The paper by Garety et al ^{Reference Garety, Fowler, Freeman, Bebbington, Dunn and Kuipers1} was an extremely important and methodologically robust examination of the impact of psychosocial interventions for schizophrenia. The editorial by Scott ^{Reference Scott2} in the same issue suggested that there has been an overpromise of CBT and the inclusion in the National Institute for Health and Clinical Excellence (NICE) ³ guideline might have been oversold as there was a lack of evidence of efficacy in schizophrenia. There are several points which need to be added to those discussed in the paper and in the editorial.

The hypothesis used to calculate power was based on the primary outcome of relapse from a non-affective psychosis (ICD–10 category F20–29, and not F2 as reported in the paper), using TAU, CBT for psychosis and family intervention as comparison interventions. It is therefore important to focus on this outcome and it is surprising that this was not analysed in greater detail.

The published relapse rates after full remission and from full/partial remission in the no-carer pathway were 35.4% and 37% respectively for TAU and 46.8% and 54.6% respectively for CBT; in the carer pathways they were 21.4% and 25.9% for TAU, 27.3% and 28% for CBT, 22.2% and 20.8% for family intervention. It would have been important to analysis the pathways separately as the no-carer pathway shows a trend for an increase in relapse rates. This was indeed the statistical evaluation in the seminal personal therapy/family therapy 3-year study by Hogarty et al, ^{Reference Hogarty, Kornblith, Greenwald, DiBarry, Cooley, Ulrich, Carter and Flesher4} where offering therapeutic intervention in a no-carer pathway led to significantly increased rates of psychotic relapse. The discussion in the published paper was thus incorrect in the assertion that the effect of having a carer during psychological intervention had not been reported before.

The second table of results showed the mean number of relapses in the no-carer pathway: 0.79 for TAU and 1.17 for CBT; for the carer pathway this was 0.31 for TAU, 0.63 for CBT and 0.96 for family intervention. The relapse rates point towards an increase in hypothesised outcome and the risk of harm or hazard ^{Reference Marlowe5} needs to have been discussed in greater detail, to give balance to what has already been acknowledged to be an oversold intervention.