Maintenance Notice

Due to necessary scheduled maintenance, the JMIR Publications website will be unavailable from Monday, March 11, 2019 at 4:00 PM to 4:30 PM EST. We apologize in advance for any inconvenience this may cause you.

Who will be affected?


Currently submitted to: JMIR Formative Research

Date Submitted: Feb 16, 2020
Open Peer Review Period: Feb 16, 2020 - Apr 4, 2020
(closed for review but you can still tweet)

NOTE: This is an unreviewed Preprint

Warning: This is a unreviewed preprint (What is a preprint?). Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn (a note "no longer under consideration" will appear above).

Peer-review me: Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period (in this case, a "Peer-Review Me" button to sign up as reviewer is displayed above). All preprints currently open for review are listed here. Outside of the formal open peer-review period we encourage you to tweet about the preprint.

Citation: Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author).

Final version: If our system detects a final peer-reviewed "version of record" (VoR) published in any journal, a link to that VoR will appear below. Readers are then encourage to cite the VoR instead of this preprint.

Settings: If you are the author, you can login and change the preprint display settings, but the preprint URL/DOI is supposed to be stable and citable, so it should not be removed once posted.

Submit: To post your own preprint, simply submit to any JMIR journal, and choose the appropriate settings to expose your submitted version as preprint.

Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.

Internet-delivered vs face-to-face cognitive behavior therapy for anxiety disorders: systematic review and meta-analysis

  • Narges Esfandiari; 
  • Mohammad Ali Mazaheri; 
  • Vahid Sadeghi-Firoozabadi; 
  • Mona Cheraghi; 



Over the last twenty years, Internet-delivered cognitive behavior therapy (ICBT) has been tested in a large number of randomized controlled trials, often with positive results. However not widely known about the efficacy of ICBT as compared to face-to-face cognitive behavior therapy (CBT). In the present systematic review and meta-analysis, ICBT for anxiety disorders were directly compared to face- to-face CBT within the same trial. This study aimed to reinvestigate the efficacy of ICBT compared to face-to-face cognitive behavior therapy (CBT) for anxiety disorders. A total of 8 studies Out of the 236 articles screened, met all inclusion criteria. Results showed a pooled effect size at post-treatment of Hedges g = .01 (95% CI, -0.16 to 0.18), indicating that ICBT and face-to-face treatment created equivalent overall effects. While the overall results indicate equivalence, there have been few studies of the anxiety disorders so far, and for the majority, ICBT has not been compared against face-to-face treatment. Therefore, more research is needed to establish the general equivalence of the two treatment formats.


