Medical researchers are turning to social media to gather insights into patient experiences, The Wall Street Journal reported April 10.
Data drawn from social media posts have the ability to gather insights into patients' experiences that are often overlooked or difficult to attain when relying mainly on data from medical reports and physicians' charts.
Medical researchers are utilizing machine-learning algorithms to sift through social media posts to obtain information about how clinicians can improve care.
According to the report, social media data is being used to examine the effects of buprenorphine, breast cancer care, adverse side effects of HPV vaccines and suicide rates.
Social media data is also being combined with EHRs to improve care.
In an Oct. 7, 2019, paper published in Nature, medical researchers trained algorithms to sort through 50,000 Facebook posts from about 50 young adult and adolescent psychosis patients who suffered relapses and re-hospitalizations.
The algorithms detected distinct changes in the months before a relapse, including changes in language patterns, changes in the number of posts between midnight and 5 a.m. and the frequency of tagging or friending.
This allowed the researchers to predict psychosis relapses and hospitalizations about 71 percent of the time using Facebook posts.