Can big data fill gaps in epidemic awareness, responses? Researchers say yes, with caveats

By on .

With social media, cell phone logs, online news reports and electronic medical records, researchers say, “big data” is getting bigger, and can go where traditional infectious disease surveillance can’t . . .

The more that was learned about how and where Ebola was spreading during the 2014-2015 crisis in West Africa, the more quickly it was contained, with infections stemming from exposure at funerals and in hospitals peaking early on, and declining as the patterns became known, an article in a Journal of Infectious Diseases supplement notes this week.

But with the traditional means of gaining that knowledge far slower than the rate at which the disease was spreading, the cost was high. So, in a decade that has seen outbreaks of Zika, Middle East Respiratory Syndrome, SARS, pandemic flu, and the spread of antimicrobial resistance outpace responses, epidemiologists are increasingly taking a lesson from meteorologists and marketers and widening their scope of sources. The result, the authors of the lead article in the series note, is “the dawning of the big-data era in infectious diseases.”

Even the definition of big data is growing, they write, highlighting how health insurance records of U.S. patients seeking care for flu-like symptoms aligned with reports of the illness tracked by the Centers for Disease Control and Prevention, how cell phone logs can link travel and disease transmission, and how internet search term surges can indicate an emerging public health issue faster than physicians’ reports.

In West Africa, the article Elucidating Transmission Patterns from Internet Reports: Ebola and Middle East Respiratory Syndrome as Case Studies points out, while information from detailed contact tracing efforts during the Ebola outbreak remains scarce more than two years later, online news and informational reports provided information that has been validated since. Future use of online reports could be improved with more technologically sophisticated tools to “scour” internet and social media platforms. The authors, led by Gerardo Chowell of Georgia State University’s School of Public Health, note, news reports and other online sources come with their own limitations, including greater coverage of sensational and large scale events and anecdotes. Still, they write, the picture pieced together from internet-based information, was “well in line” with the statistics that followed.

3 thoughts on “Can big data fill gaps in epidemic awareness, responses? Researchers say yes, with caveats

  1. Muguwa Joseph

    Big/Small data, fast/slow in dissemination, large/ small coverage, the cardinal issue should be the optimal utilisation and data content. It should be in a manner that it is easily understood in the way it is sent, disseminated and received by the beneficiaries, how many can easily access it. Do the intended receipients actually benefit and more importantly learn from such data e,g the Ebola?, God forbid, but I bet if it struck again and in the same region, the response might not be any different from when it first struck!. We have also known from the recent US Presidential elections how data or info. was being manipulated for selfish interests. So, the issue is being non-complacent and in a sustained manner; not the mod, manner, amount or ease in dissemination. Concluding, I entirely agree with researchers when they say “yes, with caveats”

  2. John Spencer

    The real potential isn’t so much about “big data” it’s about the methods and techniques to obtain and analyze data and then communicate the findings. In other words, it’s not about the size of the data but how one uses it. “Big data” can add valuable context that was previously missing, but how does one do the hard work of making use of it? When you’re dealing with large, complex data that exists at multiple scales and across different periods of time, traditional analysis techniques are often inadequate. The real game changer for epidemic response and global health in general is data science. Data science is an approach that provides tools and methods that can make use of big data and effectively communicate findings. It’s how companies like Facebook or Google handle big data, it’s time for global health to start taking advantage of it too.

  3. Jim Thomas

    I agree that emerging types of big data should be explored for their use in detecting and responding to trends, including epidemics. The emergence is from sources like cell phones, search engines, social apps, and more. But before blindly using their data, we will need research to understand their biases and other shortcomings when applied to public health.

    There is another type of big data that is not emerging, but is well-established: routine health information systems (RHIS). These systems accumulate data from virtually all clinic visits and all health-related occurrences. These routine systems are invaluable for understanding the true distribution of needs and resources in a population. Exploration of the emerging big data sources should be regarded as a complement to RHIS, not a replacement. This is not a situation like cell phone technology, which famously leap frogged land lines in less developed countries. It is more like microwave ovens, that gained a spot next to the regular ovens in our kitchens. We must continue to invest in the big data system we have known and needed for a long time.

    Jim Thomas, MPH, PhD
    Associate Professor of Epidemiology
    Director of the MEASURE Evaluation project
    University of North Carolina


Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.