The National Service Framework for Mental Health (Department of Health, 1999) provides seven standards for delivery of care in the National Health Service (NHS). The aim of Standard Two is that any service user who contacts their primary health care team with a mental health problem should:

1. • • have their mental health needs identified and assessed, and

2. • • be offered effective treatments, including referral to specialist services for further assessment, treatment and care if they require it.

This gives local primary care trusts (PCTs) responsibility to take a lead in implementing shared care protocols to support more consistent, effective and acceptable care between primary and secondary services.

Although the acceptability of clinical guidelines and their derivatives in primary care remains controversial (Reference KendrickKendrick, 2000), there is evidence that they can be effective in influencing professional practice. Recent interest has been generated in collaborative initiatives across the primary-secondary care interface (Reference JankowskiJankowski, 2001). Since those guidelines endorsed or developed by more regional or nationally-representative groups are more likely to be valid than those developed by local groups (Reference Grimshaw and RussellGrimshaw & Russell, 1993), it has been proposed that local groups should concentrate on adapting these to local circumstances (Reference Littlejohns, Cluzean and BaleLittlejohns et al, 1999). In this project, a ‘protocol’ is defined as ‘a local adaptation of a well-designed guideline’ (Scottish Clinical Resources and Audit Group, 1993). Although their numbers are increasing, appropriately well-designed guidelines for mental health care are few and far between, and the National Institute for Clinical Excellence (NICE) has only recently commended two for use in the NHS. What methods, then, are appropriate for developing local protocols? What happens when protocols are needed for practice where no national recommendations exist?

The Liaison at the Interface of Care project was established in 1999 to improve the communication between primary and secondary care in the Paddington area and was commissioned by two local primary care groups (that later merged to form Westminster Primary Care Trust) for the work. The team consists of two project workers and a project lead, and they report to a multi-disciplinary steering group comprising representatives from Westminster Primary Care Trust (previously Kensington and Chelsea and Westminster Health Authority, Marylebone and Westway primary care groups (approximately 110 general practitioners)) and Brent, Kensington, Chelsea and Westminster NHS Mental Health Trust. This project set out to develop local protocols to guide patient and practitioner decisions about the assessment, management referral and shared care of patients presenting with mental health needs in primary care settings for each of the National Service Framework priority areas. Here, we report the outcome of the work on a protocol for working with people with psychotic illness.

Method

The project team comprised a project worker (H.M.) and a consultant psychiatrist (S.M.), and their work was guided by a carefully selected group of advisors.

The development of the protocol involved several stages. In scoping the work, it was agreed by the steering group that clinical recommendations should be separated into two parts, addressing the immediate management of acute psychotic illness and the management of patients with established diagnosis of a chronic or recurring condition. The scoping exercise also identified that the protocols were relevant to the management of people fulfilling the diagnostic criteria for any psychotic illness (excluding bipolar affective disorder and those with an organic brain syndrome), aged between 10-65 years and presenting in a primary care setting.

Box 1. Describing the project's structure and tasks

Steering Group	Clinical and managerial representatives from primary care trust, trust and health authority: guides the work of the Project Team; scopes the protocol topic; agrees guideline sources and inclusion criteria; endorses final protocol
Project Team	Project worker guided by experienced clinician: retrieves and appraises well-designed guidelines; develops protocol template; facilitates consensus/consultations process; formats, prints and disseminates final protocol
Shared Care Consensus Group (SCCG)	Local general practitioners, a psychiatrist and facilitator (max. 8 members): modulates recommendations to suit local circumstances
Community Mental Health Team (CMHT) Advisory Panel	Local multi-disciplinary CMHT members: advise SCCG about relevant aspects of referral criteria and joint working
Project Advisors	Local academic, clinician, manager and user group: respond to Project Team's technical or practical difficulties between Steering Group meetings

Relevant articles were identified using two evidence-based electronic databases (EBMH and TRIP) and the Royal College of Psychiatrists' Research Unit's Clinical Guidelines Bibliography (2000). Both electronic databases were searched using relevant key words, and the printed bibliography was hand searched. Documents relevant to the clinical topic, published in the English language by an international, national or regional health care organisation or consortium where included when they were relevant (in part or in full) to the care of people between the ages of 16-65 years. Finally, a clinician (S.M.) reviewed the list of articles retrieved and suggested any that had been very recently published or otherwise omitted. All articles fulfilling these criteria were appraised using an adapted version of the Appraisal Instrument for Clinical Guidelines (Reference Cluzeau, Littlejohns and GrimshawCluzeau et al, 1999), using 11 questions that focused on the rigour of the development process, disregarding issues of presentation, clarity and implementation. Those documents that scored at least six were then read in detail.

The included documents were scrutinised and their clinical recommendations entered onto a template describing the principal clinical decision points, i.e. assessment, indications for treatment, effective treatments, first-line treatment, further treatment, indications for referral within the Primary Health Care Teams or to the Community Mental Health Teams. The template's rows captured a succession of recommendations made by each of the source documents about clinical actions, allowing them to be compared and contrasted.

