Applying linear mixed models to estimate reliability in clinical trial data with repeated measurements