Twitter: A Good Place to Detect Health Conditions

Abstract

With the proliferation of social networks and blogs, the Internet is increasingly being used to disseminate personal health information rather than just as a source of information. In this paper we exploit the wealth of user-generated data, available through the micro-blogging service Twitter, to estimate and track the incidence of health conditions in society. The method is based on two stages: we start by extracting possibly relevant tweets using a set of specially crafted regular expressions, and then classify these initial messages using machine learning methods. Furthermore, we selected relevant features to improve the results and the execution times. To test the method, we considered four health states or conditions, namely flu, depression, pregnancy and eating disorders, and two locations, Portugal and Spain. We present the results obtained and demonstrate that the detection results and the performance of the method are improved after feature selection. The results are promising, with areas under the receiver operating characteristic curve between 0.7 and 0.9, and f-measure values around 0.8 and 0.9. This fact indicates that such approach provides a feasible solution for measuring and tracking the evolution of health states within the society.

Social media users have different experiences, motivations, and quality of life

“…Social networking sites are used as an information sources but also for individuals to disseminate personal health information, experiences and knowledge (Scanfeld et al, 2010;Prieto et al, 2014). Use of these online sites has translated to significant psychological and physical well-being for users.…”

“…Historically, internet users have searched the web for medical information. However, with the growth of social network sites such as Facebook, Twitter, Instagram, and Google þ, the Internet is now used not only as an information source but also for individuals to disseminate personal health information, experiences and knowledge (Scanfeld et al, 2010;Prieto et al, 2014). Research conducted in the area of emotional disclosure (Pennebaker et al, 1988), an intervention in which individuals write or talk about emotionally stressful materials reveals beneficial effects in some but not all participants (Vedhara et al, 2010).…”

Detecting themes of public concern: A text mining analysis of the Centers for Disease Control and Prevention's Ebola live Twitter chat

“…2 Social media platforms, such as Twitter, allow health communicators to gather large samples of user-generated content to garner insights to inform early response strategies. [3][4][5][6] Text mining of social media data has been used to detect and track diseases, 7,8 natural disasters, 3,9 and life-threatening events, 6,10 as well as to estimate public knowledge concerning health issues. 5,11,12 During crises, social media can serve as crucial communication channels or catalysts for public panic with a proliferation of unmoderated comments.…”

Characterizing Tweet Volume and Content About Common Health Conditions Across Pennsylvania: Retrospective Analysis (Preprint)

“…Social media channels like Twitter offer a new opportunity to track regional health trends by observing health related communication generated by the public and for the public. [1][2][3][4][5][6][7] There is an opportunity to determine how emerging digital data sources are complementary (i.e., social media data has similar findings to traditional health data sources) and augmentative (i.e., social media provides new real-time information about health not available in data collected through traditional means). To better quantify the value added of social media for public health surveillance, needed is an understanding of how much data exists about different health conditions.…”