Internet Interventions (Dec 2024)
Crowdsourcing integrated into a digital mental health platform for anxiety and depression: A pilot randomized controlled trial
Abstract
Background: Anxiety and depression are major public health concerns. Digital mental health interventions (DMHIs) are effective at reducing anxiety and depression, especially when they leverage human support. However, DMHIs that rely on human supporters tend to be less scalable. “Crowdsourced peer support,” in which a “crowd” of many peers provides users support via structured and focused interactions, may enable DMHIs to provide some of human support's unique benefits at scale. Objective: To conduct a pilot trial of two versions of a digital mental health intervention for anxiety and depression: one with crowdsourced peer support and one without. Methods: We conducted a two-armed pilot randomized controlled trial examining two versions of the novel “Overcoming Thoughts” platform: crowdsourced (intervention) vs. non-crowdsourced (control). The crowdsourced version allowed participants to view and interact with other users' content. We randomly assigned 107 participants to use the crowdsourced (n = 56) or non-crowdsourced (n = 51) platform for 8 weeks. Participants completed assessments at baseline, 4 weeks, 8 weeks, and 16 weeks. At each time point, these assessments included measures of anxiety and depression, including the Depression, Anxiety, and Stress Scale (DASS, primary outcome), the Patient Health Questionnaire (PHQ-9, secondary outcome), and the Generalized Anxiety Disorder Questionnaire (GAD-7, secondary outcome). We also collected usage information, including the number of exercises started, and safety data. Results: Using mixed models controlling for demographic factors, we compared the conditions' effectiveness in reducing depression and anxiety over time. Although we found significant drops over time in the DASS at both Week 8 and Week 16 (ps 0.10). Conclusions: Neither version of the “Overcoming Thoughts” platform (crowdsourced or non-crowdsourced) reduced anxiety or depression significantly more than the other. Future work should investigate how digital platforms can better leverage crowdsourced support, and if crowdsourced support may be especially useful in certain kinds of systems, populations, or target areas. Optimizing intervention engagement and obtaining the large sample sizes needed for appropriate statistical power will be key challenges for similar studies.NCT: 04226742
