“…MRI, histology, arthroscopy and other outcome scores will be compared using one-way analysis of variance, if available using preoperative values as covariates. For the continuous outcome measures that are collected repeatedly over time (such as the Lysholm knee score), repeated measures analyses will be performed on the change from the baseline score, using standard multilevel modeling techniques [3]. Such analyses have the advantages of being able to combine results from different time-points to maximize power, and also to investigate the precise form of any benefit (whether, for example, any treatment benefit, should one exist, increases or decreases with time).…”