Strategy for early detection and assessment of frank psychosis

The pathway to care in first-episode psychosis can be long and traumatic, and most people have active psychotic symptoms for 1–2 years before treatment (e.g. Larsen et al, 1996). Long duration of untreated psychosis is associated with ineffective and demoralising help-seeking and a variety of traumatic events, including high rates of involuntary hospital admission (Reference Johnstone, Crow and JohnsonJohnstone et al, 1986).

A strategy identifying pathways from the onset of psychosis to successful engagement with specialist treatment is vital: this may reveal circuitous and undesirable routes and opportunities for earlier intervention.

In the UK, most cases of first-episode psychosis pass through primary care, so the interface with secondary care is of particular interest. However, British data on the treatment of first-episode psychosis in general practice are sparce. Studies of schizophrenia in primary care describe problems with reviewing psychiatric status and difficulties of communication with secondary services (Reference Lang, Johnstone and MurrayLang et al, 1997). Overseas data suggest that general practitioners (GPs) have a low index of suspicion of the presence of psychosis and little confidence in making the diagnosis and that they see mental health services as slow to respond to referrals (Reference Lincoln and McGorryLincoln & McGorry, 1995).

A number of strategies have been used to strengthen the links between primary and secondary care for people with serious mental illness. One such strategy has been the introduction of standards for and audit of communication between the two services and the development of shared care. These could become part of contractual arrangements (Reference Bindman, Johnson and WrightBindman et al, 1997). Another has been to site a liaison CPN within primary care, taking care to ensure that the appropriate patient group is prioritised (Reference Lang, Johnstone and MurrayLang et al, 1997). The Early Psychosis Prevention and Intervention Centre (EPPIC) programme in Melbourne, Australia, has published on the internet guidelines for GPs on the early detection and management of psychosis (http://home.vicnet.net.au/~eppic/).

The Birmingham Early Intervention Service is developing a shared-care programme with local GPs that involves GP training, increased communication regarding client progress and treatment and shared care of selected clients after a period of intensive management by the early intervention service. The service also maintains good relationships with the local police and has conducted some training with them in the recognition of psychosis. Regular audits of pathways to care are conducted.

Focus on therapeutic engagement

Engagement, or the formation of a therapeutic alliance, is an independent predictor of treatment retention rates and a range of good of symptomatic and functional outcomes in psychosis. However, it is difficult to achieve (Reference Frank and GundersonFrank & Gunderson, 1990). In particular, young people doubt the usefulness of professional help, and their negative stereotypes of mental illness and fear of mental health services form barriers to self-referral (Reference Lincoln and McGorryLincoln & McGorry, 1995).

In the face of the failure of clients with first-episode psychosis to attend out-patient appointments or to take antipsychotic medication, the temptation to ‘wait and see’ is considerable. However, this approach often ends in a crisis and preventive opportunities are lost. For example, after a first episode of psychosis, clients who cease contact with mental health services are more likely to be readmitted under the Mental Health Act (Reference McGovern and CopeMcGovern & Cope, 1991).

There has been little systematic study of engagement of clients with first-episode psychosis. Opinion leaders in the field (e.g. Birchwood et al, 2000a), however, agree that it is fostered by a search for common ground with the client, an avoidance of a premature confrontation of their explanatory model of illness and the delivery of treatment in as flexible a manner as possible. If a client does not accept that he or she is ill, practical assistance and an initial focus on the resolution of a problem of importance to the individual, but unrelated to the psychotic disorder, may be helpful. Studies of substance misuse suggest that aspects of the service-delivery process, such as home visits, short waiting lists and frequent client contact with a single worker, can also maximise client engagement (e.g. Henggeler et al, 1996).

For the client to develop a positive, trusting relationship with a mental health worker, a single keyworker needs to be allocated as early as possible, regardless of whether the client is perceived as having ‘complex needs’. This worker should preferably be assigned for the entire ‘critical period’. Failure to take prescribed medication or attend appointments should not lead to reduced contact or discharge; in such cases contact should be increased. Staffing levels must take into consideration the fact that frequent contact is often needed to engage such clients.

Furthermore, since most people suffering from first-episode psychosis are young, best-practice services for early intervention must also be ‘youth attractive’. In their approach to the engagement of clients and in the interventions offered they should reflect youth culture and the legitimate aspirations of young people concerning work and autonomy.

