Using the 3 most popular Web search engines, Google, Bing, and Yahoo [38,39], a keyword search for Web-based intervention programs for anxiety was performed in March 2015. UK versions of the search engines were used (.co.uk). Before starting the search, existing search history and cookies were deleted and future tracking and cookies were disabled in the browser. A list of the 9 search term combinations used can be found in Textbox 1. Primarily simple and lay keywords were used to simulate a Web search that was relatively likely to be conducted by an individual searching for Web-based programs. It has been shown that most individuals rarely consider more than the first 20 links generated by a search engine [40]. As featured links placed at the top and bottom were also considered, we chose to assess the first 25 links. This resulted in 675 hyperlinks being screened (3 search engines × 9 search terms × 25 hyperlinks).

All 675 hyperlinks were screened for eligible Web-based programs for anxiety by the first author. The screening process consisted of two stages. The first stage involved screening all 675 hyperlinked websites to eliminate clearly irrelevant websites. All hyperlinks were screened and organized into 1 of the following 3 categories: websites with Web-based programs, websites linking to websites with Web-based programs, and websites with irrelevant content. Irrelevant content included, among others, e-counselling websites, mental health information websites, support groups/forums, online mental health screening/assessment, therapist or mental health clinic websites, scholarly articles, blogs, Facebook pages, Wikipedia, videos, and broken links.

For all websites categorized as “websites of Web-based programs” and “websites linking to websites of Web-based programs,” duplicates were removed. All remaining websites entered the second stage of screening and were screened according to the following criteria by the first author: (1) designed for anxiety symptoms (although they did not need to be focused on anxiety only), (2) currently publically accessible on the Internet (via registration, application, General Practitioner (GP) referral), (3) online component, (4) structured treatment plan (information only), and (5) available in English. Programs were excluded if they were (1) not publically accessible; (2) Web-based counseling only (Skype, email, or instant message contact with a counselor only, with no structured program associated); (3) purely informational (psychoeducation only); or (4) exclusively part of a research study.

A data extraction form was created containing 4 main categories and 9 subcategories. Table 1 provides an overview of the items in each main category and subcategories. The 4 main categories are based on the 4 specified study interests (website characteristics, program characteristics, intervention characteristics, and empirical evidence). The 9 subcategories of the 4 main categories were established by incorporating the 12 key facets of a framework designed for evaluating and reporting Internet intervention studies [41]. The extraction was undertaken by the first author in March 2015. Screenshots of all programs and websites were taken in case the program changed during the rating period. To ensure that the programs could be evaluated thoroughly, all program authors were contacted to request free access.

Several validated and widely accepted scales are available to evaluate the methodology of published studies. However, there are currently no validated criteria available for evaluating actual Web-based interventions as found on the web. Renton et al [31] created a program scoring system to evaluate Web-based depression interventions. With permission from the authors, the scale was adapted to fit the 4 specified study interests. The adapted version consists of 16 yes or no close-ended questions that are outlined in Table 2. Consistent with the scoring system used by Renton et al [31], a score of 1 was awarded if the answer was yes, and a score of 0 if the answer was no or the question could not be evaluated. Scores were converted into percentages, with higher scores indicating a larger number of met criteria on the scale.

A search log outlines the number of hits per search, as well as the number of included and excluded hyperlinks (see Multimedia Appendix 1). Most program websites were identified when the search terms “online treatment anxiety” (Program websites: n=14, websites with links to program websites: n=2) and “online cognitive behavioural therapy anxiety” (program websites: n=12, websites with links to program websites: n=4) were entered into Google. “Web therapy anxiety” entered in Yahoo did not identify any program websites. All 675 links were assessed for inclusion. Figure 1 displays the flowchart for the two-stage selection process of included programs and reasons for exclusion. In total, 176 of the 675 (25.3%) assessed hyperlinks led either directly to program websites (133/675, 19.7%) or to websites containing links to program websites (43/675, 6.4%). The first stage of the screening identified 35 potentially eligible program websites that were subsequently assessed for inclusion. Of those, 19 websites and 34 programs met the inclusion criteria. For 5 programs, the authors did not grant access and some aspects of those programs could therefore not be evaluated (Beating the Blues, Changing States, eCentreClinic, FearFighter, Social Anxiety Institute).

