Have the flu? Tweet it. Diarrhea? Review it. Help identify outbreaks.
When it comes to social media, oversharing is the new norm. When #flu season comes around, everyone on Twitter becomes inundated with talk of runny noses, strangers sneezing on their morning commute bus ride, and avoidance of the airport. People update their statuses on their interactions, frustrations, struggles and even locations. When a friend is not feeling well, they may post about it on social media and get an instant influx of well-wishes, commiserations and treatment recommendations in return. Some may search the Internet for their symptoms when they start feeling poorly. Some turn to Yelp or other online review website to share their personal health troubles like food poisoning after their visit.
As physicians, we may frown on that type of patient behavior, especially those patients who scour the Internet for a diagnosis. However, we shouldn’t. These kinds of tweets, status updates, reviews, and search queries may help predict and monitor diseases across the country.
Think of a traditional outbreak scenario. An individual comes into contact with an infected person, and subsequently, develops symptoms, and likely see a doctor. The doctor would then examine the patient, performing a history and physical, and eventually, obtain a blood sample for the laboratory to test and confirm the diagnosis. Once confirmed, sometimes taking days to weeks, the positive laboratory test is reported to local health authorities. The information is typically sent along through some sort of hierarchical structure. Local health authorities then report the results to the next appropriate health facility. The diffusion and transfer of information from the initial health visit to the central health authority for decision making and then to the public can take weeks to months.
The use of digital data can speed up this sometimes lengthy process. The analysis of digital data, such as Google search queries, has been used to monitor communicable and non-communicable diseases, as well as mental health, illegal drug use, health policy impact, and behaviors with potential health implications. In fact, the United States alone generates 8 million search queries for health-related information daily. As an easily approachable, highly cost-effective and interoperable system, social media provides real-time online data that can be systematically mined, aggregated and analyzed to inform public health agents and a real-time source of epidemic intelligence. Several studies have confirmed the potential of this epidemiology of online information or “infoveillance” to improve epidemic intelligence.
Case in point: HealthMap, a program that mines social media, local news reports, and government websites to search for evidence of disease outbreaks, identified a “mystery hemorrhagic fever” in Guinea, almost two weeks before the World Health Organization (WHO) announced the Ebola epidemic. Similarly, in 2003, another system called Global Public Health Intelligence Health Network (GPHIN), detected SARS more than two months before the first publications by the WHO.
Here is a closer look at how some are handling the new streams of Internet data:
When someone is ill, it is not rare for someone to turn to Google, WebMD, or Wikipedia for more information about their symptoms or diagnosis. Analyzing the frequency of search terms, by location, for example, can give potential signals for disease. Researchers at Los Alamos National Laboratory use the behavioral tendency to look online for answers as a foundation to predict outbreaks of influenza, Dengue fever and more. In addition, they claim that their algorithms that connect Wikipedia searches with information from the Centers for Disease Control and Prevention could lead to disease predictions in real-time.
Sickweather combines self-reported information with geolocated data from sites like Twitter to provide information on the spatial spread of disease, overlaid on a map. In 2011, McHenry County in Illinois was hit with an outbreak of whooping cough, starting with the high school cheerleaders, athletes and band members. Students, families, and citizens began talking about the illness on Facebook and Twitter. Sickweather LLC spotted the surge in comments online, identifying a potential outbreak, about two weeks before an official health statement was released on the whooping cough outbreak in that area.
A night on the town may prove to a night in the bathroom and then eventually on the business review website Yelp to share the experience. Last October, when a shigella outbreak at a San Jose, California, seafood restaurant sickened dozes, right alongside public health officials were Yelp reviewers.
“PLEASE DO NOT EAT HERE!!!!” Pauline A. wrote in her Oct. 18 review of the Mariscos San Juan #3 restaurant. “My sister-in-law … and brother-in-law along with his parents ate here Friday night, and all four of them ended up in the hospital with food poisoning!!!”
Later that day, Santa Clara County Public Health Department shut down that restaurant, and two days later, officials announced that over 80 patrons of the restaurant had become severely ill, with 12 requiring intensive care.
Some 10 percent of all reviews of restaurants on Yelp are related to food-borne illness. That can serve as a powerful tool for health departments, allowing them to plan inspections of restaurants that seem to have a great number of customers falling ill. This line of thought made the New York Department of Health and Mental Hygiene look into utilizing Yelp as a means of detection. Additionally, the public health agency of Chicago has an app called Foodborne Chicago which automatically sends information to the health agency when Twitter users complain of food-poisoning.
There have been around 50 digital surveillance systems, dubbed Event-Based Internet surveillance systems, created over the last 15 years. They use information on events impacting human health or the economy from Internet sources, instantaneously incorporating diverse streams of data. However, the use of this data isn’t without limitations, such as how to reduce online noise to deduce a signal, and the moral and ethical dilemmas of using such data and privacy implications. Yet, these systems are meant to supplement existing public health initiatives, not replace them.
So, the next time you have a patient that comes to your office with a Wikipedia found diagnosis already printed in hand, or a friend shares on Facebook their recent bout with the flu, think twice before rolling your eyes; they may actually be helping track an outbreak.
Originally published at https://www.kevinmd.com on September 25, 2016.