The developing story
Significant changes in the delivery of cognitive behaviour therapy (CBT) clinical services are underway. Services are re-thinking what they offer – and how they offer it. Until the late 1990s both referrers and CBT therapists knew what to expect if a referral was made for cognitive behaviour therapy. The patient would attend (usually after a wait) a specialist CBT service and (if accepted for therapy) receive between 12 to 20 hour-long sessions with a specialist who had typically attended a one-year postgraduate training course in CBT (BABCP, 2008).
During this first phase of CBT development, high quality services developed that were often led by practitioners who had attended trail-blazing courses such as the Newcastle, Institute of Psychiatry and Oxford Courses. The first phase was marked by the wider identification of CBT as an effective treatment for depression and a growing range of other disorders, increasing demands for CBT and notably for increasing waiting lists. These resulted from the mismatch between growing demand and limited supply (practitioners) to deliver CBT.
During the opening years of the new century a number of questions were raised about the adequacy of this model of CBT delivery. These included:
1. The recognition there could never be enough therapists to provide sufficient specialist-led CBT for all who could benefit (Lovell and Richards, Reference Lovell and Richards2000).
2. A realization that the existing service model was failing as it produced marked inequalities. For example, Shapiro, Cavanagh and Lomas (Reference Shapiro, Cavanagh and Lomas2003) identified a 20-fold discrepancy in availability of accredited CBT practitioners between the best and least well-served populations. In some areas of the UK waiting lists for CBT reached over 2 years.
3. An increasing recognition that many practitioners who were trained on expensive and intensive postgraduate CBT courses failed to continue to utilise these skills after the courses (Ashworth, Williams and Blackburn, Reference Ashworth, Williams and Blackburn1999). Although 24% of course attendees subsequently moved into dedicated CBT posts, around one in five did not use their CBT skills in their current job. This raised questions as to whether the traditional CBT model offering extended sessions transferred well into everyday practice outside specialist CBT centres.
4. An appreciation that non-attendance, poor engagement, and drop-out early in CBT therapy varied between 20 and 50%, raising important questions about which sorts of patients could really make use of CBT approaches (Bados, Balaguer and Saldana, Reference Bados, Balaguer and Saldana2007; Keijers, Kampman and Hooduin, Reference Keijers, Kampman and Hooduin2001).
5. Some critics of the “traditional” expert-delivered model argue that because of significant waiting lists those people who might make the greatest gains from CBT (i.e. those with mild to moderate depression, anxiety and panic) are least likely to be referred. This was because clinical services may raise the entry criteria for therapy and only take on the most complex cases when waiting lists exist.
In their seminal paper Lovell and Richards argued that CBT services had opted for the traditional ways of delivering expert-level CBT whilst overlooking the evidence-base for simpler (low intensity) interventions. This raises the key question: “Is there any clear evidence that one to one expert-delivered CBT is the only evidence-based way of delivering CBT?” Their paper pointed to a range of other ways of delivering CBT, all of which have some supporting evidence and laid the foundations for the current “new ways of working” initiatives that are revolutionizing the delivery of CBT.
These characterize the second phase of CBT service delivery and training. In these, CBT delivery has increasingly moved away from specialist CBT services into mainstream mental health services. Now CBT is available in all sorts of sizes and shapes, and the challenge is how to deliver and support these interventions. The locations where CBT is offered have also diversified: no longer delivered only in specialist services, increasingly workers are based in primary care, the voluntary sector, back to work programmes, and in all sorts of other settings (NIMHE, 2003). The approach of diversifying delivery of CBT fits theoretically within the so-called “stepped care models” (Bower and Gilbody, Reference Bower and Gilbody2005).
For some this is an exciting opportunity, for others it raises concerns of the de-professionalization of delivery, and a dilution of the hard-won skills of CBT, which are watered down into a poorer (and cheaper) second-best. What is certainly true is that there are high expectations on new services to deliver – and also bring cost-savings (notably in terms of helping people back to work). It is by no means certain that this will be the case.
