Wikipedia articles are a powerful tool for monitoring and prognosis of disease worldwide, according to a study published today in the journal PLoS Computational Biology. Scientists at Los Alamos National Laboratory (New Mexico), predicted several cases up to 28 days in advance of the data recorded respective health systems using Wikipedia.
We compared the number of visits to the page of dengue in Portuguese and all the links on this page to the Ministry of Health.Sara del Valle
One case dealt with an outbreak of dengue in Brazil. “We compared the number of visits to the page of dengue in Portuguese and all the links on this page to the Ministry of Health,” says Sara del Valle, head of research at Los Alamos.
The relationship is because people seek information about the disease in Wikipedia long before seeing a doctor, so that “our results show that we can predict dengue fever in Brazil with up to four weeks in advance” through increased visits Articles that talk about the symptoms and treatment of this epidemic and contagious fever.
Other cases of flu in the US, Poland and Japan; tuberculosis in China; and a triple analysis of dengue, influenza and tuberculosis in Thailand show similar results and indicate that “Wikipedia can be used not only to monitor but also to predict the prevalence of disease around the world.”
The impact of this information to health departments in the respective countries are facing tremendous improvement planning. “Forecasting helps public health to have sufficient resources to treat patients diseases,” says Sara del Valle, and thus “save lives.”
Also, one of the most interesting findings of this research is that computational models can be trained and transfer from one country to another. This means that data from one place scientists can create a model and use it to track and prognosis of disease in another. This aspect is especially interesting because “some countries do not collect data on the prevalence of disease,” says the specialist.
However, not all cases analyzed were used to test the ability of Wikipedia to predict diseases. The lack of reliable data on the ministries of some countries was the main obstacle, but there are other factors that cause “noise” when performing the studies. For example, “many people can visit the website of Ebola” after the alarm generated by the epidemic in Africa, “and that does not mean there is an outbreak in the US,” said Sara del Valle.
The choice of Wikipedia for this study was not accidental, as this popular encyclopedia provides information about access to your articles every time and also free. Thus, all the information has been validated by other researchers and the work has scientific validity. Other channels, such as Twitter and Google Trends, offered a more incomplete and biased information and, therefore, were excluded for the study.
Regardless of the outcome, Sara del Valle believes that people seek health information online has positive and negative consequences. In his view, the danger is that patients self-diagnose and wrongly not seek help in time. However, generally considered “good report that the population” and even in some situations, “the Internet has played a very important role in diagnosing rare cases.” In addition, if patients are reported, they may have “prepared questions” when to seek medical advice.
Following the publication of this article, scientists are already working with the Wikimedia Foundation, responsible for the Wikipedia project, to distribute data more accurate geographic information, and so far have used the language as a reference to the country. Once you get this data, researchers hope to create a system capable of predicting disease globally and can be found on a website.
This archive content was originally published November 16, 2014 (www.betawired.com)