Clinical improvement and engagement phenotypes with Ash: evidence from two real-world studies
At SBM 2026, Caitlin Stamatis presented two real-world studies of Ash showing reductions in depression and anxiety, improved social support, and engagement-linked outcomes.
Across two real-world studies of Ash, users showed clinically meaningful reductions in depression and anxiety (PHQ-9 d=.80, GAD-7 d=.77 at 6 weeks, sustained at 10 weeks) with improved social functioning, and higher-engagement users generally improved more.
Study type
Real-world evidence synthesis (two studies)
Published in
Society for Behavioral Medicine 2026 Annual Meeting
Partner
New York University
In a session on real-world evidence in digital mental health at the Society for Behavioral Medicine 2026 Annual Meeting, Dr. Caitlin Stamatis presented findings synthesizing two real-world studies of Ash. A prospective cohort (N=305, baseline PHQ-9 or GAD-7 ≥10) showed significant reductions in depression and anxiety at 6 weeks, sustained at 10 weeks, alongside gains in loneliness and perceived social support; a second study derived five engagement phenotypes from roughly 100,000 users and linked higher engagement to greater symptom improvement.
Safety guardrails flagged about 1% of sessions for potential crisis content, each with clinical review and escalation, and 38.6% of conversations occurred between 9pm and 5am — underscoring the value of always-available support beyond traditional care hours.
The talk drew three lessons for real-world evidence in digital health: model engagement as multidimensional phenotypes rather than simple dose counts; link passively collected behavioral metrics with repeated standardized outcomes; and treat safety as continuous risk monitoring with clinician escalation rather than one-time pre-launch validation.