Introduction
Despite extensive evidence for its effectiveness, cognitive behavioural therapy (CBT) is difficult to access because of insufficient numbers of trained clinicians, both in primary and secondary care (Hoifort etal. Reference Hoifort, Strom, Kolstrup, Eisemann and Waterloo2011). CBT's inaccessibility may be especially pertinent in Ireland because of our underdeveloped services (Department of Health and Children, 2006). For example, a 2004 survey of 231 South Western Area Health Board General Practitioners (GPs) found that less than a third of GPs had postgraduate training in psychological therapies and that 85% referred less than 5% of their service users with mental health difficulties to mental health specialists (Copty, Reference Copty2004). Moreover, data extracted from a 2008 Economic and Social Research Institute national survey found that just 20.4% of those who had consulted their GP about mental health problems in the previous year (n = 255) had subsequently attended secondary care (Tedstone Doherty etal. Reference Tedstone Doherty, Moran and Kartalova-O'Doherty2008). A recent survey of GP adult attendees in a rural area (n = 273) also found that although one in three attendees registered as having varying degrees of psychological distress, just 11% were in receipt of mental health services (Hughes etal. Reference Hughes, Byrne and Synnott2010). It therefore appears that many in Ireland who seek help with their mental health difficulties are not receiving CBT in either primary or secondary care.
To increase access to psychological therapies, particularly CBT, the United Kingdom's National Health Service (NHS) began to roll out the Improving Access to Psychological Therapies (IAPT) initiative in 2008. This entailed creating extra psychological therapist posts using a stepped-care model whereby low-intensity interventions are provided as a first option, before referral to higher intensity interventions (Reference O'Shea and ByrneO'Shea & Byrne, in press). This stepped-care model is in line with the United Kingdom's National Institute for Health and Clinical Excellence (NICE) best-practice guidelines and represents optimum usage of limited resources (NICE, 2009, 2011). In Ireland, the Health Service Executive (HSE) has also increased funding for psychological therapy provision recently. Five million euro has been allocated to the National Counselling Service to provide time-limited counselling to adults with medical cards in primary care. However, contrary to the IAPT initiative and NICE's (NICE, 2009, 2011) best-practice guidelines, a stepped-care model of service provision will only be trialled in one pilot site.
In the light of our underdeveloped services and the push to decrease the number of public sector employees, other ways to increase access to CBT must be explored, at least in the short term. One such avenue, and examined in this article, is CBT delivered over the internet, or computerised CBT (cCBT). Whether delivered in self-help or in a therapist-assisted format, cCBT has many advantages such as its convenience, low-cost and confidential nature (Andrews etal. Reference Andrews, Cuijpers, Craske, McEvoy and Titov2010). There is substantial evidence for its efficacy for a range of mental health difficulties including anxiety (Spek etal. Reference Spek, Cuijpers, Nyklicek, Riper, Keyzer and Pop2007; Andrews etal. Reference Andrews, Cuijpers, Craske, McEvoy and Titov2010) depression/low mood (Andersson & Cuijpers, Reference Andersson and Cuijpers2009; Andrews etal. Reference Andrews, Cuijpers, Craske, McEvoy and Titov2010) and alcoholism (Rooke etal. Reference Rooke, Thorsteinsson and Karpin2010). Moreover, it has been shown to be effective in primary care (Hoifort etal. Reference Hoifort, Strom, Kolstrup, Eisemann and Waterloo2011), and its growing evidence base means that it has substantial potential to assist primary care staff in delivering effective low-intensity but high-throughput psychological interventions (Wade, Reference Wade2010). NICE recommends cCBT for the treatment of mild-to-moderate anxiety and depression, delivered as part of a stepped-care model (NICE, 2006).
Owing to the largely unregulated nature of the internet, and the proliferation of internet therapy websites in recent years, the quality of services provided through such websites is likely to vary (Christensen etal. Reference Christensen, Murray, Calear, Bennett, Bennett and Griffiths2010). Clinicians therefore need to become familiar with high-quality cCBT programmes so that they can refer service users to these as part of a best-practice stepped-care model. Accordingly, the aim of this review is to highlight evidence-based cCBT programmes that potentially can be availed of by clinicians and service users in Ireland.
