Magnitude and Moderators of Depressive Symptom Improvement in Digital Placebo Arms: Systematic Review and Meta-Analysis.
Takashi Hosono, Rinka Tsutsumi, Yuki Niwa et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review and meta-analysis examined blinded randomised controlled trials using digital placebos as comparators and reporting PHQ-9 outcomes. Across 30 trials involving 5680 participants, the pooled within-control-arm improvement was Hedges g=0.44 (95% CI 0.29-0.59), with high heterogeneity (I 2 =93.2%, τ²=0.14). Primary psychiatric disorder status and baseline PHQ-9 score were identified as major moderators; together, they accounted for 51.5% of the heterogeneity in multivariable analysis. The findings indicate that symptom improvement in digital placebo arms is substantial enough to consider when designing and powering digital therapeutic trials.
Key findings
- Across 30 trials and 5680 participants, pooled PHQ-9 improvement in digital placebo arms was Hedges g=0.44 (95% CI 0.29-0.59).
- Heterogeneity in control-arm improvement was high (I 2 =93.2%, τ²=0.14).
- Primary psychiatric disorder status and baseline PHQ-9 score were major moderators, accounting for 51.5% of heterogeneity in multivariable analysis.
Why it matters
Improvement in digital placebo arms can contribute meaningfully to observed changes in digital therapeutic trials, complicating interpretation of intervention effects. The identified moderators suggest that control-arm outcomes may vary with participant characteristics, particularly baseline symptom severity and whether depression is the primary psychiatric disorder. Accounting for these factors may improve trial planning and interpretation.
What to keep in mind
The pooled estimate had high heterogeneity, and the identified moderators explained only 51.5% of it, indicating substantial variation remained unexplained.