Authors' reply

We agree with Dr Goh that the 37.5% reduction in monthly admissions cannot be directly attributed to the implementation of the crisis resolution and home-based treatment service. Only high-quality randomised controlled trials can make unbiased assertions about the effectiveness of interventions without raising doubts about confounding variables. Such trials have already been conducted using operational definitions of ‘crisis’ and have demonstrated favourable results (Johnson et al, 2005). Unfortunately these studies often lack external validity. We believe that the strength of our study relates to its naturalistic design, as all patients were included. It is essentially a service evaluation which demonstrates the effectiveness of crisis resolution services in everyday clinical practice.

Dr Goh highlights potential confounding variables and we agree that many of these factors warrant further investigation. We are aware of no significant changes in the factors identified, in particular, the use of independent hospitals is often carefully regulated and their use is minimal given the financial implications. Unfortunately very few services introduce 24 h crisis resolution services without simultaneously closing in-patient beds. Our study must be considered in the context of recent randomised trials. When taking these into account we believe that crisis resolution and home-based treatment services reduce admission rates, although we accept that other variables may have an effect, and this requires further research.