June 18 (UPI) -- The words people use on social media can help doctors determine their health status, a new study says.
Researchers have figured out a way to predict mental health and diabetes status based on the language of a Facebook post, according to research published Monday in PLOS ONE.
"This work is early, but our hope is that the insights gleaned from these posts could be used to better inform patients and providers about their health," the study's lead author Raina Merchant, the director of Penn Medicine's Center for Digital Health, said in a news release. "As social media posts are often about someone's lifestyle choices and experiences or how they're feeling, this information could provide additional information about disease management and exacerbation."
The study included an analysis of 1,000 patients who allowed the researchers to connect medical records with their Facebook posts. The researcher used models to analyze the language, age gender and other demographic factors of those posts.
Related
"Our digital language captures powerful aspects of our lives that are likely quite different from what is captured through traditional medical data," the study's senior author Andrew Schwartz, an assistant professor of computer science at Stony Brook University, said in a news release. "Many studies have now shown a link between language patterns and specific disease, such as language predictive of depression or language that gives insights into whether someone is living with cancer. However, by looking across many medical conditions, we get a view of how conditions relate to each other, which can enable new applications of AI for medicine."
The researchers found, for example, Facebook users who write posts containing the words "drink" and "bottle" were more likely to abuse alcohol. However, other connections weren't so direct.
Users who posted words such as "God" or "pray" were at 15 times higher risk of having diabetes than others who didn't use those words as much. Also, people who used words like "dumb" followed by curse words were more likely to abuse drugs or suffer psychoses.
In all, the researchers looked at 21 different conditions. and discovered all 21 were predictable just through analyzing Facebook.
"One challenge with this is that there is so much data and we, as providers, aren't trained to interpret it ourselves -- or make clinical decisions based on it," Merchant said. "To address this, we will explore how to condense and summarize social media data."