An overview of intervention program characteristics for each program can be found in Table 3 and screenshots of all programs can be found in Multimedia Appendix 2.

A summary of the types of published research evaluations for each program and the respective references can be found in Table 4. For 3/34 programs (9%), indirect research evidence was identified. The two e-couch programs are based on the MoodGYM program, for which research evidence is available; however, the e-couch program’s efficacy was not specifically empirically evaluated. The AI-Therapy program has only been evaluated for social anxiety in adults who stutter using a pre-post study without a control group. For 17/34 programs (50%), empirical studies evaluating efficacy or effectiveness were found. Studies ranged from case series and small to mid-sized pre-post interventions without comparison groups to controlled and randomized controlled trials (RCT). Both MoodGYM and This Way Up have been evaluated through 9 RCTs each. The efficacy of Beating the Blues was demonstrated by 2 RCTs, FearFighter by 2 RCTs, and myCompass by 1 RCT. For 14/34 programs (41%), no research evidence of the efficacy or effectiveness of the intervention was found.

Results were examined for programs for which anxiety symptoms were evaluated in RCTs. Beating the Blues was found to lead to a significant reduction of anxiety both at the end of treatment and at 6 months’ follow-up compared to treatment as usual [45]. For FearFighter, both the face-to-face and online program group had reduced anxiety at post-treatment and 1-month follow-up [59]. MoodGym was evaluated in adolescents and university students and levels of anxiety were found to be lower in the intervention group compared to the waitlist control group after the intervention [63,66]. In addition, combined face-to-face and online CBT was more effective in treating anxiety symptoms than either face-to-face or online as a standalone [70]. Compared to control subjects, participants in the myCompass intervention had reduced anxiety symptoms after the program and symptom scores remained at near normal levels at 3-month follow-up [78]. For This Way Up, a program targeted at generalized anxiety disorder, the intervention group participants showed significantly reduced symptoms of panic [81] and anxiety [82] at post-treatment compared to the control group. Symptom reduction was the same for technician and clinician-assisted versions of the treatment [83].

Program evaluation scores for each program and each evaluation criteria can be found in Multimedia Appendix 3. Scores for each program ranged from 69% (CBT 7 Step Self Help Course) to 100% (AI-Therapy) with an average score of 81% (SD 7%). Concerning the evaluation criteria, all program websites specified for which patient group or symptoms the program was designed, defined or stated the utilized model of change, presented program author names and credentials, and provided contact details. About half of the program websites had been empirically evaluated (19/34, 55.9%), specified which information was covered in the intervention modules (18/34, 52.9%), and provided evidence for the program to the user (eg, attrition data, success rate, completion rate, number of users in the program, testimonials) (14/34, 41.2%). Only 4 program websites (11.8%) specified whether the intervention was tailored to the user or generic for all users. This question could not be evaluated for one website (Social Anxiety Institute).

To our knowledge, this is the first review of publically available Web-based programs for anxiety that showcases what individuals seeking such treatment options might find if they search the Web. The review aimed at providing consumers, practitioners, and researchers with a summary of the availability, characteristics, and efficacy of currently freely available Web-based interventions for anxiety. The review identified a wide variety of programs for anxiety, specific anxiety disorders, or anxiety in combination with stress, depression, or anger with treatments based predominantly on CBT techniques. The majority of websites were found to be credible and accessible. Of the programs reviewed, the majority required that users register and/or pay a program access fee. Half of the programs offered some form of paid therapist or professional support. Programs varied in treatment length and number of modules and employed a variety of presentation modes. Relatively few were evaluated in terms of efficacy. In particular, this review highlights two key issues: the large number and diversity of program formats and the lack of empirical evidence of efficacy for many of the identified programs. These will be discussed in more detail and results will be compared with a similar review of Web-based depression programs available on the Web [37].

