School of Informatics and Computing researchers using complex networks analysis have found that Instagram, a growing social media platform among teens, can be used to uncover drug-drug interactions (DDI) and adverse drug reactions (ADR). The work shows that this popular social media service is a very powerful source of data with great promise in the public-health domain.
The study, "Monitoring Potential Drug Interactions and Reactions via Network Analysis of Instagram User Timelines," was supported by an R01 grant from the National Institutes of Health as well as a gift from Persistent Systems, Inc., founded and led by SoIC alumnus Anand Deshpande. It was recently published and presented at the Pacific Symposium on Biocomputing (PSB 2016) in Hawaii (PubMed). The results are based on almost 7,000 user timelines associated with depression drugs which combined have 5 million-plus posts.
“We build knowledge networks from what people are talking about on their public timelines,” said Rion Brattig Correia, a doctoral student in the School of Informatics and Computing at Indiana University, who is first author on the study. “This enables us to visually graph and inspect how different drugs, symptoms and even natural products are connected to each other.”
Through his study, Correia was the recipient of two travel awards to present the work at PSB 2016. One award was provided by the NIH to support noteworthy graduate student work accepted for publication at the Pacific Symposium on Biocomputing, while the other was from the IU Graduate and Professional Student Government (GPSG).
"The universe of social media provides a very promising source of large-scale data that can help identify DDI and ADR in ways that have not been possible," said the study’s senior author Luis M. Rocha, professor of informatics, director of the Complex Networks and Systems Ph.D. Program, and a member of the IU Network Science Institute. "Given the large number of users, analysis of social media data may be useful to identify under-reported, population-level pathology associated with DDI, thus further contributing to improvements in population health. Moreover, tapping into this data allows us to infer drug interactions with natural products – including cannabis – which constitute an array of DDI very poorly explored by biomedical research thus far."
The study of social media for public health monitoring and surveillance has been receiving increasing attention in the biomedical domain. The topic was allocated its own session at PSB, where the IU researchers presented their work. This session focused on novel approaches to text and data mining methods that respond to the specific requirements of social media and can prove invaluable for public health surveillance and precision medicine.
The authors have also developed a social media drug explorer tool that enables public health monitoring scientists to quickly access and discover important Instagram user timelines and/or relevant connections in the knowledge network.
“The tool can assist the discovery of new drug interactions or symptoms based on what users are saying,” Correia said. “We are really looking at what people are describing when taking a specific drug.”