We acknowledge Dr Raven's point about not having considered the 5-year follow-up data fully. Although we noted in our discussion that the evidence for long-term follow-up was limited, ^{Reference Bird, Premkumar, Kendall, Whittington, Mitchell and Kuipers1} we thank Dr Raven for bringing these recent studies to our attention. At the time of submitting the review, the Bertelsen et al paper ^{Reference Bertelsen, Jeppesen, Petersen, Thorup, Øhlenschlaeger and le Quach2} was the only one to have examined the effects of early intervention services at 5-years following randomisation. The study showed no beneficial effect of such services over standard care in terms of positive, negative and general functioning symptoms, making its unique finding tentative. Furthermore, as nearly 50% of participants were not included in the analysis, ^{Reference Bertelsen, Jeppesen, Petersen, Thorup, Øhlenschlaeger and le Quach2} we felt it would be best to include the lack of evidence as a limitation in the Discussion.

As highlighted by Dr Raven, the more recent paper by Gafoor and colleagues ^{Reference Gafoor, Nitsch, McCrone, Craig, Garety and Power3} was published after our review had been submitted. This paper also suggests that the beneficial effects of early intervention services at 5-year follow-up are not sustained in terms of number of readmissions, giving more certainty to the view that the beneficial effects of these services may not be sustained once the treatment is ended. It is worth noting that in both studies, the intensive early intervention services were phased out after the end-point data collection period. In our review we concluded that the available evidence ‘raises the possibility that comprehensive services comparable to those described here as early intervention services, which include a full range of evidence-based psychological interventions, should be considered for people with established psychosis’. ^{Reference Bird, Premkumar, Kendall, Whittington, Mitchell and Kuipers1} The fact that the effects of early intervention services were not sustained once individuals were referred back to standard care, as demonstrated in the two studies, we think supports this idea. We did not think that it appropriate to delay the paper, as we feel that our conclusion is consistent with that reported by Gafoor and colleagues, who note: ‘Aside from limited statistical power, the absence of a difference in outcome between the two groups at 5 year follow up may reflect the withdrawal of the specialised intervention after 18 months (when there was a significant group difference), further investigation of this issue will require trials involving longer duration of specialised treatment’. ^{Reference Gafoor, Nitsch, McCrone, Craig, Garety and Power3}

It is useful that Dr Raven has brought these papers into the discussion and we feel that, on balance, the evidence from our review is still supported. Although there is now some evidence that the long-term effects of early intervention services in their present format may not be sustained once treatment is removed, a meta-analysis of long-term outcomes would still not be possible, as the papers do not share any common measures of outcome. Therefore, we still believe, as do Gafoor and colleagues, ^{Reference Gafoor, Nitsch, McCrone, Craig, Garety and Power3} that further research examining all these outcomes is warranted. Furthermore, research is needed to assess the effectiveness of services akin to the early intervention services that we studied, namely ones that provide a high level of support and a full range of interventions for all individuals at any stage of psychosis.