https://t.co/n8Al3t2Kj3 Should just google <epi term> + sander greenland more often.
" subtle problems associated with matched case-control studies that do not arise / are minor in matched cohort studies: (1) matching, even for non-confounders, can create selection bias; (3) unbiased estimation requires accounting for the actual matching"
@aztezcan @EpiEllie You do need to adjust for the matching factor in the model. As said in this paper "case-control matching superimposes this selection bias... case-control matching does not and cannot remove confounding..." (pg 7) https://t.co/xKrFkPzRy3
@stephensenn Thanks for sharing. I just did some searching and found a more recent paper by Mansournia https://t.co/thewZwUz7H
Case–control matching: effects, misconceptions, and recommendations https://t.co/T6fPLcUh9M
Case-control matching: effects, misconceptions, and recommendations. - PubMed - NCBI https://t.co/Fe1WIrzFqF