Researchers from the University of Washington and elsewhere say they can forecast the rise and fall of dengue fever outbreaks as much as three weeks in advance by analyzing the patterns of phone calls to public-health hotlines.
Dengue fever is spread by a mosquito-borne virus that infects millions of people a year. The infection causes a flu-like illness and can produce potentially lethal complications. There’s no specific cure or vaccine available, but early detection and proper medical care can reduce fatality rates below 1 percent, the World Health Organization says.
The researchers’ experiment, described today in a paper published by the journal Science Advances, focused on cases of the disease reported in Lahore, Pakistan. A telephone hotline was set up in Lahore in the wake of a 2011 dengue epidemic to help the public deal with the disease. Researchers analyzed more than 300,000 calls made to the hotline during 2012 and 2013 to see how many came from which locale, on a block-by-block basis.
They found that call volumes spiked in advance of two outbreaks, in August of 2012 and 2013, and rose at the same time as a dramatic increase in hospital cases in the fall of 2013. “The appeal of our model is its usefulness despite its sheer simplicity,” the authors wrote.
A simple, low-cost forecasting system is particularly important for developing countries like Pakistan.
“Developing worlds face challenges in tackling major outbreaks due to limited resources,” co-author Fahad Pervaiz, a UW doctoral student in computer science and engineering, said in a news release. “Our technique will equip public officials with tools to inform them about where to apply these resources in advance and hopefully save millions of lives.”
One of the challenges was to compensate for widely varying levels of detail in the descriptions that callers gave when they contacted the Pakistani hotline – not only about their symptoms, but about their location as well.
“Extracting the indicators that you can use to reliably predict where the disease would emerge two to three weeks in advance is hard,” Pervaiz said. “Our innovation is to build models that can deal with such noisy data and still give you reasonable accuracy at a micro-level of where the dengue patients are going to start appearing.”
The researchers said their forecasting system has already helped public health officials in Lahore take early steps to contain the spread of dengue fever and provide hospitals with early warning of outbreaks in their vicinity.
“We believe that this system can also be used for a broad array of diseases beyond dengue, and can easily be replicated in other developing countries at low costs,” they wrote.
The first author of the Science Advances paper, “Fine-Grained Dengue Forecasting Using Telephone Triage Services,” is Nabeel Abdur Rehman of Lahore’s Information Technology University and New York University. In addition to Pervaiz, co-authors include NYU’s Shankar Kalyanaraman, Talal Ahmad, Lakshminarayanan Subramanian and Umar Saif of ITU and the Punjab Information Technology Board.