Methods
The authors conducted computer-based literature searches of the Psych ARTICLES, PsychINFO and Academic Search Premier databases. They used Boolean operators (OR, AND) and various search terms related to cCBT such as: ‘online CBT’/‘internet therapy’/‘computerised CBT’/‘internet-delivered treatment’/‘self-help’/‘computer’/‘self-guided’/‘web’/‘cyber’. They added these to search terms for various psychological problems (e.g. ‘anxiety’/‘depression’/‘stress’/‘insomnia’) and research designs (e.g. ‘random/'controlled’/‘RCT’). In addition, they performed manual searches of various academic journals to locate articles that were included in reference lists of previously identified articles. The authors also consulted the Australian National University-developed web portal Beacon that provides a comprehensive database of cCBT programmes, how to access them, as well as the published evidence behind them (Beacon, homepage on the internet; Christensen etal. Reference Christensen, Murray, Calear, Bennett, Bennett and Griffiths2010).
To ensure that programmes included in this review were understandable, accessible and not prone to experimental bias and/or error (Spring, Reference Spring2007), the authors used the following inclusion criteria: (a) English-language-only programmes; (b) those that are currently delivered via the internet; and (c) those that had at least one published randomised controlled trial (RCT) demonstrating their efficacy. Furthermore, to enable valid comparison, the authors included only those RCTs with waitlist controls and/or face-to-face treatment groups that were not receiving a computerised intervention (Andersson & Cuijpers, Reference Andersson and Cuijpers2009).
Findings
Twenty-five cCBT programmes met the inclusion criteria (Fig. 1). Taken together, these programmes target a range of psychological difficulties in adolescents and children, and adults: generalised anxiety (GA), panic/phobia, social anxiety (SA), post-traumatic stress disorder (PTSD), depression (or low mood), eating problems, stress, insomnia, pain and alcohol misuse. Notably, and usefully, some of the programmes (e.g. ‘Beating the Blues’, ‘Mood Gym’) can treat more than one difficulty at a time.
RCT-Supported cCBT Programmes delivered, in English, on the internet.
The next section of this article profiles the identified cCBT programmes, categorised by the mental health difficulty they target. First, those cCBT programmes for GA are examined.
cCBT programmes for GA
Seven cCBT programmes for GA met this article's inclusion criteria (Table 1). Six of these programmes were developed in Australia and one was developed in the United Kingdom (‘Beating the Blues’). Currently, two programmes are freely accessible online by residents of Ireland – ‘MoodGym’ and ‘Online Anxiety Prevention’. However, ‘Beating the Blues’ is provided via primary care by the United Kingdom's NHS and can be purchased online. Four of the programmes are delivered with therapist assistance but three (i.e. ‘MoodGym’, ‘Beating the Blues’ and ‘Online Anxiety Prevention’) can be completed on a self-help basis. The total number of sessions for the programmes ranges from 5 to 10.
cCBT programmes for GA
GA, Generalised anxiety; SA, social anxiety; PTSD, post-traumatic stress disorder; SH, self help; TA, therapist assisted; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
aOne other RCT for BRAVE-ONLINE has been published, but the cCBT intervention condition consisted of half of treatment-as-usual sessions delivered via BRAVE-ONLINE. Thus it was not possible to determine whether BRAVE-ONLINE led to additional treatment benefits (Spence etal. Reference Spence, Holmes, March and Lipp2006).
bFour other RCTs were conducted on MoodGym. One was excluded from the table because the study's outcome measures measured general psychological distress rather than a specific difficulty (e.g., GA) (Hickie etal. Reference Hickie, Davenport, Luscombe, Moore, Griffiths and Christensen2010). Another was excluded because it examined stigmatising attitudes towards depression and had mixed results (Griffiths etal. Reference Griffiths, Christensen, Jorm, Evans and Groves2004). The other two were excluded because they used MoodGym in conjunction with a computerised psychoeducation intervention (Farrer etal. Reference Farrer, Christensen, Griffiths and Mackinnon2011; Lintvedt etal. Reference Lintvedt, Griffiths, Sørensen, Ostvik, Wang, Eisemann and Waterloo2013).
