Procedures tested in clinical trials require adequate competence of one or more health professionals and an adequate environment of care. In the context of clinical trials, it is usually not possible to separate the effect of treatments studied from the effect of the proficiency levels. This work proposes a method to assess not individual proficiency but practice performance of centres participating in clinical trials. We used the LC-CUSUM test to retrospectively assess centres' proficiency. We were finally able to ascertain one date for each centres from which they were considered as proficient. In conclusion, LC-CUSUM test seems suitable for centre's performance assessment to avoid potential bias in analyses due to suboptimal achievements of studied procedures. Prospective proficiency analyses could also be considered in a preliminary stage to standard clinical trials.