8.1. INTRODUCTION

The purpose of this chapter is to review the evidence to update the section of NICE Technology Appraisal Computerised Cognitive Behaviour Therapy for Depression and Anxiety (TA97; NICE, 2006), that deals with specific phobias. The Generalised Anxiety Disorder and Panic Disorder (with or without Agoraphobia) in Adults guideline (NICE, 2011c) updated the part of TA97 that covered panic disorder and Chapter 6 of this current guideline updates TA97 for ‘social phobia’.

Specific phobias are characterised by marked and persistent fear of particular (well-defined) objects or situations. Phobic situations are avoided or endured with extreme distress, which interferes with normal functioning. Specific phobias differ from other anxiety disorders in the central role of fear response rather than anticipation (Craske et al., 2009).

Specific phobias are the most common mental disorders with a median 12-month prevalence in 27 European countries of 6.4% (Wittchen & Jacobi, 2005) and a lifetime risk of approximately 13.2% (Kessler et al., 2005a). Of people with a specific phobia, half have a fear of animals or heights (Stinson et al., 2007). Prevalence of animal phobias is 3 to 7% (Becker et al., 2007; Stinson et al., 2007), and fear of heights is the most common natural environment fear, but other environmental fears (for example, flying and enclosed spaces) are also common (Becker et al., 2007).

Specific phobias typically begin in childhood, with 50% beginning by 7 years and 75% by 12 years (Kessler et al., 2005a). Animal phobias normally begin in early childhood (Becker et al., 2007; Beesdo et al., 2009), while other phobias may begin later in life; notably, situational phobias (for example, flying) may occur in adolescence or early adulthood (Beesdo et al., 2009). They are more common in women than men (Beesdo et al., 2009; Curtis et al., 1998). Children of parents with a specific phobia are at increased risk of developing the same fear (Fyer et al., 1995). Phobias often occur with other disorders, and the other disorder is typically the focus of clinical attention. Like other anxiety disorders, comorbidity is associated with greater impairment (Magee et al., 1996). Of those with one lifetime diagnosis of specific phobia, 75.8% will have a second phobia (Curtis et al., 1998; Wittchen et al., 2007).

The aetiology of phobias has been debated for decades (Mowrer, 1947; Mowrer, 1960), but complete explanatory theories are not required for successful treatment (Marks, 1981). Different forms of exposure therapy have been used successfully for at least 40 years (Wolpe, 1968). Relaxation and other behavioural techniques may be taught as coping methods for use in stress-provoking situations, but these are probably not as beneficial as live exposure, which can be efficacious in a single prolonged session (Hellstrom & Ost, 1995; Ost et al., 2001; Ost et al., 1997). Therapist-delivered CBT is the preferred treatment for most anxiety disorders, but may not be necessary for the successful treatment of specific phobias, and access to therapists may be limited.

To increase access to care and to reduce therapists' caseload, CBT can be delivered using computers and the internet. Evidence from previous reviews of self-help for anxiety and depression is encouraging; however, reviews and meta-analyses are difficult to interpret because of inconsistent methods and conclusions, and it is not clear that results from other disorders apply to specific phobias. Lewis and colleagues provide a useful overview of the older reviews (Lewis et al., 2003), and NICE previously considered CCBT for anxiety and depression through the TA process (NICE, 2006).

TA97 found some support for CCBT in general and recommended one program, FearFighter™, for the treatment of ‘phobias’. However, the appraisal did not distinguish specific phobias from other disorders, such as social anxiety disorder (previously called ‘social phobia’) and agoraphobia. This guideline completes the update of the TA97 and has undertaken a separate analysis of CCBT for social anxiety disorder and for specific phobias, which were grouped under a general heading of ‘phobias’ in TA97.

8.2. REVIEW PROTOCOL

A systematic review was undertaken using standard NCCMH procedures as described in Chapter 3 (further information about the search strategy can be found in Appendix 6). The review protocol, including the review questions, information about the databases searched, and the eligibility criteria used for this section of the guideline, is presented in Table 23. Where appropriate, meta-analysis was used to synthesise the evidence using a random effects model. For comparison, the review protocol for TA97 is also included in Table 24.

