Cognitive-behavioural therapy (CBT) is well established as an effective intervention for people with unipolar depression, generalised anxiety disorder, panic disorder with or without agoraphobia, social phobia, post-traumatic stress disorder, bulimia, obsessive-compulsive disorder and schizophrenia (National Institute for Clinical Excellence, 2004; Reference Butler, Chapman and FormanButler et al, 2006). It is also increasingly becoming the general public's psychological treatment of choice. The availability of CBT has been severely restricted, but this is now beginning to change, with much greater interest from professionals and government, increased training opportunities and demand from patients and carers. Layard's proposals for a major expansion in CBT (Reference LayardLayard, 2006) are being taken very seriously and pilot implementation sites are being established with the increasing expectation that CBT will become available to those who can benefit (Centre for Economic Performance's Mental Health Policy Group, 2006).

However, as with other mental health interventions, a course of CBT is not a panacea for all mental health problems and there are a number of unanswered questions:

1. • Who is CBT appropriate for?

2. • Who decides and on what basis?

3. • Why do patients drop out?

4. • Why is therapy terminated prematurely?

5. • Why do some patients not derive the mental health gains that they want?

Evidence base for suitability

The evidence base for determining appropriateness for CBT is limited in contrast with the dogmatic assertions often made by therapists. For example, patients may be turned down on the basis of the use of drugs and alcohol, severity of symptoms, risk issues, lacking ‘psychological mindedness’, inability or unwillingness to engage, limited intelligence or presence of Asperger syndrome or organic brain disease. But what level of use of substances or, more importantly, effect of that use should determine suitability? And what responsibility does the therapist have to assist the individual to moderate or even compensate for their intake? Similar questions arise in relation to the other exclusion criteria often used, for example should the therapist take responsibility for developing ‘psychological mindedness’ or for proactively engaging with patients? Failure to attend for assessment or discontinuation of therapy can arise from disillusionment with waiting or with mental health services generally. Termination occurs arbitrarily because the therapist leaves, the patient does not get on with the therapist, travelling difficulties, problems with the timing of appointments, or competing social priorities. The nature of the problems themselves may be relevant, for example, agoraphobia or low motivation, being too upset to talk, too angry, too depressed, hospitalised, or the patient deciding that they have recovered sufficiently so they do not need the appointment.

Treatment resistance

There is always a risk when a treatment becomes very successful of forgetting the patients that are not initially helped by the procedure. But even the strongest advocates of CBT would not expect it to be successful for all. However, arguments are often made such as ‘she has already had therapy in the past, it didn't help’ or ‘he didn't turn up after the first four sessions, he probably couldn't be bothered’. This can be enough, without further consideration, to cast the patient into a therapeutic void where other contact options are automatically ruled out. Patients in this category are often restricted to psychopharmacological interventions and a few out-patients appointments or community visits throughout the year. Others are discharged to the general practitioner or left to their own and their community's resources.

Even when a therapy course is completed, ‘treatment resistance’ is well recognised; CBT does not work with every patient. Why do some patients find it helpful whereas others do not? Why do some patients drop out prematurely? At what point in therapy should a decision be made that CBT, as offered, is not appropriate or sufficient. Should it be at a standard review after 6-10 sessions (often all that is contracted by commissioners, especially in primary care) or after 15-20 sessions, which is more in line with the evidence base? What are the options to move treatment forward once therapy has been offered and, effectively, failed? A wide variety of reasons for ‘ failure’ can be cited, these include:

1. • misdiagnosis

2. • insufficient quantity (e.g. ‘continuation CBT’ may be indicated; Reference Petersen, Harley and PapakostasPetersen et al, 2004), quality or need for a change in the type of CBT provided (e.g. cognitive-behavioural analysis system of psychotherapy; Reference Keller, McCullough and KleinKeller et al, 2000)

3. • therapeutic relationship issues

4. • concurrent social or physical problems.

Assessment of continuing needs

However, there is very little research into an area that will become increasing important as availability of CBT extends. It is currently an issue for those individuals and their carers who have been able to access CBT but continue to have mental health difficulties. Effective comprehensive assessment tools to assess these needs have not been developed even though many relevant rating instruments could be repackaged to do so. Matching individual need with appropriate remedies may include selection from the range of psychosocial and pharmacological interventions available or yet to evolve. However, the evidence base for these judgements simply does not exist. Patients failing therapy probably make up a highly vulnerable group of individuals whose claim to treatment should be heard. To ignore this could be at the peril of both patient and society because of suicidality, costs to the National Health Service (NHS), loss of income, increased morbidity and not least human suffering. ‘Failing’ CBT may be the last straw for patients who have usually waited a considerable time for what may have been presented as the latest evidence-based and remaining effective modern therapy. Evidence indeed suggests that suicidality is increased in patients that stop therapy prematurely (Reference Dahlsgaard, Beck and BrownDahlsgaard et al, 1998). It is reasonable to assume that the therapy-refractory group are at high risk of not only suicide but also long-term illness, with consequences for themselves, their relatives and society as a whole.

Selection of appropriate interventions

Many other potentially relevant interventions are already available from the NHS or social services, including social work input, for example family or community work, socialising, building informal social networks or problem-solving, and a myriad of psychological therapies, for example cognitive-analytic (Reference Ryle and KerrRyle & Kerr, 2002), transactional analysis (Reference BerneBerne, 1970), mindfulness, art and music, assertiveness (Reference DicksonDickson, 1982) and social skills training, anger or anxiety management, psychodynamic and logotherapy (Reference FranklFrankl, 1959). It is possible that substance misuse treatment, occupational therapy, social interventions, vocational advice, peer support, medical referral (for example pain clinic), relationship counselling or hitherto unused approaches will facilitate the patient's progress following unsuccessful therapy. These can be as individual, family, partnership or group interventions. Currently, referral for these occurs because they are available locally or are strongly advocated by individual therapists or referrers. Individual characteristics may be taken into account, but often these are ill-defined or exclude some people, such as those with more severe problems, who might benefit from particular therapies. This position is becoming financially untenable, and without a more systematic basis to the use of these therapies (and supporting evidence) there is danger that many potentially valuable interventions will be cut.

Conclusion

There are therefore two separate strands to developing improved and coherent management of CBT resistance: identification and assessment of patients’ continuing needs and selection of appropriate interventions. Research into treatment resistance is limited but there is a wealth of clinical knowledge available for CBT (Reference Tarrier, Wells and HaddockTarrier et al, 1998; Reference BeckBeck, 2005) and other interventions. With such a complex scenario, consulting with patient, therapists, primary care, psychiatrists and other mental health professionals to systematically identify needs is urgently required. Subsequently the development of consensus expert opinion, including experts-by-experience, could be used to organise this information into algorithms, which can then be evaluated, both qualitatively and quantitatively, and inform the assessment process and indications for individual interventions.

Psychiatry is at the front line in the battle against human misery; many CBT-resistant patients will remain in our care long after they have left, avoided or been rejected by specialist CBT therapists. Our aim is to take the first step towards establishing evidence-based guidance from which decisions can be made about future care in collaboration with the patient. We need to gather the views of patients, therapists, psychiatrists, general practitioners and other mental health staff, and systematically evaluate these using appropriate qualitative methodologies. With CBT becoming more widely available, this debate and supporting research are vital.

Declaration of interest

None.