Internet-based cognitive behavior therapy (ICBT) is almost the same as CBT but delivered through the internet in which a client completes some materials and modules through a website (Anderson et al., 2011). The patient is also guided by an online therapist that provides support through clarifying information, monitoring the progress, giving feedback to homework and allowing the patient access to the sequential treatment steps (Anderson et al., 2011). ICBT has many advantages compared to classic treatment: it requires less therapist time, in adults about 85% less therapist time per week and it’s not limited to the office hours. It also can be performed without the distances’ limitation between therapist and patient, which can decrease the possible risk of stigma in visiting a therapist, and children and their parents can be involved in the therapy program without missing school or work (Vigerland et al., 2016). Several forms of ICBT are designed by searchers from all around the world, most have a therapist, guiding the patient during treatment. Treatments are structured, usually includes up to 15 training modules, which are roughly equivalent to the number of face-to-face CBT sessions (Lindefors & Anderson, 2016). These treatments are different from many aspects, such as technical solutions, the amount of therapist's support and diagnostic processes (Lindefors et al., 2012). They have also some features in common including all are based on cognitive-behavioral approach including components of face-to-face CBT such as exposure, psycho-education and treatment during 8 to 15 weeks. Most of these programs use text messages as the main form of communication between the patient and the therapist. They have an integrated assessment system and assignment in the internet-based therapy program (Lindefors & Anderson, 2016). Since early treatment researches on ICBT that were done in the late 1990s (Andersson, Carlbring, & Lindefors, 2016), more than 200 randomized controlled trials have been issued, often with hopeful outcomes showing that ICBT is clinically effective when compared to controls (Andersson, Carlbring, & Hadjistavropoulos, 2017). Besides a few numbers of long-term, follow-up studies have indicated that the ICBT effects are kept for if five years post-treatment (Hedman, et al., 2011). Despite promising outcomes in control trials, in which ICBT is often compared with waiting list control groups, an important question is how well guided ICBT compares against face-to-face treatment (Carlbring et al., 2017). This was studied in a meta-analysis by Andersson, Cuijpers, Carlbring, Riper, and Hedman (2014) that consisted of 13 studies (N=1053) published till June 2013. The results indicated a pooled effect size at a post-treatment of Hedges g= -.01 (95% Cl, -.13 to .12), suggesting that ICBT and face-to-face treatment result in equal overall effects. This study was updated in 2018 with 20 studies (N=1418), the results showed some effects of CBT and ICBT for the psychiatric and somatic disorders. This approach offers a reasonable alternative to clinic-based therapy with advantages of low cost, reduced therapist time and greater accessibility for families who have difficulty accessing clinic-based CBT (Boettcher, et al, 2013; Spence et al, 2011). One of the most important areas that were well researched on ICBT is anxiety disorders (Andersson, 2009). Several studies showed that internet-based treatments have the same effects as face-to-face studies for different types of anxiety disorders including, panic disorder, social anxiety, fear of public speaking, fear of spider (Hedman et al, 2011; Spence et al, 2011; Boetela et al, 2010; Bergstrom et al, 2010; Andrew et al, 2011; Anderson et al, 2009). Spence and colleagues (2011) compared the effect of online vs clinic-based CBT in 115 anxious adolescents. Results showed a great reduction in both groups in comparison to the waiting list group (online group’s d=2.12; clinic group’s d=3.42). Also, these improvements were kept at 6 and 12 months follow-up. So ICBT is equally effective as a face-to-face treatment of anxiety disorders in adolescents. In another study, Hedman and colleagues (2011), investigated the cost-effectiveness of ICBT compared to Cognitive-Behavioral Group Therapy (CBGT) in the treatment of social anxiety. Results showed that the reduction of social anxiety was the same in both treatments. They also conducted a 4-year follow-up and found that participants in both treatment groups made large improvements from baseline to 4-year follow-up on the primary outcome measure (d=1.34-1.48). Anderson and colleagues (2009), compared guided ICBT with one session of live exposure therapy in 30 patients with spider phobia. The results of the research also didn’t show any difference in both groups (d=1.84; 2.58 for the internet and live-exposure groups respectively, at post-treatment). One more study, conducted by Bergstrom and colleagues, compared the effectiveness of Internet and group CBT for panic disorder in a randomized trial. 113 patients were randomly assigned to guided ICBT group (n=53) or group CBT (n=60). Again, both treatments showed a significant reduction in symptoms. For the Internet treatment, the within-group effect size (pre-post) was d=1.73, and for the group treatment, it was 1.63. Effects were maintained at 6 months follow-up. This study suggests the equal effectiveness of both treatments. Although studies in the last decade have consistently shown that ICBT is an effective treatment for anxiety disorders (Hedman et al., 2011) with the same outcomes in comparison to face to face treatment, a lot of people never receive treatment or get it after years of pain(Anderson,2009; Vigerland, 2015). Also, according to some studies, anxiety disorders, despite being extremely prevalent, have a little possibility of being treated than other psychiatric disorders (Vigerland, 2015). On the other hand, the efficacy of ICBT in different types of disorders may vary due to their characteristics (Andresson, 2009), therefore there is a crucial need to do more researches specifically to different anxiety disorders. As this area is moving ahead quickly and many new studies have been published, also to the best of our knowledge, no earlier review has targeted anxiety disorders for comparing internet-based and face to face cognitive behavior therapy, there is a need for a systematic review and meta-analysis focusing on this area. This study aimed to compare the efficacy of internet-based vs face-to-face CBT for anxiety disorders. We conducted a systematic review and meta-analysis of studies directly comparing the two treatment formats. According to strong research literature, we hypothesized that ICBT and face-to-face CBT would produce equivalent treatment effects.


