Evidence matters (hopefully). Dr Owen (like Dr Curtis ^{Reference Curtis1} whom he cites) fails to distinguish between intervention and outcome in the OCTET trial. The intervention is the imposition of a community treatment order (CTO). The time under initial compulsion (183 v. 8 days on Section 17) demonstrates a clear and unequivocal difference. Where his figure of only 50% of CTO patients experiencing compulsion comes from baffles us. The difference in the total time under compulsion during the 12-month follow-up that he cites includes the difference between the two outcomes (which includes in-patient compulsion from readmissions in both groups). There is no evidence that recruitment and selection were biased in any way and again we fail to understand on what Drs Owen and Curtis base this criticism. We adhered to the highest research standards throughout and the study has been extensively and rigorously peer reviewed.

Dr Mustafa in his letter advances no scientific critique of our work but does articulate the common response of many clinicians - ‘I have seen it work’. We have sympathy with this - we both entered this study expecting to find improved outcomes from CTOs. However, they do not deliver them and we were as disappointed as Dr Mustafa. Psychiatry has a long history of clinicians clinging to ineffective treatments convinced that they work. This is not surprising given the variation in outcomes in psychiatry and the fluctuating natural history of psychoses. Naturalistic observational studies do not prove otherwise - they have produced contradictory results, some for, some against. ^{Reference Maughan, Molodynski, Rugkåsa and Burns2} That is why we need rigorous randomised controlled trials. OCTET is such a rigorous trial and its findings, however unpalatable to some, are robust. It is also worth remembering that the only two other trials found the same. ^{Reference Rugkåsa and Dawson3} A profession that aspires to evidence-based practice should take these results seriously.