | 0.785 (1.377) | 1.759 (1.420) | 1.506 (1.143) |
| Survival | ||||
| Intercept | 2.964 (0.344) | 2.962 (0.355) | 3.290 (0.422) | 3.441 (0.276) |
| Time | −0.023 (0.011) | −0.023 (0.012) | −0.050 (0.014) | −0.052 (0.013) |
| Age at t0 | −0.021 (0.017) | −0.021 (0.018) | −0.034 (0.020) | −0.034 (0.013) |
| Sex (males) | 0.234 (0.084) | 0.234 (0.084) | 0.268 (0.091) | 0.287 (0.097) |
| γ | 0.037 (0.002) | 0.037 (0.003) | — | 0.040 (0.009) |
| γlag | — | 0.000 (0.001) | — | — |
| γ1 | — | — | 0.035 (0.002) | 0.036 (0.004) |
| γ2 | — | — | 0.270 (0.032) | 0.418 (0.084) |
| Others | ||||
| ν | 7.235 (0.100) | 7.235 (0.100) | 8.806 (0.084) | 7.231 (0.108) |
| σu | 25.081 (0.485) | 25.081 (0.485) | — | 12.832 (1.655) |
| σ1 | — | — | 24.240 (0.523) | 20.486 (1.211) |
| σ2 | — | — | 2.152 (0.074) | 1.249 (0.108) |
| ρSGP | 0.969 (0.002) | 0.969 (0.002) | — | 0.890 (0.031) |
| ρIS | — | — | −0.066 (0.042) | 0.218 (0.086) |
| AIC | 49551.96 | 49553.96 | 49878.47 | 49522.25 |
†
The random-effects models fitted were a stationary Gaussian process SGP, a stationary Gaussian process with one time lag in the survival model, lagged SGP, a random intercept and slope model IS and a stationary Gaussian process plus random intercept and slope, SGP+IS. For each model estimated parameter values are presented with standard errors in parentheses, and the Akaike information criterion AIC.