Table 23

Review protocol for the review of CCBT for specific phobias.

Table 24

Review protocol from TA97.

8.3. CLINICAL EVIDENCE

8.3.1. Studies considered³²

A broad search was conducted to identify studies using a computerised intervention based on cognitive behavioural techniques for the treatment of specific phobias in adults. Because exposure may be the most active ingredient in the treatment of specific phobias, interventions that were mainly behavioural were not excluded.

The search identified 13 RCTs. Of these, seven were included in at least one analysis: ANDERSSON2009 (Andersson et al., 2009), GILROY2000 (Gilroy et al., 2000), GRANADO2007 (Granado et al., 2007), HASSAN1992 (Hassan, 1992), HEADING2001 (Heading et al., 2001), MÜLLER2011 (Müller et al., 2011), SMITH1997 (Smith et al., 1997). Two trials (Marks et al., 2004; Schneider et al., 2005) included in TA97 (NICE, 2006) could not be included in this review because they did not report results for people with specific phobias and the authors were unable to provide disaggregated data for people with agoraphobia, social phobia and specific phobias; one author reported that, in his view, ‘group sizes had insufficient power to detect a significant improvement for the different phobia types’ (Mark Kenwright, 13 March 2013). Four trials (Fraser et al., 2001; Johnston et al., 2011; Matthews et al., 2011; Tortella-Feliu et al., 2011) were excluded because they did not include an appropriate control (that is, they compared a computerised intervention with another computerised intervention rather than a non-computerised control).

Trials were published from 1992 to 2009 and included a total of 302 participants at baseline (range 25 to 45). Participants were on average (median of means) 32 years old, all white, and mostly (93%) female. All participants had a specific phobia of spiders. See for Table 25 further details about the characteristics of interventions.

Table 25

Characteristics of interventions.

8.3.2. Risk of bias

All included trials were assessed for risk of bias (see Figure 13). None were at low risk for sequence generation, and four were at high risk of bias for allocation concealment. Trials were considered at high risk of bias for participant and provider blinding per se, but assessor blinding was considered separately for all trials, and five were at high risk of bias. For incomplete outcome data, three trials were at high risk (for example, those that reported per protocol or completer analyses and those with very high amounts of missing data) and one was unclear. None of the trials were registered in advance and there is risk of publication bias.

Figure 13

Risk of bias summary.

8.3.3. Quantitative synthesis

Meta-analyses were conducted for all critical outcomes (recovery, symptoms of specific phobia, behavioural approach test [a test commonly used in the evaluation of treatments for a broad range of specific phobias]) for each comparison at each time point (see Table 26 for a summary of trial results). As in previous chapters, the number of participants below is the number in the treatment group represented in the analysis. For all analyses of symptoms, negative SMDs favour CCBT. Similarly, an RR of greater than 1 favours CCBT. For behavioural approach tests, positive values favour CCBT and are noted with a superscript (that is, SMD⁺). GRADE profiles are included in Tables 27 to 30.

Table 26

Summary of results.

Table 27

GRADE profile – CCBT versus waitlist for specific phobias.

Table 28

GRADE profile – CCBT versus attention control for specific phobias.

Table 29

GRADE profile – CCBT versus relaxation for specific phobias.

Table 30

GRADE profile – CCBT versus exposure in vivo for specific phobias.

Compared with waitlist

Two trials (HASSAN1992, HEADING2001) compared CCBT (23 participants) with waitlist. Neither reported recovery, but there was a large effect with wide CIs for symptoms of specific phobia at post-treatment (SMD = −1.38, 95% CI = −3.72 to 0.97) and in one trial (HEADING2001) at follow-up (SMD = −0.41, 95% CI = −1.19 to 0.37). There was a large effect with wide confidence intervals in one trial (HEADING2001) for a behavioural approach at post-treatment (SMD⁺ = 2.98, 95% CI = −2.71 to 8.66), which was not present at follow-up (SMD⁺ = 0.00, 95% CI = −0.77 to 0.77).

