Background:
Schizophrenia represents a substantial cost to the NHS and society because it is common (lifetime prevalence around 0.5–1%); it begins in adolescence or early adulthood and often causes lifelong impairment. The first 3 years are a ‘critical period’ in which the course of the illness is determined. Hence under the NHS Plan, specialist early intervention in psychosis services were established to care for people who develop psychosis between the ages of 14 and 35 years for the first 3 years of their illness. However, there has been a lack of evidence-based treatments specifically designed for the early years. This is important because emerging evidence has shown that in the critical period it is vital to avoid relapse and prevent deterioration in physical health, as both can drastically reduce the chances of a full recovery.

Objectives:
To develop and evaluate three phase-specific interventions to prevent relapse and/or deterioration in physical health in people with first-episode psychosis. The interventions were (1) cognitive remediation (CR) to improve meta-cognition and insight and enhance engagement in cognitive therapy [evaluated in the IMproving PArticipation in Cognitive Therapy (IMPACT) trial]; (2) a healthy-living intervention to control weight in people taking antipsychotic medication after a first episode of psychosis [evaluated in the INTERvention to Encourage ACTivity, Improve Diet, and Reduce Weight Gain (InterACT trial)]; and (3) integrated motivational interviewing and cognitive–behavioural therapy (MiCBT) to reduce cannabis use [evaluated in the Rethinking Choices After Psychosis (ReCAP) trial]. The trials were conducted to explore the case for larger definitive trials with relapse as a primary outcome measure. However, as small trials do not have sufficient power to detect significant reductions in relapse, each was focused on a relevant primary outcome for which there was sufficient power to detect a significant difference. In all three trials relapse was a secondary outcome in the hope of detecting trends towards lower relapse rates in the presence of effective interventions or a general trend across all three studies towards lower relapse rates.

Design:
Three exploratory randomised controlled trials (RCTs) accompanied by qualitative work employing grounded theory and framework analysis to inform the interventions and determine acceptability (InterACT and ReCAP trials).

Setting:
Five early-intervention services in the north-west of England.

Participants:
Early-intervention service users aged 16–35 years who had recently experienced a first episode of psychosis. Participants in the IMPACT trial were drawn from a waiting list of people referred for routine CBT; those in the InterACT trial were required to have a body mass index (BMI) of ≥ 25 kg/m2 (or ≥ 24 kg/m2 for service users from the South Asian community); and those in the ReCAP trial met Diagnostic and Statistical Manual of Mental Disorders – Fourth Edition (DSM-IV) criteria for cannabis abuse or dependence.

Interventions:
The IMPACT trial involved 13 sessions of CR over 12 weeks; the InterACT trial involved eight face-to-face sessions plus optional group activities over 12 months; and the ReCAP trial involved MiCBT in brief (12 sessions over 4.5 months) and long (24 sessions over 9 months) forms.

Main outcome measures:
The primary outcome in the IMPACT trial was psychotic symptoms assessed by the Psychotic Symptom Rating Scales (PSYRATS). BMI was the primary outcome in the InterACT trial and cannabis use (measured by timeline follow-back) was the primary outcome in the ReCAP trial. Relapse was a secondary outcome across all three trials.

Results:
In the IMPACT trial there was no beneficial effect of CR on psychotic symptoms; however, the amount of CBT required was significantly less after CR. In the InterACT trial a small reduction in BMI in the intervention group was not statistically significant. For participants taking olanzapine or clozapine the effect size was larger although not significant. Outcome data from the ReCAP trial are not yet available. Retention in all three trials was good, indicating that the interventions were acceptable.

Conclusions:
Early-intervention services provided a good setting to conduct trials. The IMPACT trial found that CR delivered by relatively unskilled workers improved the efficiency of subsequent CBT. Across the three trials there was little evidence that any intervention reduced relapse.