Audit of early intervention in psychosis service development in England in 2005

AIMS AND METHOD 

This study provides an overview of the development of early intervention services for psychosis across England in February 2005. A bespoke self-report audit tool was completed by key informants across the eight regional development centres of the National Institute for Mental Health in England.

RESULTS 

Out of 117 teams identified, 86 have funding, of which 63 are operational with case-managed patients (as of February 2005). Only 3 teams meet all 10 audited early intervention fidelity requirements and there are variations in service model, delivery setting and resources across teams.

CLINICAL IMPLICATIONS 

Current inequity of access and the early, fragile nature of service development means that early intervention in England has reached a critical phase requiring consolidation.

The NHS Plan (Department of Health, 2000a) promised 50 early intervention services for psychosis in England by December 2004 and their development was supported by a policy implementation guide (Department of Health, 2000b). Fifty new early intervention services would 'reduce the duration of untreated psychosis (DUP) to a service median of less than 3 months . . . and provide support for the first three years' (Department of Health, 2002).
Expert consensus opinion confirms essential elements of early intervention (Marshall et al, 2004) and international research supports the efficacy of the early intervention model (Edwards et al, 2005). Two recently published randomised controlled trials focused on providing intensive assertive outreach-based care to young people (16^30 years) during the 'critical period' (Birchwood et al, 1998). The OPUS study in Denmark found advantages in terms of readmission, symptoms and quality of life for integrated, sustained treatment over treatment as usual (Nordentoft et al, 2002). In London, the Lambeth Early Onset (LEO) study found that a team delivering specialised care for patients with early psychosis was superior to standard care for maintaining contact with services and reducing readmissions to hospital (Craig et al, 2004). Our study provides an overview of the development of specialist early intervention provision in England and assesses operational services against fidelity to the policy implementation guide.

Method
The National Early Intervention Programme audited all early intervention teams in England using a bespoke self-report audit tool. This tool was developed from a consideration of the policy implementation guide service recommendations and recommendations from early intervention specialists. Consensus was agreed by the eight National Institute for Mental Health in England (NIMHE) regional development early intervention leads on the core features to be explored and question wording (http://kc.nimhe.org.uk/upload/ei%20mapping%20 audit%20questionnaire.doc). Early intervention teams were identified from the Durham Mental Health Mapping (http://www.dur.ac.uk/service.mapping/amh/) and the NIMHE early intervention regional leads' local knowledge and networks. The audit tool was circulated to the early intervention service manager in formed teams or other relevant key informants where services were under development (e.g. primary care trust service development leads). Data were collected by post for all regions except London (n=24) where face-to-face interview was part of a larger pan-London research programme (http:// www.londondevelopmentcentre.org/index.php? topic=66).
Analysis was undertaken using the Statistical Package for the Social Sciences version 12 for Windows to provide descriptive statistics of the data-set, and the qualitative responses were coded to provide a content summary. Model fidelity was assessed using a ten-point cumulative fidelity score rating the presence of ten structural factors described in the policy implementation guide and included as essential elements for provision of early intervention (National Institute for Mental Health in England, 2003;Marshall et al, 2004). The ten-point rating was only applied to teams operational before February 2005 (63 out of 117 audit returns).

Results
The trajectory of early intervention service development (Table 1) describes cumulative progress from 1997 to February 2005. There were 117 teams identified; 86 (74%) have secured funding; 18 (15%) report fixed-term funding; 36 teams (31%) are still being planned; 28 (78%) planned teams have no identified funding. Overall, 63 teams (54%) cite funding as a major concern limiting development. Funded teams cover on average 66% and operational teams 58% of their catchment population. In the south-east of England, only 20% of teams are operational (n=15) compared with 73% (n=15) in the West Midlands and 69% (n=13) in the north-west. More teams are in earlier stages of planning in London and south-east regions where substantial populations cannot currently access specialist early intervention provision.
Reviewing capacity of the 63 operational teams, data returns from 61 reveal: 37 teams (61%) are funded to support all first episodes of psychosis in those aged 14^35; 15 teams (25%) exclude those aged under 16; 34 teams (56%) provide care for 3 years; and 39 teams (64%) are funded to support maximum case-loads of 15 per case manager. In total, 2471 early intervention patients are being case-managed: 70% male, 30% female; 7% in-patients, 93% out-patients. Thirty-two teams (53%) report that their team is structured to reduce the duration of untreated psychosis, but only 23 (37%) report data (team median 9.15 months; range 2 weeks to 24 months; 2 teams (3%) are on target, reporting a median of 3 months or less).
Forty-six operational teams (75%) describe themselves as adherent to the policy implementation guide but the degree of fidelity is variable. Only 3 operational teams (5%) have formally applied for flexibilities using Department of Health flexibilities guidance (National Institute for Mental Health in England, 2003). Greatest divergence is in the provision of designated acute beds and out-of-hours support (Table 2). Fourteen teams (23%) report access to early intervention acute beds, but only 2 (3%) have dedicated sole use of acute provision. Twenty-three teams (38%) offer extended opening hours, but only 6 (10%) open 7 days and 2 (3%) are available at all times. The mean fidelity score is 6.44 (median 6; range 1^10). Predictably, more-established teams (opening before February 2004) show greater fidelity than newer teams (mean score 6.95 (n=21) compared with 6.18 original papers  .004)). Forty-five operational teams (73%) follow a specialist service model. Variations are reported in 16 teams: 9 (15%) describe a 'hub and spoke' model with a specialist hub centre and dedicated early intervention workers embedded within in-patient and community services, 4 (7%) have dispersed early intervention specialist workers in locality teams whereas 3 (5%) describe unique service models. Forty-one teams (67%) are community based, although only 16 (26%) are in non-health settings. Fiftyseven teams (94%) describe liaison with a range of adolescent services, but only 10 (16%) have specialist part-time (0.1^0.3 whole-time equivalents) input from a child and adolescent mental health service psychiatrist (n=7), psychologist (n=5) or qualified youth worker (n=3).

