How Machine Learning Uses Social Media to Track the Pandemic’s Impact on Mental Health

Recent studies in machine learning have revealed how the conversation levels about anxiety and depression have increased since the pandemic and how these technologies can help people cope with mental health stresses.

By Tauren Dyson
December 22, 2020

Mental illnesses, such as depression, are highly prevalent and have been shown to impact an individual’s physical health. Recently, artificial intelligence (AI) methods have been introduced to assist mental health providers, including psychiatrists and psychologists, for decision-making based on patients’ historical data (e.g., medical records, behavioral data, social media usage, etc.).

As coronavirus cases continue to surge around the globe, so does the fear of billions of people gripped by the uncertainty of the pandemic. In the past year, researchers have been able to measure mental stress brought on by COVID-19 by analyzing the negative language on the popular social media platform Reddit.

As told by a team of researchers from Harvard and Massachusetts Institute of Technology (MIT), they were able to develop natural language processing (NLP) that tracks language on social media that is highly connected to suicide rates and other negative mental health outcomes. This study was conducted by using machine learning to examine over 800,000 Reddit posts between January and April 2020 with a method that tracked shifts of mood in user language on the platform as COVID-19 cases began to rapidly spread. What was found were several key changes in conversations about mental health, including an overall increase in discussion about anxiety and suicide that was becoming the common norm on social media.

“When the mental health needs of so many in our society are inadequately met, even at baseline, we wanted to bring attention to the ways that many people are suffering during this time, in order to amplify and inform the allocation of resources to support them,” said MIT researcher and co-author Laurie Rumker, when revealing the study results after they were published.

Researchers used multiple natural language processing algorithms to track how frequently words were used in conversations about isolation, substance abuse, anxiety, and death. Out of the 15 subreddit groups studied, many addressed issues such as bipolar disorder, depression, and schizophrenia. Other subgroups focused on fitness, parenting, and personal finance.

This helped the researchers to pick out similar posts that were written once the pandemic began. It also enabled them to sort through the differences of each group.

Throughout that time, the data showed a spike in conversations surrounding suicide and depression.

“We found that there were these natural clusters that emerged related to suicidality and loneliness, and the amount of posts in these clusters more than doubled during the pandemic as compared to the same months of the preceding year, which is a grave concern,” said study lead author Daniel Low, a graduate student in the Program in Speech and Hearing Bioscience and Technology at Harvard and MIT.

The rise of panic levels among social media users

The first rumblings from the Reddit subgroups about the pandemic began in January 2020, when content related to the virus started to appear in subgroups that discussed health anxiety. Two months later, the pandemic began to blanket the globe and the other mental health support subgroups on the platform started to post content related to the virus.

Pandemic-related stress also began to climb in the personal finance subgroups, where researchers observed a spike in the language related to economic strife and overall pessimism.

“When COVID hit, we were all curious whether it was affecting certain communities more than others,” Low said. “Reddit gives us the opportunity to look at all these subreddits that are specialized support groups. It’s a really unique opportunity to see how these different communities were affected differently as the wave was happening, in real time.”

Other affected subgroups by COVID-19’s wave of anxiety

The mental health groups that addressed ADHD and eating disorders expressed the most grief during the onset of the pandemic. Researchers figured that coronavirus lockdowns around the world had kept people in those subgroups from accessing the in-person social support systems that helped to manage extreme behavior.

Within the eating disorder subgroup, the researchers saw posts that fixated on the gloomy news reports surrounding the pandemic. They figured that negative news raised anxiety in those members, causing them to slip back into anorexia behavior.

“The topics within these subreddit support groups were shifting a bit, as people were trying to adapt to a new life and focus on how they can go about getting more help if needed,” said study co-author Tanya Talkar, a Speech and Hearing Bioscience and Technology graduate student at Harvard and MIT.

Another algorithm allowed the researchers to organize posts into different bunches related to loneliness and substance use disorder. The researchers monitored the changes in those groups as the pandemic plowed ahead.

The algorithm showed that between January and April 2020, posts about suicide spiked by more than double compared to before the pandemic.

The researchers found that the support subgroups for post-traumatic stress disorder and borderline personality disorder had the strongest connections to suicidal expressions.

Machine learning uses for improving mental health assistance

These findings can help mental health professionals to locate and treat people suffering from anxiety, depression, and suicidal thoughts during the pandemic, according to the researchers.

They also said the data can help to identify people coping with negative mental health outcomes during other stress-inducing events, such as other natural disasters and other catastrophic events.

“Reddit is a very valuable source of support for a lot of people who are suffering from mental health challenges, many of whom may not have formal access to other kinds of mental health support, so there are implications of this work for ways that support within Reddit could be provided,” Rumker said.

The researchers are currently using the same method for a project that examines the language posted on a social media site for veterans at risk for post-traumatic stress disorders and suicide.

If used in real-time, the researchers said the analysis shows that social media platforms, like Reddit, Facebook, and Twitter, can play a big role in supporting users with mental health challenges. That could include providing contact information to suicide prevention resources or directing people to support groups that are better equipped to treat their problems.