Dose-response relationship in digital psychological therapies for people with psychosis: a systematic review, meta-analysis, and meta-regression

Introduction: Recent digital technological advances have emerged with the aim of improving accessibility, engagement, and effectiveness of psychological interventions for psychosis. Systematic reviews have provided preliminary evidence that digital health technologies for psychosis may improve symptoms. However, little research has examined how treatment effect is related to dose of therapy. Thus, we planned to investigate the association between treatment outcome and different dose characteristics, such as session length, number of sessions and their frequency. Methods: This systematic review followed the PRISMA guidelines, including a risk of bias assessment utilizing the Cochrane Collaboration’s tool. Searches were completed in November 2023 using Embase, Ovid MEDLINE(R) and APA PsychInfo, and were limited to English language and peer-reviewed journal articles. Studies included any randomised controlled trial (including pilot/feasibility studies) in adults that reported a non-interventional control condition and included clinical symptom outcome measurement and dose information. Meta-analyses and meta-regressions were completed. Results: 19 studies were included in this review. 14 studies included web, mobile or computer-based interventions, and 5 included virtual reality interventions. Digital interventions significantly improved clinical symptoms, with a small effect size (Cohen’s d = -0.14, p < 0.001, 95% CI [-0.23 to -0.05]). Although subgroup analyses were not significant, data patterns favoured interventions focusing on clinical outcomes over cognitive outcomes, and interventions that included therapist support, over those without. Due to the small overall effect size, we were not able to explore dose predictors. Discussion: This meta-analysis provided preliminary evidence that digital mental health interventions for psychosis are effective, even when not targeting symptoms directly. Despite exploring multiple dose characteristics, no significant dose-response relationship was found. Further research is needed to understand the role of dose in digital interventions for psychosis. Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/view/CRD42023411836, identifier CRD42023411836.