Dismantling, optimising, and personalising internet cognitive behavioural therapy for depression: a systematic review and component network meta-analysis using individual participant data.

BACKGROUND

Internet cognitive behavioural therapy (iCBT) is a viable delivery format of CBT for depression. However, iCBT programmes include training in a wide array of cognitive and behavioural skills via different delivery methods, and it remains unclear which of these components are more efficacious and for whom.

METHODS

We did a systematic review and individual participant data component network meta-analysis (cNMA) of iCBT trials for depression. We searched PubMed, PsycINFO, Embase, and the Cochrane Library for randomised controlled trials (RCTs) published from database inception to Jan 1, 2019, that compared any form of iCBT against another or a control condition in the acute treatment of adults (aged ≥18 years) with depression. Studies with inpatients or patients with bipolar depression were excluded. We sought individual participant data from the original authors. When these data were unavailable, we used aggregate data. Two independent researchers identified the included components. The primary outcome was depression severity, expressed as incremental mean difference (iMD) in the Patient Health Questionnaire-9 (PHQ-9) scores when a component is added to a treatment. We developed a web app that estimates relative efficacies between any two combinations of components, given baseline patient characteristics. This study is registered in PROSPERO, CRD42018104683.

FINDINGS

We identified 76 RCTs, including 48 trials contributing individual participant data (11 704 participants) and 28 trials with aggregate data (6474 participants). The participants' weighted mean age was 42·0 years and 12 406 (71%) of 17 521 reported were women. There was suggestive evidence that behavioural activation might be beneficial (iMD -1·83 [95% credible interval (CrI) -2·90 to -0·80]) and that relaxation might be harmful (1·20 [95% CrI 0·17 to 2·27]). Baseline severity emerged as the strongest prognostic factor for endpoint depression. Combining human and automated encouragement reduced dropouts from treatment (incremental odds ratio, 0·32 [95% CrI 0·13 to 0·93]). The risk of bias was low for the randomisation process, missing outcome data, or selection of reported results in most of the included studies, uncertain for deviation from intended interventions, and high for measurement of outcomes. There was moderate to high heterogeneity among the studies and their components.

INTERPRETATION

The individual patient data cNMA revealed potentially helpful, less helpful, or harmful components and delivery formats for iCBT packages. iCBT packages aiming to be effective and efficient might choose to include beneficial components and exclude ones that are potentially detrimental. Our web app can facilitate shared decision making by therapist and patient in choosing their preferred iCBT package.

FUNDING

Japan Society for the Promotion of Science.

Informations de copyright

Copyright © 2021 Elsevier Ltd. All rights reserved.

Déclaration de conflit d'intérêts

Declaration of interests TAF reports grants from Japan Society for Promotion of Science, during the conduct of the study; grants and personal fees from Mitsubishi-Tanabe, personal fees from MSD, grants and personal fees from Shionogi, outside the submitted work; a patent 2018-177688 concerning smartphone CBT apps pending; and an intellectual properties for Kokoro-app licensed to Tanabe-Mitsubishi. AC reports personal fees from Italian Network for Paediatric Trials and CARIPLO Foundation; and grants and personal fees from Angelini Pharma, outside the submitted work. EGO reports personal fees from Angelini Pharma, outside the submitted work. PCa reports personal fees from Osmond Foundation and Sandoz, outside the submitted work. JD is co-owner of Behavioral Activation Tech LLC, a small business that develops and evaluates mobile app-based treatments for depression and co-occurring disorders. DDE has served as a consultant to or on the scientific advisory boards of Sanofi, Novartis, Minddistrict, Lantern, Schoen Kliniken, Ideamed, German health insurance companies (BARMER, Techniker Krankenkasse), and a number of federal chambers for psychotherapy; is a stakeholder of the Institute for health training online (GET.ON), which aims to implement scientific findings related to digital health interventions into routine care. NRF is an employee of AbleTo. JPK reports grants and personal fees from Servier; personal fees from Beltz, Elsevier, Hogrefe, and Springer, outside the submitted work; funding for clinical trials (German Federal Ministry of Health and Servier); payments for presentations on internet interventions (Servier); and payments for workshops and books (Beltz, Elsevier, Hogrefe, and Springer) on psychotherapy for chronic depression and on psychiatric emergencies. BM is an employee of GAIA AG. DCM reports personal fees from Apple, Pear Therapeutics, and Otsuka Pharmaceuticals and has an equity interest in Adaptive Health, outside the submitted work. JMM is supported by a Wellcome Trust Grant (104908/Z/14/Z). SN is an employee of GET.ON Institut. DR is an employee of SilverCloud Health. LBS is an employee of Influents Innovations. PZ reports grants and non-financial support from Techniker Krankenkasse (German public health insurance company), outside the submitted work. CK reports personal fees from Oberbergklinik and Servier; and grants and non-financial support from Techniker Krankenkasse, outside the submitted work. MH reports grants and non-financial support from Techniker Krankenkasse, outside the submitted work. All other authors declare no competing interests.