Four of the programmes have two RCTs demonstrating their efficacy, and three have one RCT. The strongest evidence appears to be for the ‘eCentre Clinic-Worry Programme’ and ‘BRAVE-ONLINE’ (for children) as shown by large effect sizes and the use of clinically screened samples across two RCTs each. The former programme's RCT participants were self-referred, whereas the latter's RCT participants (i.e. children) were referred from many sources (e.g. parents, teachers and clinicians). ‘MoodGym’ also has substantial evidence for GA – it has two RCTs that yielded small-to-medium effect sizes. One of these RCTs had a large, school-based and non-clinical sample and the other had a small self-referred clinically screened sample. It has the added advantage of being suitable for both adults and adolescents. The United Kingdom's ‘Beating the Blues’ has two large-scale RCTs with GP-referred clinically screened samples, demonstrating its efficacy. Effect sizes here were with small to medium. However, the latter RCT only found effectiveness in those with more severe anxiety. The evidence for each of the three programmes with one RCT is not referred to here because of word constraints but is viewable in Table 1.
cCBT programmes for panic/phobia
Three cCBT programmes for panic/phobia met this article's inclusion criteria (Table 2). Two of these programmes (eCentre Clinic's-Panic & Anxiety Programmes) were developed in Australia and one was developed in the United Kingdom (‘FearFighter’). None of the three programmes are freely accessible online by residents of Ireland. However, ‘FearFighter’ is provided via primary care by the United Kingdom's NHS. ‘FearFighter’ can be completed on a self-help basis, whereas the other two programmes are delivered with therapist assistance. The total number of sessions for the programmes ranges from 6 to 9.
cCBT programmes for panic/phobia
GA, Generalised anxiety; SA, social anxiety; PTSD, post-traumatic stress disorder; SH, self help; TA, therapist assisted; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
aFearFighter has another RCT demonstrating its efficacy (Schneider etal. Reference Schneider, Mataix-Cols, Marks and Bachofen2005). However, as the control used in the study was a computerised control, it did not enable valid comparison (Andersson & Cuijpers, Reference Andersson and Cuijpers2009), and was thus excluded.
All three programmes have one RCT, with clinically screened samples, demonstrating their efficacy. The strongest evidence appears to be for ‘FearFighter’ – its RCT has the largest sample and effect sizes of the three. Its sample was referred by health professionals and also by self-referral. However, the ‘eCentre Clinic-Panic Programme’ was shown to be effective across a wider range of measures (three as opposed to two), although with smaller effect sizes, in a self-referred sample. The other programme, ‘eCentre Clinic-Anxiety Programme’, is advantageous in that it has also been shown to be effective for GA (Table 1), and SA (Table 3). It yielded a medium effect size in a self-referred sample.
cCBT programmes for SA
GA, Generalised anxiety; SA, social anxiety; SH, self help; TA, therapist assisted; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
cCBT programmes for SA
Three cCBT programmes for SA met this article's inclusion criteria (Table 3). Two of these programmes (eCentre Clinic's-Shyness & Anxiety Programmes) were developed in Australia and one was developed in Spain (‘Talk to Me’). None of the three programmes are freely accessible online by residents of Ireland. However, ‘Talk to Me’ is available for an unspecified fee. ‘Talk to Me’ can be completed on a self-help basis, whereas the other two programmes are delivered with therapist assistance. The total number of sessions for the programmes ranges from 3 to 6.
