Matched case control study design

Question

I have a few questions about matched case-control study design. How do we handle matched case-control study design if you don’t have matched controls for all the cases? Would you remove unmatched case from the cohort? Also, how to handle if I need 1:2 case:control data but some cases have only 1 control? Thank you.

Answer 1

Sometimes people drop records when there's no controls for a case. The regression model that can handle this is called conditional logistic regression, and some software packages allow for a variable number of controls per case. I don't know about R, but look at Stata's conditional logistic module.