On the OCTET and supervised community treatment orders

Supervised community treatment orders (CTOs) were implemented in England and Wales from 3 November 2008, as part of the 2007 amendments to the Mental Health Act 1983 (MHA). CTOs require patients to comply with conditions specified by their clinicians. Failure to do so may result in compulsorily recalling patients to hospital. However, a solid evidence base for the use of CTOs is lacking. Burns et al. have recently completed a randomised controlled trial (OCTET) that evaluated the impact of CTOs on hospitalisation in patients with psychosis, compared with use of section 17 leave of MHA (which allows clinicians to grant detained patients temporary leave of absence from hospital, whilst retaining powers to recall them). They concluded that CTOs did not reduce hospitalisation or improve clinical or social functioning, and recommended an urgent review of their current high usage. Nevertheless, the OCTET has a number of potentially serious methodological problems that limit its external and internal validity, undermining the trial’s outcome, as outlined hereafter. Limitation of external validity: being considered as a clear candidate for the intervention (CTO) was an exclusion criterion, resulting in a potentially substantial selection bias. This infers that any conclusions drawn by the study, only apply to cases when it is not clear for clinicians whether or not a patient would benefit from CTO. Limitation of internal validity: recruitment for the OCTET started in November 2008, immediately after the introduction of CTOs, when clinicians had no prior experience of their use. Recruitment continued for more than two years, during which time clinicians gained relevant experience and their CTO practice evolved. This indicates that CTOs may have been used differently between patients recruited at different stages of the study, and even for individual patients during the 12-month trial period. This ‘‘learning curve’’ effect is well demonstrated by the fact that around a quarter of patients who were randomised into the non-CTO group, were placed on CTOs by their clinicians, during the course of the study. Similarly, a slightly smaller proportion of patients who were randomised into the CTO group, did not receive the intervention. The Framework for Development and Evaluation of RCTs for Complex Interventions to Improve Health, published by the Medical Research Council in 2000 states the following: ‘‘once the RCT has begun, the intervention must not evolve, as the RCT results will be unusable if later participants experience a different intervention than earlier ones’’. Furthermore, there is a lack of clarity with regard to the threshold for recalling patients to hospital under CTO, which is the component of the intervention that has a major impact on the trial’s main outcome (hospitalisation). Some patients may be recalled to hospital at an earlier stage compared with others, depending on the nature of their illness. This potential variation in the intervention was not addressed by the OCTET. It has been shown that CTO is a remarkably controversial intervention. OCTET was a nonblinded RCT and clinicians were fully in charge of delivering and modifying the intervention without any experimental restrictions. Thus, clinicians’ preconceived views may have influenced the course of the 12-month follow up. The OCTET markedly departs from the conventional ‘‘gold standard’’ RCT design, by virtue of non-blindedness, variation of the intervention, as well as excessive protocol violation, while it has lost the advantage of evaluating an inclusive sample of patients, which an observational study would have offered. It actually has been previously argued that RCTs may not be the best method of researching compulsory psychiatric interventions. Ruling out the use of CTOs based entirely on the OCTET’s findings would be dogmatic and ethically unjustifiable.