Methods Data sources and search strategy To identify studies published until 2019, systematic searches in PubMed, Scopus, ProQuest, Emerald and Science direct were conducted. The search strategy employed a combination of search terms related to anxiety disorders (anxiety, anxiety disorders, fear, panic disorder, social phobia, social anxiety, generalized anxiety disorder, post-traumatic stress disorder, specific phobia, obsessive-compulsive disorder) and internet (internet, online, web, computer, computerized, internet-based, internet-delivered, internet delivery) and CBT (behavior, cognitive, therapy, treatment). The complete search strings can be viewed in Fig 1. Study Selection Articles were included if a) compare therapist-guided ICBT to face-to-face treatment using a randomized controlled design, b) use interventions aimed at the treatment of anxiety disorders (and not, for example, prevention or mere psychoeducation), c) investigate a form of ICBT where the internet treatment was the main component and not a secondary complement to other therapies, d) the internet treatment group be supported by an online therapist, e) be written in English. There were no restrictions regarding ages. Each article was independently assessed by first and corresponding authors at title, abstract and full-text level to evaluate inclusion or exclusion according to the criteria presented above. In the cases with disagreement on inclusion decision, the full-text level was reviewed for reaching consensus. Data Extraction The following variables were extracted from the included studies for further analysis: country, first author, year published, type of disorder, participant age interval, sample size in both internet-based and face to face groups, outcome measure, Type of outcome measure (categorized as “clinician-rated”, “self-rated”, “parent-rated” or a “physiological measure”), Pre- and post-treatment means and standard deviations of outcome measures, number of modules, sessions; weeks of therapy in both groups and assessment of study quality Assessment of Study Quality All included studies were assessed by quality assessment instrument (Moncrieff, Churchill, Drummond, & Mcguire, 2001), a scale of study quality on 23 different characteristics, each assessed on a 3-point scale (0 = poor, 1 = fair, 2 = good) including ratings on appropriate sample size, study design, statistical analyses and presentation of outcomes. In the original article the mean ratings of 30 mental health trials were between 16.3(SD=6.3) and 20.9(SD=9.0) and the interrater reliability was in the great range (r=0.75 to 0.86) (Moncrieff et al., 2001). Each article was evaluated independently by two authors. The mean of these scores was considered as the quality score of study. Statistical Analysis Data were analyzed by Cochrane Review Manager (RevMan) version 5.3 (Higgins & Greens, 2011 available from Using the chosen outcome in each study, a random-effects meta-analysis was carried out, see Table1. In the primary meta-analyses, we evaluated the effect of ICBT compared to face-to-face treatment using the standardized mean difference at post-treatment (Hedges’ g) as the outcome, meaning that the difference between treatments was divided by the pooled standard deviation. If both intention-to-treat and per-protocol data were presented, the earlier estimation was used in the meta-analysis. Estimates of treatment effects were conducted using all included studies as well as separately for each anxiety disorder (e.g. social anxiety). Possible differences in dropout rates between ICBT and face-to-face treatment were analyzed using meta-analytic logistic regression. All pooled analyses were conducted in a random-effects model framework, assuming variation in true effects in the included studies and accounting for the hypothesized distribution of effects (Carlbring, et al, 2017). Forest plots were also inspected to assess variation in effects across studies.


