By studying page views made by visitors to Wikipedia, researchers at Los Alamos National Laboratory believe it may be possible to predict epidemics as they emerge.
Influenza outbreaks around the United States, Japan, Poland and Thailand were predicted by researchers before they happened, by monitoring Wikipedia page views. Tuberculosis was also predicted in China and Thailand, along with Dengue fever in Brazil and Thailand.
All of the outbreaks were predicted four weeks or more in advance, except for one outbreak of tuberculosis in China.
"A global disease-forecasting system will change the way we respond to epidemics. In the same way we check the weather each morning, individuals and public health officials can monitor disease incidence and plan for the future based on today's forecast. The goal of this research is to build an operational disease monitoring and forecasting system with open data and open source code," Sara Del Valle of Los Alamos National Laboratory said.
Del Valle and her team developed a way to "train" computer models for new regioins, based on data collected in other regions. Using these techniques, a model developed for Poland could be applied to predictions for Thailand, or other countries. This could assist researchers in creating models for lands where public health information is not readily available.
Access logs for Wikipedia are open data, readily available to the public.
Disease monitoring traditionally involves studying medical records, hospital visits, and calls made to doctor's offices. However, as people seek more data online, many are seeking healthcare information on the World Wide Web before consulting with wellness professionals. The older techniques are also much slower, and more expensive, than gathering data from Wikipedia.
In eight out of 14 cases examined by researchers, visits to Wikipedia pages on diseases spiked four weeks prior to local governments declaring an outbreak.
Languages used by visitors to Wikipedia were used as method of determining location, although researchers admit the technique only provided approximate geographical data.
Influenza takes the lives of between 3,000 and 49,000 Americans each year, and more accurate predictions could assist medical personnel and organizations in delivering health care where it is most needed.
Many medical researchers are wary of claims that Wikipedia visits could be used to predict outbreaks of disease. Only those people who have Internet access, are familiar with the illness they may be experiencing, and who chose to look up information on Wikipedia would be included in such predictions.
Study of how Wikipedia could be used to help predict outbreaks of disease and emerging epidemics was detailed in the journal PLOS Computational Biology.