The assertive outreach model provides a good structure for workers, and different aspects of the model encourage engagement. Low case-loads allow the time required for the development of a therapeutic relationship and for persistent follow-up of individuals in danger of being lost to services. The ‘team’ model, which demands that all the staff know all the clients, enables the best matching of staff skills to client needs as well as continuity of input when there are problems in particular relationships or breaks owing to holidays and so on. The use of extended hours and community outreach provides the flexibility to see patients at the most appropriate time and place. In our own service, keyworkers have an individual case-load of 15, somewhat higher than recommended in a strict assertive outreach model. All members of the team know all the clients. The service operates 7 days a week, with flexibility of hours if necessary. Workers attempt to engage the client on the basis of the latter's self-perceived needs, building on his or her strengths. Much of the work involves practical help with aspects of everyday life, including benefits and housing. Providing the majority of services within the team gives continuity, which enhances engagement. The team is involved with clients during admissions and the team psychiatrist is responsible for in-patient care. The service also employs former service users, to whom clients may better relate, as support workers. A vocational assistance programme led by an occupational therapist is available if needed. The client does not have to accept an illness model of psychosis for initial engagement, and cognitive-behavioural therapy (CBT) is available to address problems that the client perceives. The service also has a small social budget, which is very important in allowing keyworkers to engage with clients in low-stigma settings such as sporting events. This is also formalised in a weekly social recovery group. There is a staff v. clients football match most Friday afternoons, at which relatives and friends are welcomed as players or spectators.

Comprehensive assessment

The stress-vulnerability model conceptualises the development of psychosis as the action of social or psychological stressors acting on the vulnerable individual. This implies the need for a full assessment of mental state, risk and biological, psychological and social vulnerabilities and stressors.

People with first-episode psychosis often come to care after a prolonged symptomatic period that may have led to unstable living or financial circumstances. Furthermore, comorbidity with problems such as substance misuse, depression, suicidal thinking, social avoidance and phenomena similar to post-traumatic stress disorder (PTSD) are common, and need assessment and treatment both in their own right and because of their potential to act as stressors provoking relapse (Reference Birchwood, Todd and JacksonBirchwood et al, 1998).

In our service, initial assessment involves a traditional psychiatric history and mental status examination plus relevant physical investigations. However, assessment is not considered complete until information is obtained about the client's sense of stigma, his or her own explanatory model of the illness and view of the future. The individual's access to appropriate social resources, opportunities and social networks is also assessed, as are the current and ‘best ever’ levels of occupational functioning. Most clients will allow us to talk to their family, and the quality of family relationship is noted, as well as the family's attitude to, and understanding of, the client's illness. These data are summarised as the psychological and social needs and strengths of the client, the strengths and needs of the family and the client's own perception of his or her needs.

Embracing diagnostic uncertainty

The embracing of diagnostic uncertainty is crucial in work with first-episode psychosis, since many of the symptoms found in the prodromal or early stages of psychosis are sometimes found in normal teenagers (Reference McGorry, McFarlane and PattonMcGorry et al, 1995). For true early detection, referring agents must feel free to refer clients for expert assessment on the basis of a suspicion rather than a certainty of psychosis, and a low threshold for reassessment must be set.

The application of a diagnosis too early in the course of a functional psychosis not only has been found to be unreliable, but may also be harmful. The early course of psychosis is often characterised by changing symptoms, and premature diagnosis (e.g. of schizophrenia) tends to engender pessimism in clients, families and staff, which may exert a negative influence on the course of ensuing psychoses (Reference McGorryMcGorry, 1995).

Provided that organic causes of psychosis are excluded, a symptom-based approach to treatment has been advocated. This might include prescribing, for example, mood stabilisers in the presence of a full or partial affective syndrome. Decisions about diagnosis can then be deferred until some degree of symptom stability has emerged.

Our service, like most similar ones, has adopted the above practice. We have observed that clients generally find the term ‘psychosis’ more tolerable than ‘schizophrenia’ and we focus psychoeducation on risk and vulnerability factors for psychosis in general. A series of booklets have been developed for this purpose (available from M.B. upon request).