Results Systematic Review Included Studies The database search resulted in 236 articles, which after screening 8 of them met all inclusion criteria and were included in the study (see Fig. 1). One of them (Hedman et al., 2014) was a long term follow up of another study (Hedman et al., 2011) and are not reported separately. Thus, 8 studies were included in the systematic review of the meta-analyses (n=537). The total number of 8 studies investigated ICBT against some form of CBT (individual format, n = 4 or group format, n = 4). In terms of disorder studied, three of the studies targeted social anxiety (Hedman et al., 2011; Andrews et al., 2011; Hedman et al., 2014), two of them panic disorder (Carlbring et al., 2005; Bergstrom et al., 2010), one adolescence anxiety (Spence et al., 2011), one spider phobia (Andersson et al., 2009), and a fear of public speaking (Botella et al., 2010). The total number of participants was 348 in ICBT and 316 in face-to-face conditions. The studies were carried out in Australia, Spain, and Sweden. The smallest study had 47 participants, and the largest had 126. All studies were published between the years 2005 and 2014. The characteristics of each study are presented in Table 1. Duration and intensity ICBT interventions included between 4 and 15 treatment modules (M= 10, Md= 10) that were to be completed in 5 to 15 weeks (M= 10.87, Md= 10). Face-to-face treatments consist of 2 and 15 sessions that were held between 1 and 15 weeks. Five interventions have a “once a week” format with approximately the same number of treatment modules and weeks. One intervention had more weeks than modules in ICBT (Andrews et al., 2011), two had more modules than weeks (Andersson et al., 2009; Botella et al., 2010). Treatment Adherence Treatment adherence was not systematically reported. Three studies reported mean number of modules completed in ICBT and sessions attended in CBT within the treatment period (Bergstrom et al., 2010; Andrews et al., 2011; Spence et al., 2011) and three studies reported both average number of completed modules and sessions and also proportion who completed all modules (Hedam et al., 2011; Hedam et al., 2014; Carlbring et al., 2005). One study reported the number of dropouts (Botella et al., 2010) and one study presented no clear definition of treatment adherence (Andersson et al., 2009) Outcome Informant Two out of eight studies used clinician-rated symptom severity as the primary outcome (Bergstrom et al., 2010; Spence et al., 2011) and the other six, used self-rated measures (Hedman et al., 2011; Hedman et al., 2014; Carlbring et al., 2005; Botella et al., 2010; Andersson et al., 2009; Andrews et al., 2011). Overall Study Quality Involved studies had a total score on the Moncrieff rating scale ranging from 21 to 37 points (M= 33.75, SD= 36). For the total rating, each study see Table 1. One out of eight studies reported a power calculation with full details (study 6), and two studies mentioned power calculation without details. Seven of 8 studies had a specified primary outcome measure (studies 1, 2, 3, 4, 5, 6 and 8). Six of the studies used blind assessors (studies 1, 4, 5, 6, 7 and 8), but testing of blinding was reported in three of them (studies 5, 6 and 8). Half of the trials included a representative sample (for example all consecutive admissions at a clinic against volunteers; (studies 4, 5, 7 and 8). Seven of the studies presented their results from intention to treat-analyses (1, 2, 4, 5, 6, 7 and 8). Meta-Analysis Main findings: ICBT vs. face-to-face treatment All studies A forest plot displaying the effect sizes of studies is presented in figure 2. Through the results, an effect size estimate (g) below 0 favors ICBT while an effect size above 0 represents larger effects for face-to-face CBT. The pooled between-group effect size at post-treatment across all 8 studies was g = .05 (95% CI, −0.16 to 0.18), showing that ICBT and face-to-face treatment produced equivalent overall effects. Two studies (Andersson, 2009; Spence, 2011) were excluded from the meta-analysis because the SD of pre-tests were not reported. The results below are presented separately. Social anxiety disorder The pooled-between group effect size in the three studies targeting social anxiety disorder (Andrews et al., 2011; Hedman, Andersson, et al., 2011; Hedman et al., 2014) was g= .04 (95% CI, −0.19 to 0.27) in line with the notion of equivalent effects. Panic disorder The pooled-between group effect size in the two studies targeting panic disorder (Carlbring et al., 2005; Bergstrom, Andersson et al., 2010) was g= -0.17 (95% CI, −0.48 to 0.14) which represents a non-significant effect in favor of ICBT, however, the sample size in the first study(Carlbring et al., 2005) was small. Fear of public speaking The pooled-between group effect size in one study targeting fear of public speaking(Botella et al., 2010) was g= 0.25 (95% CI, −0.16 to 0.67) in favor of face-to-face treatment, but given the smaller size of the face to face group, the finding was not significantPublication bias Figure 3 displays a funnel plot relating effect sizes on the main outcomes of the studies to the standard errors of the estimates. As shown in figure 3, the effect size was evenly distributed around the averaged effect. The lower left section of the funnel plot includes studies suggesting that there is no major bias of the pooled effect estimate. This is due to small, unpublished studies with results favoring internet treatment. Test of heterogeneity Tests showed significant heterogeneity (????2 = 8.97; I2 = 44%; p = .11) as shown in Table 2. This heterogeneity was largely driven by the study of ICBT vs. face-to-face CBT for panic disorder by Carlbring et al. (2005). If this study was removed from the analysis, I2 dropped from 44 to 27 and heterogeneity would decrease significantly. The pooled effect size across all studies changed marginally from g = .01 (95% CI, −0.16 to .18) to g = −0.06 (95% CI, −0.11 to 0.24) if this study was removed from the analysis.Figure 3. Funnel plot to assess for publication bias by relating effect sizes of the studies to standard errors abbreviation: SMD, the standardized mean difference


