Estimated to affect around one in five adults in the United States, with young people more likely to be affected, the problem of loneliness is a public health issue that urgently needs addressing.
Although many argue that technology, particularly social media, is making people more lonely, researchers from the University of Pennsylvania Medical School believe that Twitter could be a useful tool for predicting loneliness.
By applying linguistic analytic models to tweets, researchers were able to gain an insight into the topics and themes that could be associated with loneliness.
As part of the study, published in the journal BMJ, researchers analysed public accounts from users based in Pennsylvania and found that 6,202 accounts used words such as “lonely” or “alone” more than five times between 2012 and 2016.
According to the researchers, users who posted about loneliness had a higher association with anger, depression, and anxiety, when compared to the “non-lonely” group. The lonely group was also more likely to tweet about relationship problems, substance abuse, use expletives and use phrases such as “want somebody”, “no one to” or “I can’t”.
Lonely users were also more likely to tweet at night and tweeted nearly twice as often as users not identified as lonely. “Non-lonely” users were also more likely to engage directly with other Twitter users.
Even if users do not explicitly tweet about their feelings, by analysing themes and linguistic markers, the researchers believe they may be able to construct a “loneliness prediction system”, which could make it easier to identify those who are experiencing loneliness and go about providing support.
“This could be very powerful and have long-lasting effects on public health”
“Loneliness can be a slow killer, as some of the medical problems associated with it can take decades to manifest,” said the lead author Sharath Chandra Guntuku, PhD, a research scientist at the University of Pennsylvania Medical School’s Center for Digital Health.
“If we are able to identify lonely individuals and intervene before the health conditions associated with the themes we found begin to unfold, we have a chance to help those much earlier in their lives. This could be very powerful and have long-lasting effects on public health.”
Although having some insight into users’ mental health may help identify those who may need support, the study’s senior author Raina Merchant, MD, the director of the Center for Digital Health, explains that loneliness is not always expressed or solved in the same way:
“It’s clear that there isn’t a one-size-fits-all model. Some interventions include buddy systems, peer-to-peer networks, therapy, and skill development for navigating day-to-day interactions with others.”
It is also important that the technology is used wisely, with tracking users’ mental state without their knowledge having implications for privacy, and potential for misuse if the information gets into the wrong hands.
In the future, the researchers hope to develop a better measure of the different aspects of loneliness that online users expressing. A predictive model they developed as a result of this study has demonstrated accuracy in predicting loneliness in a group that opted-in to share their Twitter data and took a validated loneliness survey.
This could be used to develop tools to identify lonely patients receiving care in hospitals and help them accordingly.