Policy initiatives: doing well by people with depression and Increasing Access to Psychological Therapies (IAPT)
In both England and Scotland there have been moves towards making CBT more accessible. Significant funding has led first to the development of 13 pilot sites in Scotland (SEHD 2006), and later the Doncaster and Newham pilot sites in England (the Increasing Access to Psychological Therapies – IAPT) Project (CSIP, 2008). This latter programme is being significantly expanded with the development of 11 new Pathfinder sites announced in July 2007, and a £170 million pound investment in psychological therapies delivery by 2010/2011 announced in October 2007.
Critics of these initiatives have been vocal and a number of websites question the focus on wider access to CBT approaches (e.g. Critics, 2008a, b, c). These, however, miss a key benefit in delivering CBT in different ways. Although the Treasury hopes to fund IAPT services through cost-savings from helping people back to work, the most important part of the approach from the user point of view is to increase access to evidence-based psychological treatments. The approach therefore challenges practitioners to ask questions about how they use their time to provide such evidence-based treatments. A useful summary of the main issues is Summerfield and Veale (Reference Summerfield and Veale2008).
Delivering CBT self-help: what do we know?
A seminal overview by Gellatly et al. (Reference Gellatly, Bower, Hennessy, Richards, Gilbody and Lovell2007) has helped identify how services should deliver CBT self-help for depression.
a) CBT supported/guided self-help is effective and far more effective than psycho-educational resources, whose impact is clinically negligible.
b) Support is required. CBT self-help offered without support/guidance is not usually effective. Unsupported (pure) self-help has an effect size of 0.06 – compared to 0.8 for supported/guided self-help.
c) Supportive monitoring/encouragement is all that is required. There is no need to add in “therapy”. An implication is that it may be unhelpful and inaccurate to define low intensity high capacity CBT self-help interventions as being “psychotherapy”.
d) Fewer sessions are as effective as offering many sessions of support.
e) The type of support can be either face-to-face or by telephone. It is not known whether e-support such as online forums, live chat or email will be as effective.
f) The qualifications and professional background of the person offering the support does not matter. What seems to be more important is how they can engage and support the person.
g) There is no difference in effectiveness of book-based self-help (bibliotherapy – effect size 0.48) and computerized CBT (CCBT – effect size 0.38). This suggests that CCBT is just another way of delivering CBT self-help.
These findings suggest that two important current initiatives require some additional thought. First, book prescription schemes (Book Prescription Schemes, 2008) where delivery is provided without support are likely to have far less impact than supported models such as those in the South-West of England (Farrand, Confue, Byng and Shaw, Reference Farrand, Confue, Byng and Shaw2008). Second, the creation of the separate NICE technology appraisal of Computerized CBT (CCBT – NICE, 2006) has had a distorting effect on low intensity working by requiring investment in CCBT at the expense of other forms of CBT self-help.
Patient selection for CBT self-help
As there are no differences in effectiveness between main forms of CBT self-help, the choice of approach will be best decided by patient expectations and preference, (DoH, 2001). Are people sufficiently informed to make such choices? For example, Mitchell and Gordon (Reference Mitchell and Gordon2007) found that the credibility of CCBT was initially very poor amongst their sample. However, after a short informational presentation the proportion of people expressing a preference for CCBT increased from below 10% to 30%. Significant thought therefore needs to be put into communication when introducing such low intensity interventions. Providing “taster” examples of different approaches e.g. short excerpts of different materials, screenshots, handouts or short DVDs that explain what can be expected of the different treatments can be used. The positioning of interventions in the clinical service also affects take-up and use. For example, services that positively introduce self-help as the “first step of treatment” (e.g. Bara-Carril et al., Reference Bara-Carril, Williams, Pombo-Carril, Reid, Murray, Aubin, Harkin, Treasure and Schmidt2004) tend to have far higher take-up (78%) than situations where self-help is offered simply as an option to patients who already have the offer of face-to-face therapy with a practitioner (26%, e.g. Whitfield, Hinshelwood, Pashley, Campsie and Williams, Reference Whitfield, Hinshelwood, Pashley, Campsie and Williams2006).
