In their trial of cognitive–behavioural therapy (CBT) and family intervention for relapse prevention in psychosis, Reference Garety, Fowler, Freeman, Bebbington, Dunn and Kuipers1 Garety et al state: ‘There were no differences between the groups, in either [the no-carer or carer] pathway, in the primary outcomes of patterns of remission and relapse’. However, data in their Table 1 indicates that more patients who received CBT relapsed than those who received treatment as usual (TAU) (CBT 60/122, TAU 41/119 for all the patients randomised to CBT or TAU). A statistical analysis (logistic model) for the proportion of relapses reveals a significant reduced relapse frequency for TAU.
The differences remain significant (P=0.0153) when only patients in the no-carer pathway are considered (CBT 53/97, TAU 34/92), but there are no differences for those in the carer pathway (CBT 7/25, TAU 7/27), although here the numbers are small.
It is possible that differences in gender and age distribution between the CBT and TAU arms of the trial, or even differences between centres, could have led to different results in the statistical analyses performed by the authors. However, randomisation should have minimised such differences and the authors make no mention of them in the paper.
Hence, on the basis of the results reported, CBT appears to have a detrimental effect on relapse in non-affective psychosis.