Information Epidemiology and Surveillance in the Google Era

Abstract

Information epidemiology (infodemiology) approaches are increasingly employed in exploring online behavior and in predicting/forecasting diseases/epidemics, providing real time information and the revealed instead of the stated users’ interests/preferences that are not otherwise accessible, thus tackling issues of traditional data collection and monitoring. This Thesis examines how users’ Google behavior towards health topics can be useful in public health epidemiology and surveillance. Studying the state of the art in 2017, gaps identified included an up-to-date systematic review, a methodology framework for rigorous data collection and reporting, as well as limited number of approaches in predictions/forecastings and several public health topics that had not been studied before. To fill the gaps and advance the topic, this Thesis, consisting of 8 interconnected papers, includes: a systematic review of Google Trends in health/medicine categorized by methodology approaches; a methodology framework for rigorous data collection and reporting; six research papers in public health topics, namely COVID-19, STIs, Measles, and asthma, employing basic statistical tools to explore associations, predictability, and forecastings. Several factors limiting the applicability of this approach were also identified and discussed, e.g., lack of small interval health data, misspellings, sudden events. The collective results could have significant implications for effective policy making, suggesting how multidisciplinary approaches in public health epidemiology and surveillance could make full use of the information and web tools that are available. The latter was especially evident during the COVID-19 pandemic -with open access to real time data- when such approaches were employed for epidemiology and surveillance. During chaotic conditions like in pandemics/epidemics, when policy makers are required to make fast and important decisions, it is vital to proceed with a statistical understanding of Google Trends time series and the users’ behavior in accordance with its real determinants, combining medical and non-medical parameters from a variety of research fields, that also take into account the public’s awareness and online behavior towards the explored topics.