Authors' reply

We believe that Dr Alam has misunderstood the difference between efficacy and effectiveness research. The national guidelines on the clinical management of schizophrenia (National Institute for Clinical Excellence, 2002) confirmed CBT to be an evidence-based treatment for persistent symptoms of schizophrenia. However, that decision was based almost entirely on efficacy trials where CBT was given by expert therapists to highly selected samples of people with schizophrenia without comorbidities and using an active comparator such as befriending or supportive counselling (e.g. Sensky et al, 2000). Expert therapists and uncomplicated patients are rare in clinical psychiatric practice. Therefore the next step was to design an effectiveness trial to see whether mental health nurses, without prior experience of CBT could be trained over a short period and then supervised to effectively and safely deliver brief CBT to large numbers of people with schizophrenia in the community. As this involved raters being masked to group allocation, this was therefore not an ‘open-label’ trial.

In relation to the effect size, it is certainly true that when an antipsychotic is compared with a placebo in drug-naive patients a much larger effect is demonstrable. The patients recruited to this trial were, however, almost entirely stabilised on antipsychotics and had already achieved such improvement from them. The effect size with any psychological treatment added to antipsychotics is always likely to be less than that initially achieved by the medication. We acknowledge that the effect size on symptoms at follow-up is modest but the impact on relapse is significant, clinically and in terms of resource savings, for such a brief intervention.