Regression analysis with follow-up

Hi everyone! For my thesis, I am doing research on social participation and health. However, I am a bit unsure about how to build the regression model. I am planning on doing several analyses with different independent variables that represent social participation scores. To keep it simple, I will give an example with the analysis of the score: total frequency of participation, a score with a range of 0-3.

Independent variable: total frequency of participation (measured at baseline, 2 weeks before hospitalization)
Dependent variable: self-rated health (measured at follow-up, 3 months post-discharge)
Covariates: Age, gender, civil status, memory problems, fatigue

I am interested in whether there is an association between how socially embedded an individual is (before hospitalization) and health status 3 months after hospitalization. I would not like to measure change scores between baseline and follow-up, but just the overall health status at follow-up to see if people that are more socially embedded have a better self-rated health at that moment. My suggestion is to do a regression analysis with self-rated health measured at follow-up as dependent variable, and then add self-rated health at baseline as a covariate.

However, I am not sure if this is the correct way to do it. Can anyone help me out with this? :)


Fortran must die
How is your dependent variable scored? On ordinal, interval or nominal data. How you do regression depends on this.

If you are asking someone to assess themselves before and after some event then a test retest methodology (ANOVA) might be a way to go. Problems with bias of the rater is going to be a major issue your methodology needs to address.