Impact of the Hawthorne effect in a longitudinal clinical study: the case of anesthesia