The Pentagon Wants You to Help Them Find the Next Pandemic

Ever heard of Chikungunya? It’s a mosquito-borne virus that causes joint pain and fever and can be debilitating. It’s also spreading fast, having hit the Americas for the first time in decades at the end of last year and new cases were reported in Florida this last month. There is no official cure, yet, but recent research into a vaccine shown promise.

If you can build a model for the Defense Advanced Research Projects Agency, or DARPA, explain where it’s going next, and do it by the end of September, they’ll give $150,000.

The initial submission, the entry exam if you will, should “contain a detailed description of the planned data sources and model applicability.” That has to be in by Sept. 1. It should include “predictions for the next six months, followed by five monthly update submissions, due on the first of each subsequent month, with predictions for the remaining period of the challenge.”

A forecast of peak new cases per country or territory is due Oct.1.

The final submission is due Feb. 1, 2015 and should include a full account of the methods used to reach the predictions. You can read the announcement here.

How DARPA, which was originally thought up to conceive of next-generation weaponry, got into the business of epidemiology funding says a lot about the changing nature of threats since the agency was founded years ago.

The recent spread of the Ebola virus across Africa surprised the World Health Organization (WHO)  and others. What many found far more surprising was that an experimental ‘cure” the ZMAPP cocktail, had been in the works for years, courtesy of funding by DARPA.

 Chikungunya (CHIKV) isn’t the deadliest thing you can catch. So why is DARPA funding research to model its spread? “A substantial number of patients develop chronic joint pain (weeks and months, perhaps even years) after their infection. There are substantial impacts on health care systems and economic productivity in communities with large outbreaks, both acutely and then with the long-term health impacts,” Arny Col. Matthew Hepburn, DARPA Program Manager told Defense One.

So far, as of July 31, some 473,167 people were infected with the illness in the Caribbean and 284 have been infected across continental North America. “Scientists anticipated transmission of chikungunya in the U.S. since the 2005-2007 outbreak in Asia, Africa, and Europe,” Dr. Rajeev Vaidyanathan, Director of Environmental Science at Clarke, which focuses on mosquito-borne illnesses said in a press release.

“Unlike West Nile, which is maintained in birds, humans are the reservoir for CHIKV—if a mosquito bites an infected human, it can become infected and can spread the virus to other humans,” Vaidyanathan says.  “Managing and implementing a mosquito control program is critical to lowering the risk of infection. Because the mosquitoes responsible for transmitting CHIKV breed in temporary containers, such as pots and tires, reducing the availability of breeding sites and controlling adult populations are essential to controlling the spread of the virus.”

More importantly, Chikungunya provides a testing ground for modeling the spread of other, similar illnesses. The thing about RNA viruses is that they mutate very rapidly, a phenomenon called antigenic drift.  Chikungunya means that it could become more deadly, which is what happened in 2009 with the H5N1 influenza strain. That means that understanding how it’s moving is key preventing a future, more lethal Chikungunya from hitting human populations.

The good news is that modeling disease spread is evolving faster as well because of information technology (recent disappointment about the performance of Google Flu Trends not withstanding.) Earlier this year, a lesser known modeling engine called HealthMap was able predict the current outbreak before the WHO made an official announcement.

Some of this prediction has become strangely specific. Back in 2010, a researcher named Adam Sadelik showed that with a dataset of 630,000 geo-tagged tweets he could predict, in certain instances, a future flu case with 90 percent accuracy up to 8 days before he subject showed signs of flu.

We’re also able to chart not only where flu goes, but how it’s changing on the way there. Databases like ProMED-mail and Global Infectious Diseases and Epidemiology Network (GIDEON) have given epidemiologists and healthcare workers in remote clinics around the world new tools for logging and broadcasting infectious disease incidents. And visualization engines like SupraMap allow agencies to chart out where antigenic rifts might occur, so you can begin to see new superbugs emerging in closer and closer to real-time. The most recent challenge is a part of that and builds off of the insight gained from DARPA's 2010 Prophecy (Pathogen Defeat) program. But whereas Prophecy explored virus mutation this challenge is more focused on spread.

“We anticipate learning a great deal about new techniques for modeling infectious diseases.  We hope the knowledge gained through this challenge can support capabilities for informing decision-making during public health emergencies, and may have even broader applicability in emergency response,” Hepburn said.

Modeling Chikungunya spread is a dress rehearsal for something more dangerous down the road.

This post has been updated.