Once completed, the template listed a large number of recommendations, grouped together according to practice area. These fell into two groups; firstly, there were those that were consistent across guidelines and a single summary statement of these recommendations were made. Secondly, there were a smaller number of areas where no such consensus emerged and two or more contrasting summary statements were needed.

Over one half-day session, a facilitated shared care consensus group met to consider the appropriateness of the protocol's recommendations, particularly those areas where recommendations from different sources conflicted. The group used an informal consensus method and invited the advice of a local psychiatrist, community psychiatric nurse and clinical psychologist in matters relating to referral to secondary services, and shared working practices. Since the final protocol was long, it was also summarised to a briefer version that contained the protocol's core recommendations and these were presented as a desktop, A4-sized flow-chart. The final draft was circulated widely by post to local Primary Health Care Teams, Community Mental Health Teams and user and carers' groups for comments on presentation and clarity before completion. The work of developing the six protocols began in October 2000, and was completed within 6 months.

Results

The systematic search identified 22 articles, of which nine reached the criteria for inclusion. Both protocols included a framework describing the respective roles of the general practitioner and Primary Health Care Teams compared with psychiatrist and Community Mental Health Teams in the patient's shared care. The ‘Management of patient with an acute episode of psychosis’ protocol provided a comprehensive range of recommendations regarding assessment of risk and psychological and social functioning, indications for immediate or routine referral to specialist services, and prescribing advice. ‘Managing a patient with a chronic or recurring psychotic illness’ also included recommendations about other aspects of ongoing primary care and provided criteria for liaising with, or referring to the Community Mental Health Teams, including when immediate referral was indicated. The protocol was formatted both as a full text version, and as desktop flow chart (see Figures 1 and 2 ‘Desktop decision flow charts’).

Figure 1 Managing a patient with symptoms of an acute psychosis illness.

Figure 2 Managing a patient with symptoms of recurring or chronic psychotic illness.

Discussion

We have described a methodology for developing a shared care protocol for the management of people with psychotic illness in local primary care teams. A total of six National Service Framework protocols were developed using this method during the 6-month project. The development process was systematic, rapid and responsive to local circumstances. It encouraged wide participation from the local mental health constituency, and proved efficient and feasible.

Establishing local mechanisms to adapt national guidance into local protocols has been promoted as one method for implementing evidence-based practice more widely in the NHS (Reference Littlejohns, Cluzean and BaleLittlejohns et al, 1999). At the time this project began, no clinical guidelines had been endorsed by NICE for use in the NHS, although Treatment Choices in Psychological Therapies and Counselling (Department of Health, 2001) was published during the project and NICE (2002) prescribing recommendations have since been published. In the absence of documents fulfilling the definition of an evidence-based guideline, we had to adopt less stringent criteria to capture the ‘best available evidence’. Additionally, we wanted to develop protocols covering both clinical care, such as prescribing and service delivery, such as referral criteria and shared care. We recognised that many of the documents we considered contained recommendations that were suggestions for, rather than definitions of, good practice and these provided a valuable starting point for discussion about approaches that might work best in our localities. If our inclusion criteria had been too stringent, a great deal of useful material would have been ignored. However, as more evidence-based material is published, the protocols' inclusion criteria can be easily adjusted so that they are more discerning for evidence-based material.

It is tempting to regard a protocol as the end product of a development project, but we have viewed it as one aspect of an ongoing and broad dialogue between health care workers, service users and managers. This involves exploration not only of the care locally delivered and the evidence that justifies this, but also the way in which it is delivered and where. It is often tempting to regard a protocol as an opportunity to define best evidence, implement change and improve practice. However, unless just as much (and usually much more) time and energy is invested in talking with practitioners about their existing practice and the local circumstances that support it, these aspirations will not be achieved. At each review date, the current protocols can be reviewed, new evidence retrieved and the feasibility of changes considered. This can then be incorporated as appropriate.

Is it reasonable to have each locality or trust developing its own protocols for what is a universal issue? As well as identifying clinical recommendations already endorsed by expert groups, the consultation and consensus methods ensured these were combined with a style of shared care that was valued locally as feasible. For instance, local audits in our area reflected national findings that at least 30% of people assessed by Primary Health Care Teams as having severe and enduring mental health needs have no current contact with specialist mental health teams. In the light of this knowledge, the protocol recommended the Community Mental Health Teams' involvement in assessment and management planning for all patients presenting with symptoms of an acute psychotic illness. However, in more chronic but stable conditions, the importance of the primary care team's role in providing continuity, recognising subtle changes in the patient's mental state or their tolerance to treatments and remaining vigilant to their physical health is made explicit. These arrangements were considered appropriate in a setting where the quality of primary mental health and physical care was assured and specialist services were easily accessible. This allows secondary care services to prioritise work with those whose needs are most pressing. The project's steering group is now considering the role of information technology templates and other decision support tools, audit and training programmers for local Primary Health Care Teams and Community Mental Health Teams to support this protocol's implementation and those of the other five protocols covering depression, anxiety, substance misuse, post-natal depression and referral for psychological treatment that were developed at the same time. Their clinical recommendations will be reviewed in the light of recently published guidelines in 2003.