|
|
Individually time-varying measurements |
|
Message/Author |
|
|
In a study with a large dataset we are setting time for each patient's baseline measurement to zero and then each follow-up time is in months since the baseline measurement. There are 8 potential time points with lots of individual variation on number of time points and time of follow-up. The whole study is four years. Is this the best way to specify the LGC model or would I be better off with different model assumptions about time: TSCORES = months2-months8 ; Analysis: TYPE = RANDOM MISSING; ESTIMATOR = ML; ALGORITHM=INTEGRATION; MCONVERGENCE = .001; Model: i | d1@0; i s | d2-d8 AT months2-months8; |
|
|
I think you should simply have i s | d1-d8 AT months0-months8; where everyone has months0=0. |
|
Back to top |
|
|