What 32 Million Tweets Tell Us About Health and the Twitterverse


Jan 26, 2017

Group of women exercising and using their cell phones

Photo by sanjeri/Getty Images

This commentary originally appeared on The Health Care Blog on January 25, 2017.

How can we gauge whether America is prioritizing health and well-being? Since public attitudes toward health-related topics are widely shared on social media, we gazed into the mirror that is Twitter and tried to answer that question by sifting through 32 million health-related tweets, one of the largest social media samples ever collected for health research.

Posts and conversations on Twitter have the potential to shed light on the public's views about a seemingly endless array of health-related topics—obesity, exercise and fitness, safe sex, alcohol use, medication adherence and mental health. Accordingly, researchers have turned to social media to better understand these topics.

The set of health-related tweets was built from those originating in the U.S. over the course of one year, 2014. A tweet made its way into the sample if it contained a predetermined health-related keyword such as “well-being,” “fitness” or “medical.” Discussions about acute care and illness slightly edged out those about well-being. Tweets about well-being often seemed like an electronic bulletin board for good behaviors, such as eating right or exercising. Those about illness often mentioned health care services and medical occupations.

RAND conducted this research in collaboration with the Robert Wood Johnson Foundation, which is promoting and developing a “Culture of Health,” a framework (PDF) to encourage public engagement in well-being, and partnerships to support innovative health-related initiatives. The aim is to build healthier, more-equitable communities.

In a Culture of Health, public discourse would emphasize health-promotion activities (such as exercise) and well-being (for example, stress management), rather than illness and treatment. To assess public attitudes about these subjects—as reflected in Twitter conversations—we developed a measure to analyze discussion of these topics.

The more often people discussed well-being, the more important we assumed it was to them. For example, these two tweets illustrate that engaging in health-promoting activities seems to improve well-being:

Just completed a 5.16 mi walk — Enjoying this beautiful evening weather nice fitness walk! :)
im so proud of myself my doctor said ive been losing weight and that im all healthy and im so happy :)))))

Large-scale events can also shift online conversation. Some of the notable spikes and dips in the ratio of health discussion during this period appear to have been driven by recurring events (such as a New Year's Day spike in well-being tweets) or major health events (like a spike in acute care tweets during the Ebola outbreak in the United States).

In addition, a fair number of health-related tweets were commercial (such as spam or advertising) and did not reflect the personal perspective of an individual tweeter. However, these commercial tweets still may shed light on other aspects of the social media environment (such as exposure to unhealthy advertising) that may influence health behaviors.

Mining the 2014 tweets for health-related posts and discussions has given us a baseline to understand the evolution of online health discourse going forward. Over time, this baseline information will help test the hypothesis that wellness-related discussion will increase relative to that of acute care, especially in communities or populations where Culture of Health-style interventions or programs are put in place.

Why do we expect this to happen? As well-being and healthy living are given increased emphasis in society, the public should increasingly prioritize wellness and health promotion over illness and treatment.

Social media is still in relative infancy, so developing ways to better understand what people choose to say and share with others online is likely to only grow in importance. Additional analysis of online discourse could seek further insight about community or civic engagement on health-related matters or other topics related to well-being. There also may be opportunities to test and extend well-established findings that people with healthier in-person social networks tend to be healthier themselves.

The ability to find links between social discourse and better health raises the possibility that interventions focusing on online social networks could be an avenue for improving health attitudes and behavior in the future. As a pervasive outlet for personal expression and networking, social media should continue to gain power as a tool for understanding and influencing health as a shared priority.

Douglas Yeung is a behavioral scientist and Linnea Warren May is a project associate at the nonprofit, nonpartisan RAND Corporation. Matthew Trujillo is a program officer at the Robert Wood Johnson Foundation.