The investigation of public Web searches allows us to expand the boundaries of public health research. We call this tool infodemiology. This portmanteau of informatics and epidemiology can be defined as the science of how information is distributed throughout the internet and what the determinants of this may be, with the aim of informing public health policies.
Tools exist that allow us to make sense of public interest in topics, for example Google Trends, which quantifies Google searches for specific key words, and Global Public Health Intelligence, which monitors news and other media on the internet to gather information on disease outbreaks.
There are, however, other applications of infodemiology and Peru, in its state of epidemiological transition, presents a particularly interesting case study for the application of infodemiological tools in public health. For instance, infodemiology can be used to detect increased public interest in a particular public health matter at a time of increased media coverage. One study (1) reported that the number of searches for the key word “tuberculosis” in Peru increased in March, during the week in which World Tuberculosis Day is celebrated. Such research can help public health officials to understand whether mass health education campaigns spark the interest of the general population.
Another strong driver of public interest in health issues in Peru are celebrities. At CRONICAS Center of Excellence in Chronic Disease, we hypothesised that media scandals involving public figures and illegal substances would spark the public’s attention in these substances. Indeed, a recent study (2) from CRONICAS found that the portrayal of steroid use in reality television and in sports media was a strong driver of the public’s online interest in anabolic androgenic steroids. Peaks in search volumes for key words related to anabolic steroids occurred concomitantly with the publication of news articles in Peruvian media portraying their use.
Infodemiology, however, is still imperfect. The accuracy of various infodemiology tools have been criticised, for example when Google Trends was condemned for overestimating the number of Flu cases in the United States in 2012-13. Further, trending media reports about illnesses or fear-inducing portrayal of certain epidemics may cause a confounding effect in the volume of searches for different keyword queries. Similarly, in-context semantic analysis is hard to perform. For instance, if someone is performing a Google query for the term “flu”, it may be possible that person is experiencing influenza symptoms, or it may be a student writing a term paper for a health class, or is simply someone who has read an article on H1N1 and may be interested in the topic. Crude estimates for peaks in flu may be unable to discern between each of these cases, and it may be easy to fall in a correlation/causation fallacy. Thus, setting up the proper research question and hypotheses to be tackled by infodemiology is necessary before analysing peaks in information-seeking behaviour.
Whilst infodemiology has much to offer, the challenge for Peru is not one of accumulation of data, but its collation and analyses. New hypotheses can now be tested with these easy-to-collect metrics in Peru and elsewhere. However, it is important to remember that internet searchers can be a result of a number of factors and may not necessarily reflect the outcome of interest; combining infodemiology with other epidemiological practices will strengthen this area of research and its applications in public health.
References
- Sebastián Tapia-Villarreal, Álvaro Taype-Rondán, Impact of the World Tuberculosis Day in searches on Google in South American countries, Rev Med Chile 2016; 144: 406-408
- J.L. Avileza, A. Zevallos-Moralesc, A. Taype-Rondana, Use of enhancement drugs amongst athletes and television celebrities and public interest in androgenic anabolic steroids. Exploring two Peruvian cases with Google Trends, Public Health, 2017.
Jose L. Avilez is affiliated with the CRONICAS Center of Excellence in Chronic Diseases – University Peruana Cayetano Heredia and the Faculty of Mathematics, University of Waterloo, Ontario, Canada.
Leave a Reply