Digital Psychotherapies for Adults Experiencing Depressive Symptoms: Systematic Review and Meta-Analysis (Preprint)

Joanna Omylinska Thurston, Supritha Aithal, Shaun Liverpool, Rebecca Clark, Zoe Moula, January Wood, Laura Viliardos, Edgar Rodríguez-Dorans, Fleur Farish-Edwards, Ailsa Parsons, Mia Eisenstadt, Marcus Bull, Linda Dubrow-Marshall, Scott Thurston, Vicky Karkou

Research output: Working paperPreprint

Abstract

Background:
In the UK depression affects 5% of adults and the National Health Service’s (NHS) Long Term Plan highlighted the need for developing digital psychotherapies to address this issue. This systematic review examines an available range of digital psychotherapies for adults with depression considering both effectiveness and user perspectives, addressing gaps in previous reviews.
Objective:
The systematic review focused on identifying (i) the most common types, (ii) helpful and unhelpful aspects and (iii) effectiveness of digital psychotherapies for adults with depression.
Methods:
A mixed-methods protocol was developed using PRISMA guidelines (PROSPERO ID: CRD42021238462). The search strategy used PICOS guidelines covering 2010-2021 timeframe. Seven data bases were searched and an Excel form was developed to gather information. Thirteen authors extracted data. Randomised controlled trials (RCTs) were evaluated using a risk of bias assessment tool. Studies with non-RCT designs were evaluated using the Mixed Methods Appraisal Tool (MMAT). Qualitative narrative synthesis presented helpful and unhelpful factors using a modified Behaviour Change Model.  Meta-analyses of depression outcomes were conducted using the standardised mean difference (SMD, calculated as Hedges' g) of post-intervention change between digital psychotherapy and control groups.
Results:
Of 2891 initial records, 150 records (126 studies) were included for analysis. Quantitative studies (100) with RCT design (73) were most common. The overall sample size included 50,209 participants (72.3% (F) and 23.5% (M)) MoodGYM was the most popular named digital intervention (13) followed by Beating the Blues (6). Digital interventions included: 1. ‘standalone’/ non-human contact interventions (55) 2. ‘human contact’ interventions (4) 3. ‘blended’ including standalone and human contact interventions (67) Helpful factors included: motivation and accessibility (standalone), explanation of tasks (human contact) and reminders/resources, plus learning skills to manage symptoms (blended). Unhelpful factors included problems with usability and lack of direction/explanation. 67 studies with 14,564 participants were used in a meta-analysis that revealed a moderate effect on depression (Hedges g = -0.56). Analysis of studies with blended approaches revealed a large effect size (Hedges g = -0.79) in comparison to interventions involving human contact (Hedges g = -0.33) or no human contact (Hedges g = -0.39). Subgroup analysis of 63 studies showed that once a week was sufficient to achieve a large effect (Hedges g = -0.88).
Conclusions:
The review examined systematically a range of digital forms of psychotherapy for depression, which is a new contribution to the existing evidence base. Blended interventions were found to be most common/ helpful and meta-analysis demonstrated that they were superior to other digital interventions offered. These interventions were especially helpful for those from diverse ethnic groups and young women. Future research should focus on understanding how to achieve a greater effect size for these groups in particular.
Original languageEnglish
PublisherJMIR Mental Health
ISBN (Electronic)2368-7959
DOIs
Publication statusPublished - 14 Dec 2023

Publication series

NameJMIR Mental Health
PublisherJMIR Publications

Keywords

  • digital psychotherapies
  • Psychotherapy
  • depressive symptoms
  • Depression - therapy
  • systematic review

Research Centres

  • Research Centre for Arts and Wellbeing

Fingerprint

Dive into the research topics of 'Digital Psychotherapies for Adults Experiencing Depressive Symptoms: Systematic Review and Meta-Analysis (Preprint)'. Together they form a unique fingerprint.

Cite this