The strongest evidence here is clearly for the ‘eCentre Clinic-Shyness Programme’. It has four RCTs demonstrating its efficacy in clinically screened samples, compared with the other two programmes that have one RCT behind them. Moreover, for the treatment of SA, it yielded large effect sizes. Three of its RCT's participants were self-referred but, notably, the other one had participants who were GP referred (although with a small sample size). As mentioned above, the ‘eCentre Clinic-Anxiety Programme’ is advantageous in that it has also been shown to be effective for three mental health difficulties – GA (Table 1), panic/phobia (Table 2) and SA (Table 3). For SA, it yielded a moderate effect size in a self-referred, clinically screened sample. The other programme ‘Talk to Me’ had moderate effect sizes for SA in a self-referred, clinically screened sample. However, it is geared towards public speaking difficulties to a larger extent than SA.
cCBT programmes for PTSD
One cCBT programme for PTSD, the ‘eCentre Clinic-PTSD Programme’, met this article's inclusion criteria (Table 4). This therapist-assisted programme has seven sessions. It was developed in Australia and is not currently accessible online by residents of Ireland. It has one RCT showing its efficacy within a self-referred, clinically screened sample. It yielded a moderate effect size. The RCT is limited by its relatively small sample size.
cCBT programmes for PTSD
PTSD, Post-traumatic stress disorder; TA, therapist assisted; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
cCBT programmes for depression (or low mood)
Six cCBT programmes for depression met this article's inclusion criteria (Table 5). Three of these programmes were developed in Australia. The other three were developed in the United Kingdom, Germany and the United States, respectively. Currently, two of these programmes, ‘MoodGym’ and ‘MoodHelper’, are freely accessible online by residents of Ireland. However, ‘Beating the Blues’ is provided via primary care by the United Kingdom's NHS and can be purchased online. Similarly, ‘Deprexis’ can be purchased online. Whereas ECentre Clinic's ‘Sadness’ and ‘Well-being’ programmes are both delivered with therapist assistance, the other three programmes can be completed on a self-help basis. The total number of sessions for the programmes ranges from 5 to 10, although ‘MoodHelper’ is a diary-style intervention completed at one's own pace.
cCBT programmes for depression
GA, Generalised anxiety, SA, social anxiety; SH, self help; TA, therapist assisted; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
aFour other RCTs were conducted on MoodGym (see note in Table 1).
One of the programmes, ‘MoodGym’, has three RCTs showing its efficacy, two programmes have two RCTs, and three programmes have one RCT. Although ‘MoodGym’ has the most RCTs, one of these had a small, self-referred sample and one had a school-based non-clinical sample. Furthermore, in the latter RCT, ‘MoodGym’ was found to be effective for males only. However, it was found to be effective in an RCT with a large, self-referred, clinically screened sample. Overall, it seems that the strongest evidence is for ‘Beating the Blues’ and ‘eCentre Clinic-Sadness’. The former has two large-scale RCTs in GP-referred, clinically screened samples that yielded medium effect sizes. The latter has two RCTs in self-referred, clinically screened samples that both yielded large effect sizes. However, one of the RCTs for ‘eCentre Clinic-Sadness’ had a small sample size and both RCTs are limited by little to no follow-up data. The three programmes with one RCT are not elaborated on here because of word constraints, but the evidence behind each is detailed in Table 5.
cCBT programmes for eating problems
Three cCBT programmes for eating problems met this article's inclusion criteria (Table 6). These programmes were developed in the United States, United Kingdom and Switzerland, respectively. None of these are freely accessible online by residents in Ireland. However, ‘Overcoming Bulimia Online’ (UK) can be purchased online for ∼€75 and ‘Student Bodies’ can be purchased via private contract by institutions and individuals. The other programme ‘Salut BN’ (Switzerland) is a long-term fee-based intervention that requires substantial service integration. Overcoming ‘Bulimia Online’ and ‘Student Bodies’ can be completed on a self-help basis, but ‘Salut BN’ is delivered with therapist assistance and integrated face-to-face sessions. The former two programmes have eight sessions each, but ‘Salut BN’ has seven multifaceted ‘modules’ that are completed over 6 months.
cCBT programmes for eating problems
TA, Therapist assisted; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
aStudent Bodies has another RCT which (partially) shows its efficacy (Celio etal. Reference Celio, Winzelberg, Wilfley, Eppstein-Herald, Springer, Dev and Barr Taylor2000). However, it was excluded because the intervention group completed Student Bodies alongside face-to-face and group therapy sessions. Thus it was not possible to determine whether Student Bodies led to additional treatment benefits.
