Using Social Data to Build Our Evidence Base

Jul 16, 2015, 2:22 PM, Posted by Alonzo L. Plough, Lori Melichar

Social media offers an exciting opportunity to innovate in health research—but the social data sandbox could use more players to conduct research, share datasets, and generate ideas about what we should be studying.

What do our tweets reveal about our health? What can we learn from Twitter about the health of those in our community? Can analysis of Twitter activity help predict an epidemic like the flu weeks before a community is inundated with cases?

Nicholas Christakis, director of the Human Nature Lab at Yale University and Thomas Keegan, Deputy Director of the Yale Institute for Network Science are conducting pilot studies in San Francisco and Boston to explore these questions and more. With funding from RWJF, Christakis’s lab uses Twitter posts that include mention of flu symptoms to map how the virus spreads outward from individuals to family, friends, and others in their social networks. This mapping method, which identifies central “influential” individuals, offers the possibility of early detection of the flu and therefore early intervention to prevent its spread. In addition to giving health officials and medical personnel a valuable head-start in responding to and preventing the spread of contagious illness, this kind of insight could also help people make decisions about their own behavior, including getting flu vaccines and being more diligent about hand washing.

Christakis’s and Keegen’s work on influenza is a great example of how investigators are increasingly wielding research tools and methodologies that can extract valuable health-related insights from the expanding sea of digital information. Other innovators we have supported and/or admired such as James Fowler, Dev Roy, Daniel Zoughbie, Tom Valente, John Brownstein, Raina Merchant, Damon Centola, are also doing cutting-edge work to advance our ability to derive important health insights from social media data. Others experimenting in this realm include those who consider themselves part of the “quantified self” movement by seeking to derive insights from their own personal data, as well as those applying predictive analytics to datasets that contain millions of records. Because they are frontrunners in an emerging field, these pioneers have had to create datasets and develop novel tools and methods to generate results. In some cases, that has meant adopting the high tech tools from fields outside of clinical or behavioral health science such as agile software development and machine learning.

Our focus on the analyses of social networks and digital information has been evolving for several years now. We have supported research on how offline social networks contribute to obesity and other health outcomes, and explored how personal health data can be used in the clinic, and for understanding public health issues. These methods have shown promise; now we are interested in applying them directly to the challenges and opportunities of building a Culture of Health—including making health a shared value and engaging with a broad array of partners.

Here are some current examples of some cutting-edge work we have already funded in this emerging area of research:

  • The Health Data Exploration (HDE) network seeks to bring together companies making wearable devices and smartphone apps that collect and store personal health data (Think FitBit, Jawbone and RunKeeper) with researchers interested in mining their data for patterns and trends. We expect that personal health data can reveal the ways that everyday activities promote health or lead to disease, and can yield insights about the long-term, cumulative health effects of environment and lifestyle. The HDE Network is undertaking multiple projects that seek to leverage this anonymous and aggregated personal health data in ways that will ultimately transform our understanding of individual and population health.
  • The Atlas of Caregiving Pilot is a yearlong research project exploring the individual challenges and experiences of family caregivers. The project combines traditional research methods, such as surveys and caregiver interviews, while also employing wearable technologies to capture and track detailed information in the home. The data collected from the wearable devices will enrich our understanding of the daily routines and experiences of family caregivers, helping us understand what causes the most stress for caregivers, which caregiving duties come naturally to some but not to others, and what kind of supports would be most useful. With a better understanding of family caregiving in the United States we can develop technologies, services, and policies that improve this experience for all involved.
  • Genetic Alliance’s Platform for Engaging Everyone Responsibly (PEER) is a digital tool that enables individuals to share their health information with researchers and each other on their own terms, advancing the understanding of health and disease. Using PEER, people are able to set their own data sharing, privacy, and access preferences. With support from RWJF, Genetic Alliance is creating a so-called “white label” version of PEER that will allow disease advocacy and community organizations to customize and use the platform with their own members, developing condition-specific disease registries and surveys that can help accelerate research.

Leaders are emerging in this field, but the social data sandbox could use more players, including those interested in doing the research, those with interesting datasets to share, and those with ideas about what we should be studying. Some avenues for exploration include such questions as: How can social media help us identify other kinds of contagion? How might we uncover novel health fixes, including strategies people use to sleep more? Eat better? Get more exercise? How are attitudes about healthy communities changing? How can we detect progress in social determinants of health such as poverty, housing, and transportation?

In the end, we recognize that tackling some of America’s most complex health and policy issues will require investing in the most advanced research tools and evaluation methods—including those that are emerging and unfamiliar. We urge researchers, individuals, and companies to consider joining forces with the Health Data Exploration network to share emerging, innovative ideas for using social media and other digital information to build an evidence base for our Culture of Health strategy. Visit our website to hear about programs and new funding opportunities like our Evidence for Action program and stay connected with us on twitter. It is through knowledge sharing and partnerships that we can accelerate the discovery of powerful solutions to the nation’s biggest health challenges.

Are you using novel approaches to measuring health determinants and outcomes? Apply for funding through our Evidence for Action program today >>