Discussion This meta-analysis has strong findings and is the first study of anxiety disorders, comparing the internet and face-to-face treatments. The meta-analysis found out that dropouts did not systematically favor one treatment format over the other. Although there were no special effects in favor of either ICBT or face-to-face treatment, there were some fascinating non-significant results in reverse directions, depending on disorders. Most noticeable was the non-significant superiority of internet over the face to face therapy. A possible cause for this is that in one study with g = -0.17 the sample size is face to face group is smaller than the internet group (Andrews et al., 2011) and in another study with g= -0.51 the sample size is small in general (Carlbring et al., 2005). In two studies there were effect sizes in favor of face to face treatment which could be explained by the high effective of the gold standard face to face treatment (Carlbring et al, 2018). The other important difference was that in the treatment of social anxiety disorder, the results were in moderate favor of ICBT (Andrews et al., 2011; Hedman et al., 2014). Probably this non-significant finding can be accepted in light of the possibility that the therapist could be a phobic object itself and the therapeutic relationship would increase the anxiety in these patients. Also, in face-to-face treatment, the patient’s self-focus will be raised and therefore his ability to completely concentrate on the therapy might be prevented (Carlbring et al, 2018). The quality of the studies differed both in terms of accuracy and sample size. However, study quality did not influence outcomes significantly, although the number of studies was too small to conclude that with confidence. Moreover, there was no major bias of the pooled effect estimate. The current meta-analysis had some strong points, including a persistent result across studies concerning the effectiveness of ICBT compared to face-to-face CBT and the reasonably high quality of the trials included. Nevertheless, the study also had limitations. First, there was a possible problem with heterogeneity. Since that was driven by a single extreme value on the panic disorder (Carlbring et al., 2005) that favored ICBT, we decided to report the findings both with and without that study included in the analysis. However, the relative effect of that study on the pooled effect size across all studies was insignificant, with g = .01 changing to g = −.06 if it was excluded. Second, It has been pointed out that internet interventions have high potential for reducing emotional distress, enhancing mental health, and promoting well-being, but there could also be negative impacts related to treatment, though a meta-analysis found that the conditions of 5.8% of participants involved in ICBT worsened (Rozental, Magnusson, Boettcher, Andersson, & Carlbring, 2017). That number compares well with the 5–10% found in face-to-face treatments (Hansen, Lambert, & Forman, 2002) and is much lower than the 17.4% of control group participants in internet-delivered trials (Rozental et al., 2017). In a recent study that investigated the remission rates in ICBT, symptom severity seems to be predictive of outcome inversely, while having more symptoms and being a female increase the chance of improvement (Andersson, Carlbring, Rozental, 2019). Third, the treatment review is based only on CBT, which makes it hard to generalize to other treatment approaches. Whereas the great majority of present treatments are based on cognitive-behavioral platforms, there are other forms of internet-delivered psychotherapy, such as psychodynamic psychotherapy (Johansson et al., 2012), physical exercise (Nyström et al., 2017), and different forms of ICBT, including attention bias modification (Carlbring et al., 2012), problem- solving therapy (Warmerdam, van Straten, Twisk, Riper, & Cuijpers, 2008), and acceptance and commitment therapy (Ivanova et al., 2016). These internet-based intervention programs were