Treatment for first-episode psychosis has emerged as an important field in psychiatry because of the preventive possibilities (Reference McGorry, Birchwood, Fowler and JacksonMcGorry, 2000). Recovery from psychotic symptoms is common after the first episode, with 75–90% achieving remission from positive symptoms 1 year after treatment (Reference Lieberman, Jody and GeislerLieberman et al, 1993; Reference Edwards, Maude and McGorryEdwards et al, 1998; Reference Addington, Leriger and AddingtonAddington et al, 2003a ). However, even in the case of ‘best practice’ there are some limitations to biological treatments. Adherence rates to medication are notoriously low in patients with first-episode psychosis (Reference Coldham, Addington and AddingtonColdham et al, 2002). Some patients are characterised as ‘slow responders’, others are at risk of treatment resistance even when adherence is addressed and even with ideal biological interventions, relapse rates are very high after the first year of follow-up (Reference Lieberman, Jody and GeislerLieberman et al, 1993; Reference Robinson, Woerner and AlvirRobinson et al, 1999; Reference Edwards, Maude and Herrmann-DoigEdwards et al, 2002). Most importantly, functional recovery (e.g. social, vocational, interpersonal) remains a major challenge. The illness remains disabling and problematic for patients and their families, as symptom improvement is not always matched with functional improvement (Reference Tohen, Strakowski and ZarateTohen et al, 2000; Reference Addington, Young and AddingtonAddington et al, 2003b ). It is therefore critical that we develop treatment approaches to complement pharmacotherapy to achieve improved functional outcome. In addition, such treatment needs to focus on limiting psychosocial damage by offering sustained treatment during this critical early period when vulnerability is at its peak and ‘we have the best opportunity to provide a degree of damage control’ (Reference McGorry, Lieberman and MurrayMcGorry, 2001, p. 156).

CBT FOR PSYCHOSIS: OUTCOME

One particular psychological therapy, cognitive–behavioural therapy (CBT), is gaining recognition as a potentially effective treatment for improving outcome among patients with schizophrenia (Reference Cormac, Jones and CampbellCormac et al, 2002). Four separate randomised controlled trials (RCTs) have demonstrated the effectiveness of CBT. In a comparison of CBT v. routine care (pharmacotherapy) v. supportive therapy significantly more patients in the CBT group demonstrated improvement in positive symptoms (Reference Tarrier, Yusopoff and KinneyTarrier et al, 1998). There was an advantage for CBT at the 1-year follow-up over routine care, but at the 2-year follow-up both CBT and supportive counselling demonstrated an advantage over routine care (Tarrier et al, Reference Tarrier, Wittkowski and Kinney1999, Reference Tarrier, Kinney and McCarthy2000). A second trial (Reference Garety, Fowler and KuipersGarety et al, 1997; Reference Kuipers, Garety and FowlerKuipers et al, 1997) comparing CBT with routine care demonstrated a greater reduction in symptoms for 21% of the CBT group v. 3% of routine care participants. A third study (Reference Sensky, Turkington and KingdonSensky et al, 2000) demonstrated that CBT was effective in treating positive as well as negative symptoms in schizophrenia compared with a supportive therapy called ‘befriending’. Both forms of therapy led to significant clinical improvement in symptoms at the end of treatment, but at the 9-month follow-up only the CBT group had sustained improvements. In a fourth study of acutely ill in-patients (Drury et al, Reference Drury, Birchwood and Cochrane1996, Reference Drury, Birchwood and Cochrane2000), those receiving CBT demonstrated a significantly greater reduction in positive symptoms over the first 12 weeks, a difference that was maintained at 9 months.

Very few CBT trials have focused on populations with first-episode psychosis. The SoCRATES trial used a large representative sample (n=315; 83% first episode) to compare a 5-week treatment package of CBT plus routine care with supportive therapy plus routine care and with routine care alone during the acute phase of the psychotic illness. The aim of the SoCRATES trial (Reference Lewis, Tarrier and HaddockLewis et al, 2002a ) was to determine the impact of CBT during the acute phase of the psychotic illness on accelerating remission from acute symptoms. At 70 days there were trends towards faster improvement of positive symptoms in the CBT group compared with supportive therapy and routine care (Reference Lewis, Tarrier and HaddockLewis et al, 2002a ). At 18 months’ follow-up, CBT demonstrated significant advantages in outcome over routine care and some advantages over supportive therapy (Reference Lewis, Tarrier and HaddockLewis et al, 2002b ).