Choosing and introducing CBT self-help resources
Research has been completed on only a small proportion of self-help materials. It seems likely that any reasonably executed CBT self-help written or computerized resources are likely to be effective. Should delivery therefore only focus on materials that have been shown to be effective in randomised controlled studies – or should the basis for decision be based on a judgement whether resources credibly reflect the CBT model?
Accessibility and CBT self-help
Sticking to effective CBT principles is not enough for high effectiveness. An issue in CBT in general, and CBT self-help in particular is accessibility. The average reading age of traditional CBT models is over 17 (Williams and Garland, Reference Williams and Garland2002). To place this in context, the reading age of the Sun newspaper is age 11 and the Times 17 (Martinez, Whitfield, Dafters and Williams, Reference Martinez, Whitfield, Dafters and Williams2008). Around 20% of the population struggle to read at reading age 11. The reading age of all but one commonly used or recommended CBT self-help materials is over 13. Materials should be chosen to reflect a range of abilities and ways of working. Non-written formats such as DVD, video and audiotapes could be used where people struggle to read or prefer not to read.
Supported interventions
New forms of training need to be developed to equip new IAPT workers with adequate skills for offering support. Such training includes university accredited courses that train practitioners in the use of CBT self-help materials (Williams, Dafters, Ronald, Martinez and Garland, Reference Williams, Dafters, Ronald, Martinez and Garland2008). Attending training in using CBT self-help leads to higher confidence and more positive attitudes towards self-help approaches (Keeley, Williams and Shapiro, Reference Keeley, Williams and Shapiro2002). The competencies required to do low intensity working are different from high intensity working (DH/IAPT, 2008a and b).
Some problems with stepped care
Matched care or stepped care?
The evidence base for low intensity interventions has been growing, with high quality trials supporting the effectiveness of CBT self-help for depression, low-intensity forms of behavioural activation (e.g. Hopko, Lejuez, LePage, Hopko and McNeil, Reference Hopko, Lejuez, LePage, Hopko and McNeil2003; Veale, Reference Veale2008), collaborative care approaches that focus on organizational changes and can be used within low intensity work (Bower, Gilbody, Richards, Fletcher and Sutton, Reference Bower, Gilbody, Richards, Fletcher and Sutton2006; Richards, Lovell, Gilbody, Gask, Torgerson et al., Reference Richards, Lovell, Gilbody, Gask, Torgerson, Barkham, Bower, Bland, Lankshear, Simpson, Fletcher, Escott, Hennessy and Richardson2008) and group treatments based on CBT principles (Brown, Elliott, Boardman, Ferns and Morrison, Reference Brown, Elliott, Boardman, Ferns and Morrison2004; White, Reference White1998).
However, there remains a perception that low-intensity treatments remain second-best to high intensity working. For example, the language of stepped care assumes a hierarchy. Phrases such as “stepping up” and “stepping down” imply that certain treatments (e.g. face-to-face) are “better”. Many practitioners report they believe that one to one treatments are superior to self-help. Keeley et al. (Reference Keeley, Williams and Shapiro2002) surveyed 500 BABCP accredited practitioners about their attitudes towards the use of self-help approaches. Most respondents considered self-help to be less effective than therapist intervention in terms of potential benefits to the client (68.5%), client compliance (72.6%), client satisfaction (73.1%), and client expectancy of success (67.6%).
It is unclear, however, to what extent such beliefs are correct. To date few comparisons have been completed. Ghosh, Marks and Carr (Reference Ghosh, Marks and Carr1988) and Cuijpers, van Straten and Smit (Reference Cuijpers, van Straten and Smit2006) suggest that both self-help and one to one may be equally effective. A comparison of group and individual treatment as part of the NICE depression review (NICE, 2004) concluded “It was not possible to make comparisons with either individual CBT, antidepressants or no active treatment.” It is clear that for many of the important questions, answers based upon the research data are not yet possible.
