A new study in the Journal of Clinical Medicine looks beyond a single Supportiv session, examining the longer-term emotional trajectories of repeat live-moderated digital peer support users.

This retrospective longitudinal study analyzed 4,099 eligible chat sessions using GPT-4o-mini generated sentiment scores that were plotted to individual user-level ordinary least squares trajectories across accumulated conversation time.

Supportiv conversations providing emotional support for depression, grief, loneliness, suicidality, and work-related concerns were included in the analyses, and for all topic-emotion pairs, repeated digital peer support engagement was associated with declines in negative sentiment.

For trajectories that reached statistical significance, expressed sentiment scores declined for:

  • Despair in 90.9% of trajectories
  • Depression in 86.1% of trajectories
  • Grief in 85.7% of trajectories
  • Work-related stress in 82.4% of trajectories
  • Loneliness in 72.7% of trajectories

Other key findings:

  • Approximately 40–55% of all negative trajectories reached statistical significance.
  • Repeat users returned for subsequent sessions 10.4 days on average.
  • Users discussing depression returned sooner, and those discussing suicidality (18.6 days) or loneliness (11.8 days) returned later.
  • Suicidality-related sessions were approximately 20% longer and included ~35% more resources than average.

The service’s flexible access, low cost, and low stigma format may help facilitate users’ likelihood to engage with care over time as needed (for example, more frequent use for chronic conditions like depression vs episodic instances of suicidality or loneliness). 

These findings suggest that anonymous, 24/7 accessible, and real-time moderated DPS may be a scalable support tool for a variety of clinically relevant emotional states.

Read the full paper here.