Cognitively Orientated Psychotherapy for Early Psychosis (COPE; Reference Jackson, Edwards, Hulbert, McGorry and JacksonJackson et al, 1999) aims to facilitate adjustment after a first episode of psychosis. In an open trial those receiving COPE demonstrated improved illness adaptation as assessed by an integration and sealing-over scale (Reference McGlashan, Wadeson and CarpenterMcGlashan et al, 1977) compared with those who had not participated (Reference Jackson, McGorry and HenryJackson et al, 2001). It has been demonstrated (Reference Thompson, McGorry and HarriganThompson et al, 2003) that sealing-over/integration is an important factor related to recovery which is malleable over time.

In the RCTs that compared CBT with varied forms of supportive therapy, the positive impact of CBT was inconsistently diminished relative to the supportive therapy, although never outperformed (Lewis et al, Reference Lewis, Tarrier and Haddock2002a ,Reference Lewis, Tarrier and Haddock b ; Reference Tarrier, Wittkowski and KinneyTarrier et al, 1999; Sensky et al, 2002). In a comprehensive review, Penn et al (Reference Penn, Mueser and Tarrier2004) consider the potential mechanisms behind this reported effectiveness of supportive therapy and conclude that there is an important role for ‘meaningful social contact’ with others. The clear implication from the review is that CBT should target social needs and goals, placing a greater emphasis on the interpersonal context and social consequences of relationships, including the therapeutic relationship.

CBT IN EARLY PSYCHOSIS PSYCHOSIS PROGRAMMES

The results from clinical trials of CBT for psychosis and the need to develop psychosocial interventions for patients with first-episode psychosis make CBT a compelling treatment to consider as an integral part of early psychosis services. Comprehensive early intervention programmes are being developed throughout the world (Reference Edwards and McGorryEdwards & McGorry, 2002). Despite diversity in services that usually reflects the need for congruency with local settings, services usually offer similar components in order to integrate biological, psychological and social aspects of treatment. These typically are ongoing optimal pharmacotherapy and psychiatric case management, plus a range of psychosocial treatments that may include psychoeducation, individual CBT, phase-of-illness-specific groups, vocational services and a family component (Reference Addington and AddingtonAddington & Addington, 2001; Reference Addington, Burnett, McGorry and GleesonAddington & Burnett, 2004). In implementing a CBT component for an early psychosis service several issues must be addressed: (a) the theoretical model on which the CBT component is based; (b) goals of CBT for early psychosis; (c) a clearly defined outline of the approach; and (d) a plan for training and delivery of CBT.

THEORETICAL BASIS OF CBT FOR EARLY PSYCHOSIS

Cognitive models of psychopathology have demonstrated the effectiveness of well-described interventions and of training models that include methodologies for assessing therapist adherence and competence (Reference Vallis, Perris and McGorryVallis, 1998). Furthermore, since the early work of Beck et al (Reference Beck, Rush and Shaw1979) a ‘second generation’ of models of cognitive therapy has developed giving attention to constructs, such as affect, early development, attachment, interpersonal processes and the therapeutic relationship (Reference Guidano and LiottoGuidano & Liotto, 1983; Reference Safran, Vallis and SegalSafran et al, 1986; Reference Greenberg and SafranGreenberg & Safran, 1987; Reference Safran and SegalSafran & Segal, 1990; Reference Ryle and KerrRyle & Kerr, 2002). These ‘second-generation’ models have been incorporated into a meta-model of CBT (Reference Howes, Parrott, Vallis, Howes and MillerHowes & Parrott, 1991; Reference Howes, Vallis, Dobson and CraigHowes & Vallis, 1996), which places five major models of CBT on a dimension from rationalism to constructivism. These different models focus specifically on: (a) automatic thoughts; (b) faulty processing styles and dysfunctional assumptions regarding the self; (c) core cognitions or self-schema; (d) emotional and cognitive development; and (e) interpersonal and interactional factors in addition to cognitions. They are well described elsewhere (Reference Beck, Rush and ShawBeck et al, 1979; Reference Guidano and LiottoGuidano & Liotto, 1983; Reference Safran, Vallis and SegalSafran et al, 1986; Reference Greenberg and SafranGreenberg & Safran, 1987; Reference Safran and SegalSafran & Segal, 1990; Reference Howes, Parrott, Vallis, Howes and MillerHowes & Parrott, 1991; Reference Howes, Vallis, Dobson and CraigHowes & Vallis, 1996; Reference Ryle and KerrRyle & Kerr, 2002). With the exception of Beck (Reference Beck and RectorBeck & Rector, 2000) and Perris & McGorry (Reference Perris and McGorry1998), theorists of cognitive therapy do not address psychosis.

