longitudinal mixed models

[Buyadaa Oyunchimeg]

Dear Stata users,

I am working on repeated measure data of kidney markers and trying to construct a mixed model.
In order to calculate eGFR decline, I would like to include age both as a linear and quadratic term and center for modeling, using random terms for intercept and age.

I wonder if the model was constructed correctly.

sum current_age
gen current_age_c=current_age - `r(mean) '

mixed eGFR c.current_age_c i.other_variables || current_age: current_age_c,
mixed eGFR c.current_age_c##c.current_age_c i.other_variables || current_age: current_age_c##current_age_c,

Many thanks,

Oyun

[Clyde Schechter]

Not quite. I suppose that the repeated measures you have represent longitudinal data obtained from patients over time. So, hopefully, you have a patient id variable. Let's call that variable id. Then the model you want is

mixed eGFR c.current_age_c##c.current_age_c other variables || id: c.current_age_c##c.current_age_c

[Buyadaa Oyunchimeg]

As always, thanks very much, Clyde.