Are there any situations where community treatment orders are effective?

Community treatment orders (CTOs) remain controversial partly because the evidence for their effectiveness is rather weak. In general, the better the control of confounding and other bias through randomisation, matching or multivariate analyses, the less likely it is that CTOs reduce readmissions or bed-days. In particular, studies that include consideration of forensic history or dangerousness do not show an effect (Burns et al., 2013; Kisely et al., 2004).

Health service measures may, of course, not reflect important areas such as symptom severity or social functioning. Qualitative research suggests that most clinicians and relatives are positive about CTOs while, unsurprisingly, people with severe mental illness are ambivalent (O’Reilly et al., 2006). In particular, most patients describe feeling coerced even though some also believe that CTOs provide structure to their lives. Quantitative studies of wider outcomes are comparatively rare but, where available, a systematic review found no randomised evidence for effects on treatment compliance, social functioning, homelessness, mental state, quality of life or arrests (Kisely and Campbell, 2014).

Despite the lack of evidence, the use of CTOs will continue in Australia and New Zealand. This being the case, it is important to investigate whether there are any situations where CTOs may be useful. For instance, they may be more effective if applied earlier on in a person’s illness, or, conversely, following several repeat orders (Kisely and Campbell, 2007). However, the use in such circumstances does raise the question of whether CTOs are really a least restrictive alternative (Kisely and Campbell, 2007). In addition, improvements following multiple orders could either be due to long-term benefits or because people who did well were continued on the order as it was presumed responsible. Furthermore, the changes in health service use with repeated orders were restricted to readmissions and the effect on bed-days, a more critical outcome, was not reported (Kisely and Campbell, 2007).

A paper in this issue of the journal suggests that the mode of discharge from an order may also affect outcome (Vine et al., 2016). The authors compared patients who were discharged by the Mental Health Review board, those who were discharged by the treating psychiatrist and those where the order was allowed to expire. Although this was an observational study, the authors adjusted for obvious confounders such as socio-demographic characteristics, CTO duration and diagnosis. They found that people who were discharged by the treating service were less likely to be placed on a subsequent order than the other two groups. The authors suggest that unplanned or abrupt discharge arising from expiry or discharge by the review board may be associated with worse outcomes.

In terms of applicability to clinical practice, there are several caveats. First, only 12% of orders ended through expiry or discharge by the review board. Second, the authors did not adjust for previous health service use or residential care even though these were measured. Third, the study only compared different modes of CTO discharge. It did not compare CTOs with other interventions such as assertive community treatment. An alternative interpretation of the findings might therefore be that CTOs remain ineffective, but that some are even less effective than others. In addition, the worse prognosis in those whose orders expired could be due to patient-based factors such as less insight and a better ability to evade services rather than just the fact that the order was allowed to lapse. Equally, patients discharged by the review board may have been more adept at disguising their symptoms.

Having said all this, if CTOs are to be used, careful discharge planning may optimise the chances of success. Given concerns that CTOs contravene the United Nations Convention on the Rights of People with Disabilities, we need more information such as this to inform policy-makers and service providers on the most appropriate use of these orders.

See Research by Vine et al., 50(4): 363–370.