Social media is constantly evolving, and many users post tips and information on a broad range of subjects, much of which is misleading and sometimes even dangerous.
As far as medical information is concerned, a number of doctors are trying to drown out health misinformation on social media by recruiting more medical professionals to join talks and discussions on line.
Facebook data collection
On a slightly different note, a new study has found that the language used on Facebook could help identify health issues in users, including diabetes, anxiety, depression and psychosis.
The study, published in PLOS ONE by Penn Medicine and Stony Brooke University, indicates that the language used in posts could reveal the presence of disease.
If a Facebook user consents, their posts can be monitored just like physical symptoms. In the study, an automated data collection technique was used in which researchers looked at the Facebook posts of nearly 1 000 patients who consented to have their electronic medical data linked to their profiles.
Three research models were constructed by the researchers to identify medical conditions:
- The first model analysed the language used in Facebook posts.
- The second one used demographics such as age and sex.
- The third combined both data sets.
A total of 21 different conditions were considered and researchers found that all 21 could be predicted based on Facebook posts alone. The research also found that 10 conditions were better predicted through data from Facebook posts than from demographics.
Lifestyle choices and experiences
Some of the data from Facebook posts seemed instinctual. The words "drink" and "bottle" were said to be predictive of alcohol abuse while people who used religious languages or words such as "God" or "pray" were 15 times more likely to have diabetes than those who rarely used those terms. "Malicious" words such as "dumb" or other unflattering terms served as an indicator for psychosis and drug abuse.
Lead author, Raina Merchant, MD, MS, director of Penn Medicine's Center for Digital Health and associate professor of Emergency Medicine told Science Daily that "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".
Different from traditional data
Dr Andrew Schwartz, the study's senior author, added, "Our digital language captures powerful aspects of our lives that are likely quite different from what is captured through traditional medical data." He went on to say, "Many studies have now shown a link between language patterns and specific diseases, 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."
Members of the research team were able to provide proof that analysis of Facebook posts could predict a diagnosis of depression almost three months earlier than a diagnosis by a clinic. This could prove to be useful in the future, as the option to have your Facebook posts analysed can help to refine patient care.
The study received funding from a Robert Wood Johnson Foundation Pioneer Award.
Image credit: iStock