Exploiting call data of a public helpline, Indian and Pakistani scientists have created a disease forecasting model that brought down the number of dengue cases in Lahore from 21,000 to 250, within a year.
The success of the model comes out at a time when several Indian cities, including the national capital, are trying to grapple with the vector-borne disease that returns every monsoon.
The forecasting model not only alerts the public health officials on the number of patients but also gives them adequate time to clean up the trash and remove stagnant water from areas where the disease was likely to strike. The intervention drastically reduced the cases.
It was put in place following a major dengue outbreak in Lahore in 2011 that affected 21,000 people and killed 350 individuals. With the forecast model in place, the number of cases came down to 257 a year later and 1,600 in 2013.
“This can be easily adopted in many Indian cities where dengue is a big problem. The success of such a system also requires awareness to enable people to use such a hotline,” Lakshminarayan Subramanian, the lead researcher from New York University, told DH.
Subramanian and his colleague Shankar Kalyanaraman developed it with support of scientists at the Information Technology University, Lahore, and Punjab Information Technology Board, with a minimal cost. For the last five years, it is in operation and getting updated.
Its principal input is thousands of calls made to a telephone helpline used by the people not only to know about dengue and its symptoms, but also for reporting mosquito breeding, requesting for fogging, reporting water stagnation and checking bed availability in hospitals.
“Our system not only flags an outbreak but also makes an accurate forecast of both the number of patients and their locations two to three weeks ahead of time. The ability of our system to accurately forecast patients and their locations is critical for the government to mobilise and target its resources to contain an outbreak,” the researchers reported in the July 8 issue of the journal ‘Science Advances’.
The hotline enabled collection of epidemic-style symptoms directly from the users in a crowdsourced manner. As dengue fever takes between 3-14 days for the symptoms to even show up in a patient, there is a big lag between the incidence of a disease and the patient actually coming to a hospital.
