Elsevier

Psychiatry Research

Volume 256, October 2017, Pages 371-377
Psychiatry Research

Effectiveness of an individually-tailored computerised CBT programme (Deprexis) for depression: A meta-analysis

https://doi.org/10.1016/j.psychres.2017.06.081Get rights and content

Highlights

  • Computerized CBT (cCBT) programmes differ widely but have rarely been evaluated separately through meta-analysis.

  • Deprexis is a tailored cCBT programme for depression with unique features, justifying a programme-specific meta-analysis.

  • Comparisons from 8 studies demonstrated effectiveness of Deprexis for depression reduction (g=0.54, 95% CI: 0.39–0.69).

  • The weighted-average dropout rate for participants allocated to Deprexis treatment arms was relatively low (26.5%).

  • The positive findings for Deprexis add to the growing evidence-base for individually-tailored cCBT programmes.

Abstract

Computerised cognitive behavioural therapy (cCBT) programmes differ widely but have rarely been evaluated separately through meta-analysis. Through a meta-analysis of randomised controlled trials, we investigated the effectiveness of ‘Deprexis’, an unconventional and individually-tailored cCBT programme for depression. Comparisons from eight studies (N = 2402) demonstrated the effectiveness of Deprexis for depressive symptoms at post-intervention, with a medium effect size (g = 0.54, 95% CI: 0.39–0.69). Analogous results arose when study quality, screening, and randomisation procedure were taken into account. The level of guidance provided alongside Deprexis had a statistically non-significant impact upon the effect size. There was no significant difference in the performance of Deprexis in developer-led trials compared with non-developer trials, and no publication bias was detected. The weighted-average dropout rate for participants allocated to Deprexis treatment arms in included studies was 26.5%. Based primarily on trials in naturalistic community settings, the findings support the effectiveness of Deprexis for depressive symptoms. The positive findings add to the growing evidence-base for individually-tailored cCBT programmes and point to the need for further investigations of apparent systematic differences in the effectiveness of specific cCBT programmes.

Introduction

The effectiveness of psychological therapies such as cognitive behavioural therapy (CBT) for depression is well-established but accessing them remains difficult due to limited health care resources (Bower and Gilbody, 2005). One way of increasing access is to deliver the therapeutic content of therapy using Internet-based self-help interventions requiring less clinician involvement (Andrews et al., 2010). Most of these interventions are broadly based on the principles of CBT and collectively they are referred to as Internet-based or computerised CBT (cCBT). Meta-analyses indicate that cCBT is effective, on average, for depression, supporting its utility as a low-intensity and easily accessible treatment option (Andersson and Cuijpers, 2009, Andrews et al., 2010, Arnberg et al., 2014, Spek et al., 2007). There are various issues with cCBT however, including high dropout rates (Andersson and Cuijpers, 2009), better evidence for effectiveness in experimental conditions than in routine clinical settings (Andersson et al., 2009), and markedly reduced effectiveness when cCBT is provided without regular clinician guidance (Andersson and Cuijpers, 2009, Andrews et al., 2010, Arnberg et al., 2014, Spek et al., 2007).

An issue that has not received much research attention is the heterogeneity of available cCBT programmes: there exist over 50 programmes for depression which differ in their evidence-base, content, duration, and overall quality (Christensen et al., 2010, Lintvedt et al., 2013). Most cCBT programmes target one mental health problem and are highly structured yet there is growing support for the effectiveness of both transdiagnostic and individually-tailored programmes (Johansson et al., 2012, Newby et al., 2016). Despite their wide heterogeneity, cCBT programmes have rarely been evaluated separately in previous meta-analyses. Separate evaluations would add precision to the findings and could help clinicians make more informed decisions regarding the provision of specific programmes to clients.

Deprexis (Meyer et al., 2009) is a prominent cCBT programme for depression that is unconventional in its structure and content: it is therefore worthy of an evaluation through meta-analysis separate from other cCBT programmes; moreover, a separate evaluation facilitates the detailed exploration of the unconventionality of Deprexis and its implications for cCBT research. Unlike most existing programmes for depression, Deprexis is individually-tailored: users navigate through the programme by selecting among pre-determined response options and subsequent content is tailored to match expressed needs and preferences (Meyer et al., 2009). The order of displayed content can vary, depth and detail vary depending on preferences, text passages are added or omitted depending on personal suitability, and some content sequences are presented only for those with matching interests. The minimum number of main content modules presented for a given user is 6, whereas the maximum number is 10. There is no assumption or requirement that users should or need to complete all modules to derive benefit. Any module can be repeated once the user has completed the personalised module sequence. Deprexis is broadly based on CBT but – atypically – it also includes content related to acceptance, mindfulness, attachment, interpersonal processes, positive psychology, dream-work, nutrition and physical exercise (Table 1).

