Coping with a global pandemic has damaged the mental health of millions of people. A team of researchers from MIT and Harvard University have shown that they can measure those effects by analyzing the language people use to express their concerns online.
Using machine learning to analyze the text of more than 800,000 Reddit posts, the researchers were able to identify changes in the tone and content of the language people used during the first wave of the pandemic. Covid-19 progressed, from January to April 2020. Their analysis revealed a number of important changes in mental health conversations, including an overall increase in discussions about mental health. anxiety and suicide.
“We have found that there are these natural clusters emerging in connection with suicides and loneliness, and the number of posts in these clusters more than doubled during the pandemic compared to the same month of the year. Previously, it was a serious concern, ”says Daniel Low, a graduate student in the Technical Science and Speech and Listening Engineering program at Harvard and MIT and lead author of the study.
The analysis also revealed different effects for people who already had different types of mental illness. This finding could help psychiatrists, or potential moderators of Reddit forums that have been studied, identified, and helped people with mental health problems, researchers say. enjoy.
“When the mental health needs of so many people in our society are not being fully met, even at a basic level, we want to pay attention to the ways many people are suffering from time to time. This, to amplify and inform the allocation of resources to support them. ”Said Laurie Rumker, a PhD student in the PhD Program in Biomorphology and Integrated Genomics at Harvard and one of the study̵7;s authors.
Satrajit Ghosh, a chief research scientist at the McGovern Brain Research Institute of MIT, who is the lead author of the study, appeared in Journal of Internet Medicine Research. Other authors of the paper include Tanya Talkar, a graduate student in the Program in Technical Science and in Listening and Speaking Biology at Harvard and MIT; John Torous, director of digital psychiatry at Beth Israel Deaconess Medical Center; and Guillermo Cecchi, a Principal Research Officer at IBM’s Thomas J. Watson Research Center.
A wave of anxiety
New research develops from MIT class 6.897 / HST.956 (Machine Learning for Healthcare), in MIT’s Department of Electrical Engineering and Computer Science. Low, Rumker and Talkar, all of whom took the course last spring, have done some previous research on using machine learning to detect mental health disorders based on how people speak. and what they say. After the Covid-19 pandemic began, they decided to focus their class project on analyzing Reddit forums for different types of mental illness.
“When Covid was successful, we were all curious whether it would affect certain communities more than others,” said Low. “Reddit gave us the opportunity to consider all of these subreddits as specialized support groups. It’s a truly unique opportunity to see how differently these different communities are affected as the wave unfolds, in real time. “
The researchers analyzed posts from 15 subreddit groups covering a wide variety of mental illnesses, including schizophrenia, depression, and bipolar disorder. They also include several groups for topics not specifically related to mental health, such as personal finance, fitness, and parenting.
Using some sort of natural language processing algorithm, the researchers measured the frequency of words related to topics such as anxiety, death, isolation and substance abuse, and grouped Post together based on similarity in the language used. These approaches allowed researchers to identify similarities between the posts of each group after a pandemic outbreak, as well as distinctive differences between groups.
The researchers found that while people in most advocacy groups started posting about Covid-19 in March, the group for health anxiety started much earlier, in January. As the pandemic progressed, however, other mental health groups began to closely resemble the health anxiety group, in terms of the most commonly used language. At the same time, the personal finance group showed the most negative semantic shift from January to April 2020, and significantly increased the use of words related to economic and psychological stress. negative.
They also found that the mental health groups most negatively affected during the early pandemic were those associated with ADHD and eating disorders. Researchers hypothesize that without their regular social support system, due to being locked out, people with those disorders find it much more difficult to manage their condition. In those groups, researchers found posts about being overly focused on news and reverting anorexic-like behaviors because meals were not followed by others due to isolation.
Using a different algorithm, the researchers grouped the posts into clusters like loneliness or substance use, and then tracked how those groups changed as the pandemic progressed. Suicide-related posts more than doubled from pre-pandemic levels, and groups significantly related to the pandemic suicide group are support groups for boundary personality disorder and stress disorder. after injury.
Researchers also found the advent of new topics especially looking for mental health help or social interaction. “The topics in these subreddit support groups have changed a little bit, as people are trying to adapt to their new life and focus on how they can seek more help if needed,” says Talkar.
While the authors emphasized that they cannot see the pandemic as the sole cause of the language changes observed, they note that there were more significant changes between January and April. 2020 versus the same month in 2019 and 2018, showing inexplicable changes in the usual annual trends.
Source of mental health
This type of analysis could help mental health providers identify segments of the population most susceptible to mental health deterioration due to not only the Covid-19 pandemic but also stress factors. Other mental health straighteners like controversial elections or natural disasters, researchers say.
Additionally, if applied to Reddit or other social media posts in real time, this analysis can be used to provide users with additional resources, such as instructions for a Other support groups, information on how to seek mental health treatment or suicide hotline numbers.
“Reddit is a very valuable resource for a lot of people facing mental health challenges, many of whom may not have formal access to other forms of health support. Mental health is different, so this work makes sense for the ways in which Reddit support can be provided, ”says Rumker.
Researchers now plan to use this approach to study whether posts on Reddit and other social media sites can be used to detect mental health disorders. One current project involves screening social media posts for veterans about suicidal risks and post-traumatic stress disorder.
The study was funded by the National Institutes of Health and the McGovern Institute.