Researchers have found a promising new method for identifying individuals with depression by analyzing their language use on Facebook. In a study of 683 patients visiting an emergency department, the language preceding their first diagnosis of depression was used to identify depressed patients with fair accuracy. Emotional, interpersonal, and cognitive processes such as sadness, loneliness, hostility, preoccupation with the self, and rumination were found to be predictors of depression. Restricting Facebook data to only the six months before the first documented diagnosis of depression yielded higher prediction accuracy, as far out as three months before diagnosis. This unobtrusive method of depression assessment may become a feasible complement to existing screening and monitoring procedures.
The study's findings suggest that analyzing language on social media may be an effective and scalable way to extend the scope of current depression screening methods. With depression being the most prevalent mental illness and under-diagnosed and under-treated, the use of social media language analysis could provide an important new tool in identifying those who need help. By focusing on the emotional, interpersonal, and cognitive processes that are predictors of depression, this approach has the potential to improve the accuracy of depression diagnosis and treatment.
Read the research: Facebook language predicts depression in medical records
