Specific phobias

Behavioural treatments for specific phobias, with exposure to the feared situation or object, lead to clinically significant improvement in 70–85% of cases (Roth and Fonagy, Reference Roth and Fonagy2005). Perhaps due to this very high success rate there has been little investigation of cognitive techniques in specific phobias, although these are often employed clinically, particularly to maximize the impact of exposure by putting it in the context of a behavioural experiment. For specific fears very brief treatments may be employed (i.e. one prolonged session for treating spider phobia) and therapist-directed exposure is more effective than self-directed exposure (e.g. Ost, Salkovskis and Hellstrom, Reference Ost, Salkovskis and Hellstrom1991).

Panic disorder with or without agoraphobia

Clark et al.'s CT for panic disorder has been shown to be highly effective both over a 12-session protocol and in a briefer 4-session format with self-help materials (Clark et al., Reference Clark, Salkovskis, Hackmann, Middleton, Anastasiades and Gelder1994, Reference Clark, Salkovskis, Hackmann, Wells, Ludgate and Gelder1999) with 70–80% of patients remaining panic-free at 12-month follow-up. Other CBT treatments show similar results (e.g. Panic Control Treatment, see Barlow, Reference Barlow2002). The NICE guidelines state that CBT should be used as the treatment of choice for panic disorder (NICE, 2004).

Obsessive compulsive disorder

The NICE guidelines present a clear stepped care model for the treatment of OCD, which includes CBT (including exposure and response prevention) as a recommended treatment option at each step (NICE, 2005a). Cognitive models for OCD (e.g. Salkovskis, Reference Salkovskis1985) highlight that it is the appraisal of intrusions as meaning that the person may be responsible for harm unless they take action to prevent it, rather than the content of the intrusion itself, that leads to distress. Currently the evidence for the additional efficacy of cognitive techniques over that of exposure and response prevention alone is mixed. The addition of cognitive techniques may be associated with lower drop-out rates and may be particularly indicated for those who have not previously responded to treatment, or those with forms of OCD where exposure and response prevention is not possible (e.g. mental rituals) (Roth and Fonagy, Reference Roth and Fonagy2005). Severe, chronic OCD remains a clinical challenge.

Generalized anxiety disorder

Although NICE guidelines recommend CBT as the treatment of choice for GAD (NICE, Reference Barlow, Allen and Choate2004) there remains more variation within cognitive models of GAD than in cognitive models of other anxiety disorders. Models of GAD focus on themes including worry as a process of cognitive avoidance (e.g. Sibrava and Borkovec, Reference Sibrava, Borkovec, Davey and Wells2006), the role of metacognitive beliefs (e.g. Wells, Reference Wells1997), and emotion dysregulation (e.g. Mennin, Heimberg, Turk and Fresco, Reference Mennin, Heimberg, Turk and Fresco2005). Perhaps the best-established and supported model is Borkovec's avoidance theory, which proposes that worry allows individuals to approach emotional topics at an abstract level and hence avoid aversive images, autonomic arousal, and negative emotions in the short-term (Sibrava and Borkovec, Reference Sibrava, Borkovec, Davey and Wells2006).

Post-traumatic stress disorder

Trauma-focused CBT is a recommended treatment for PTSD (NICE, 2005b). Predominantly behavioural treatments that focus on imaginal and in vivo exposure in imagery have been shown to be highly effective (e.g. Foa et al., Reference Foa, Hembree, Cahill, Raunch, Riggs, Feeny and Yadin2005). The more recent Ehlers and Clark (Reference Ehlers and Clark2000) cognitive conceptualization of PTSD suggests that a wider range of cognitive behavioural strategies can be employed in order to reduce the sense of current threat experienced by people with PTSD (Ehlers et al., Reference Ehlers, Clark, Hackmann, McManus, Fennell and Grey2008). This treatment has among the highest effect sizes (d = 2.4–2.8, and 64%–75% high end state functioning) and the lowest drop-out rates (0%) seen in the literature (Ehlers et al., Reference Ehlers, Clark, Hackmann, McManus, Fennell, Herbert and Mayou2003; Ehlers, Clark, Hackmann, McManus and Fennell, Reference Ehlers, Clark, Hackmann, McManus and Fennell2005).

Social phobia

Cognitive models of social phobia highlight the role of increased self-focused attention and the subsequent use of misleading internal information to make excessively negative inferences about how one appears to others (e.g. Clark and Wells, Reference Clark, Wells, Heimberg, Leibowitz, Hope and Schneier1995). Currently no NICE guidelines exist for social phobia but meta-analyses attest to the efficacy of CBT (e.g. Fedoroff and Taylor, Reference Fedoroff and Taylor2001). Although previously provided in group format, the largest effect sizes currently reported in the literature are for individual CBT (d = 2.1–2.6, 76% clinically significant improvement; Clark et al., Reference Clark, Ehlers, McManus, Hackmann, Fennell, Campbell, Flower, Davenport and Louis2003, Reference Clark, Ehlers, Hackmann, McManus, Fennell, Grey, Waddington and Wild2006) and direct comparisons indicate that a group format is not advantageous (e.g. Mortberg, Clark, Sundin and Wistedt, Reference Mortberg, Clark, Sundin and Wistedt2006).

