Table 3.
Results of the model defined by Equations 1 and 2 for Z(A) = (1,A,G,E,S)T fitted on the Medical Research Council Cognitive Function and Ageing Study data.
| Covariate | Regression coefficient | Estimate | 95% CI |
|---|---|---|---|
| Model intercepts | β0.12 | − 8.184 | ( − 12.028, − 5.245) |
| β0.13 | − 11.150 | ( − 12.006, − 10.106) | |
| β0.23 | − 7.472 | ( − 10.168, − 5.265) | |
| Age (years) | βA.12 | 0.051 | (0.013, 0.104) |
| βA.13 | 0.101 | (0.089, 0.111) | |
| βA.23 | 0.065 | (0.039, 0.097) | |
| Gender (men versus women) | βG.12 | 0.340 | ( − 0.050, 0.744) |
| βG.13 | 0.298 | (0.159, 0.427) | |
| βG.23 | 0.412 | (0.128, 0.707) | |
| Education (10 years or more) | βE.12 | − 0.025 | ( − 0.472, 0.355) |
| βE.13 | − 0.228 | ( − 0.381, − 0.077) | |
| βE.23 | 0.159 | ( − 0.144, 0.507) | |
| Smoking (current versus never/ex) | βS.12 | 0.203 | ( − 0.155, 0.591) |
| βS.13 | 0.503 | (0.383, 0.644) | |
| βS.23 | 0.347 | (0.124, 0.647) | |
| Time spent in state 2 (years) | γ | − 0.001 | ( − 0.021, 0.021) |
A = age, G = gender, E = education, S = smoking. We assume transition intensities qij (i,j) ∈ {(1,2),(1,3),(2,3)} to be piecewise constant and introduce history in the process by fitting the time spent in state 2 as a time dependent covariate in q23. We based the confidence intervals on 450 bootstrap samples.