Multivariable multivariate linear regression analysis for outcomes at different time pooints

[Cinzia Del Giovane]

Hi,
I want to assess the association between an exposure (present/no present) and depressive symptoms (as continuous) in a cohort study. In the cohort there are patients in which the last follow-up data varies, i.e. for one participant I have the outcome at 1 year, for another at 6 year. As I have multiple time points for the outcome I was thinking to use a multivariate linear regression. Then I need also to adjust the model for some confounders. Which is the correct command in STATA for this model?

The other solution I thought to apply is to use the last observation for each individual as dependent variable and then adjust the model for the length of follow-up. Does it make sense also?

Thanks
Cinzia

[Carlo Lazzaro]

Cinzia:
welcome to this forum.
If your data represent measures of dependent variable and predictors on the same sample of patients taken at different time points, you may want to consider an -xt- command availbale in Stata for panel data analysis and add time as a predictor..

[Phil Bromiley]

You may want to consider treatment models as well.

[Cinzia Del Giovane]

Thanks for your answers. I will try with the xt first.

Best,
Cinzia

[Carlo Lazzaro]

Cinzia:
before plugging in -time- as a predictor in your regression model, take a look at -help fvvarlist-.