Could social networking be the next public health research tool? Quite possibly, according to information gathered in a study by two Johns Hopkins University computer scientists. The scientists, Mark Dredze and Michael J. Paul, sorted through 2 billion Twitter posts and uncovered intriguing patterns about allergies, cancer, obesity, insomnia, flu cases and many other ailments.

Twitter, often referred to as a mini-blog, enables people to “tweet” messages of 140 characters or fewer in real time. Tweets can be about anything, from restaurant recommendations to the latest news and health topics.

Dredze, a researcher at the university’s Human Language Technology Center of Excellence, and Paul, a doctoral student, used software to filter out the 1.5 million messages referring to health matters posted between May 2009 and October 2010.

“Most of the health patterns were not especially surprising by themselves because we were specifically looking for patterns that we were already aware of, such as flu and allergy seasons,” says Paul. “We wanted to confirm that the information extracted from Twitter indeed matches what we know to be true about the real world.”

While the scientists scoured posts for confirmation of known patterns, however, several discoveries surprised them. “For example, thousands of people mentioned that they were using antibiotics to treat influenza,” says Paul. “Antibiotics are ineffective against the flu, so it appears that there is some misinformation out there.” In addition, they found that users living in states with lower rates of healthcare access were more likely to tweet about being sick by a statistically significant margin.

In about 200,000 of the health-related tweets the researchers were able to track some trends by time and region. This enabled them to see differences in the parts of the country where allergy seasons began, for example. They found that allergy season starts earlier in the warmer states and later in the Midwest and the Northeast.

The researchers believe health professionals will eventually be able to use the data to monitor seasonal and regional trends regarding different illnesses such as the flu, similar to what is offered by Google Flu Trends, says Paul. “Twitter could be a great tool for health practitioners to learn about and communicate with their patients outside of the office, in a more personal way than the anonymous data mining that we have focused on in our study.”

CONNECT THE DOTS

For more information on health topics and trends in the U.S. check out the Centers for Disease Control and Prevention (CDC); for global trends go to The World Health Organization.

It’s an aggregator of Twitter content relating to various diseases and medical conditions. Created by Pixels & Pills, it uses Petri dishes—visually and metaphorically—to capture and convey dialogue on a number of ailments.

In short, each Petri dish represents an illness (e.g. sleeping disorders or diabetes), and each cell within a dish represents a tweet. Clicking on a Petri dish offers a look at what’s being said about that particular condition.

The most tweeted topics are color-coded so you can see what’s popular. When I was on the site the other day, “psychiatry” was leading, with 280 tweets. A particularly interesting tweet led to a study on colors and feelings, where doctors used color to gauge the moods of patients with verbal challenges.

Health Tweeder is a nice snapshot of real-time discussions in the world of health. It turns hashtags and key issues into groupings that are easy to navigate and peruse, making it the sort of experience that fosters new learning and encourages conversation. So…dig in.