Compared with attention control

Three trials (GRANADO2007, MÜLLER2011, SMITH1997) compared CCBT (61 participants) with an attention control. One study (GRANADO2007) reported little difference in recovery at follow-up (RR = 1.15, 95% CI = 0.40 to 3.31). Combining all trials, there was evidence of a medium-sized effect for symptoms of specific phobia at post-treatment (SMD = −0.58, 95% CI = −0.94 to −0.21) and in one trial (GRANADO2007) evidence of a large effect for a behavioural approach (SMD = −0.83, 95% CI = −1.65 to 0.00). One trial (SMITH1997) reported a small effect at follow-up (SMD = −0.21, 95% CI = −0.77 to 0.35).

Compared with relaxation

One trial (GILROY2000) compared CCBT (15 participants) with a behavioural relaxation intervention. Recovery was not reported. There was evidence of a large effect with wide CIs on symptoms at post-treatment (SMD = −1.19, 95% CI = −1.97 to −0.41), but which decreased at follow-up (SMD = −0.65, 95% CI = −1.39 to 0.08). There was evidence of large effects on a behavioural approach at post-treatment (SMD⁺ = 0.93, 95% CI = 0.17 to 1.69) and at follow-up (SMD⁺ = 1.23, 95% CI = 0.44 to 2.02).

Compared with exposure in vivo

Four trials (ANDERSSON2009, GILROY2000, HASSAN1992, HEADING2001) compared CCBT with exposure in vivo. Combining all trials, the effect favoured exposure in vivo (SMD = 0.34, 95% CI = −0.04 to 0.71), and was maintained at follow- up (SMD = 0.35, 95% CI = −0.42 to 1.11). There was evidence of a medium-sized effect in three trials (GILROY2000, HASSAN1992, HEADING2001) for a behavioural approach at post-treatment, favouring exposure in vivo (SMD⁺ = −0.63, 95% CI = −1.09 to −0.18) with no important heterogeneity, but the effect was smaller at follow-up (SMD⁺ = −0.29, 95% CI = −0.84 to 0.27).

8.4. CLINICAL EVIDENCE SUMMARY

Systematic searches identified seven trials of computerised interventions for specific phobia, compared with no treatment or exposure. The review for CCBT for specific phobias identified interventions for spider phobias only. No evaluations of interventions for any other specific phobia were suitable for inclusion in the review. Trials were generally assessed as being of low quality and at high risk of bias, including selective outcome reporting and publication bias, and the numbers of participants in the trials were small. Comparisons with a waitlist or an attentional control produced medium to large effects on symptoms and on the behavioural approach test. In contrast when compared with the results of therapist-delivered exposure treatment, CCBT did not appear to be efficacious, with the direction of the effect favouring exposure in vivo on both symptoms and the behavioural approach test.

8.5. FROM EVIDENCE TO RECOMMENDATIONS

In developing recommendations the GDG was mindful of the low quality of the evidence and the high risk of bias. Trials also focused only on one specific phobia (there are a significant number of other common phobias including snakes, heights, flying and needles). The GDG considered that therapist-delivered single session exposure therapy (Davis et al., 2012) is an efficacious treatment for specific phobias and the four trials included in this chapter suggest that it is probably superior to CCBT. The GDG also noted that none of the interventions evaluated is available in the UK. Given the very low quality of the evidence and the narrow focus of the interventions (spider phobia only), the GDG felt that the available data do not provide sufficient evidence to suggest that CCBT is an efficacious treatment for specific phobias. The GDG was aware that other efficacious treatments are available (but not the subject of this review) and in these circumstances decided not to recommend CCBT for the treatment of specific phobias.

8.6. RECOMMENDATIONS

8.6.1.1.

Do not routinely offer computerised CBT to treat specific phobias in adults.