Results of OCTET study comparing use of S17 leave and CTO has been published (see paper). The rate of readmission was not reduced by CTO compared to use of S17 leave. Other studies have also shown no reduction in readmission. As might have been expected, this finding was despite the period of supervised community treatment being on average more than three times longer on CTO than by using S17 leave.
CTOs were actually introduced because it was believed they would reduce death by suicide and homicide, supported by fantasy estimates of how many lives would be saved (see my unpublished paper). Three people died in the CTO group (two by suicide and one by accidental death) and two people died in the S17 leave group (one by suicide and one by natural causes). As death is a rare event, it's not going to be possible to demonstrate in a randomised controlled trial whether CTO reduces death. However, as the authors of the study say, because of the restrictions on patients' liberty, the costs and benefits of CTOs do need to be assessed.