Adjusting for sample size imbalance in a predictor variable in Cox PH model

Hi,
Does the Cox PH model adequately adjust the sample size imbalance of a predictor variable when model fitting? For example, I’m fitting a Cox PH model using ‘cph()’ function using rms for a cohort study data. Overall sample size is 1200 with 674 events. I noticed that one of the predictors (categorical variable with two levels) has a large imbalance with 1050 in one group and 150 in the other. Will this imbalance have an impact to the parameter estimates? If so, what strategies are there to adjust for the imbalance?

Thank you.

As I understand, predictor imbalance does not matter beyond precision, which is solved by having enough data points from each of the predictor categories. – i.e., sample size. Curious to see if others have counterpoints, but that doesn’t seem to be much of a problem in your case.

2 Likes

Thank you for your input @ giuliano-cruz.