I have just read the following paper. It is sometimes quoted as showing that CBT or GET have been shown to restore the ability to work. See underlined bit. Free full text: http://archinte.ama-assn.org/cgi/content/full/164/10/1098 or http://archinte.ama-assn.org/cgi/reprint/164/10/1098 When one reads the full paper, there is not enough evidence to suggest that the underlined bit deserves to be in the abstract. It seems to come from Table 6: Here's the text: With regard to GET, there were 66 people in the trial. The actual figures in the Fulcher paper are: 31/66 is where the 47% comes from. However, remember that half the group got another treatment initially (flexibility exercises and relaxation therapy) so one can't say that the GET was the reason for the increase. And there was no control group - these were Oxford criteria patients so maybe they would have improved anyway.