Statistical analyses
Random-effects linear regression models were used to find the association between the clinical outcomes and time due to the longitudinal nature of the data. The model included a fixed term for time and random terms for the slope and intercept. The data were first checked for normality and homoscedasticity over time. Linear contrast statements were used to find mean differences in clinical outcomes at various time points. Each individual clinical outcome was run on a separate regression model. All analyses were run by the study statistician, Gary Philips, using Stata 13.1, (StataCorp, College Station, TX, USA).