We are living in a time when we have access to more data than ever before—and quite frequently, more information than we know what to do with. The internet and social media have changed the ways we conduct business, find dates, make friends, and express ourselves. These are obvious changes that for the most part we have accepted in our daily lives. However, what people may not immediately recognize is that the information superhighway has changed the ways we understand our health and seek medical treatment. With publicly available services such as WebMD’s Symptom Checker and Google Health, people can independently research their symptoms and come to a more educated self-diagnosis.
Determining whether or not this behavior change is good for health outcomes is not the purpose of this post, but I want to highlight that this behavior is occurring and is on the rise. As more people globally are connecting to the internet, and as online databases become more sophisticated and widespread, more people are turning to the world wide web for answers. What epidemiologists are beginning to realize is that these searches and online queries may have implications for disease outbreak surveillance. By monitoring internet searches and website traffic (which is being constantly tracked, don’t kid yourself), it may be possible to catch a disease outbreak faster, meaning quicker response and more effective intervention.
Typical epidemiological surveillance relies on disease registries. The process being: patient comes to doctor with presenting symptoms, disease is confirmed, case report written and then submitted to a local registry which then files with state and federal registries so that patterns can be monitored. This approach is effective and accurate, but it is slow and self-diagnosis is potentially making it slower. Because people are loath to go the doctor (it is often expensive and inconvenient), they will wait a few days after presenting symptoms to make sure it is an illness that merits a trip to the office.
Instead, many people search online during the first few days of symptoms to help determine the severity of their condition and for possible home remedies. Only if symptoms persist or the web services suggest a visit to the physician immediately do people make the trip. The new method of tracking diseases online, coined “Webidemiology,” has the potential to catch disease cases before they present to the doctor or the disease registry. It may seem like a far-stretch, but Google flu trends—estimating incidence of flu-like illness based on web searches–from 2004 to 2010 very closely matches the CDC’s registry. When I first saw the graph (available at the npr.com link below) I was pretty surprised. I didn’t expect that at it’s infancy, web surveillance could be so accurate. It is both hopeful and promising, but has a few drawbacks and limitations.
First of all, such tracking will only prove useful for a few diseases with uncommon symptoms. Because many illnesses begin with similar or overlapping indicators, it will be difficult to determine which disease is being presented. However, I believe that as the field of webidemiology progresses, computer programs will become increasingly capable at accurately identifying diseases with similar symptoms. Media coverage of a particular outbreak or potential outbreak may increase web searches and create muddled web surveillance, but these cases should be easily identified as such, and web surveillance of such incidences will still prove valuable to public health officials.
One of the major issues that I foresee with webidemiology is the perceived breach of privacy that occurs when outside sources view search trends. What many people don’t realize is that search histories and trends are publicly available. It is naïve to believe that there is anonymity on the internet, and people who require such levels of privacy should avoid the internet entirely. It is empowering to have access to such information. I understand the fear of having personal information tapped for government programs, but for public health epidemiology, I must say that I am excited for this new use of information mining. I look forward to watching as the methods become more sophisticated and reliable markers for disease outbreak.
Original article on NPR:
Statistics and estimates regarding online health information by the CDC:
Google’s Flu Trends: