Scientists from the Information Technology University (ITU), Punjab, and New York University used call data to build a statistical model — described July in a paper in Science Advances — that accurately predicts dengue patient numbers in 10 sub-regions of the city 2—3 weeks in advance.
“Most studies develop methods to predict infectious diseases but rarely develop front-end ways to access the predictions.”
Ben Althouse, Santa Fe Institute
"The exciting thing compared to previous disease surveillance systems is that this doesn't just tell you trends, it actually forecasts the number of dengue patients you will get up to three weeks into the future," says Umar Saif, vice-chancellor of ITU and chairman of the Punjab Information Technology Board, which implemented the system.
“Any developing country without access to sophisticated surveillance can set up a helpline and use the statistical methods, described in this paper, to predict cases before it gets out of hand,” says Saif. The paper shows that an accurate, location-specific disease forecasting system can be built by analysing call data from a public health hotline.
The free helpline, set up in 2011 following a major epidemic that killed 350 people, allows citizens to check disease symptoms and treatment options, and also report stagnant water where mosquitoes breed. Medically-trained operators can then identify suspected dengue cases and pinpoint dengue hotspots. Public health workers equipped with smartphone apps can also geo-tag houses of confirmed dengue cases and deploy containment efforts based on correlating this data and call volumes. Since there is no vaccine or cure for dengue fever containment efforts rely on disease surveillance and vector control.
Rafiq Khanani, president, Infection Control Society, Pakistan, says: “The method described in this paper is a practical way of prediction, keeping in view the resources and socio-cultural milieu of the country.”
Since the helpline was first set up dengue cases in Punjab have dropped from 21,000 in 2011 to under 500 a year now.
Ben Althouse, a disease modelling expert at the Santa Fe Institute in the US, says the value of the model lies in allowing public health officials to access data on the real-time dashboard. “Most studies develop methods to predict infectious diseases but rarely develop front-end ways to access the predictions.”
This piece was produced by SciDev.Net’s South Asia desk.