First, the great variety and large number of identified programs for anxiety is noteworthy. Programs differed in their level of support, accessibility, and presentation. A similarly great variability among identified programs was also found for Web-based depression interventions [37]. Concerning accessibility, more than half of the programs required an access fee. Considering the high costs and waiting times for psychotherapy in many countries, paid Web-based programs may provide an affordable alternative. However, programs often could only be purchased for a limited period. Many users may not be able to finish the program in the allotted time, and being able to receive treatment at one’s own pace might be an important reason for choosing Web-based treatment over face-to-face therapy.

Overall, most programs used a multimedia presentation for the intervention delivery. With the current rapid pace of advances in technology, more engaging ways of translating therapeutic techniques into interactive techniques could be created for Web-based interventions to distinguish them from traditional self-help material. Increased engagement through interactivity may increase adherence and effectiveness [94,95], especially when considering reports of low utilization and high dropout rates of Web-based interventions [19,96]. As individuals may differ in their preferred style of therapy and time and resources available for treatment, trying different programs and considering the access period is recommended before choosing a program.

The number of identified anxiety programs was similar to the number listed in the Beacon directory [35]. In total, 33 distinct programs were found in the directory. About half of those programs were also identified by this review and some of those identified in this review were not listed on Beacon. It is important to note here that the Beacon website is not updated very often; for example, some reviews of anxiety programs were last updated in 2009. The difference may also be a result of the keywords used and the way search engines are designed and work. Search engines are often referred to as “information gatekeepers,” as they are able to include and exclude websites and influence the ranking of websites in the search results [39]. These results suggest that even though a multitude of Web-based programs exists, it may be difficult for interested consumers to identify and compare all options. Having specialized services like the Beacon directory and keeping them up-to-date is therefore important to provide consumers with knowledge about program differences, credibility, and effectiveness. This will in turn help consumers to be able to compare programs and choose the one most suitable for them. A review of Web-based depression interventions identified a similar number (n=32) of programs on the Web and 12 of those programs were also included in this review [37]. Those were mostly programs that offered interventions targeting both anxiety and depression issues.

To ensure that consumers access programs of appropriate quality and safety, national and/or international platforms are needed that provide consumers with reliable guidance on evidence-based and effective Web-based intervention options. For example, the E‐Mental Health Strategy for Australia [97] outlines the development of an e-mental health portal that provides reliable information and accessible pathways for consumers and caregivers to navigate and use evidence-based Web-based mental health support. In addition, there is little consistent guidance on necessary quality standards, as well as legal and ethical issues regarding Web-based interventions for professionals. There is, for example, the Suggested Principles for the Online Provision of Mental Health Services by the International Society for Mental Health Online (ISMHO); however, this document mainly addresses online counselling and there are no guidelines specifically addressing Web-based programs. In the context of mental health apps, a review has also highlighted the need for standards and guidelines for developers to follow and frameworks for consumers to assess credibility and legitimacy [32].

Concerning the evidence base of the included programs, all were found to be based on CBT principles. This is consistent with prior reviews, which found that some form of CBT or other behavioral therapy was included in most Web-based interventions [98], as well as in publically available Web-based intervention programs for depression [37]. In general, research evidence indicates that CBT is an effective treatment for anxiety disorders (eg, [99,100]. This suggests that all reviewed programs were to some extent developed using an evidence-based approach; however, this does not guarantee that the evidence-based approach used is necessarily effective in the program.

