Authors' reply

The 36-month outcomes of the REACT trial ¹ that ACT shows no clinical advantage over support from standard CMHTs bemuses ACT proponents. Grewal & Cowan report reductions in in-patient service use for ACT patients in Sandwell, but their results are confounded by the national reduction in bed use since the implementation of the National Service Framework for Mental Health, a common problem with before-and-after studies of ACT in the UK. Glover et al ² showed how crisis resolution teams reduced admissions across the country, but ACT teams failed to impact further on this. The Sandwell ACT team's provision of a crisis service may therefore have influenced their outcomes. Nevertheless, it is noteworthy that although in-patient service use was the primary outcome in the REACT study, there were no statistically significant differences in other outcomes between the two treatments, including social functioning, symptoms, needs, attitudes towards medication, adverse events, substance misuse and quality of life. ³

The lack of efficacy for ACT in the UK appears to be related to the degree to which comparison services replicate critical aspects of ACT. ⁴ In the REACT study, the CMHTs shared with the ACT teams four of the seven key components: primary clinical responsibility; community-based; team leader doing clinical work; time-unlimited service (the others being daily team meetings, sharing of case-loads, and operating 24 hours a day).

A consistent finding in studies of ACT is that it is more acceptable to ‘difficult to engage’ patients than standard care, but although UK ACT services are engaging patients, as Shetty rightly states, they are not building on this to deliver the evidence-based interventions likely to improve clinical outcomes. In some cases this is due to inadequate specialist staffing, although this was not an issue in the REACT study. A survey of 222 English ACT teams in 2003 found that only half had a psychiatrist, a fifth had a psychologist and very few had a substance misuse or vocational rehabilitation specialist. In addition, only 12% were operating with high model fidelity and many did not operate outside office hours (C. Wright, personal communication, 2009). A comparison of ACT in London and Melbourne, Australia, found that London teams had around a quarter of the input from a psychiatrist, only half operated outside office hours (versus most Melbourne teams), only a third made the bulk of their contacts away from the office (versus the majority of Melbourne teams), and they scored lower for case-load sharing (C. Harvey, personal communication, 2009).

Inadequate implementation of the ACT model, inadequate delivery of evidence-based interventions, and similarities between key elements of ACT and standard care therefore appear to explain the variation in its effectiveness reported in the international literature. In the UK, ACT teams need to be staffed appropriately and operate with the critical components likely to result in improved outcomes. Otherwise, their lack of cost-effectiveness ⁵ will make them vulnerable to closure.