Efficacy

While current CBT treatments for anxiety disorders have demonstrated efficacy, there remains room for improvement as many patients do not achieve high end state functioning, particularly at longer term follow-up. An important conceptual advance in CBT for anxiety disorders has resulted from the identification of “safety-seeking behaviours” (Salkovskis, Reference Salkovskis1991). These are behavioural and mental strategies the person uses in order to prevent a feared outcome (e.g. sitting down and trying to relax to prevent a heart attack in panic disorder). Several experimental studies have demonstrated the role of safety behaviours in maintaining anxiety (e.g. McManus, Sacadura and Clark, in press), and have shown that dropping safety behaviours enhances the effectiveness of CBT for anxiety (e.g. Salkovskis, Hackmann, Wells, Gelder and Clark, Reference Salkovskis, Hackmann, Wells, Gelder and Clark2007). This identification of the role of safety behaviours in preventing disconfirmation of feared predictions has led to much more extensive use of behavioural experiments in CBT for anxiety disorders.

General developments in the theory and practice of CBT have impacted on the efficacy of CBT for anxiety disorders. There is an emerging consensus that fear-related information is multiply represented; in both a more sensory/experiential system as well as in a verbal system (e.g. Wells and Matthews, Reference Wells and Matthews1994). This has led to the explicit activation of the more sensory, experiential level in current CBT treatments for anxiety, such as using imagery re-scripting techniques (Wild, Hackmann and Clark, Reference Wild, Hackmann and Clark2007) or transforming the meaning of traumatic memory hotspots within imaginal exposure (Grey, Young and Holmes, Reference Grey, Young and Holmes2002). Both of these techniques have been shown to produce belief change and alleviate emotional distress in the short-term but their longer term impact remains to be evaluated.

A further influence has come from the self-proclaimed “third wave” cognitive-behavioural approaches that suggest that emotion regulation may be better achieved by counteracting maladaptive response-focused emotion regulation strategies (e.g. discouraging emotional suppression) than by the traditional CBT approach that focuses on promoting adaptive antecedent-focused emotion regulation strategies (e.g. reappraisal of the emotional stimuli) (Hofmann and Asmundson, in press). In other words, such approaches focus on changing the patient's relationship with and response to their thoughts (metacognitive processes) rather than on evaluating the content of their thoughts (e.g. Acceptance and Commitment Therapy; Hayes, Strosahl and Wilson (Reference Hayes, Strosahl and Wilson1999) and Mindfulness-based Cognitive Therapy; Teasdale, Segal and Williams (Reference Teasdale, Segal and Williams1995)). This raises interesting questions about the theoretical mechanisms that underlie change in CBT. Case series data suggest that such approaches may have benefits for anxious patients that are maintained at 3-year follow-up (Miller, Fletcher and Kabat-Zinn, Reference Miller, Fletcher and Kabat-Zinn1995). Whether the incorporation or adoption of such developments increases the overall efficacy of treatments remains to be seen. Because CBT develops from the interplay of theory, experimental studies, clinical practice and outcome research it is able to incorporate developments from other fields, such as experimental and cognitive psychology, in order to further increase its efficacy.

Effectiveness

It has been estimated that fewer than 30% of patients treated in routine clinical settings currently receive evidence-based treatments (Goisman, Warsaw and Keller, Reference Goisman, Warsaw and Keller1999). Furthermore, even when evidence-based treatments are applied the reported effect sizes and drop-out rates are often less favourable than those reported in the original research trials (e.g. Addis et al., Reference Addis, Hatgis, Krasnow, Jacob, Bourne and Mansfield2004). Hence, a future challenge for CBT for the anxiety disorders is its successful dissemination. Two issues to be tackled in disseminating CBT for anxiety disorders into routine clinical practice are, first, whether CBT protocols delivered in RCTs can generalize to the patients seen in routine clinical practice and, second, whether it is possible to train therapists to the standards necessary to achieve the same effects as seen in clinical trials.

