Multivariate logistic regression for repeated data

[Luigi Vanvitelli]

Dear all,

I have data on a group of individuals with a binary outcome (use of all three main preventive measures for COVID-19) evaluated at three time period (before the vaccination, after the primary cycle, after the booster dose). My aim is to test the change in the use of all preventive measures in the different periods and which variables may be associated with the correct use. Then I would perform a multivariate regression analysis for repeated data. Any suggestions how can I approach this analysis and the commands in Stata?

Thanks for your suggestions.

[Carlo Lazzaro]

Luigi:
welcome to this forum.
Pedantic as it may seem, if you have one categorical regressand only but >=2 predictors, you are interested in a multiple regression, as multivariate regression implies >1 regressand and >=2 predictors.
That said, you seem to have a panel data too.
Hence, you may want to take a look at -xtlogit-.

[Luigi Vanvitelli]

Carlo:
thank you.
In the data, I have the outcome for 800 individuals that has been measured in the 800 subjects in three different period of time and 7 predictors.
The xtlogit can be used also for repeated measures?

Kind regards,
Luigi

[Carlo Lazzaro]

Luigi:
if the same sample of 800 individuals (N or cross-sectional dimension) were measured three times (T or time-series dimension) you actually have a panel dataset.
If your regressand is binary and you have a panel dataset, -xtlogit- is the way to go.
The role of predictors is to give afair and true view of the data generating process that you're investigating (but this holds for any regression).