Discussion
The study has several limitations. The audit tool represents consensus interpretation by specialist practitioners of early intervention and service development leads, although it had not been subjected to a formal validation process. There is no accurate central database of early intervention teams (operational and planned) and thus the study sample was built through local intelligence networks and existing early intervention studies; some services may have been missed. The audit window was 3 months. These methodological considerations should be recognised when reviewing the presented overview or 'snapshot' of development of early intervention in England.
This study is the first comprehensive national audit of early intervention, complementing data collected through the adult mental health services mapping projects, which reported 109 teams in July 2005 (Commission for Healthcare Audit and Inspection, 2005).
It is clear that there has been rapid growth in early intervention teams, but many are fragile and coverage is inequitable showing marked regional variations. Out of 86 funded teams in February 2005, 63 are operational. On average, one-third of the population has no local early intervention service.
The majority of teams follow a specialist stand-alone service model recommended by the policy implementation guide, but only three teams meet all ten audited fidelity requirements and there are variations in staffing, funding arrangements and delivery setting. These inequities are further heightened by variations in provision: a third do not cover the full age range; even where local early intervention teams exist a quarter of those under 16 years have no access; nearly half do not provide care for 3 years; longer-established teams tend towards more staff and case-load capacity. Of the anticipated 7500 cases of first-episode psychosis annually (Department of Health, 2000a), only 2471 are currently case managed by an early intervention team.
Some differences in provision appear to relate to the stage of service development: newer teams are still accumulating cases and lack dedicated psychiatrist input. Indicators that are commonly absent include out-of-hours support, designated acute beds and input from the child and adolescent mental health service. Despite the importance of reducing the duration of untreated psychosis, almost half of operational teams are only funded to case manage, with no capacity for early detection. Even fewer teams measure the duration of untreated psychosis and reported ranges beginning at 2 weeks suggest inconsistencies in standardised measurement. A quarter of teams deviate from a specialist service model recommended by the policy implementation guide, although few have formally applied for flexibilities. The audit did not detect regional variations in choice of service model, but those services describing themselves as specialist were modelled on an existing early intervention service, have a dedicated responsible medical officer, and are engaged in research activity. Where teams choose novel service models, formal evaluation is important to ensure quality and to demonstrate effectiveness of alternative approaches. Variation may be acceptable if original papers Pinfold et al Early intervention in psychosis services patient outcomes and cost-effectiveness are equivalent. However, when variability in adherence to the policy implementation guide is combined with inequity of access and funding concerns, a picture emerges where few teams are sufficiently resourced to provide a comprehensive service for the patient and the family. Although the original brief for 50 services each covering 1 million population is translating into a larger number of teams covering smaller areas, any celebration of early intervention provision should be tempered by an awareness of the current inequity of access and the early, fragile nature of service development. On-going national programmes of research, including the First Episode Research Network and the Pan-London Research Network, will aid understanding of the impact of early intervention for patients and families, but key structural challenges remain. The main goal should be to build and consolidate services. This includes improving capacity and geographical coverage to ensure equitable access for all patients and families, particularly those under 16 years, improved measurement of the duration of untreated psychosis using standard methodology, effective engagement with child and adolescent mental health services, early detection strategies and capacity to undertake psychosis education and promotional work. Variation in adherence to the policy implementation guide per se does not indicate a poorer service model but innovative early intervention teams should document the impact of their practices on patient and family outcomes.

Declaration of interest
None.