Indeed, as Perris & McGorry (Reference Perris and McGorry1998) suggest, when developing cognitive interventions for psychosis it is clearly an advantage to expand other theoretical ideas to broaden and deepen the approach. In fact, Perris & McGorry go as far as anticipating the challenge as not only integrating different models of CBT but integrating the meta-theories underpinning CBT with theories of cognitive neuropsychology and another neuroscientific aspect of psychosis, such as the neuroimaging paradigms.

What is important for the development of CBT interventions in early psychosis is that this meta-perspective allows the CBT therapist to maintain both a conceptual and theoretical fidelity and yet blend models to maximise flexibility and pragmatism to address a range of problems in a range of cases (Howes & Parrott, 1992; Reference Howes, Vallis, Dobson and CraigHowes & Vallis, 1996). A further rationale for this broader approach is aptly stated by John Strauss:

Thus, CBT for early psychosis must address functional outcome and not just psychotic symptoms. Adopting only one model of CBT is too narrow and is deficient in addressing both the range of difficulties following the first episode as well as the heterogeneity of psychosis. Alternatively, without this meta-perspective the different models of CBT could become so eclectic as to develop into an amalgam of techniques and strategies that appear cognitive or behavioural in nature and which are adopted and used without any real understanding of their implications. Thus, CBT for early psychosis can and should be based on sound theoretical models of cognitive therapy and psychopathology that lend themselves to further refinement and testing.

GOALS OF CBT IN EARLY PSYCHOSIS

The goals of CBT need to take into account not only the symptoms of the illness, but the impact of the illness on an individual. This includes isolation from families and friends, damage to social and working relationships, depression and demoralisation and an increased risk of self-harm, aggression and substance misuse. Persistent symptoms that remain after the early recovery phase are an additional problem and add to the already disrupted developmental trajectory. The goals for CBT are to increase the understanding of psychotic disorders, to promote adaptation to the disorder, to increase self-esteem, coping and adaptive functioning, to reduce emotional disturbance and secondary morbidity and to prevent relapse. When used as a treatment for delusions and hallucinations CBT aims to reduce the distress that these symptoms can cause and to provide the person with strategies and skills to manage residual symptoms in everyday life. The overall goal of CBT in early psychosis is to enhance both symptomatic and functional recovery and as such should be available to every one in an early psychosis programme.

MODULAR APPROACH TO CBT FOR EARLY PSYCHOSIS

We describe a modular approach of CBT for first-episode psychosis. These modules include: engagement, education, addressing adaptation, treating coexisting anxiety or depression, coping strategies, relapse prevention, and treating positive and negative symptoms, and have been guided by a wide range of texts and manuals of empirically supported treatment models that offer both unique and complementary perspectives of CBT for psychosis.

Three texts offer both a theoretical basis for and a systematic guide to the therapy (Reference Kingdon and TurkingtonKingdon & Turkington, 1994; Reference Fowler, Garety and KuipersFowler et al, 1995; Reference Chadwick, Birchwood and TrowerChadwick et al, 1996); Nelson offers a detailed description of CBT for symptoms as a practice manual (Reference NelsonNelson, 1997); and others offer a range of useful case studies (Reference Kingdon and TurkingtonKingdon & Turkington, 2002; Reference MorrisonMorrison, 2002). Drawing from the work of several of the above texts, Systematic Treatment of Persistent Psychosis (STOPP; Reference Hermann-Doig, Maude and EdwardsHermann-Doig et al, 2003) is the only manual that has a specific focus on CBT for first-episode psychosis. Additional works address the individualised formulation (Reference Fowler, Birchwood, Flower and JacksonFowler, 2000), adaptation to the illness (Reference Jackson, Edwards, Hulbert, McGorry and JacksonJackson et al, 1999), coping strategy enhancement (Reference Tarrier, Birchwood and TarrierTarrier, 1992), strategies for hallucinations (Reference Haddock and SladeHaddock & Slade, 1996), relapse prevention strategies (Reference Birchwood and SpencerBirchwood & Spencer, 2001; Reference Gumley, O'Grady and McNayGumley et al, 2003; Reference Gleeson, Gleeson and McGorryGleeson, 2004) and CBT strategies to enhance adaptive functioning (Reference Penn, Mueser and TarrierPenn et al, 2004).

