Harrison-Read's observation of the impact of intensive case management on hospitalisation in highly selected heavy service users confirms a clinical observation with which we would generally agree. The UK700 trial (Reference Burns, Creed and FahyBurns et al, 1999) generated considerable controversy and consequently we were reluctant to perform post hoc analyses. Subsequent work has generally confirmed and, to some extent, explained the UK 700 findings (Reference Burns, Catty and WattBurns et al, 2002). It is quite possible that a low case-load size has particular advantages for some groups with severe mental illness, and the UK700 study did find such benefit for those with mild or moderate intellectual disability (Reference Hassiotis, Ukoumunne and ByfordHassiotis et al, 2001); groups with very heavy service use may contain more of such service users.

We did not, however, conclude in our current paper (Burns et al, 2007) that ‘… there is no overall clinical advantage associated with any particular case-load size within the approximate range 1:10 to 1:20’ as Harrison-Read states. Our conclusions are more limited, namely that there is a change in practice across this range but we make no claims about its impact on outcome. Indeed, we make it clear that we cannot draw such conclusions because of the way in which our two proxy measures were constructed. If anything, our findings confirm the likely importance of case-load size by demonstrating that different levels are associated with change in practice. The importance of our findings are that they challenge a strongly held belief that there is a predetermined case-load level at which intensive case management ‘switches’ to assertive community treatment. This view was frequently advanced to discount the UK700 trial's results, claiming that the intensive case management case-load (1:15) was above this critical threshold.

We agree wholeheartedly with Harrison-Read that clarity and precision about case-load size, content of care and effective targeting of the patient population are all necessary for both good clinical care and for meaningful research. We hope that researchers will move on from trying to explain away differences in outcome studies to exploring differences to obtain a better understanding of which components are effective. Our original conclusion that ‘how extra resource is used is more important than how it is organised’ (Reference Burns, Creed and FahyBurns et al, 1999) remains valid.