Johannes Eichstaedt, PhD
Martin Seligman, PhD
Daniel Stokes, MSc
Lyle Ungar, PhD
H. Andrew Schwartz, PhD
National Institute on Minority Health and Health Disparities
Robert Wood Johnson Foundation
National Heart, Lung, and Blood Institute
National Institute on Drug Abuse
Templeton Religion Trust
Once Upon a Time Foundation
Edna Bennett Pierce Prevention Research Center, Pennsylvania State University
More than 3 billion people use social media worldwide. Whether posting a status about plans, sharing a photo, or logging a review, everyday users leave a trail of information in the form of digital footprints.
We've set out to explore what we can learn about a person's health by mining their digital footprints. Similar to how a genome – a collection of an organism's genetic information – can reveal information about a person's health status and risk factors for certain conditions or diseases, we believe that an individual's Social Mediome can reveal important insights about health and health care.
Our work in this area falls into two categories.
Population-level data mining: We conduct natural language processing on publicly available sets of population health and social media data to map correlations and identify trends.
Data mining paired with human subjects research: We explore the link between social media activity and health behaviors and outcomes at the individual level. All participants explicitly consent to share data from their social media accounts and electronic health records (EHRs).
We believe that harnessing data from social media in the ways described above can help us determine new ways to predict, measure, and intervene in multiple disease profiles such as diabetes, heart disease, and mental health in the future.
To date, more than 6,000 contributors have donated their data to our Social Mediome – sharing more than 60 million data points. Learn about some of our key findings below.
Many patients are willing to share and link their social media data with the data in their EHR. Learn more.
Facebook status updates can predict health conditions such as diabetes and depression better than demographic information. Learn more.
Language from Facebook posts can help predict an individual’s 10-year risk for atherosclerotic cardiovascular disease. Learn more.
Facebook users change their language before visiting emergency departments. Learn more.
Facebook language can predict depression in medical records. Learn more.
Analyzing narrative posts and comments on Reddit can help researchers learn about the experiences of patient subgroups, such as COVID-19 long haulers. Learn more.
This work is ongoing. Ultimately, our goal is to provide both patients and providers with a deeper understanding of how social media can be used as a diagnostic and descriptive tool.