This review found promising evidence for intensive public awareness campaigns to enhance early detection of first-time psychosis. The review was well conducted and the authors' cautious conclusions are likely to be reliable.

To evaluate interventions designed to improve the identification of people with first episode psychosis.

The Cochrane Database of Systematic Reviews, DARE, Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, PsycINFO, CINAHL and IBSS were searched to May 2009 for relevant studies; search terms were reported. Searches were undertaken for grey literature with the Zetoc, SIGLE, NHS NRR, HMIC, PsycEXTRA and United Kingdom Clinical Research Network databases. The reference lists from included studies and relevant review articles were checked to identify additional studies. Reviewers contacted experts to locate further unpublished studies.

Studies that compared early detection interventions against standard service procedure for the early detection of first episode psychosis were eligible. In particular, interventions to reduce delays in treatment and facilitate access to treatment were used. Populations with at-risk mental states or those already treated for psychosis were excluded, as were studies that evaluated specific treatments or the impact on patient outcomes of early interventions.

The studies were conducted in the UK, Ireland, Australia, Norway, Canada and Singapore. In the cluster randomised controlled trials (CRCTs) general practices were provided with educational workshops and follow-up contact compared with general practitioner (GP) practices with no workshop. Other initiatives were multi-focus interventions including face-to-face training for GPs, school visits and large-scale public awareness campaigns. Intervention duration ranged from eight months and four years. Outcomes evaluated included duration of untreated psychosis, severity of symptoms at presentation with psychosis, referrals to mental health services by GPs and incidence of first-time psychosis.

Study selection was performed by one of two reviewers and a subsample of 20 studies were checked by an additional reviewer. Any queries about inclusion were resolved by a third reviewer.

Methodological quality was assessed with the evaluation tool designed by Thomas et al using selection bias, allocation bias, masking, confounders, data collection methods and withdrawals that occurred before and after treatment. Studies were judged to be strong, moderate or weak.

It was not clear how many reviewers performed the quality assessment.

Data were extracted by two independent reviewers on the basis of target groups that the intervention was designed to influence and the mediums of interventions. Any disagreements were resolved by discussion or by a third reviewer.

The results were summarised in a narrative synthesis with tables.

Eleven studies were included in the review: two CRCTs, two prospective two-group natural experiments and seven studies that retrospectively compared two groups or one group with a historical comparison. The CRCTs used appropriate allocation procedures and stated data collection methods but had high losses to follow-up (54 to 64% for the analyses of duration of untreated psychosis). Six studies obtained data from an adequate number of eligible patients. The nine non-randomised studies could not measure or adjust for differences between the intervention and comparison areas.

Multi-focus initiatives (two studies) found significant reductions in duration of untreated psychosis with the intervention compared with controls. Two other multi-focus intervention studies found no differences.

The two CRCTs found no differences in general practice referrals to specialist mental health services. There were no differences in referral source compared with the control group. One CRCT found that general practice patients who had received the intervention were less likely to present at accident and emergency departments.

There was some evidence of an impact on GP referral behaviour with the education initiatives. One trial found a shorter time from initial contact with a patient to referral to mental health services for GPs who received educational interventions. Two studies (one CRCT and one retrospective two-group comparison study) found that GPs who had received the intervention were more likely to refer patients to mental health services than GPs in no-intervention groups.

Two large-scale multi-focus interventions reported similar rates of treated cases of first episode psychosis for the intervention and comparison. One multi-focus intervention found that patients were more likely to self-refer and were less likely to be referred by police than patients in the historical comparison period.

Severity of symptoms at referral was examined in five studies and two found that patients exposed to multi-focus interventions presented with less severe symptoms than comparison groups.

It was unclear how early detection of first episode psychosis could be achieved but evidence was most promising for intensive public awareness campaigns. These initiatives required organisation and resourcing.

The review addressed a question that was broad in scope. Criteria for study inclusion were defined. A number of appropriate databases were searched for relevant studies and attempts were made to identify unpublished literature. It was unclear if there were any language restrictions, which meant it was difficult to ascertain language bias risk. Steps were taken to minimise errors and biases for the performance of study selection and data extraction, but were not stated for quality assessment.

The authors' decision to summarise the results in a narrative synthesis was justified given the heterogeneity of the interventions. Caution was required when interpreting the results because of the inclusion of non-randomised studies, the results of which were associated with a number of potential biases.

The review was well conducted and the authors' cautious conclusions were likely to be reliable.

Practice: The authors stated that large initiatives that involved public awareness campaigns may be more efficient and effective with organisational leadership and resources from regional or NHS planners.

Research: The authors stated that more good-quality studies were required to address gaps in the knowledge base. Campaigns that target community professionals or young people were of value, but these initiatives lacked current evidence of effectiveness.

National Institute for Health Research.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.