For the past year, I’ve had the pleasure of serving on the communications team of the National Collaborative for Bio-Preparedness (NCB-P), of which SAS is a partner and the analytics provider. That experience was heavy on my mind as I recently watched the film Contagion.
I may drop some minor spoilers here but without giving much away, the disease in the movie is extremely aggressive, with high reproductive and mortality rates. As the disease spreads and societal structure breaks down, the film shows what can happen in the face of a crisis that is poorly understood and overwhelming for government and public health responders, despite their best efforts.
The film attempts to frighten the viewer, and in my case, mission accomplished. However, I found myself thinking about all I’ve learned since joining NCB-P. Several times, in the midst of the panic and horror, I actually felt excited. NCB-P has developed innovative analytic techniques that can improve early detection, putting public health and government officials ahead of the curve.
It will take a lot of work to create true national biosurveillance. The movie uses the phrase “50 states, 50 health departments” to encapsulate the diverse data challenge. I say it’s actually more complex than that. Proper biosurveillance involves data sources beyond health departments. For instance, EMTs often make first contact with a sick person and will enter notes about the patient’s condition. The contents of the preceding 911 call are also a valuable data source. Both of these generate unstructured data that can be analyzed with text analytics to identify spikes in the amount of “chatter” about certain symptoms. That can be overlaid with geography. A sudden rise in the number of influenza-like illness complaints in EMT notes and 911 calls in a single zip code, for instance, can indicate the beginnings of an outbreak.
Using SAS Analytics technology and experts, NCB-P has demonstrated the ability to perform such analyses, by analyzing actual data from an outbreak. (Not like the one in Contagion, thankfully.) By adding additional data sources such as poison control, agricultural including livestock and crops, food supply chain, weather and climate, veterinary and others, we can increase situational awareness and detect a syndrome even earlier, no matter where in the biosphere it first shows up.
The film is also a study of modern society. One of the main characters is a blogger and conspiracy theorist that helps fuel the hysteria. Social and traditional media become a major influence in how the country responds. However, what the movie did not delve into was how the outbreak manifested in social media channels like Twitter and Facebook. The outbreak was gearing up over a holiday weekend and health officials correctly anticipated a rise in cases once people returned from trips and went back to work and school.
Social media are treasure troves of data. When people are sick they will often post about it, and even include their symptoms. Plus, mobile technology enables people to post wherever they are and associate locations with posts. Much like the EMS and 911 data sources, this unstructured data can be analyzed with social media analytics to reveal the geographic areas where symptoms are showing up most frequently. The volume of “chatter” could also give advance indication of the magnitude of a problem or outbreak, even before a holiday weekend ends.
While certain elements of the film take dramatic license, there is no doubt that such a pandemic could occur. Contagion touches on other challenges facing pandemic response but all of those can be alleviated – though not solved – by earlier detection. Getting ahead of the curve improves response and containment and reduces the demand on local, state and federal resources – and ultimately, saves lives.