Time*predictor interaction term for longitudinal study analysis


New Member
Hello Forum,
Currently I am analyzing results of longitudinal study evaluating the effects of possible dichotomous predictor on continious outcome variable over time. For the analysis I used GEE model:

outcome~possible predictor+established predictors+ outcome at baseline

Using this model, the GEE p value of effect of possible predictor on outcome was less than 0.05.

In another model I used the "possible predictor*time" interaction term:

outcome~"possible predictor*time"+established predictors+ outcome at baseline

In this model the effect of "possible predictor*time" interaction term was not significant.
How would you interpret these results?
As I found in the literature, both types of models are being used for the analysis of longitudinal data.
Thank you very much in advance.