One of the programmes, ‘Student Bodies’, has two RCTs showing its efficacy, and the other two programmes have one RCT each. All four RCTs have screened, self-referred, female-only samples. The two RCTs on ‘Student Bodies’ have large sample sizes and yielded medium effect sizes. However, screening ensured that the sample was non-clinical in nature. The RCTs on ‘Overcoming Bulimia Online’ and ‘Salut BN’ yielded medium to large effect sizes within comparatively smaller samples. However, as both had clinically screened samples, there is arguably stronger evidence behind them than ‘Student Bodies’.
cCBT programmes for stress
One cCBT programme for stress, ‘Stress & Mood Management’, met this article's inclusion criteria (Table 7). This self-help programme has four sessions. It was developed in the United States and is currently only available if purchased by organisations on behalf of their employees. It has one RCT showing its efficacy, but the significance level (p-value) was relatively weak and the effect size was low. Furthermore, participants were not clinically screened and it was a self-referred (although large) sample.
cCBT programmes for stress
SH, Self help; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
cCBT programmes for insomnia
One cCBT programme for insomnia, ‘SHUTi’, met this article's inclusion criteria (Table 8). This self-help programme has six sessions. Developed in the United States, it is freely accessible by residents of Ireland through research participation only (via the website). It has one RCT with a clinically screened sample showing its efficacy. The effect size here was large but the study is limited by its small sample in which participants were self-referred.
cCBT programmes for insomnia
SH, Self help; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
cCBT programmes for pain
Three cCBT programmes for pain met this article's inclusion criteria (Table 9). Two of these programmes were developed in the United States, and the other was developed in Canada. One of the programmes, ‘Pain-ACTION’ (i.e. for chronic back pain management), is freely accessible online by residents of Ireland. However, the other two programmes are freely available in indirect ways – ‘Help Yourself Online’ can be accessed by contacting its authors and ‘Web-MAP’ can be availed of via research participation. ‘Help Yourself Online’ is delivered with therapist assistance, but the other two programmes can be completed on a self-help basis. ‘Help Yourself Online’ and ‘Web-MAP’ have seven and eight sessions, respectively, and ‘Pain ACTION’ is an unstructured programme.
cCBT programmes for pain
SH, Self help; TA, therapist assisted; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
All three programmes have one RCT demonstrating their efficacy. ‘Pain ACTION's’ RCT has the largest sample and effect sizes, and its sample was clinically screened. However, significant improvements were not made on several of its measures and its sample had self-referred participants. Although each yielded medium effect sizes, the other two programmes’ RCTs are limited by the absence of clinical screening, the use of non-standardised measures (e.g. pain diaries), and small sample sizes. The RCT on ‘Web-MAP’ had a physician-referred sample whereas the RCT on ‘Help Yourself Online’ had a self-referred sample.
cCBT programmes for alcohol misuse
Two cCBT programmes for alcohol misuse met this article's inclusion criteria (Table 10). Both of these programmes were developed in the Netherlands, and are accessible online by residents of Ireland. However, whereas ‘Jellinek’ is freely accessible, ‘Look at Your Drinking’ must be purchased for a negotiated fee. Moreover, the former can be completed on a self-help basis, whereas the latter is delivered with therapist assistance. ‘Jellinek’ is an unstructured, journal-style programme, whereas ‘Look at Your Drinking’ has nine sessions.
cCBT programmes for alcohol misuse
SH, Self help; TA, therapist assisted; PI, post-intervention; FU, follow-up; d, Cohen's d (between-group).
