329488
“Guess who just had a #PanicAttack”: Using Twitter to inform an early warning system for mental illness
Objective: This presentation explains current efforts to use Twitter as part of a larger project to develop and implement an early warning system for mental health and substance use issues. The questions we ask are (1) What are trends in the content of Tweets that mention specific mental illness terminology? (2) What proportion of Tweets express positive versus negative sentiments? (3) How viable is data mining of Twitter as part of an early warning system?
Methods: We imported 1000 Tweets including 33 keywords related to mental illness on a weekly basis for eight weeks using open-source NodeXL, prepared and cleaned the data for analysis, and interpreted data by coding for general trends and determining frequency of words/word pairs and positive/negative sentiments.
Results: The presentation will share the findings of the research, including which keywords are used by Twitter users to express personal symptoms of mental illness, whether these symptoms are posted in real time, as well as other notable contexts in which keywords are used such as stigma, promoting awareness, and resource sharing. Postings of symptom experience were found for a majority of the keywords.
Discussion: While Twitter is easily and quickly accessible, stigma may limit individuals’ expression of symptoms and generational differences in technology use may impact the utility of this method’s contribution to an early warning system across the entire population. Regardless, Twitter can be a useful tool for real-time tracking of mental illness experience on a national scale. We will discuss other recommendations for and limitations of the use of Twitter as a component of early warning systems for behavioral health.
Learning Areas:
Communication and informaticsEpidemiology
Public health or related research
Social and behavioral sciences
Learning Objectives:
Describe how to mine Twitter for mental health-related topics with open source software.
Explain how Twitter users describe mental illness symptoms.
Evaluate whether data mining of social media such as Twitter is useful for an early warning system for behavioral health.
Keyword(s): Mental Health, Surveillance
Qualified on the content I am responsible for because: After earning my PhD in Sociology (specializing in health and mental health), I became the postdoctoral fellow for the CAPE Project, a national study of community behavioral health. The current phase of the project, for which I am a co-PI, aims to create a replicable, sustainable, low-cost early warning system for behavioral health (mental health and substance abuse).
Any relevant financial relationships? Yes
| Name of Organization | Clinical/Research Area | Type of relationship |
|---|---|---|
| Substance Abuse and Mental Health Services Administration | Mental Health | Postdoctoral fellow subcontracted through Michigan State University |
| United States Dept. of Agriculture | Mental Health | Postdoctoral fellow subcontracted through Michigan State University |
I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.