Identifying and Understanding Communities Using Twitter to Connect About Depression: Cross-Sectional Study

Area of Science:

• Social Sciences
• Public Health
• Digital Communication

Background:

• Depression is a leading global health issue, often linked to social isolation.
• The rise of social media presents new avenues for individuals seeking support for depression.
• Twitter users are increasingly using the platform to find online connections related to depression.

Purpose of the Study:

• To identify online communities on Twitter using the #MyDepressionLooksLike hashtag.
• To understand community characteristics associated with users seeking web-based connections for depression.
• To analyze the correlation between community factors and online engagement with depression-related topics.

Main Methods:

• Collected 104 tweets from May 25 to June 1, 2016, in the northeastern US and Washington DC.
• Employed Poisson multilevel regression to predict tweet counts per county, adjusting for demographic and socioeconomic factors.
• Calculated tweeting index values (TIVs) to assess undertweeting and overtweeting, and examined correlations with community characteristics.

Main Results:

• Tweet counts correlated significantly with the proportion of females, single-person households, and the 15-44 age group.
• Undertweeting (TIV < 1) was observed in eastern, seaboard areas, while expected tweeting (TIV ≈ 1) occurred in western, inland areas.
• Overtweeting (TIV > 1) showed upward gradients with nonprofit presence, but U-shaped relationships with many community factors.

Conclusions:

• Web-based communities may complement or act as proxies for offline social connections, not replace them.
• A higher number of nonprofits correlated with increased online engagement for depression discussions.
• Further research with broader spatiotemporal scope is recommended to validate these findings.