Abstract
Drs Karagianis, Hastings, and Gray disagree with our assessment of the evidence on the effectiveness of community treatment orders (CTOs). In our reviews of studies from Canada and abroad, we apply recognised principles of evidence-based medicine (EBM) to assess the quality of the research. We found some studies show therapeutic benefits for CTOs. These include studies commissioned by governments, often conducted by nonacademic consultancies and not subject to peer review, and a number of small-scale, uncontrolled studies. These are the kinds of studies conducted in Canada. They do not provide robust evidence for efficacy of CTOs, especially when few high-quality studies report positive outcomes, and all the randomized studies and the best nonrandomized ones find no evidence of therapeutic benefit. 1,2
The letter by Karagianis illustrates the need for EBM. He cites numerous studies showing many seriously ill patients disengage from treatment, and this is a serious problem. We share that concern. But studies showing the existence of a problem cannot establish that a certain remedy will fix that problem. On the question of remedy, Karagianis simply asserts that CTOs will address engagement and will reduce hospitalisation. He does not engage at all with the evidence we reviewed on whether CTOs have those effects. The premise of EBM is, however, that mere assertions of efficacy do not suffice.
Hasting and Gray repeat criticisms levelled at the OCTET trial that have been addressed fully elsewhere. 3 –8 That trial did not merely compare 2 forms of community compulsion: controls were randomised to voluntary status via leave from hospital, which lasted on average 8 days. Those randomised to CTO were on an order for an average of 180 days. 9 If CTOs are effective, why was no difference in outcomes found between these 2 groups? 9,10 Why, similarly, was no difference found in the 2 other randomised controlled trials, both from the United States, where controls received entirely voluntary care? 11,12
Patients were not cherry-picked for OCTET: the clinicians who recruited accepted that genuine equipoise existed between the CTO and the control condition. We are puzzled by the assertions that patients who did not understand the study and so could not provide informed consent (n = 9) would represent the majority of patients on CTOs in Canada 13 and that those who declined to participate (the figure is 20%, not 33%)—when taking part would have given them a 50% chance of not going on a CTO—were those CTOs “are designed for.” Where is the evidence in support? Crossovers between arms are not ideal but common in effectiveness trials. In this case, they are less of a worry given that the per-protocol analysis gave identical results and there was “not an inkling” of difference in outcomes between the 2 groups. 9,10,14
Kargianis is right that absence of evidence is not evidence of absence. But we examined the current evidence, using scientific criteria, and the weight does not favour CTOs.