A new study is exploring whether it is possible to use data from social media to spot the early stages of a severe psychiatric mental illness.
The Feinstein Institutes for Medical Research and IBM have come together to develop AI that aims to predict a patient’s psychiatric diagnosis more than a year before they are diagnosed officially.
The algorithm examines images and language posted to platforms such as Facebook through private messages and posts to assess whether they are likely to have schizophrenia or bipolar disorder.
“There was an interest in harnessing these ubiquitous, widely used platforms in understanding how we could improve the work that we do,” stated Michael Birnbaum, assistant professor at Feinstein Institutes’ Institute of Behavioural Science. “We wanted to know what we can learn from the digital universe and all of the data that’s being created and uploaded by the young folks that we treat. That’s what motivated our interest.”
One participant in the study was Christian Herrera Gaton, a former student at Jay John College of Criminal Justice who was diagnosed with bipolar disorder after going missing for three days in 2016. In August 2020, he was admitted to Zucker Hillside Hospital after experiencing a bout of the illness alongside stress regarding the Covid pandemic. While undergoing treatment, he was approached by researchers who asked him to join the study.
Gaton, who offered almost 10 years’ worth of Facebook and Instagram data to aid research, said, “It’s been a difficult experience to deal with [Covid] and to go through everything with the hospitals and losing friends because of doing stupid things during manic episodes. It’s not easy, but at least I get to join this research study and help other people.”
The study feeds the data into a machine that learns to predict which group an individual belongs to. It makes these predictions using previous information it has been taught and via patterns it has learnt to recognise. The machine learning algorithms were developed by IBM, and Feinstein Institutes’ researchers managed data collection and participation recruitment and assessment.
Researchers took care to anonymise the data provided, and stripped names and addresses from written posts. Furthermore, they offered assurance that images were not analysed closely, and the software instead based its assessment on “shape, size, height, contrast and colours.”
Over 3.4 million messages and 142,390 images from 223 participants were analysed during the study. Researchers assessed posts from the 18 months prior to the individual’s first psychiatric hospitalisation and found that those with schizophrenia and mood disorders were “more prone to discuss anger, swearing, sex and negative emotions” in their Facebook posts.
Birnbaum believes there is an opportunity to use data from social media to deliver better healthcare. Gaton has stated that he believes he could have “avoided time in the hospital” if he had received an earlier diagnosis.
“Harnessing social media platforms could be a significant step forward for psychiatry, which is limited by its reliance on mostly retrospective, self-reported data,” the study stated.
