Twitter can predict rates of coronary heart disease, study says

Chuck Bednar for – Your Universe Online

Twitter could serve as a dashboard indicator of a community’s overall psychological well-being, and can even predict rates of coronary heart disease, researchers from the University of Pennsylvania report in a new study.

Building on previous research that identified various risk factors that can contribute to cardiovascular disease (including low income, smoking, or psychological stress), graduate student Johannes Eichstaedt and his colleagues demonstrated that Twitter can actually provide more data about heart disease risk than a combination of traditional warning signs.

Furthermore, the researchers found that the microblogging website also characterizes the overall psychological atmosphere of a community. Expressions of negative emotions (anger, stress, and fatigue) in a county’s tweets were linked to an increased risk of heart disease, while positive ones (excitement and optimism) were typically associated with a reduced risk.

Measuring emotional state

While investigators have long believed that psychological well-being is important to a person’s physical health, it has proven difficult to measure. Eichstaedt, who studies in the Department of Psychology, and his associates believe that Twitter could be an effective epidemiological tool to measure and analyze the collective mental state of a community.

Writing in the journal Psychological Science, the study authors explained that “a cross-sectional regression model” constructed using Twitter posts did a better job of predicting atherosclerotic heart disease-related mortality than a model which combined 10 different common demographic, socioeconomic, and health-related risk factors, including diabetes, hypertension, and obesity.

Since there is no way to directly measure a person’s innermost emotions, the team turned to traditional psychological research practices that can discern this information from the words used by people as they write or speak. Previous studies had indicated that this type of analysis can be as effective as traditional questionnaires in assessing an individual’s personality.

“Getting this data through surveys is expensive and time consuming, but, more important, you’re limited by the questions included on the survey. You’ll never get the psychological richness that comes with the infinite variables of what language people choose to use,” Eichstaedt said.

Having previously observed the correlation between language and emotional state, he and his fellow investigators attempted to see if they could demonstrate the connections between those emotional states and the physical outcomes rooted in them. They considered coronary heart disease–the leading cause of death worldwide–as an ideal candidate for the experiment.

“Psychological states have long been thought to have an effect on coronary heart disease,” said co-author Margaret Kern, an assistant professor at the University of Melbourne in Australia. “For example, hostility and depression have been linked with heart disease at the individual level through biological effects. But negative emotions can also trigger behavioral and social responses; you are also more likely to drink, eat poorly and be isolated from other people which can indirectly lead to heart disease.”

Turning tweets into data

Using data collected by public health officials and a collection of public tweets posted in 2009 and 2010, the researchers used established emotional dictionaries and automatically-generated word clusters that reflected behaviors and attitudes to analyze a random sample of Twitter posts from people who had made their locations public.

In all, they had enough tweets and health data to analyze people living in nearly 1,300 counties that were home to 88 percent of the country’s population. They matched risk factors and cause of death information with digital epidemiological data obtained by Twitter, and found that negative emotional language and topics (including expletives and the word “hate”) were strongly linked to heart disease mortality, even after accounting for variables like education and income.

On the other hand, the opposite correlation was found with positive emotional language (such as the words “wonderful” or “friends”), which suggests that being optimistic and having positive experiences could help protect people from heart disease. The results match up with existing sociological research that suggests that combined characteristics of communities can actually be more effective at predicting physical health than data linked to any lone individual.

“The relationship between language and mortality is particularly surprising,” said H. Andrew Schwartz, a visiting assistant professor in the university’s School of Engineering and Applied Science’s Department of Computer and Information Science, “since the people tweeting angry words and topics are in general not the ones dying of heart disease. But that means if many of your neighbors are angry, you are more likely to die of heart disease.”

“We believe that we are picking up more long-term characteristics of communities,” added Lyle Ungar, a professor of computer and information science. “The language may represent the ‘drying out of the wood’ rather than the ‘spark’ that immediately leads to mortality. We can’t predict the number of heart attacks a county will have in a given timeframe, but the language may reveal places to intervene.”


Follow redOrbit on TwitterFacebookInstagram and Pinterest.