In this context, another major finding was that several programs did not provide any research evidence or provided only limited evidence of the efficacy of the treatment. This is similar to findings from the review of Web-based depression interventions, which showed that 63% had not been evaluated using RCTs [37]. This finding is interesting considering the numerous systematic reviews and meta-analyses of Web-based interventions [25,30,101]. However, some may currently be in the process of being evaluated and not yet published. In addition, the absence of evidence of efficacy in terms of RCTs for a particular intervention also does not necessarily mean that the intervention does not work, especially if it is based on evidence-based approaches such as CBT. For treatment efficacy, the predominant model has been “empirically supported treatments” [102]. However, recently it has been proposed that clinical treatment decisions should be based on the best available research evidence, a clinician’s expertise, and patient characteristics [103]. It has also been argued that RCTs evaluating interventions should focus on evaluating intervention principles rather than each actual implementation [104]. However, unlike therapists who require accreditation to practice an evidence-based approach such as CBT, no such accreditations currently exist for Web-based programs. Hence, any Web-based program can claim to be based on CBT, but may not fulfil all requirements and therefore not work, which is especially problematic for programs requiring an access fee. Therefore, programs should ideally undergo appropriate empirical evaluation before being made available online. The development of an accreditation service for Web-based interventions may help improve this issue and enable consumers to make more informed decisions. It is also important to acknowledge competing interests within the eHealth space. Developers with a commercial focus may not be as concerned about treatment efficacy and researchers developing programs may not have the resources to sustain a publically available program. For programs for which published empirical studies were identified in this review, there was a large variety of study designs and quality of evidence. Only This Way Up programs, MoodGYM, myCompass, Beating the Blues, and FearFighter underwent rigorous evaluation through RCTs.

In regard to this study, a few limitations have to be noted. The representativeness and comprehensiveness of the search and identified programs may be affected by various characteristics of the Web, search engines, and search terms. First, the ranks of websites vary by location on commercial search engines. The search for this review was performed in the UK and it is likely that the same search in another country may have yielded different results. The Web and search engines are also dynamic. Results of search engines vary over time, meaning searches conducted several months before or after the current search could present a different set of programs. In addition, currently existing programs may change or be discontinued and new programs may be released. It is also possible that some individuals may not use the three search engines and would have therefore received different results. However, a considerable strength of this review is that the three most popular search engines were used.

Secondly, the first 25 hyperlinks from the search were included in this review. It is likely that more programs are available, which at the time did not have the page ranking to be identified by the search. This may especially be the case for recently created services [105]. However, it has been suggested that most people rarely consider more than the first 20 links [40]; thus the identified sample of this review is believed to be representative of what an average Internet user might discover when searching the Web for Web-based intervention options. Page ranking is also influenced by various search engine optimization techniques, algorithms of the search engines, as well as cookie settings of the browsers [106], and thereby impacts the results of the search. To combat this, we removed search engine histories and cookies were disabled on all three browsers. While this may not be a complete list of currently available Web-based programs for anxiety, it is a comprehensive snapshot of programs found in March 2015.

Third, the program evaluation scale used was an adapted version of the scoring system used by Renton et al [37]. However, the summative scores do not account for the fact that items within this scale may not be equivalent in terms of importance. Using different weightings based on importance would add great value to the rating. The development of such a scale was considered beyond the scope of this review; however, it would be important to develop a weighted scale for similar future reviews.

Lastly, it is important to acknowledge that the definition of Web credibility is complex and consists of multiple dimensions [107]. In the case of this paper, only a limited number of credibility dimensions that focused on trustworthiness rather than expertise, were assessed.

This review found that individuals searching for Web-based intervention programs for anxiety are presented with a large number and variety of potential programs to choose from. For consumers with limited knowledge about intervention quality criteria it may be challenging to choose an appropriate program. With the number of people using the Internet increasing, it is likely that more individuals will search for information about treatments options in general and, specifically, online interventions. It is therefore important for health professionals working with mental health clients to be aware of the diversity of Web-based interventions and that not all have had their efficacy tested in robust research trials. Directories such as Beacon can assist clinicians, as well as individuals in this task; however, it is important to keep services like these up-to-date. There is a definite need for consistent guidelines and standards on developing and providing Web-based mental health intervention programs for professionals and a platform with reliable up-to-date information for professionals and consumers about the quality and accessibility of Web-based interventions. This review is the first to identify and review Web-based anxiety interventions available on the Web. Therefore, research is needed in reviewing and evaluating Web-based intervention programs for other mental health related issues. There is also a need to develop standardized evaluation scales for publically available Web-based intervention programs to facilitate the rating process and ensure its rigor. For future research, it may also be interesting to explore health professionals’ and consumers’ experiences and perceptions of those programs.