One of the advantages of offering a modular approach is that there is a range of interventions to meet many of the needs of clients with first-episode psychosis and thus there is less need for ‘exclusion criteria’. Phases of illness include acute inpatient, acute out-patient, in recovery, in remission and in prolonged recovery. It is recommended that CBT be introduced to patients with first-episode psychosis once medication, stabilisation and symptom remission has begun, in order to enhance the goal and expectation of optimal recovery. Typically, the length of treatment is approximately 20 sessions over 6 months. This allows strategies to be offered to those who may be experiencing a prolonged symptomatic recovery. Only a brief description of the modules is possible.

DELIVERY OF AND SUPERVISION IN CBT

There are several options for delivery of CBT in an early psychosis service (Herrmann-Doig et al, 2003). First, the overall configuration of services needs to be considered in determining where in the process of treatment the CBT occurs. Cognitive–behavioural therapy can be offered by individual therapists as well as the case manager. This allows case management to continue independently but runs the risk that patients may become overwhelmed by multiple contacts with a range of service providers. Alternatively, CBT can be incorporated into the role of the case manager (cf. Reference Turkington and KingdonTurkington & Kingdon, 2000). This may be practical provided that there are adequate resources. The case managers can offer some CBT with the option of more intensive CBT, if required, being offered by a more experienced therapist. In this context, case managers require adequate supervision from an experienced CBT clinician.

To enhance the delivery of CBT in an early psychosis programme, there needs to be clarification of the different roles of those providing ongoing treatment as well as communication and agreement among team members about treatment plans. Second, clinicians offering CBT need to have access to quality training and quality ongoing supervision which focus on both general therapist skills as well as skills specific to CBT. Third, the service not only requires adequate resources to support the implementation of CBT but must have a clearly stated philosophy and operational policy consistent with the CBT that is being delivered.

DISCUSSION

Offering a specifically designed therapy as described here is based on the fact that the needs of patients with first-episode psychosis are not the same as those with more established forms of schizophrenia and the belief that individually tailored therapeutic input in the early years may have an important long-term impact. Here, we have suggested a theoretically driven modular approach to CBT that places more emphasis emphasis than some of the earlier described models on the interpersonal and social context and the social consequences of symptoms than on symptoms of the illness (Reference Penn, Mueser and TarrierPenn et al, 2004).

There is a risk as early intervention programmes develop that having a CBT component is seen as desirable and may develop atheoretically out of a pragmatic need to offer some of the well-described interventions. First, an early psychosis service needs to carefully consider the nature and philosophy of their programme and how CBT will best fit in terms of treatment delivery, training and supervision. The CBT component then needs to be based on firm theoretical grounds to optimally offer the course of treatment most needed by these individuals.

Considerably more research both conceptual and empirical is required to evaluate the effectiveness of CBT in the treatment of psychotic disorders at this early stage. Despite recent criticism of CBT for psychosis (Reference Turkington and McKennaTurkington & McKenna, 2003), the very few quality RCTs that have been completed (only one with patients with first-episode psychosis) have, in fact, paved the way for future endeavours in CBT. They have demonstrated effectiveness not only for CBT in both chronic and early stages of psychotic disorders but also for intervention in the interpersonal context. The fact that the ‘supportive therapies’ demonstrated effectiveness, although never outperforming the CBT (Reference Tarrier, Wittkowski and KinneyTarrier et al, 1999; Reference Sensky, Turkington and KingdonSensky et al, 2000; Lewis et al, Reference Lewis, Tarrier and Haddock2002a ,Reference Lewis, Tarrier and Haddock b ), is a valuable addition to our knowledge in attempts to develop the best treatments for these individuals. Schizophrenia and other psychotic disorders remain disorders of interpersonal functioning more so than other psychiatric disorders and it makes sense to add the interpersonal focus as in the development of CBT for other disorders. We have a lot more to learn about the phase of illness at which to intervene most effectively (Reference Lewis, Tarrier and HaddockLewis et al, 2002a ).

Thus, before CBT is ‘consigned…to history’ (McKenna in Reference Turkington and McKennaTurkington & McKenna, 2003, p. 478) there is a need for future research to determine the effectiveness of CBT in early psychosis services by including an interpersonal component, examining optimal timing of delivering interventions and considering a wider range of clinical and functional outcomes. Our goal is to find the most effective strategies to help these individuals and we need to learn from these early trials to advance and develop our psychological treatments (Reference Penn, Mueser and TarrierPenn et al, 2004) and then to test their effectiveness.

Clinical Implications and Limitations