not included in the analysis since they generally do not make a direct comparison with face-to-face psychotherapy. Although those clinical approaches are not as popular as CBT in online interventions, since psychodynamic treatment has been found to work in a few trials, it is suggested to study that intervention as well. However for most conditions such as anxiety disorders and health problems they are very few if any other psychotherapy orientations tested and CBT is unchallenged (Andersson, 2014). Fourth, we have compared ICBT to face-to-face therapy regardless of whether it was delivered in an individual or group setting. To resolve the relative efficacy of individual and group settings, head-to-head comparisons need to be conducted (Carlbring et al, 2018). Fifth, we analyzed only the primary outcome measures in the studies (e.g, LSAS: Leibowitz Social Anxiety Scale); we did not include secondary outcomes (e.g, The Spence Children’s Anxiety Scale (SCAS). We cannot, at this point and with the few studies available for each condition, conclude that ICBT and face-to-face therapy are equally effective in all conditions. For example, there are very few studies on gaining knowledge following CBT and even fewer on ICBT (Andersson, Carlbring, Furmark, & on behalf of the SOFIE Research Group, 2012), and the therapy formats may vary in this regard. Moreover, patient characteristics, such as cognitive flexibility (Lindner et al., 2016), have not been considered. This is possibly important since studies are suggesting that different predictors of outcome are relevant when comparing face-to-face treatment vs. Internet treatment (Ebert et al., 2016). Therefore we suggest future reviews to analyze secondary outcomes as well and also consider different variables such as cognitive flexibility in the study. Finally, most of the studies recruited participants only across self-referral or using a combination of self-referral and clinical recruitment. It has been proposed that recruiting through sources that suggest more active treatment-seeking behaviors (e.g. Google searches, viewing postings on mental health websites) leads to participants with more severe anxiety than those recruited through more passive sources of information (Carlbring et al., 2018). So It is suggested for future research to recruit through the digital footprint of users all over the world to find out who is seeking therapy. The results of our meta-analysis are interesting both from theoretical and practical standpoints. In terms of theories about change in psychotherapeutic interventions, the findings suggest that the role of a face-to-face therapist may not be as essential as suggested in the previous literature for producing large treatment effects (Wampold, 2001). Even if factors such as therapeutic alliance are established in guided ICBT, they are hardly important for its outcome (Andersson, paxling, et al, 2012). Indeed, understanding what makes ICBT work is a challenge for future research as only a few studies to date have examined mediators of outcome (e.g. Hedman et al., 2013; Hesser, Westin, & Andersson, 2014; Karyotaki et al., 2015). In conclusion, the purpose of this systematic review and meta-analysis was to collect and analyze studies in which ICBT had been directly compared with face-to-face CBT. The results show that two treatment formats are equally effective in treating social anxiety disorder, adolescent anxiety, panic disorder, spider phobia and fear of public speaking. Clinical Trial: The authors declare no conflict of interest. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors


Please cite as:

Esfandiari N, Mazaheri MA, Sadeghi-Firoozabadi V, Cheraghi M

Internet-delivered vs face-to-face cognitive behavior therapy for anxiety disorders: systematic review and meta-analysis

JMIR Preprints. 16/02/2020:18275

DOI: 10.2196/preprints.18275


Download PDF

Request queued. Please wait while the file is being generated. It may take some time.

© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.