It is also unclear whether there are any differences in effectiveness between the various types of low intensity treatments. Computer and book based interventions for depression appear equivalent in outcome (Gellatly et al., Reference Gellatly, Bower, Hennessy, Richards, Gilbody and Lovell2007). Interestingly, overall there are more credible high quality outcome studies supporting bibliotherapy than computerized CBT. However, it may be that group CBT is more effective than parentally supported bibliotherapy for anxiety in adolescents (Rapee, Abbott and Lyneham, Reference Rapee, Abbott and Lyneham2006). There are currently no other studies that directly compare low intensity CBT with high intensity CBT.
Perhaps the current most helpful approach is to consider the issue of how people like to work and learn. CBT is in essence a self-help form of psychotherapy that shares much in essence with adult learning models. In particular, some people are likely to find it easier to work using self-help approaches than face-to-face. This is not really an issue of superiority at all but of choice. For these reasons there are advantages in referring to “matched care” rather than stepped care. Here, the person is offered the type of intervention and type of support they need and wish to work with. Our initial assessment therefore needs to incorporate an assessment of prior learning and learning styles as well as diagnosis and formulation.
Stepped/matched care supervision issues
In traditional expert level CBT clinical supervision is a key to effective therapy. However, low intensity delivery implies that different competencies (and hence training and type of supervision) are required. Support needs to be appropriate for the task in hand and the person delivering the intervention. For example, for low intensity high capacity interventions the needs will depend on the practitioner and the type/complexity of the patient being supported. Generally few difficulties come up when clear patient selection occurs. However, sometimes new information comes to light, another diagnosis may be present, or for other reasons progress slows, worsens or a crisis occurs. In this case the practitioner needs to detect this, be able to quickly seek advice from a mental health expert, and know how best to quickly access support for the patient. To address such needs, the University of York has developed an online monitoring questionnaire based on the original Seattle approach that automatically alerts online supervisors when “worrying” developments are rated during the completion of online questionnaires (http://www.pc-mis.co.uk/).Similar processes can be put into services using paper and pencil approaches with no need for automation (but a high need for organization).
Criticisms of CBT self-help approaches
The main criticisms of CBT self-help approaches are concerns that some patients require more intensive or flexible approaches than can be attained by a manualized approach. There may also perhaps be a dilution of the potential benefits of higher intensity CBT working if a patient has had prior exposure to low intensity treatment (Graham, Reference Graham2003). Such concerns are testable, and point to the need for future research to record the possibility of side effects/unintended consequences of providing low intensity treatment, including impact on the take-up and use of subsequent high intensity CBT.
To date, what evidence there is suggests that the offer of CBT self-help to patients on a waiting list to see an expert CBT practitioner leads to a significant reduction in subsequent face-to-face therapy with the practitioner (Kenwright, Reference Kenwright2005). Perhaps, however, the greatest concern is raised in the conclusions in Gellatly et al. (Reference Gellatly, Bower, Hennessy, Richards, Gilbody and Lovell2007) that CBT self-help is only really effective when supported by a practitioner. This raises the possibility that much of the effect may be non-specific therapist factors rather than the impact of the materials themselves. Ideally, future studies will include active controls (placebos) in their design to address this – as has been suggested for trials of CBT in psychosis (Turkington and McKenna, Reference Turkington and McKenna2003).
Conclusions
Training and service delivery models in CBT are moving from a first to a second phase of delivery – characterized by different sizes and shapes of CBT and focused on accessibility. However, such accessibility must not come at the expense of quality. New low intensity services should use routine audit tools at each patient contact. At the same time high quality studies of low intensity interventions are required with large numbers of representative patients, clear questions, and credible control groups. In particular, accurate pictures of take-up, retention and completion should be recorded.
Implementation of the new services will be challenging. Sometimes change will be resisted by referrers, or by team members unaccustomed, unfamiliar and sceptical of new ways of working. However, many have found that the effort is worthwhile. Crucially, phase two CBT working does not replace phase one. How the two approaches complement each other and compare will be two of the interesting questions to be addressed over the next few years.
Conflict of interest declaration
CW is author of several CBT self-help books, booklets, CD Roms, CCBT packages and training courses for CBT for depression, bulimia, anorexia, and anxiety.
Comments
No Comments have been published for this article.