Facebook posts may assist identify conditions similar to diabetes, nervousness, depression, and psychosis say scientists who recommend that language, phrases, and words on social media posts with patient consent could be monitored identical to physical signs.
Employing an automated data assortment technique, the researchers from the University of Pennsylvania and Stony Brook University in the US analyzed the entire Facebook post history of almost 1,000 sufferers who agreed to have their digital medical record information linked to their profiles.
The researchers then constructed three modules to analyze their predictive power for the sufferers: one module just analyzing the Facebook post language, another that used demographics like age and sex, and the last that mixed the two datasets.
Looking into 21 different circumstances, researchers discovered that all 21 were predictable from Facebook alone. Ten of the situations were higher predicted by the use of Facebook data instead of demographic data.
“This work is early, but we hope that the insights discovered from these posts could be used to better inform sufferers and providers regarding their health,” stated Raina Merchant, an associate professor at the University of Pennsylvania.
A few of the Facebook information that was discovered to be more predictive than demographic data appeared spontaneous.
For instance, “drink,” and “bottle” were proven to be more predictive of alcohol abuse. But others weren’t as easy.
For example, the people that often used religious words like “God” or “pray” in their posts were 15X more vulnerable to diabetes than those who used these terms the least.
Also, words expressing hatred — like “dumb” and other expletives — served as signs of drug abuse and psychoses.