Refers to bias arising from inappropriate accounting for patient dropouts, withdrawals, and deviations from trial protocols.