Including fixed practice effects in the models contributes little to model performance. Adding practice effects to models with only age andgenderoronly age,genderand deprivation increases model R2D by at most 0.01. Dropping all the practice fixed effects from the full EDC count model reduces the R2D by 0.02.