Dropout rates in clinical trials of smartphone apps for depressive symptoms: A systematic review and meta-analysis

Background: Low engagement and attrition from app interventions is an increasingly recognized challenge for interpreting and translating the findings from digital health research. Focusing on randomized controlled trials (RCTs) of smartphone apps for depressive symptoms, we aimed to establish overall dropout rates, and how this differed between different types of apps. Methods: A systematic review of RCTs of apps targeting depressive symptoms in adults was conducted in May 2019. Random-effects meta-analysis were applied to calculate the pooled dropout rates in intervention and control conditions. Trim-and-fill analyses were used to adjust estimates after accounting for publication bias. Results: The systematic search retrieved 2,326 results. 18 independent studies were eligible for inclusion, using data from 3,336 participants randomized to either smartphone interventions for depression (n = 1,786) or control conditions (n = 1,550). The pooled dropout rate was 26.2%. This increased to 47.8% when adjusting for publication bias. Study retention rates did not differ between depression vs. placebo apps, clinically-diagnosed vs. self-reported depression, paid vs. unpaid assessments, CBT vs. non-CBT apps, or mindfulness vs. non-mindfulness app studies. Dropout rates were higher in studies with large samples, but lower in studies offering human feedback and in-app mood monitoring. Discussion: High dropout rates present a threat to the validity of RCTs of mental health apps. Strategies to improve retention may include providing human feedback, and enabling in-app mood monitoring. However, it critical to consider bias when interpreting results of apps for depressive symptoms, especially given the strong indication of publication bias, and the higher attrition in larger studies.