We congratulate Luo et al. (Reference Luo, Law, Wang, Shi, Zeng, Ma, Chow, Liu, Zhao, Liu, Yao and Phillips2018) on the results of their randomised-controlled trial of assertive community treatment in mainland China. They reported statistically significant differences between the assertive community treatment group and the control group with less admissions, less relapses and shorter duration of relapses and increased employment in the assertive community treatment group. They rightly stated that further research with a larger sample will be necessary before firm conclusions can be drawn. Apart from sample size we consider two other issues important:

• • The very low treatment intensity in the control group

• • Whether an intermediate flexible model would bring similar gains and offer greater value and opportunity for spread in such a large country

The added value of orthodox assertive community treatment (ACT) in developed countries is under scrutiny. Although there is a lot of evidence in favour of ACT from studies conducted in the United States (Dixon, Reference Dixon2000), these results were not replicated in Europe. Randomised-controlled trials in the United Kingdom (Killaspy et al., Reference Killaspy, Bebbington, Blizard, Johnson, Nolan, Pilling and King2006) and the Netherlands (Sytema et al., Reference Sytema, Wunderink, Bloemers, Roorda and Wiersma2007) failed to show a reduction in bed use. A likely explanation for this difference is the increasingly marginal gap between the quality of care and treatment in the control and intervention group.

Luo et al. (Reference Luo, Law, Wang, Shi, Zeng, Ma, Chow, Liu, Zhao, Liu, Yao and Phillips2018) mention that the control group in their study should get at least an outpatient appointment in a psychiatric clinic every 3 months, and a home visit by a public health professional also every 3 months, but that this does not always happen due to limited resources, hence patients in the control group received a visit every 3 months and a monthly telephone call. Patients in the control arm of studies in ACT with a negative outcome (Killaspy et al., Reference Killaspy, Bebbington, Blizard, Johnson, Nolan, Pilling and King2006; Sytema et al., Reference Sytema, Wunderink, Bloemers, Roorda and Wiersma2007) received more input, typically averaging fortnightly contact. We would have liked to have seen Luo et al. report the actual number of contacts in both the ACT group and the control group. On the face of it, it would seem that they have tested a ‘Ferrari’ model of community treatment against a bicycle and found the Ferrari superior.

The basic idea behind specialist ACT is offering intensive contact to individuals with high levels of relapse and service use, but it could well be that standard care models can be enhanced flexibly by adjusting the number of contacts during crisis situations. This is the idea behind flexible assertive community treatment (FACT), namely offering routine care when possible and more intensive care when needed. There are currently no randomised-controlled trials for FACT but in a number of non-randomised studies FACT was effective (Nugter et al., Reference Nugter, Engelsbel, Bähler, Keet and van Veldhuizen2016; Sood et al., Reference Sood, Owen, Onyon, Sharma, Nigriello, Markham and Seabrook2017; Firn et al., Reference Firn, Alonso-Vicente, Hubbeling, Holley, White and Jones2018). Though we do not have knowledge of the Chinese context, it could be that FACT or other intermediate models represents the mid-range Toyota in our simple analogy, reliable and designed for mass consumption rather than for the lucky few.