Dr Bindman and Dr Kripalani have suggested an analysis of the association between DUP and outcomes in subgroups by specific first-line treatment modalities. Unfortunately, it was not possible to analyse this in our meta-analysis, since none of the original studies had used only one treatment modality, but a combination of them in the early phases of treatment. As Bindman & Kripalani point out, and based on current knowledge of the efficacy of treatments in the early phase of schizophrenia, it would not be ethical to study treatment without antipsychotic medication in a first-episode clinical sample. ¹ Also, DUP is usually defined as ending at the initiation of antipsychotic medication, which in clinical practice usually occurs about the same time as other treatment modalities begin; therefore, the included studies give only a little information on the effects of different treatments. However, it is interesting to note that de Haan et al ^{Reference de Haan, Linszen, Lenior, de Win and Gorsira2} investigated the effect of delay in intensive psychosocial treatment by comparing this effect with delay in treatment with antipsychotic medication; and found that delay in psychosocial treatment may be a more important predictor of negative symptoms than delay in antipsychotic treatment.

The discussion about the possible effects of antipsychotics has been rather intense recently. However, the current guidelines for treatment of psychosis and schizophrenia clearly indicate that antipsychotic medications are effective and recommended treatment for active psychotic symptoms, ¹ though there is not so much evidence for the long term (i.e. several years of antipsychotic treatments ^{Reference Harrow, Jobe and Faull3} ). Additionally, the clinical use of these medications is not always straightforward because of their known side-effects and the fact that, in all psychiatric disorders and other illnesses in medicine, there are always patients who do not want to take the recommended treatment. This seems to have been the case in the trial pointed out by Bindman & Kripalani. ^{Reference Morrison, Turkington, Pyle, Spencer, Brabban and Dunn4} When considering the long-term effects of antipsychotics, it is evident that the long-term treatment of psychosis needs to be developed further.

We agree that it would be dangerous to see different treatments as alternatives to each other, and it has been shown that in psychiatry a combination of different treatments is, in general, more effective than any of them alone. ^{Reference Cuijpers, van Straten, Warmerdam and Andersson5} Psychotherapy in the early phase of illness could be effective not only in preventing psychosis at prodromal phase, but also in enhancing adherence to antipsychotic medication. ¹ Current treatment guidelines do not suggest that treatment of first-episode psychosis should include only antipsychotic medication without psychosocial treatment, but rather state that medication is one of the cornerstones of psychosis treatment. We believe there is still a lot to do in developing both medication and psychosocial treatments for schizophrenia, and hopefully active research can support this development.