Refers to systematic differences between comparison groups in withdrawals (drop-outs) from a study, loss to follow-up, or other exclusions of patients and how these losses are analyzed.