Deprexis was developed by a multidisciplinary team of clinical psychologists, physicians, software engineers, graphic artists, and others at Gaia AG, Hamburg, Germany, a small to medium enterprise (SME) specialising on research and development in e-Health. The programme is currently available in five languages (German, English, Swedish, Portuguese, and Italian) and can be used with or without guidance by a clinician. When used in a guided format, clinicians can view individual users’ progress through the programme and symptom trajectory, provided that the user has granted permission to share these data. Programme cost varies depending on region and usage circumstances (e.g., free to participants in research studies; cost per license in other contexts). Deprexis is currently used within the German health care system and other countries. The programme can be accessed for 90 days after initial registration with an individual access voucher via a secure, encrypted website (https://deprexis.broca.io). It is recommended that users complete 1–2 sessions of about 30 min per week, although this can vary based on personal preferences. Using responsive design, Deprexis is optimized for use on any computer with internet access, including tablet computers and smartphones.

The aim of this meta-analysis was to investigate the effectiveness of Deprexis for depressive symptoms. Several randomised controlled trials (RCTs) conducted on Deprexis with large sample sizes could be examined. An additional benefit of the current investigation relates to the tailored format of Deprexis: relatively few evaluations of tailored cCBT have so far been undertaken and thus the findings may have important implications for clinical practice.

Section snippets

Eligibility criteria for study selection

  • (i)

    Participants: Adults with elevated depressive symptoms or adults seeking some form of treatment for depression.

  • (ii)

    Intervention: Deprexis, with or without clinician or technician guidance.

  • (iii)

    Comparison conditions: Waiting lists, delayed treatment, no treatment, care as usual (CAU), or computerised control conditions.

  • (iv)

    Outcome: Self-report or clinician-rated measures of depression. We did not undertake analysis of the (heterogeneous) secondary outcomes investigated in included studies.

  • (v)

    Study design:

Study selection

The literature search flow is displayed in Fig. 1. After removing duplicates, we screened the abstracts of 28 studies. There was an 89% agreement rate between authors CT and GOR in screening outcomes. Fourteen studies were assessed for eligibility at the ‘full-text’ level. Eight studies ultimately included in the meta-analysis (Beevers et al., 2017, Berger et al., 2011, Fischer et al., 2015, Klein et al., 2016, Meyer et al., 2009, Meyer et al., 2015, Moritz et al., 2012, Schröder et al., 2014).

Study characteristics

Summary of main findings

This meta-analysis of eight studies demonstrated the effectiveness of Deprexis for depressive symptoms at post-intervention, with a medium effect size (g = 0.54, 95% CI: 0.39–0.69). Analogous results arose when study quality, depression screening, and randomisation procedure were taken into account; effect sizes in these sensitivity analyses ranged from g = 0.42 to g = 0.59. Interestingly, the level of clinician or technician guidance provided alongside Deprexis had minimal (and statistically

Declaration of interest

Björn Meyer is part of the development team of Deprexis and is employed by Gaia AG, Hamburg, the owner and operator of Deprexis, but he was not involved in the study selection or analysis.

Funding

None.

Acknowledgments

None.

References (33)

  • T. Berger et al.

    Internet-based treatment of depression: a randomized controlled trial comparing guided with unguided self-help

    Cogn. Behav. Ther.

    (2011)
  • T. Berger et al.

    Effects of a transdiagnostic unguided Internet intervention (‘velibra’) for anxiety disorders in primary care: results of a randomized controlled trial

    Psychol. Med.

    (2016)
  • P. Bower et al.

    Stepped care in psychological therapies: access, effectiveness and efficiency

    Narrat. Lit. Rev. Br. J. Psychiatry

    (2005)
  • H. Christensen et al.

    Beacon: a web portal to high-quality mental health websites for use by health professionals and the public

    Med. J. Aust.

    (2010)
  • J. Cohen

    Statistical Power Analysis for Behavioral Sciences

    (1988)
  • S. Duval et al.

    Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis

    Biometrics

    (2000)
  • Cited by (78)

    • Content validity and perceived utility of a self-help online acceptance and commitment therapy program focused on repetitive negative thinking

      2022, Journal of Contextual Behavioral Science
      Citation Excerpt :

      Most of the existing studies feature online adaptations of acceptance and commitment therapy (ACT; Hayes et al., 2011; Wilson & Luciano, 2002) or other acceptance-based interventions. These OPIs are usually self-help-oriented and arranged in short modules of multimedia content (Brown et al., 2016; Cavanagh et al., 2014; Kelson et al., 2019; Spijkerman et al., 2016; Twomey et al., 2017). However, to our knowledge, there are no adaptations of ACT that can be used for broad treatment and prevention purposes in Spanish-speaking Latinos.

    • Chatbot-Delivered Psychotherapy for Adults With Depressive and Anxiety Symptoms: A Systematic Review and Meta-Regression

      2022, Behavior Therapy
      Citation Excerpt :

      Chatbots offer a solution to the shortage of mental health workers, as they are able to function autonomously while providing guided psychotherapy via automated script-based dialogue (Oh et al., 2017), and are suggested to form therapeutic alliances with individuals similar to those formed with human therapists (Kiluk et al., 2014). Previous meta-analyses (Twomey et al., 2017; Wahle et al., 2017) that included chatbots as part of the intervention reported similar effect sizes as did our review, which demonstrates the consistency of effect sizes when chatbots are adopted as part of the intervention. The meta-analyses of Moman et al. (2019) and Bennett et al. (2019) that included various technological interventions (discussion forums, online workbooks, video games) but did not involve chatbots reported lower effect sizes.

    View all citing articles on Scopus
    View full text