Diabetes, anxiety and other diseases can be predicted by the language of your Facebook posts
New Delhi: Facebook is one of the most widely used social media networks all around the world. The virtual world of social media has made life easier by helping us connect with a lot of people at once, and by facilitating communication between our long lost connections. However, a recent study has found that Facebook may be capable of more than just that – it may also be able to predict mental and physical conditions like diabetes, anxiety, depression and psychosis in patients.
According to ScienceDaily, a study from Penn Medicine and Stony Brook University researchers have found that the language used in Facebook posts may be able to identify these disease in a person. Researchers found that certain words and the kind of language used could be an indicator of a disease, and with the patient’s consent, it can also be monitored like other regular and physical symptoms. The study was published in Plos One.
The Facebook post history of around a thousand patients who agreed to link their medical records with their profiles was analyzed, with the help of an automated data collection technique. Then, three models were created to analyze the Facebook post language, the demograph, and the third one which combined the two datasets. The study then looked into 21 different health conditions and found that all 21 were predictable from Facebook alone, and 10 of them were predicted better through the use of Facebook data instead of demographic information.
Use of different words was intuitively linked with some health issues and disorders. The use of the words “bottle” and “drink” were associated with alcohol abuse, and the people who mentioned religious language like “God” and “Pray” were 15 times more likely to have diabetes than those who used these terms the least. Use of other words like “dumb” and other expletives was recorded as indicators of drug abuse and psychoses.
Last year, the basis of this study showed that Facebook data analysis could predict depression three months earlier than a diagnosis in a clinic. This study built on the same finding to find out if analysing the data could help in detecting any other health issues and disorders.