A common criticism of CBT for anxiety disorders is that the effects seen in RCTs do not generalize to routine clinical settings, because the patients in RCTs are a highly selected, more treatment-responsive group. There is some evidence to support this claim, with dissemination studies carried out in clinical settings often reporting smaller effect sizes and/or higher drop out rates. There are also instances where the effect sizes and drop out rates seen in RCTs are replicated in routine clinical services (e.g. Foa et al., Reference Foa, Hembree, Cahill, Raunch, Riggs, Feeny and Yadin2005). In addition, the most common reason for excluding patients from RCTs is in fact failing to meet a minimum severity or duration criteria (Stirman, DeRubeis, Crits-Christoph and Rothman, Reference Stirman, DeRubeis, Crits-Christoph and Rothman2005). However, there may be some important differences between the populations studied in academic and clinical settings. Most treatment trials focus on patients with one main anxiety disorder and exclude those with co-morbid anxiety disorder diagnoses of equal severity or with the diagnosis of ADNOS. This may present a problem for the dissemination of CBT for anxiety disorders as co-morbidity is the norm rather than the exception in anxiety disorders (Brown, Campbell, Lehman, Grisham and Mancill, Reference Brown, Campbell, Lehman, Grisham and Mancill2001), and a significant proportion of those with clinical anxiety problems meet criteria only for ADNOS (Zimmerman, McDermut and Mattia, Reference Zimmermann, McDermut and Mattia2000). There is very little data on how to treat either ADNOS or co-morbid anxiety disorders, with some studies showing counter intuitive results. For example, the study by Craske et al. (Reference Craske, Farchione, Allen, Barrios, Stoyanova and Rose2007) on patients with panic disorder and a co-morbid disorder reported that CBT, which focused only on panic disorder, rather than CBT that focused on the panic but also attempted to address the most severe co-morbid disorder, had more benefits both for the panic disorder and for the co-morbid disorder. The challenge of conceptualizing and effectively treating co-morbidity and ADNOS has led to calls for the development of a transdiagnostic approach to treating anxiety disorders (e.g. Barlow, Allen and Choate, Reference Barlow, Allen and Choate2004). Such transdiagnostic approaches attempt to identify the common maintaining mechanisms across the different anxiety disorders and may have the potential to strike the balance between completely idiosyncratic formulations and rigid diagnostic models. However, more research is needed to establish the efficacy, and especially the comparative efficacy of transdiagnostic approaches to treating anxiety disorders.

The second challenge for the successful dissemination of CBT for anxiety disorders is whether it is possible to train therapists in clinical settings to a similar standard and to achieve similar effects as achieved in RCTs. Historically, the dissemination of therapy skills literature has highlighted the difficulty in achieving change in therapists' clinical skills, let alone their patients' outcomes, by providing training (King et al., Reference King, Davidson, Taylor, Haines, Sharp and Turner2002; Walters, Matson, Baer and Ziedonis, Reference Walters, Matson, Baer and Ziedonis2005). However, four recent studies have demonstrated that it is possible to impact both CBT skills (Mannix et al., Reference Mannix, Blackburn, Garland, Gracie, Moorey, Reid, Standart and Scott2006; Sholomskas et al., Reference Sholomskas, Syracuse-Siewert, Rounsaville, Ball, Nuro and Carroll2005) and patient outcomes with relatively brief training programs (Grey, Salkovskis, Quigley, Ehlers and Clark, Reference Grey, Salkovskis, Quigley, Ehlers and Clark2006; Westbrook, Sedgwick-Taylor, Bennett-Levy, Butler and McManus, Reference Westbrook, Sedgwick-Taylor, Bennett-Levy, Butler and McManus2008). One common factor to these four studies reporting successful dissemination of CBT skills is the use of ongoing case supervision following workshop based training.

Whilst some progress has been made in generalizing CBT for anxiety disorders into clinical settings and in implementing CBT protocols in routine clinical practice, further dissemination and generalization research is needed in order to determine how best to disseminate CBT for anxiety disorders from academic to clinical settings, on the scale required.

Cost-effectiveness

A final challenge for the future of CBT relates to its cost-effectiveness. If, as recommended by Lord Layard (Layard, Reference Layard2006), CBT is to be delivered to the majority of patients with anxiety disorders who may benefit from it, it will be necessary to take a stepped care approach, working up from the lower intensity interventions to higher intensity treatments. At the lowest level of intervention, there are a vast number of self-help books for anxiety disorders. While such popular texts are seen as helpful by therapists and clients alike, there is little evidence for their efficacy as a stand-alone treatment for any anxiety disorder. Similarly, there have been a number of computer programs developed to provide CBT for anxiety. A NICE technology appraisal reports that only “Fearfighter” has sufficient evidence to support its recommendation for the treatment of panic and phobias (NICE, 2006). It should be used in a stepped care manner, and therapist provided CBT remains a more efficacious treatment. Similarly, studies evaluating other forms of CBT self-help for anxiety disorders suggest that “augmented self-help” with some therapist contact is more effective that pure self-help (e.g. Rapee, Abbott, Baillie and Gaston, Reference Rapee, Abbott, Baillie and Gaston2007). There have also been successful attempts to deliver CBT protocols in large scale group formats (e.g. White, Reference White1998) or in briefer formats with more homework (Clark et al., Reference Clark, Salkovskis, Hackmann, Wells, Ludgate and Gelder1999) and to deliver full CBT protocols in more flexible formats to suit patients (Ehlers, Reference Ehlers2006).