Treatment in the least restrictive setting using low-dose medication

McGorry et al (1991) found high levels of symptoms of PTSD following hospitalisation for an acute psychotic episode. It is therefore recognised that exposure of treatment-na¨ve patients to the adverse experiences associated with acute psychiatric wards must be minimised. While some services have tried to reduce the trauma for clients of an in-patient admission for first-episode psychosis by providing special nursing, single rooms and accommodation for family members, others have found treatment at home to be a viable alternative to hospitalisation.

For example, Fitzgerald & Kulkarni (1998) successfully treated 22 of 37 first-episode psychosis patients entirely at home, using home visiting as frequent as twice daily. The success of home-based treatment depended not on the degree of initial psychopathology, but on the degree of family support available to the client. In our service, we have used the 24-hour trust-wide home-treatment team to provide out-of-hours assistance to clients, with good results.

The aim of pharmacotherapy in first-episode psychosis should be to maximise the therapeutic benefit while minimising adverse effects. A number of guidelines for this can be drawn from the literature. These are listed in Box 3 and discussed in detail below.

1. 1. If possible, have an antipsychotic-free observation period during which the diagnosis of psychosis can be confirmed and organic causes can be excluded. During this time benzodiazepines may be used for tranquillisation, as an alternative to antipsychotic medication.

2. 2. At present, there is no methodologically sound evidence that novel antipsychotics (excluding clozapine) are more effective than conventional ones in treating the positive and negative symptoms of psychosis in unselected first-episode psychosis populations. However, they have been repeatedly found to cause fewer side-effects (e.g. Emsley, 1999), with resultant improved adherence to treatment. The escalating cost of atypical antipsychotics and their impact on service budgets have led to greater scepticism about their use, especially since trials have not compared their side-effects with those of very low-dose regimes of conventional antipsychotics. If conventional antipsychotics are chosen as a first-line treatment, a low threshold for change to atypical antipsychotics is needed in the face of extrapyramidal side-effects, poor response or non-compliance. Similarly, factors that may increase the likelihood of the early development of tardive dyskinesia (e.g. organic brain disease) should prompt the early use of atypical antipsychotics, which it is hoped will lead to a lower incidence of this side-effect (see Bebbington, 2000, for a review of the area).

3. 3. The positive symptoms of most patients with first-episode psychosis will respond to 2–3 mg haloperidol equivalents daily, but they may take 2–4 weeks to do so. There appears to be no response advantage in exceeding doses of 6 mg haloperidol daily and doses above this are likely to cause unpleasant side-effects and potential non-compliance. Similarly, 2–4 mg risperidone daily should be adequate in most cases. There may be some advantage in prescribing higher doses of antipsychotics in the treatment of hostility, but benzodiazepines are to be preferred for this purpose owing to their absence of extrapyramidal side-effects (see Remington et al, 1998, for a review of the area).

4. 4. With sufficiently assertive treatment involving changes in antipsychotics as necessary, symptoms in the vast majority (Lieberman et al, 1993, report 83% in their study) of patients with first-episode psychosis will remit. Therefore, remission of symptoms, not adjustment to them, should be the aim of treatment.

5. 5. Studies have shown that, even with changes in antipsychotics (excluding clozapine), remission rates of positive and negative symptoms tend to plateau after 3–6 months of treatment following a first episode of psychosis (Reference Edwards, Maude and McGorryEdwards et al, 1998). Failure of symptoms to remit after 6 months of treatment with adequate doses of two antipsychotics (preferably including one atypical agent) and psychosocial input as appropriate should trigger prompt action. For example, review for unsuspected organic factors, mood disorder, substance misuse or non-compliance with medication: then consider using CBT and/or clozapine for persistent symptoms.

6. 6. It is not clear for how long people with first-episode psychosis should continue maintenance antipsychotic medication. Current consensus guidelines suggest ceasing antipsychotics after 1–2 years of remission following a first episode of non-affective psychosis. Future guidelines may recommend a longer period of treatment, in light of Robinson et al's (1999) 5-year follow-up of people with first-episode schizophrenia, which showed that relapse rates rose quickly to 51% over the first 2 years after index admission, then plateaued to a certain extent, reaching 78% by 5 years. Some people with first-episode psychosis stay well with shorter periods of antipsychotic treatment (Reference Crow, MacMillan and JohnsonCrow et al, 1986). Identifying this group is a challenge for future research, but it is unlikely to include those who have had psychotic symptoms for more than 1 year prior to treatment (Reference Crow, MacMillan and JohnsonCrow et al, 1986).