Multiple imputation was a valid approach to estimate absolute risk from a prediction model based on case-cohort data.