Both programmes have one RCT demonstrating their efficacy, with large, self-referred samples. Participants in each RCT were screened for alcohol use levels. ‘Look at Your Drinking’ yielded large effect sizes, whereas ‘Jellinek’ only yielded small-to-medium effect sizes. Both RCTs are limited in that they rely primarily on retrospective self-report for ascertaining alcohol consumption levels.
Conclusions
This article identified 25 RCT-supported cCBT programmes that are delivered in English, via the internet. The highest number of programmes were for GA (n = 7) and depression (n = 6). Similarly, the highest number of RCTs (across the programmes) were for GA (n = 11) and depression (n = 10). These findings are in line with research that shows that most of the evidence for cCBT's effectiveness is for anxiety and depression (Spek etal. Reference Spek, Cuijpers, Nyklicek, Riper, Keyzer and Pop2007; Andrews etal. Reference Andrews, Cuijpers, Craske, McEvoy and Titov2010). These findings also complement NICE's best-practice guidelines that recommend the use of cCBT for mild-to-moderate anxiety and depression (NICE, 2006). Nevertheless, this article demonstrated that a wide variety of cCBT programmes for a range of mental health difficulties can potentially be availed of in Ireland, although for a set or negotiated fee for most programmes.
Various issues should be taken into account when interpreting the positive findings concerning cCBT. First, just 20% of the 35 indentified RCTs had participants who were referred to cCBT programmes by health professionals (mainly GPs), with the rest self-referred. Some research has indicated that self-referees to cCBT have better outcomes than those referred by mental health professionals, but worse outcomes than those referred by GPs (Mataix-Cols etal. Reference Mataix-Cols, Cameron, Gega, Kenwright and Marks2006). Therefore, it is likely that the referral source influenced the findings, although in an unclear manner. Second, little or no follow-up data were included in several of the RCTs. If such data are not available for a particular cCBT programme, clinicians should perhaps only refer service users to it as a precursor or adjunct to ‘treatment as usual’. Third, 80% of the identified programmes were for adults only, and some presentations (e.g. PTSD, stress and insomnia) had only one programme each. Thus, although it has wide applicability, cCBT may not be suitable for various client groups. Finally, as the authors of the cCBT programmes were also authors on most of the RCTs, as has been reported in studies concerning face-to-face CBT (Cuijpers etal. Reference Cuijpers, Smit, Bohlmeijer, Hollon and Andersson2010), it is possible that a publication bias may have led to negative results for particular cCBT programmes not being published.
Irish clinicians and service users may also find it interesting to note that various cCBT programmes have also been developed in Ireland in recent years. For example, the online mental health promotion project, ‘Headsup’ that is run by the Rehab Group, provides a cCBT skills programme (HeadsUp). The Technology Enhanced Therapy project set up collaboratively by the National Digital Research Centre (Trinity College Dublin) and the charity Parents Plus also provides online programmes (entitled ‘SilvercloudHealth’) for depression and eating problems for adolescents and young adults (SilverCloud). However, unlike the programmes detailed in this review, these programmes have as yet no published RCTs demonstrating their effectiveness, though two RCTs are reportedly underway.
Looking to the future, the HSE is funding a 2-year collaborative stepped-care and high-throughput service in Roscommon for adults with mild-to-moderate mental health difficulties. A key part of this pilot is the planned development by the psychology services of HSE-owned cCBT programmes for common mental health difficulties. In-house ownership of such programmes will substantially reduce the cost of future provision of such programmes to our service users, and will facilitate adapting these over time to better meet evolving or emerging clinical needs.
To conclude, NICE (2006) recommends that cCBT programmes such as those identified in this article are best administered by clinicians as part of a stepped-care model alongside low-intensity interventions such as bibliotherapy and brief CBT (Twomey & Byrne, Reference Twomey and Byrne2012). Referring service users to these programmes could be beneficial, especially in the light of our underdeveloped mental health services and the limited availability of one-to-one CBT. Before choosing a particular programme, clinicians and service users are advised to examine its effectiveness, particularly as profiled in this article's tables. It is time for us to log on to cCBT and to ‘step up’ or expand our mental health services.
